1 2013-11-06 Nick Clifton <nickc@redhat.com>
3 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
4 unless it is the last part of a spec substring.
5 * doc/invoke.texi (Spec Files): Document space insertion
8 2013-11-06 Christian Bruel <christian.bruel@st.com>
10 * gcc/config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
11 Factorize probabilities, Use adjust_address instead of
12 adjust_automodify_address when possible. Enable for optimize.
13 (sh_expand_strlen): New function.
14 * gcc/config/sh/sh-protos.h (sh_expand_strlen): Declare.
15 * gcc/config/sh/sh.md (strlensi): New pattern.
16 (UNSPEC_BUILTIN_STRLEN): Define.
18 2013-11-06 Jakub Jelinek <jakub@redhat.com>
21 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
22 (expand_assignment): If *bitpos is negative, set *offset
23 and adjust *bitpos, so that it is not negative.
25 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
27 * config/i386/bdver3.md : Added two additional decoder units
28 to support issue rate of 4 and remodeled vector unit.
29 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
30 architectures is set to 4.
31 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
32 lookahead is set to 4 for BD architectures.
34 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
36 * config/rs6000/rs6000.c (rs6000_option_override_internal):
37 Remove restriction against use of VSX instructions when generating
38 code for little endian mode.
40 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
42 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
43 for both big and little endian.
44 (mulv8hi3): Swap input operands for merge high and merge low
45 instructions for little endian.
47 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
49 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
50 define_insn to define_expand that uses even patterns for big
51 endian and odd patterns for little endian.
52 (vec_widen_smult_even_v16qi): Likewise.
53 (vec_widen_umult_even_v8hi): Likewise.
54 (vec_widen_smult_even_v8hi): Likewise.
55 (vec_widen_umult_odd_v16qi): Likewise.
56 (vec_widen_smult_odd_v16qi): Likewise.
57 (vec_widen_umult_odd_v8hi): Likewise.
58 (vec_widen_smult_odd_v8hi): Likewise.
59 (altivec_vmuleub): New define_insn.
60 (altivec_vmuloub): Likewise.
61 (altivec_vmulesb): Likewise.
62 (altivec_vmulosb): Likewise.
63 (altivec_vmuleuh): Likewise.
64 (altivec_vmulouh): Likewise.
65 (altivec_vmulesh): Likewise.
66 (altivec_vmulosh): Likewise.
68 2013-11-05 Mike Stump <mikestump@comcast.net>
70 * Makefile.in (mostlyclean): Remove c-family objects.
72 2013-11-05 Ian Lance Taylor <iant@google.com>
74 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
75 If possible, add .cfi directives to record change to bx.
76 * config/i386/i386.c (ix86_emit_cfi): New function.
77 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
79 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
82 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
83 immediately after a label for a tablejump pattern.
85 * config/arm/arm.c (is_jump_table): Remove.
86 (create_fix_barrier): Use tablejump_p instead.
87 (arm_reorg): Likewise.
88 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
89 (thumb2_output_casesi): Likewise.
90 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
91 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
92 casesi_shift_media, casesi_load_media): Likewise.
93 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
94 * config/microblaze/microblaze.md: Likewise.
96 2013-11-05 Tobias Burnus <burnus@net-b.de>
98 * doc/invoke.texi (-Wdate-time): Document.
100 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
102 * double-int.c (lshift_double, rshift_double): Remove
103 SHIFT_COUNT_TRUNCATED handling.
105 2013-11-05 Jeff Law <law@redhat.com>
107 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
108 * common.opt (-fisolate-erroneous-paths): Add option and
110 * gimple-ssa-isolate-paths.c: New file.
111 * gimple.c (check_loadstore): New function.
112 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
113 Verify OP is in the argument list and the argument corresponding
114 to OP is a pointer type. Use operand_equal_p rather than
115 pointer equality when testing if OP is on the nonnull list.
116 Use check_loadstore rather than count_ptr_derefs. Handle
117 GIMPLE_RETURN statements.
118 * tree-vrp.c (infer_nonnull_range): Remove.
119 * gimple.h (infer_nonnull_range): Declare.
120 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
121 * passes.def: Add pass_isolate_erroneous_paths.
122 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
123 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
124 * tree-ssa.c (struct count_ptr_d): Remove.
125 (count_ptr_derefs, count_uses_and_derefs): Remove.
126 * tree-ssa.h (count_uses_and_derefs): Remove.
128 2013-11-05 Jakub Jelinek <jakub@redhat.com>
130 PR rtl-optimization/58997
131 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
132 get_iv_value to be in iv->mode rather than iv->extend_mode.
133 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
134 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
135 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
138 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
140 * gimple.h: Move some prototypes to gimple-expr.h and add to include
142 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
143 Move to gimple-expr.h.
144 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
145 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
146 is_gimple_id, virtual_operand_p, is_gimple_addressable,
147 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
149 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
150 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
151 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
152 is_gimple_address, is_gimple_invariant_address,
153 is_gimple_ip_invariant_address, is_gimple_min_invariant,
154 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
155 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
156 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
157 gimple_decl_printable_name, useless_type_conversion_p,
158 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
160 * gimple-expr.c: New File.
161 (useless_type_conversion_p, gimple_set_body, gimple_body,
162 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
163 gimple_can_coalesce_p, extract_ops_from_tree_1,
164 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
165 is_gimple_address, is_gimple_invariant_address,
166 is_gimple_ip_invariant_address, is_gimple_min_invariant,
167 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
168 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
169 is_gimple_mem_ref_addr): Relocate here.
170 * Makefile.in (OBJS): Add gimple-expr.o.
172 2013-11-05 David Malcolm <dmalcolm@redhat.com>
174 * gengtype-parse.c (struct_field_seq): Support empty structs.
176 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
178 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
180 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
182 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
184 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
186 (LONG_DOUBLE_TYPE_SIZE): New define.
187 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
189 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
192 * doc/extend.texi [visibility ("visibility_type")]: Add example
193 about visibility attribute on namespace declaration.
195 2013-11-05 Richard Biener <rguenther@suse.de>
198 * passes.def (all_passes): Start with pass_fixup_cfg again.
200 2013-11-05 Richard Biener <rguenther@suse.de>
202 PR tree-optimization/58955
203 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
206 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
208 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
210 (vec_pack_ufix_trunc_v2df): Likewise.
212 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
215 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
216 pattern, instead of word_mode.
218 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
220 (set_storage_via_setmem): Likewise.
222 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
224 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
225 gimple.h and the rest of the condition in eliminate_build.
226 (eliminate_build): Call new routine.
227 * gimple.h (phi_ssa_name_p): Delete.
229 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
231 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
232 case of no prefix and reserving zero slots, because when that's the
233 case we'll never get here.
234 * vec.h (va_heap::reserve): Don't try and handle
235 vec_prefix::calculate_allocation returning zero because that should
238 2013-11-05 Richard Biener <rguenther@suse.de>
241 * tree-dfa.c (get_ref_base_and_extent): Merge common code
242 in MEM_REF and TARGET_MEM_REF handling. Make sure to
243 process trailing array detection before diving into the
244 view-converted object (and possibly apply some extra offset).
246 2013-11-05 Joseph Myers <joseph@codesourcery.com>
248 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
250 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
252 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
254 PR tree-optimization/58958
255 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
256 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
258 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
260 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
261 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
263 2013-11-05 Jakub Jelinek <jakub@redhat.com>
265 PR tree-optimization/58984
266 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
267 set *SIZE_P if non-NULL on success.
268 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
270 (ipcp_transform_function): Likewise. Punt if size of access
271 is different from TYPE_SIZE on v->value's type.
273 2013-11-05 Tobias Burnus <burnus@net-b.de>
275 * doc/invoke.texi (-fopenmp-simd): Document new option.
276 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
277 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
278 * tree.c (attribute_value_equal): Ditto.
280 2013-11-04 Wei Mi <wmi@google.com>
282 * sched-rgn.c (add_branch_dependences): Keep insns in
283 a SCHED_GROUP at the end of BB to remain their location.
285 2013-11-04 Wei Mi <wmi@google.com>
287 * gcc/config/i386/i386.c (memory_address_length): Extract a part
288 of code to rip_relative_addr_p.
289 (rip_relative_addr_p): New Function.
290 (ix86_macro_fusion_p): Ditto.
291 (ix86_macro_fusion_pair_p): Ditto.
292 * gcc/config/i386/i386.h: Add new tune features about macro-fusion.
293 * gcc/config/i386/x86-tune.def (DEF_TUNE): Ditto.
294 * gcc/doc/tm.texi: Generated.
295 * gcc/doc/tm.texi.in: Ditto.
296 * gcc/haifa-sched.c (try_group_insn): New Function.
297 (group_insns_for_macro_fusion): Ditto.
298 (sched_init): Call group_insns_for_macro_fusion.
299 * gcc/target.def: Add two hooks: macro_fusion_p and
302 2013-11-04 Kostya Serebryany <kcc@google.com>
304 Update to match the changed asan API.
305 * asan.c (asan_function_start): New function.
306 (asan_emit_stack_protection): Update the string stored in the
307 stack red zone to match new API. Store the PC of the current
308 function in the red zone.
309 (asan_global_struct): Update the __asan_global definition to match
311 (asan_add_global): Ditto.
312 * asan.h (asan_function_start): New prototype.
313 * final.c (final_start_function): Call asan_function_start.
314 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
317 2013-11-04 Wei Mi <wmi@google.com>
319 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): Separate
320 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
321 * gcc/config/i386/i386.c (ix86_option_override_internal): Ditto.
322 (ix86_issue_rate): Ditto.
323 (ix86_adjust_cost): Ditto.
324 (ia32_multipass_dfa_lookahead): Ditto.
325 (ix86_sched_init_global): Ditto.
326 (get_builtin_code_for_version): Ditto.
327 * gcc/config/i386/i386.h (enum target_cpu_default): Ditto.
328 (enum processor_type): Ditto.
329 * gcc/config/i386/x86-tune.def (DEF_TUNE): Ditto.
331 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
333 PR rtl-optimization/58967
334 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
335 !lra_in_progress for mode sizes bigger word.
337 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
339 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
340 arguments to merge instruction for little endian.
341 (vec_widen_umult_lo_v16qi): Likewise.
342 (vec_widen_smult_hi_v16qi): Likewise.
343 (vec_widen_smult_lo_v16qi): Likewise.
344 (vec_widen_umult_hi_v8hi): Likewise.
345 (vec_widen_umult_lo_v8hi): Likewise.
346 (vec_widen_smult_hi_v8hi): Likewise.
347 (vec_widen_smult_lo_v8hi): Likewise.
349 2013-11-04 Ian Lance Taylor <iant@google.com>
351 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
352 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
353 builtins that take pointers.
354 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
355 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
356 (run_gcc): Pass -fnon-call-exceptions.
358 2013-11-04 Jakub Jelinek <jakub@redhat.com>
360 * optabs.c (expand_vec_perm): Revert one incorrect line from
363 PR tree-optimization/58978
364 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
365 use_stmt by single_imm_use directly. Only call single_imm_use
368 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
370 PR rtl-optimization/58968
371 * lra-spills.c (return_regno_p): New function.
372 (lra_final_code_change): Use it.
374 2013-11-04 Joseph Myers <joseph@codesourcery.com>
376 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
377 * target.def (float_exceptions_rounding_supported_p): New hook.
378 * targhooks.c (default_float_exceptions_rounding_supported_p): New
380 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
381 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
383 * doc/tm.texi: Regenerate.
384 * config.gcc (powerpc*-*-linux*): Set extra_objs.
385 * config/rs6000/rs6000-linux.c: New file.
386 * config/rs6000/rs6000-protos.h
387 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
388 * config/rs6000/linux.h
389 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
390 * config/rs6000/linux64.h
391 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
392 * config/rs6000/t-linux (rs6000-linux.o): New rule.
393 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
395 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
397 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
398 Replace the define_insn_and_split with a define_insn and two
399 define_splits, with the split after reload re-permuting the source
400 register to its original value.
401 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
402 (*vsx_le_perm_store_v8hi): Likewise.
403 (*vsx_le_perm_store_v16qi): Likewise.
405 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
407 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
410 2013-11-04 Jakub Jelinek <jakub@redhat.com>
412 PR tree-optimization/58946
413 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
414 bbs with bbinfo[idx].op != NULL before all blocks with
415 bbinfo[idx].op == NULL.
417 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
419 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
420 (avr_dump_double_int_hex): Likewise.
421 (avr_log_vadump): Remove %D and %X handling.
422 * config/avr/avr.c (avr_double_int_push_digit): Delete.
423 (avr_map_op_t): Change map from double_int to unsigned int.
424 (avr_map_op): Update accordingly.
425 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
426 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
427 unsigned ints rather than double_ints.
429 2013-11-03 Marek Polacek <polacek@redhat.com>
431 Implement -fsanitize=vla-bound.
432 * opts.c (common_handle_option): Handle vla-bound.
433 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
434 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
435 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
437 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
439 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
442 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
444 * config/i386/constraints.md (Ts, Tv): New address constrains.
445 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
446 constraint for address_no_seg_operand.
447 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
448 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
449 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
450 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
451 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
452 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
453 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
454 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
455 constraint for vsib_address_operand.
457 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
459 * gcse.c (pre_delete): Remove references to regmove from comments.
460 * recog.c: (validate_replace_rtx_1): Likewise.
461 * config/rl78/rl78.c: Likewise.
462 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
463 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
464 * common/config/mmix/mmix-common.c: Likewise.
466 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
468 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
469 * gimplify.c (gimplify_compound_lval): Likewise.
470 * graphite-clast-to-gimple.c (gloog): Likewise.
471 * graphite-dependences.c (loop_is_parallel_p): Likewise.
472 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
473 (limit_scop); Likewise.
474 (build_scops): Likewise.
475 (dot_scop): Likewise.
476 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
477 (build_scop_drs): Likewise.
478 (insert_stmts): Likewise.
479 (insert_out_of_ssa_copy): Likewise.
480 (remove_phi): Likewise.
481 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
482 * hw-doloop.c (discover_loop): Likewise.
483 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
484 * tree-dfa.c (dump_enumerated_decls): Likewise.
485 * tree-if-conv.c (if_convertable_loop_p): Likewise.
486 * tree-inline.c (tree_function_versioning): Likewise.
487 * tree-loop-distribution.c (build_rdg): Likewise.
488 (rdg_flag_vertex_and_dependent): Likewise.
489 (distribute_loop): Likewise.
490 * tree-parloops.c (loop_parallel_p): Likewise.
491 (eliminate_local_variables): Likewise.
492 (separate_decls_in_region): Likewise.
493 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
494 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
495 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
496 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
497 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
498 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
499 (vectorizable_condition): Likewise.
501 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
503 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
504 * configure: Regenerate.
505 * config/i386/i386.md (*movdi_internal): Change
506 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
507 (*movdf_internal): Ditto.
508 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
509 * config/i386/sse.md (vec_concatv2di): Output interunit movq
510 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
512 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
514 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
517 2013-10-31 David Malcolm <dmalcolm@redhat.com>
519 Automated part of renaming of symtab_node_base to symtab_node.
521 Patch autogenerated by rename_symtab.py from
522 https://github.com/davidmalcolm/gcc-refactoring-scripts
523 revision 58bb219cc090b2f4516a9297d868c245495ee622
524 with ChangeLog entry fixed up by hand.
526 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
528 (cgraph_node_for_asm): Likewise.
529 * cgraph.h (symtab_node_base): Likewise.
530 (cgraph_node): Likewise.
531 (varpool_node): Likewise.
532 (is_a_helper <cgraph_node>::test): Likewise.
533 (is_a_helper <varpool_node>::test): Likewise.
534 (symtab_nodes): Likewise.
535 (symtab_register_node): Likewise.
536 (symtab_unregister_node): Likewise.
537 (symtab_remove_node): Likewise.
538 (symtab_get_node): Likewise.
539 (symtab_node_for_asm): Likewise.
540 (symtab_node_asm_name): Likewise.
541 (symtab_node_name): Likewise.
542 (symtab_insert_node_to_hashtable): Likewise.
543 (symtab_add_to_same_comdat_group): Likewise.
544 (symtab_dissolve_same_comdat_group_list): Likewise.
545 (dump_symtab_node): Likewise.
546 (debug_symtab_node): Likewise.
547 (dump_symtab_base): Likewise.
548 (verify_symtab_node): Likewise.
549 (verify_symtab_base): Likewise.
550 (symtab_used_from_object_file_p): Likewise.
551 (symtab_alias_ultimate_target): Likewise.
552 (symtab_resolve_alias): Likewise.
553 (fixup_same_cpp_alias_visibility): Likewise.
554 (symtab_for_node_and_aliases): Likewise.
555 (symtab_nonoverwritable_alias): Likewise.
556 (availability symtab_node_availability): Likewise.
557 (symtab_semantically_equivalent_p): Likewise.
558 (fixup_same_cpp_alias_visibility): Likewise.
559 (symtab_prevail_in_asm_name_hash): Likewise.
562 (varpool_first_variable): Likewise.
563 (varpool_next_variable): Likewise.
564 (varpool_first_static_initializer): Likewise.
565 (varpool_next_static_initializer): Likewise.
566 (varpool_first_defined_variable): Likewise.
567 (varpool_next_defined_variable): Likewise.
568 (cgraph_first_defined_function): Likewise.
569 (cgraph_next_defined_function): Likewise.
570 (cgraph_first_function): Likewise.
571 (cgraph_next_function): Likewise.
572 (cgraph_first_function_with_gimple_body): Likewise.
573 (cgraph_next_function_with_gimple_body): Likewise.
574 (symtab_alias_target): Likewise.
575 (symtab_real_symbol_p): Likewise.
576 (symtab_can_be_discarded): Likewise.
577 * cgraphbuild.c (mark_address): Likewise.
578 (mark_load): Likewise.
579 (mark_store): Likewise.
580 * cgraphunit.c (decide_is_symbol_needed): Likewise.
582 (enqueue_node): Likewise.
583 (referred_to_p): Likewise.
584 (cgraph_process_same_body_aliases): Likewise.
585 (analyze_functions): Likewise.
586 (handle_alias_pairs): Likewise.
587 (output_weakrefs): Likewise.
589 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
590 * ipa-inline-analysis.c (inline_write_summary): Likewise.
591 * ipa-prop.c (remove_described_reference): Likewise.
592 (try_decrement_rdesc_refcount): Likewise.
593 (ipa_edge_duplication_hook): Likewise.
594 * ipa-ref.c (ipa_record_reference): Likewise.
595 (ipa_maybe_record_reference): Likewise.
596 (ipa_clone_ref): Likewise.
597 (ipa_clone_references): Likewise.
598 (ipa_clone_referring): Likewise.
599 (ipa_find_reference): Likewise.
600 (ipa_remove_stmt_references): Likewise.
601 (ipa_clear_stmts_in_references): Likewise.
602 * ipa-ref.h (symtab_node_base): Likewise.
604 (ipa_record_reference): Likewise.
605 (ipa_maybe_record_reference): Likewise.
606 (ipa_clone_references): Likewise.
607 (ipa_clone_referring): Likewise.
608 (ipa_clone_ref): Likewise.
609 (ipa_find_reference): Likewise.
610 (ipa_remove_stmt_references): Likewise.
611 (ipa_clear_stmts_in_references): Likewise.
612 * ipa-reference.c (ipa_reference_write_optimization_summary):
614 * ipa.c (enqueue_node): Likewise.
615 (process_references): Likewise.
616 (walk_polymorphic_call_targets): Likewise.
617 (symtab_remove_unreachable_nodes): Likewise.
618 (address_taken_from_non_vtable_p): Likewise.
619 (comdat_can_be_unshared_p_1): Likewise.
620 (comdat_can_be_unshared_p): Likewise.
621 (can_replace_by_local_alias): Likewise.
622 (function_and_variable_visibility): Likewise.
623 * is-a.h: Likewise (within example in comment).
624 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
625 (lto_symtab_encoder_encode): Likewise.
626 (lto_symtab_encoder_delete_node): Likewise.
627 (lto_symtab_encoder_in_partition_p): Likewise.
628 (lto_set_symtab_encoder_in_partition): Likewise.
629 (output_refs): Likewise.
630 (compute_ltrans_boundary): Likewise.
631 (output_symtab): Likewise.
632 (input_node): Likewise.
633 (input_ref): Likewise.
634 (input_edge): Likewise.
635 (input_cgraph_1): Likewise.
636 (input_refs): Likewise.
637 (output_cgraph_opt_summary): Likewise.
638 (input_node_opt_summary): Likewise.
639 (input_cgraph_opt_section): Likewise.
640 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
642 * lto-streamer-out.c (lto_output): Likewise.
643 (output_symbol_p): Likewise.
644 (produce_symtab): Likewise.
645 * lto-streamer.h (lto_encoder_entry): Likewise.
646 (lto_free_function_in_decl_state_for_node): Likewise.
647 (lto_symtab_encoder_encode): Likewise.
648 (lto_symtab_encoder_delete_node): Likewise.
649 (lto_symtab_encoder_in_partition_p): Likewise.
650 (lto_set_symtab_encoder_in_partition): Likewise.
651 (lto_symtab_encoder_lookup): Likewise.
652 (lsei_node): Likewise.
653 (lto_symtab_encoder_deref): Likewise.
654 * symtab.c (symtab_hash): Likewise.
655 (assembler_name_hash): Likewise.
656 (symtab_nodes): Likewise.
657 (hash_node): Likewise.
659 (hash_node_by_assembler_name): Likewise.
660 (eq_assembler_name): Likewise.
661 (insert_to_assembler_name_hash): Likewise.
662 (unlink_from_assembler_name_hash): Likewise.
663 (symtab_prevail_in_asm_name_hash): Likewise.
664 (symtab_register_node): Likewise.
665 (symtab_insert_node_to_hashtable): Likewise.
666 (symtab_unregister_node): Likewise.
667 (symtab_get_node): Likewise.
668 (symtab_remove_node): Likewise.
669 (symtab_initialize_asm_name_hash): Likewise.
670 (symtab_node_for_asm): Likewise.
671 (symtab_add_to_same_comdat_group): Likewise.
672 (symtab_dissolve_same_comdat_group_list): Likewise.
673 (symtab_node_asm_name): Likewise.
674 (symtab_node_name): Likewise.
675 (dump_symtab_base): Likewise.
676 (dump_symtab_node): Likewise.
677 (dump_symtab): Likewise.
678 (debug_symtab_node): Likewise.
679 (verify_symtab_base): Likewise.
680 (verify_symtab_node): Likewise.
681 (verify_symtab): Likewise.
682 (symtab_used_from_object_file_p): Likewise.
683 (symtab_node_availability): Likewise.
684 (symtab_alias_ultimate_target): Likewise.
685 (fixup_same_cpp_alias_visibility): Likewise.
686 (symtab_resolve_alias): Likewise.
687 (symtab_for_node_and_aliases): Likewise.
688 (symtab_for_node_and_aliases): Likewise.
689 (symtab_nonoverwritable_alias_1): Likewise.
690 (symtab_nonoverwritable_alias): Likewise.
691 (symtab_semantically_equivalent_p): Likewise.
692 * value-prof.c (init_node_map): Likewise.
693 * varasm.c (find_decl): Likewise.
694 * varpool.c (varpool_node_for_asm): Likewise.
695 (varpool_remove_unreferenced_decls): Likewise.
697 2013-10-31 David Malcolm <dmalcolm@redhat.com>
699 Manual part of renaming of symtab_node_base to symtab_node.
701 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
702 clashes with the preferred name for the base class.
703 (const_symtab_node): Remove redundant typedef.
705 2013-10-31 Jakub Jelinek <jakub@redhat.com>
707 * optabs.c (expand_vec_perm): Avoid vector mode punning
709 * expmed.c (store_bit_field_1): Likewise.
710 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
711 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
712 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
713 * config/i386/i386.c (ix86_expand_vector_move_misalign,
714 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
715 ix86_expand_sse_unpack, ix86_expand_args_builtin,
716 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
717 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
718 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
719 expand_vec_perm_vpshufb2_vpermq,
720 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
721 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
722 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
723 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
726 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
728 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
729 Use gcc_unreachable for unhandled modes. Do not check results of
730 expand_simple_binop. If not expanded to target, move the result.
732 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
733 Shiva Chen <shiva0217@gmail.com>
735 * config.gcc (nds32*-*-*): Add nds32 target.
736 * config/nds32/nds32.c: New file.
737 * config/nds32/nds32.h: New file.
738 * config/nds32/nds32.md: New file.
739 * config/nds32/constants.md: New file.
740 * config/nds32/constraints.md: New file.
741 * config/nds32/iterators.md: New file.
742 * config/nds32/nds32-doubleword.md: New file.
743 * config/nds32/nds32-intrinsic.md: New file.
744 * config/nds32/nds32_intrinsic.h: New file.
745 * config/nds32/nds32-modes.def: New file.
746 * config/nds32/nds32-multiple.md: New file.
747 * config/nds32/nds32.opt: New file.
748 * config/nds32/nds32-opts.h: New file.
749 * config/nds32/nds32-protos.h: New file.
750 * config/nds32/nds32-peephole2.md: New file.
751 * config/nds32/pipelines.md: New file.
752 * config/nds32/predicates.md: New file.
753 * config/nds32/t-mlibs: New file.
754 * common/config/nds32: New directory and files.
756 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
757 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
758 * doc/install.texi (Cross-Compiler-Specific Options): Document
759 --with-nds32-lib for nds32 target.
760 * doc/extend.texi (Function Attributes, Target Builtins): Document
761 nds32 specific attributes.
763 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
765 * lra-constraints (process_alt_operands): Use the result
766 elimination register for operand when matching constraints.
768 2013-10-31 Jakub Jelinek <jakub@redhat.com>
770 * tree-vrp.c (maybe_set_nonzero_bits): New function.
771 (remove_range_assertions): Call it.
773 * tree.c (tree_ctz): New function.
774 * tree.h (tree_ctz): New prototype.
775 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
776 first argument from tree to const_tree.
777 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
778 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
779 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
781 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
782 * tree-vect-loop.c (vect_analyze_loop_operations,
783 vect_transform_loop): Don't force scalar loop for bound just because
784 number of iterations is unknown, only do it if it is not known to be
785 a multiple of vectorization_factor.
786 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
788 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
789 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
790 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
791 (set_nonzero_bits, get_nonzero_bits): New prototypes.
792 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
793 a default def isn't partially constant.
794 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
795 is known to be partially zero.
796 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
797 to see if a default def isn't partially constant.
798 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
799 creation of a range, if VR_RANGE, try to improve nonzero_bits from
801 (set_nonzero_bits, get_nonzero_bits): New functions.
803 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
804 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
805 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
806 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
807 uses but in the condition and ASSERT_EXPR and the other successor of
808 the predecessor bb is __builtin_unreachable (), set_range_info of the
809 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
811 2013-10-31 Martin Jambor <mjambor@suse.cz>
813 PR rtl-optimization/58934
815 2013-10-30 Martin Jambor <mjambor@suse.cz>
816 PR rtl-optimization/10474
817 * ira.c (find_moveable_pseudos): Do not calculate dominance info
819 (interesting_dest_for_shprep): New function.
820 (split_live_ranges_for_shrink_wrap): Likewise.
821 (ira): Calculate dominance info and df analysis. Call
822 split_live_ranges_for_shrink_wrap.
824 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
825 Yury Gribov <y.gribov@samsung.com>
828 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
829 shadow pointer to avoid clobbering the main one.
831 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
833 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
835 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
838 * ira-color.c (update_costs_from_copies): Add new parameter. Use
839 it for calling update_costs_from_allocno.
840 (assign_hard_reg): Call restore_costs_from_copies only for
841 !retry_p. Pass new argument to update_costs_from_copies.
842 (color_pass): Pass new argument to update_costs_from_copies.
843 (ira_mark_allocation_change): Ditto.
845 2013-10-30 Sharad Singhai <singhai@google.com>
848 * opts.c (common_handle_option): Remove deprecated option
849 -ftree-vectorizer-verbose.
850 * doc/invoke.texi (Debugging Options): Ditto.
851 * opts-global.c (handle_common_deferred_options): Ditto.
852 (dump_remap_tree_vectorizer_verbose): Delete.
853 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
855 2013-10-30 DJ Delorie <dj@redhat.com>
857 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
858 that take no arguments.
860 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
863 * reload1.c (update_eliminables_and_spill): New function, broken
865 (reload): Use it. Check for frame size change after frame size
866 alignment, and call update_eliminables_and_spill first if continue-ing.
868 2013-10-30 Cong Hou <congh@google.com>
871 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
872 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
873 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
875 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
877 PR rtl-optimization/58369
878 * reload1.c (compute_reload_subreg_offset): New function.
879 (choose_reload_regs): Use it to pass endian-correct
880 offset to subreg_regno_offset.
882 2013-10-30 Tobias Burnus <burnus@net-b.de>
885 * gcc/tree-cfg.c (replace_loop_annotate): Replace warning by
888 2013-10-30 Jason Merrill <jason@redhat.com>
890 * configure.ac (loose_warn): Add -Wno-format if
891 --disable-build-format-warnings.
893 2013-10-30 David Malcolm <dmalcolm@redhat.com>
895 * cgraphunit.c (analyze_functions): Split symtab_node declarations
896 onto multiple lines to make things easier for rename_symtab.py.
898 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
899 (symtab_semantically_equivalent_p): Likewise.
901 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
904 * lra.c (check_rtl): Remove address check before LRA work.
906 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
908 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
909 POINTER_PLUS_EXPR in the defining statement.
911 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
914 * tree-pass.h (make_pass_regmove): Remove.
915 * timevar.def (TV_REGMOVE): Remove.
916 * passes.def (pass_regmove): Remove.
917 * opts.c (default_options_table): Remove entry for regmove.
918 * doc/passes.texi: Remove regmove pass description.
919 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
921 (-fdump-rtl-regmove): Ditto.
922 * common.opt (foptimize-register-move, fregmove): Ignore.
923 * Makefile.in (OBJS): Remove regmove.o.
925 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
927 (struct ira_allocno) New member allocno_prefs.
928 (ALLOCNO_PREFS): New macro.
929 (ira_prefs, ira_prefs_num): New external vars.
930 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
931 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
932 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
934 (ira_add_allocno_copy_to_list): Remove prototype.
935 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
936 (ira_pref_iterator): New type.
937 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
938 (FOR_EACH_PREF): New macro.
939 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
940 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
942 (ira_setup_alts): New function.
943 (decrease_live_ranges_number): New function.
944 (ira): Call the above function.
945 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
946 (ira_create_allocno): Initialize allocno prefs.
947 (pref_pool, pref_vec): New static vars.
948 (initiate_prefs, find_allocno_pref, ira_create_pref): New
950 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
951 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
952 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
953 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
954 (ira_add_allocno_copy_to_list): Make static. Rename to
955 add_allocno_copy_to_list.
956 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
957 swap_allocno_copy_ends_if_necessary.
958 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
959 ira_remove_allocno_prefs.
960 (ira_flattening): Ditto.
961 (ira_build): Call initiate_prefs, print_prefs.
962 (ira_destroy): Call finish_prefs.
963 * ira-color.c (struct update_cost_record): New.
964 (struct allocno_color_data): Add new member update_cost_records.
965 (update_cost_record_pool): New static var.
966 (init_update_cost_records, get_update_cost_record): New functions.
967 (free_update_cost_record_list, finish_update_cost_records): Ditto.
968 (struct update_cost_queue_elem): Add member from.
969 (initiate_cost_update): Call init_update_cost_records.
970 (finish_cost_update): Call finish_update_cost_records.
971 (queue_update_cost, get_next_update_cost): Add new param from.
972 (Update_allocno_cost, update_costs_from_allocno): New functions.
973 (update_costs_from_prefs): Ditto.
974 (update_copy_costs): Rename to update_costs_from_copies.
975 (restore_costs_from_copies): New function.
976 (update_conflict_hard_regno_costs): Don't go back.
977 (assign_hard_reg): Call restore_costs_from_copies. Add printing
979 (pop_allocnos): Add priniting more debug info.
980 (color_allocnos): Remove prefs for conflicting hard regs.
981 Call update_costs_from_prefs.
982 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
983 (get_dup_num): Rename, modify, and move to ira.c
984 (process_regs_for_copy): Add prefs.
985 (add_insn_allocno_copies): Put src as first arg of
986 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
987 * ira-costs.c (record_reg_classes): Modify and move code into
988 record_operands_costs.
989 (find_costs_and_classes): Create prefs for the hard reg of small
991 (process_bb_node_for_hard_reg_moves): Add prefs.
993 2013-10-30 Richard Biener <rguenther@suse.de>
996 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
997 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
999 (pre_and_rev_post_order_compute): ... which now wraps it.
1000 * graph.c (draw_cfg_nodes_no_loops): Use
1001 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
1004 2013-10-30 Christian Bruel <christian.bruel@st.com>
1006 * gcc/config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
1007 (sh_expand_cmpstr): Handle known align and schedule improvements.
1008 * gcc/config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
1009 * gcc/config/sh/sh.md (cmpstrnsi): New pattern.
1011 2013-10-30 Martin Jambor <mjambor@suse.cz>
1013 PR rtl-optimization/10474
1014 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1016 (interesting_dest_for_shprep): New function.
1017 (split_live_ranges_for_shrink_wrap): Likewise.
1018 (ira): Calculate dominance info and df analysis. Call
1019 split_live_ranges_for_shrink_wrap.
1021 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1024 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
1026 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1028 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
1029 * tree.h (POINTER_BOUNDS_P): New.
1030 (BOUNDED_TYPE_P): New.
1032 (pointer_bounds_type_node): New.
1033 * tree.c (build_common_tree_nodes): Initialize
1034 pointer_bounds_type_node.
1035 * gimple.h (gimple_call_get_nobnd_arg_index): New.
1036 (gimple_call_num_nobnd_args): New.
1037 (gimple_call_nobnd_arg): New.
1038 (gimple_return_retbnd): New.
1039 (gimple_return_set_retbnd): New
1040 * gimple.c (gimple_build_return): Increase number of ops
1041 for return statement.
1042 (gimple_call_get_nobnd_arg_index): New.
1043 * gimple-pretty-print.c (dump_gimple_return): Print second op.
1045 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1047 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
1048 with "chkp ctor" and "bnd_legacy" attributes.
1049 * gimplify.c (gimplify_init_constructor): Avoid infinite
1050 loop during gimplification of bounds initializer.
1052 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1054 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
1055 (handle_bnd_legacy): New.
1056 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
1057 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
1060 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
1062 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
1063 (BT_FN_PTR_CONST_PTR): New.
1064 (BT_FN_CONST_PTR_CONST_PTR): New.
1065 (BT_FN_PTR_CONST_PTR_SIZE): New.
1066 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
1067 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
1068 (BT_FN_VOID_CONST_PTR_SIZE): New.
1069 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
1070 * chkp-builtins.def: New.
1071 * builtins.def: include chkp-builtins.def.
1072 (DEF_CHKP_BUILTIN): New.
1073 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
1074 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
1075 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
1076 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
1077 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
1078 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
1079 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
1080 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
1081 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
1082 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
1083 * common.opt (fcheck-pointer-bounds): New.
1084 * toplev.c (process_options): Check Pointer Bounds Checker is
1086 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
1088 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1090 * target.def (builtin_chkp_function): New.
1091 (chkp_bound_type): New.
1092 (chkp_bound_mode): New.
1093 (fn_abi_va_list_bounds_size): New.
1094 (load_bounds_for_arg): New.
1095 (store_bounds_for_arg): New.
1096 * targhooks.h (default_load_bounds_for_arg): New.
1097 (default_store_bounds_for_arg): New.
1098 (default_fn_abi_va_list_bounds_size): New.
1099 (default_chkp_bound_type): New.
1100 (default_chkp_bound_mode): New.
1101 (default_builtin_chkp_function): New.
1102 * targhooks.c (default_load_bounds_for_arg): New.
1103 (default_store_bounds_for_arg): New.
1104 (default_fn_abi_va_list_bounds_size): New.
1105 (default_chkp_bound_type): New.
1106 (default_chkp_bound_mode); New.
1107 (default_builtin_chkp_function): New.
1108 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
1109 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
1110 (TARGET_STORE_BOUNDS_FOR_ARG): New.
1111 (TARGET_BUILTIN_CHKP_FUNCTION): New.
1112 (TARGET_CHKP_BOUND_TYPE): New.
1113 (TARGET_CHKP_BOUND_MODE): New.
1114 * doc/tm.texi: Regenerated.
1115 * langhooks.h (lang_hooks): Add chkp_supported field.
1116 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
1117 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
1119 2013-10-29 Andrew Pinski <apinski@cavium.com>
1121 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
1122 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
1123 fails, combining the branches anyways.
1124 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
1125 increases the number of combinings.
1126 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
1128 2013-10-29 Mike Stump <mikestump@comcast.net>
1130 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
1131 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
1132 (make_vector_mode): Increase namebuf to 16.
1133 (emit_insn_modes_h): When processing BImode, don't
1134 also match partial int modes.
1135 (emit_class_narrowest_mode): Likewise.
1137 * config/bfin/bfin-modes.def: Add precision to PDI.
1138 * config/m32c/m32c-modes.def: Add precision to PSI.
1139 * config/msp430/msp430-modes.def: Add precision to PSI.
1140 * config/rs6000/rs6000-modes.def: Add precision to PTI.
1141 * config/sh/sh-modes.def: Add precision to PSI and PDI.
1143 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
1146 * config/sh/sh.md (*addc): Rename existing variations to ...
1147 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
1148 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
1149 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
1150 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
1152 2013-10-29 Teresa Johnson <tejohnson@google.com>
1155 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
1156 insanities when updating probabilities.
1158 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1160 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
1161 prettyprinter for cgraph_node to reflect the conversion of the
1162 symtable types to a C++ class hierarchy: it now *is* a
1163 symtab_node_base, rather than having one (named "symbol").
1165 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
1167 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
1168 __cilkrts_pop_frame. If matched, then return true for built-in
1170 (expand_builtin): Added BUILT_IN_CILK_DETACH and
1171 BUILT_IN_CILK_POP_FRAME case.
1172 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
1173 (lhs_cilk_detect_spawn): Likewise.
1174 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
1175 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
1176 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
1177 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
1178 (LANG_HOOKS_CILKPLUS): Likewise.
1179 * tree.h (CILK_SPAWN_FN): Likewise.
1180 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
1181 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
1182 (OBJS): Added cilk-common.o.
1183 (BUILTINS_DEF): Added cilk-builtins.def.
1184 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
1185 (lhd_cilk_detect_spawn): Likewise.
1186 * langhooks.h (lang_hooks_for_cilkplus): New struct.
1187 (struct lang_hooks): Added new field called "cilkplus."
1188 * cilk-common.c: New file.
1190 * cilk-builtins.def: Likewise.
1191 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
1192 "__cilk" macro and set it to 200.
1193 * function.h (struct function::cilk_frame_decl): New field.
1194 (struct function::is_cilk_function): Likewise.
1195 (struct function::calls_cilk_spawn): Likewise.
1196 * gimplify.c (gimplify_call_expr): Added a check if the function call
1197 being gimplified is a spawn detach point. If so, then add pop_frame
1198 and detach function calls.
1199 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
1200 for gimplifying _Cilk_spawn and _Cilk_sync statements.
1201 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
1202 function. If so, added a _Cilk_sync and gimplified it.
1203 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
1204 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
1205 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
1207 (can_inline_edge_p): Prevent inling of spawnee function.
1208 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
1209 for functions that use Cilk keywords.
1210 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
1211 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
1212 CILK_SYNC_STMT cases.
1213 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
1215 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
1216 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
1217 * passes.texi (Cilk Keywords): New section that describes the compiler
1218 code changes for handling Cilk Keywords.
1220 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1222 Patch autogenerated by refactor_symtab.py from
1223 https://github.com/davidmalcolm/gcc-refactoring-scripts
1224 revision 58bb219cc090b2f4516a9297d868c245495ee622
1226 * asan.c (asan_finish_file): Update for conversion of symtab types to
1227 a true class hierarchy.
1228 * cfgexpand.c (estimated_stack_frame_size): Likewise.
1229 * cgraph.c (cgraph_get_body): Likewise.
1230 (cgraph_get_create_real_symbol_node): Likewise.
1231 (verify_cgraph_node): Likewise.
1232 (verify_edge_corresponds_to_fndecl): Likewise.
1233 (verify_edge_count_and_frequency): Likewise.
1234 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
1235 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1236 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
1237 (cgraph_node_cannot_return): Likewise.
1238 (cgraph_set_pure_flag_1): Likewise.
1239 (cgraph_set_const_flag_1): Likewise.
1240 (cgraph_set_nothrow_flag_1): Likewise.
1241 (cgraph_make_node_local_1): Likewise.
1242 (cgraph_for_node_and_aliases): Likewise.
1243 (cgraph_for_node_thunks_and_aliases): Likewise.
1244 (cgraph_node_can_be_local_p): Likewise.
1245 (cgraph_node_cannot_be_local_p_1): Likewise.
1246 (cgraph_function_body_availability): Likewise.
1247 (dump_cgraph_node): Likewise.
1248 (cgraph_rtl_info): Likewise.
1249 (cgraph_mark_address_taken_node): Likewise.
1250 (cgraph_remove_node): Likewise.
1251 (cgraph_release_function_body): Likewise.
1252 (cgraph_update_edges_for_call_stmt_node): Likewise.
1253 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
1254 (cgraph_make_edge_direct): Likewise.
1255 (cgraph_resolve_speculation): Likewise.
1256 (cgraph_speculative_call_info): Likewise.
1257 (cgraph_turn_edge_to_speculative): Likewise.
1258 (cgraph_create_edge_1): Likewise.
1259 (cgraph_set_call_stmt): Likewise.
1260 (cgraph_node_for_asm): Likewise.
1261 (cgraph_add_thunk): Likewise.
1262 (cgraph_same_body_alias): Likewise.
1263 (cgraph_create_function_alias): Likewise.
1264 (cgraph_create_node): Likewise.
1265 (cgraph_create_empty_node): Likewise.
1266 (record_function_versions): Likewise.
1267 (used_from_object_file_p): Likewise.
1268 * cgraph.h (symtab_can_be_discarded): Likewise.
1269 (symtab_real_symbol_p): Likewise.
1270 (cgraph_mark_force_output_node): Likewise.
1271 (cgraph_edge_recursive_p): Likewise.
1272 (symtab_alias_target): Likewise.
1273 (varpool_all_refs_explicit_p): Likewise.
1274 (varpool_can_remove_if_no_refs): Likewise.
1275 (cgraph_only_called_directly_or_aliased_p): Likewise.
1276 (cgraph_next_function_with_gimple_body): Likewise.
1277 (cgraph_first_function_with_gimple_body): Likewise.
1278 (cgraph_function_with_gimple_body_p): Likewise.
1279 (cgraph_next_function): Likewise.
1280 (cgraph_first_function): Likewise.
1281 (cgraph_next_defined_function): Likewise.
1282 (cgraph_first_defined_function): Likewise.
1283 (varpool_next_defined_variable): Likewise.
1284 (varpool_first_defined_variable): Likewise.
1285 (varpool_next_static_initializer): Likewise.
1286 (varpool_first_static_initializer): Likewise.
1287 (varpool_next_variable): Likewise.
1288 (varpool_first_variable): Likewise.
1289 (varpool_node_name): Likewise.
1290 (varpool): Likewise.
1292 (is_a_helper <varpool_node>::test): Likewise.
1293 (is_a_helper <cgraph_node>::test): Likewise.
1294 (varpool_variable_node): Likewise.
1295 (cgraph_function_or_thunk_node): Likewise.
1296 (varpool_alias_target): Likewise.
1297 (cgraph_alias_target): Likewise.
1298 (cgraph_node_name): Likewise.
1299 (varpool_node_asm_name): Likewise.
1300 (cgraph_node_asm_name): Likewise.
1301 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
1302 (cgraph_rebuild_references): Likewise.
1303 (rebuild_cgraph_edges): Likewise.
1304 (record_eh_tables): Likewise.
1305 (build_cgraph_edges): Likewise.
1306 (mark_store): Likewise.
1307 (mark_load): Likewise.
1308 (mark_address): Likewise.
1309 (record_type_list): Likewise.
1310 (record_reference): Likewise.
1311 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
1312 (cgraph_materialize_clone): Likewise.
1313 (cgraph_function_versioning): Likewise.
1314 (cgraph_copy_node_for_versioning): Likewise.
1315 (update_call_expr): Likewise.
1316 (cgraph_find_replacement_node): Likewise.
1317 (cgraph_create_virtual_clone): Likewise.
1318 (cgraph_clone_node): Likewise.
1319 * cgraphunit.c (compile): Likewise.
1320 (output_weakrefs): Likewise.
1321 (output_in_order): Likewise.
1322 (expand_function): Likewise.
1323 (assemble_thunks_and_aliases): Likewise.
1324 (expand_thunk): Likewise.
1325 (mark_functions_to_output): Likewise.
1326 (handle_alias_pairs): Likewise.
1327 (analyze_functions): Likewise.
1328 (walk_polymorphic_call_targets): Likewise.
1329 (varpool_finalize_decl): Likewise.
1330 (process_function_and_variable_attributes): Likewise.
1331 (cgraph_process_same_body_aliases): Likewise.
1332 (analyze_function): Likewise.
1333 (cgraph_add_new_function): Likewise.
1334 (cgraph_finalize_function): Likewise.
1335 (referred_to_p): Likewise.
1336 (cgraph_reset_node): Likewise.
1337 (cgraph_process_new_functions): Likewise.
1338 (enqueue_node): Likewise.
1339 (decide_is_symbol_needed): Likewise.
1340 * coverage.c (coverage_compute_profile_id): Likewise.
1341 * dbxout.c (dbxout_expand_expr): Likewise.
1342 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
1343 (reference_to_unused): Likewise.
1344 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1345 * gimplify.c (unvisit_body): Likewise.
1346 (unshare_body): Likewise.
1347 * ipa-cp.c (ipcp_generate_summary): Likewise.
1348 (ipcp_decision_stage): Likewise.
1349 (identify_dead_nodes): Likewise.
1350 (decide_whether_version_node): Likewise.
1351 (decide_about_value): Likewise.
1352 (perhaps_add_new_callers): Likewise.
1353 (create_specialized_node): Likewise.
1354 (update_profiling_info): Likewise.
1355 (ipcp_propagate_stage): Likewise.
1356 (estimate_local_effects): Likewise.
1357 (good_cloning_opportunity_p): Likewise.
1358 (devirtualization_time_bonus): Likewise.
1359 (propagate_constants_accross_call): Likewise.
1360 (initialize_node_lattices): Likewise.
1361 (ipcp_cloning_candidate_p): Likewise.
1362 (determine_versionability): Likewise.
1363 (print_all_lattices): Likewise.
1364 (print_lattice): Likewise.
1365 (ipcp_discover_new_direct_edges): Likewise.
1366 * ipa-devirt.c (ipa_devirt): Likewise.
1367 (likely_target_p): Likewise.
1368 (update_type_inheritance_graph): Likewise.
1369 (possible_polymorphic_call_target_p): Likewise.
1370 (dump_possible_polymorphic_call_targets): Likewise.
1371 (devirt_variable_node_removal_hook): Likewise.
1372 (record_binfo): Likewise.
1373 (maybe_record_node): Likewise.
1374 (build_type_inheritance_graph): Likewise.
1375 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1376 (inline_generate_summary): Likewise.
1377 (inline_analyze_function): Likewise.
1378 (do_estimate_growth): Likewise.
1379 (simple_edge_hints): Likewise.
1380 (estimate_node_size_and_time): Likewise.
1381 (estimate_edge_devirt_benefit): Likewise.
1382 (compute_inline_parameters): Likewise.
1383 (estimate_function_body_sizes): Likewise.
1384 (compute_bb_predicates): Likewise.
1385 (initialize_inline_failed): Likewise.
1386 (dump_inline_summary): Likewise.
1387 (dump_inline_edge_summary): Likewise.
1388 * ipa-inline-transform.c (inline_transform): Likewise.
1389 (preserve_function_body_p): Likewise.
1390 (save_inline_function_body): Likewise.
1391 (inline_call): Likewise.
1392 (clone_inlined_nodes): Likewise.
1393 (can_remove_node_now_p): Likewise.
1394 (can_remove_node_now_p_1): Likewise.
1395 * ipa-inline.c (early_inliner): Likewise.
1396 (early_inline_small_functions): Likewise.
1397 (inline_always_inline_functions): Likewise.
1398 (ipa_inline): Likewise.
1399 (flatten_function): Likewise.
1400 (inline_small_functions): Likewise.
1401 (speculation_useful_p): Likewise.
1402 (recursive_inlining): Likewise.
1403 (update_caller_keys): Likewise.
1404 (reset_edge_caches): Likewise.
1405 (update_edge_key): Likewise.
1406 (edge_badness): Likewise.
1407 (relative_time_benefit): Likewise.
1408 (want_inline_self_recursive_call_p): Likewise.
1409 (want_inline_small_function_p): Likewise.
1410 (want_early_inline_function_p): Likewise.
1411 (num_calls): Likewise.
1412 (can_early_inline_edge_p): Likewise.
1413 (can_inline_edge_p): Likewise.
1414 (report_inline_failed_reason): Likewise.
1415 * ipa-profile.c (ipa_profile): Likewise.
1416 (ipa_propagate_frequency): Likewise.
1417 (ipa_propagate_frequency_1): Likewise.
1418 (ipa_profile_generate_summary): Likewise.
1419 * ipa-prop.c (ipcp_transform_function): Likewise.
1420 (read_replacements_section): Likewise.
1421 (ipa_prop_read_section): Likewise.
1422 (ipa_modify_call_arguments): Likewise.
1423 (ipa_print_node_params): Likewise.
1424 (propagate_controlled_uses): Likewise.
1425 (update_indirect_edges_after_inlining): Likewise.
1426 (remove_described_reference): Likewise.
1427 (ipa_make_edge_direct_to_target): Likewise.
1428 (ipa_analyze_node): Likewise.
1429 (ipa_analyze_params_uses): Likewise.
1430 (ipa_compute_jump_functions): Likewise.
1431 (ipa_get_callee_param_type): Likewise.
1432 (ipa_print_node_jump_functions): Likewise.
1433 (ipa_initialize_node_params): Likewise.
1434 (ipa_populate_param_decls): Likewise.
1435 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
1436 (write_agg_replacement_chain): Likewise.
1437 (ipa_write_node_info): Likewise.
1438 (ipa_edge_duplication_hook): Likewise.
1439 (try_decrement_rdesc_refcount): Likewise.
1440 * ipa-pure-const.c (propagate_nothrow): Likewise.
1441 (propagate_pure_const): Likewise.
1442 (pure_const_read_summary): Likewise.
1443 (pure_const_write_summary): Likewise.
1444 (analyze_function): Likewise.
1445 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
1446 (ipa_ref_referring_ref_list): Likewise.
1447 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
1448 (ipa_remove_stmt_references): Likewise.
1449 (ipa_find_reference): Likewise.
1450 (ipa_dump_referring): Likewise.
1451 (ipa_dump_references): Likewise.
1452 (ipa_record_reference): Likewise.
1453 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
1454 (ipa_reference_write_optimization_summary): Likewise.
1455 (write_node_summary_p): Likewise.
1456 (propagate): Likewise.
1457 (read_write_all_from_decl): Likewise.
1458 (generate_summary): Likewise.
1459 (analyze_function): Likewise.
1460 (propagate_bits): Likewise.
1461 (ipa_reference_get_not_written_global): Likewise.
1462 (ipa_reference_get_not_read_global): Likewise.
1463 * ipa-split.c (execute_split_functions): Likewise.
1464 (split_function): Likewise.
1465 * ipa-utils.c (ipa_merge_profiles): Likewise.
1466 (dump_cgraph_node_set): Likewise.
1467 (ipa_reverse_postorder): Likewise.
1468 (ipa_edge_within_scc): Likewise.
1469 (ipa_get_nodes_in_cycle): Likewise.
1470 (ipa_free_postorder_info): Likewise.
1471 (ipa_reduced_postorder): Likewise.
1472 (searchc): Likewise.
1473 (recursive_call_p): Likewise.
1474 * ipa.c (ipa_cdtor_merge): Likewise.
1475 (record_cdtor_fn): Likewise.
1476 (function_and_variable_visibility): Likewise.
1477 (varpool_externally_visible_p): Likewise.
1478 (cgraph_externally_visible_p): Likewise.
1479 (comdat_can_be_unshared_p): Likewise.
1480 (comdat_can_be_unshared_p_1): Likewise.
1481 (address_taken_from_non_vtable_p): Likewise.
1482 (ipa_discover_readonly_nonaddressable_vars): Likewise.
1483 (symtab_remove_unreachable_nodes): Likewise.
1484 (walk_polymorphic_call_targets): Likewise.
1485 (process_references): Likewise.
1486 (enqueue_node): Likewise.
1487 (has_addr_references_p): Likewise.
1488 (cgraph_non_local_node_p_1): Likewise.
1489 * is-a.h (varpool_analyze_node): Likewise.
1490 * lto-cgraph.c (input_symtab): Likewise.
1491 (merge_profile_summaries): Likewise.
1492 (input_cgraph_1): Likewise.
1493 (input_edge): Likewise.
1494 (input_varpool_node): Likewise.
1495 (input_node): Likewise.
1496 (input_overwrite_node): Likewise.
1497 (compute_ltrans_boundary): Likewise.
1498 (output_refs): Likewise.
1499 (lto_output_varpool_node): Likewise.
1500 (lto_output_node): Likewise.
1501 (reachable_from_other_partition_p): Likewise.
1502 (referenced_from_other_partition_p): Likewise.
1503 (lto_output_edge): Likewise.
1504 (output_node_opt_summary): Likewise.
1505 (add_node_to): Likewise.
1506 (reachable_from_this_partition_p): Likewise.
1507 (lto_set_symtab_encoder_in_partition): Likewise.
1508 (lto_symtab_encoder_in_partition_p): Likewise.
1509 (lto_set_symtab_encoder_encode_initializer): Likewise.
1510 (lto_symtab_encoder_encode_initializer_p): Likewise.
1511 (lto_set_symtab_encoder_encode_body): Likewise.
1512 (lto_symtab_encoder_encode_body_p): Likewise.
1513 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1515 * lto-streamer-in.c (lto_read_body): Likewise.
1516 (fixup_call_stmt_edges): Likewise.
1517 (fixup_call_stmt_edges_1): Likewise.
1518 * lto-streamer-out.c (produce_symtab): Likewise.
1519 (output_symbol_p): Likewise.
1520 (write_symbol): Likewise.
1521 (lto_output): Likewise.
1522 (copy_function): Likewise.
1523 (output_function): Likewise.
1524 * passes.c (function_called_by_processed_nodes_p): Likewise.
1525 (ipa_write_optimization_summaries): Likewise.
1526 (ipa_write_summaries): Likewise.
1527 (do_per_function_toporder): Likewise.
1528 (do_per_function): Likewise.
1529 (dump_passes): Likewise.
1530 * symtab.c (symtab_semantically_equivalent_p): Likewise.
1531 (symtab_nonoverwritable_alias): Likewise.
1532 (symtab_nonoverwritable_alias_1): Likewise.
1533 (symtab_for_node_and_aliases): Likewise.
1534 (symtab_resolve_alias): Likewise.
1535 (fixup_same_cpp_alias_visibility): Likewise.
1536 (symtab_alias_ultimate_target): Likewise.
1537 (symtab_used_from_object_file_p): Likewise.
1538 (verify_symtab_base): Likewise.
1539 (dump_symtab_base): Likewise.
1540 (symtab_node_name): Likewise.
1541 (symtab_node_asm_name): Likewise.
1542 (symtab_dissolve_same_comdat_group_list): Likewise.
1543 (symtab_add_to_same_comdat_group): Likewise.
1544 (symtab_unregister_node): Likewise.
1545 (symtab_insert_node_to_hashtable): Likewise.
1546 (symtab_register_node): Likewise.
1547 (unlink_from_assembler_name_hash): Likewise.
1548 (insert_to_assembler_name_hash): Likewise.
1549 (eq_assembler_name): Likewise.
1550 (hash_node_by_assembler_name): Likewise.
1551 (eq_node): Likewise.
1552 (hash_node): Likewise.
1553 * toplev.c (wrapup_global_declaration_2): Likewise.
1554 * trans-mem.c (ipa_tm_execute): Likewise.
1555 (ipa_tm_transform_clone): Likewise.
1556 (ipa_tm_transform_transaction): Likewise.
1557 (ipa_tm_transform_calls_redirect): Likewise.
1558 (ipa_tm_insert_gettmclone_call): Likewise.
1559 (ipa_tm_insert_irr_call): Likewise.
1560 (ipa_tm_create_version): Likewise.
1561 (ipa_tm_create_version_alias): Likewise.
1562 (ipa_tm_mark_forced_by_abi_node): Likewise.
1563 (ipa_tm_mark_force_output_node): Likewise.
1564 (ipa_tm_diagnose_tm_safe): Likewise.
1565 (ipa_tm_mayenterirr_function): Likewise.
1566 (ipa_tm_scan_irr_function): Likewise.
1567 (ipa_tm_note_irrevocable): Likewise.
1568 (ipa_tm_scan_calls_clone): Likewise.
1569 (get_cg_data): Likewise.
1570 * tree-eh.c (tree_could_trap_p): Likewise.
1571 * tree-emutls.c (ipa_lower_emutls): Likewise.
1572 (create_emultls_var): Likewise.
1573 (lower_emutls_function_body): Likewise.
1574 (gen_emutls_addr): Likewise.
1575 (emutls_decl): Likewise.
1576 (new_emutls_decl): Likewise.
1577 * tree-inline.c (tree_function_versioning): Likewise.
1578 (optimize_inline_calls): Likewise.
1579 (expand_call_inline): Likewise.
1580 (estimate_num_insns): Likewise.
1581 (copy_bb): Likewise.
1582 (delete_unreachable_blocks_update_callgraph): Likewise.
1583 * tree-nested.c (gimplify_all_functions): Likewise.
1584 (create_nesting_tree): Likewise.
1585 (check_for_nested_with_variably_modified): Likewise.
1586 * tree-pretty-print.c (dump_function_header): Likewise.
1587 * tree-profile.c (tree_profiling): Likewise.
1588 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
1589 (modify_function): Likewise.
1590 (convert_callers): Likewise.
1591 (convert_callers_for_node): Likewise.
1592 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
1593 (associate_varinfo_to_alias): Likewise.
1594 (create_variable_info_for): Likewise.
1595 (get_constraint_for_ssa_var): Likewise.
1596 * tree-vectorizer.c (increase_alignment): Likewise.
1597 * tree.c (find_decls_types_in_var): Likewise.
1598 (find_decls_types_in_node): Likewise.
1599 (free_lang_data_in_decl): Likewise.
1600 * value-prof.c (gimple_ic_transform): Likewise.
1601 (gimple_ic): Likewise.
1602 (check_ic_target): Likewise.
1603 (init_node_map): Likewise.
1604 * varasm.c (decl_binds_to_current_def_p): Likewise.
1605 (default_binds_local_p_1): Likewise.
1606 (dump_tm_clone_pairs): Likewise.
1607 (assemble_alias): Likewise.
1608 (find_decl): Likewise.
1609 (mark_decl_referenced): Likewise.
1610 * varpool.c (varpool_for_node_and_aliases): Likewise.
1611 (varpool_extra_name_alias): Likewise.
1612 (varpool_create_variable_alias): Likewise.
1613 (add_new_static_var): Likewise.
1614 (varpool_finalize_named_section_flags): Likewise.
1615 (varpool_remove_unreferenced_decls): Likewise.
1616 (enqueue_node): Likewise.
1617 (varpool_assemble_decl): Likewise.
1618 (assemble_aliases): Likewise.
1619 (varpool_analyze_node): Likewise.
1620 (cgraph_variable_initializer_availability): Likewise.
1621 (varpool_add_new_variable): Likewise.
1622 (ctor_for_folding): Likewise.
1623 (dump_varpool_node): Likewise.
1624 (varpool_remove_initializer): Likewise.
1625 (varpool_remove_node): Likewise.
1626 (varpool_node_for_decl): Likewise.
1627 (varpool_create_empty_node): Likewise.
1628 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
1629 (ix86_get_function_versions_dispatcher): Likewise.
1631 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1633 * cgraph.h (symtab_node_base): Convert to a class;
1634 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
1635 chain_next/prev from symtab_node_def.
1636 (cgraph_node): Inherit from symtab_node; add GTY option
1637 tag ("SYMTAB_FUNCTION").
1638 (varpool_node): Inherit from symtab_node; add GTY option
1639 tag ("SYMTAB_VARIABLE").
1640 (symtab_node_def): Remove.
1641 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
1642 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
1643 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
1644 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
1646 * ipa-ref.h (symtab_node_def): Drop.
1647 (symtab_node): Change underlying type from symtab_node_def to
1649 (const_symtab_node): Likwise.
1651 * is-a.h: Update examples in comment.
1653 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
1654 (assembler_name_hash): Likewise.
1656 2013-10-29 Martin Liska <marxin.liska@gmail.com>
1658 * doc/tree-ssa.texi (gimple_phi_result): Document.
1659 (gimple_phi_num_args, gimple_phi_arg): Likewise.
1660 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
1661 (PHI_RESULT, PHI_NUM_ARGS): Remove.
1662 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
1664 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1666 * expr.h: Revert change and include tree-core.h.
1667 * rtl.h: Revert change and don't include tree-core.h.
1669 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1671 * config/darwin.c: Include gimple.h.
1672 * config/i386/winnt.c: Likewise.
1674 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
1676 PR tree-optimization/19831
1677 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
1679 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1681 * tree-outof-ssa.h: Remove include files.
1682 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
1684 * tree-ssa-coalesce.c: Likewise.
1685 * cfgexpand.c: Likewise.
1686 * tree-ssa-ter.c: Likewise.
1687 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
1688 * lto-streamer.h: Likewise.
1689 * cgraphbuild.c: Add gimple.h to include list.
1690 * data-streamer-in.c: Likewise.
1691 * ipa-cp.c: Likewise.
1692 * tree-streamer.c: Likewise.
1693 * lto-compress.c: Likewise.
1694 * ipa-reference.c: Likewise.
1695 * data-streamer-out.c: Likewise.
1696 * lto-cgraph.c: Likewise.
1697 * cgraphclones.c: Likewise.
1698 * ipa-utils.c: Likewise.
1699 * data-streamer.c: Likewise.
1700 * ipa-split.c: Likewise.
1701 * lto-section-in.c: Likewise.
1702 * tree-streamer-out.c: Likewise.
1703 * ipa-prop.c: Likewise.
1704 * tree-streamer-in.c: Likewise.
1705 * symtab.c: Likewise.
1706 * opts-global.c: Likewise.
1707 * lto-opts.c: Likewise.
1708 * lto-section-out.c: Likewise.
1709 * lto-streamer.c: Likewise.
1710 * rtl.h: Add tree-core.h to include list.
1711 * expr.h: Remove tree-core.h from include list.
1712 * gimple.h: Likewise.
1713 * ipa-utils.h: Likewise.
1714 * streamer-hooks.h: Likewise.
1715 * streamer-hooks.c: Include input.h.
1717 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1719 * config/arm/arm.c (cortexa7_extra_costs): New table.
1720 (arm_cortex_a7_tune): New.
1721 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
1723 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
1725 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
1727 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1729 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
1730 to use autogenerated markers for a class-hierarchy, every class
1731 must have a GTY marker.
1732 * gengtype.h (struct type): Add linked list of subclasses to
1733 the "s" member of the union.
1734 (add_subclass): New decl.
1735 * gengtype-state.c (read_state_struct_type): Set up subclass
1737 * gengtype.c (get_ultimate_base_class): New.
1738 (add_subclass): New.
1739 (new_structure): Set up subclass linked list.
1740 (set_gc_used_type): Propagate usage information to subclasses.
1741 (output_mangled_typename): Use get_ultimate_base_class.
1742 (walk_subclasses): Use the subclass linked list, avoiding an
1743 O(N^2) when writing out all types.
1744 (walk_type): Issue an error if the base class is missing a tag,
1745 rather than generating bogus C code. Add a gcc_unreachable
1746 default case, in case people omit tags from concrete subclasses,
1747 or get the values wrong.
1748 (write_func_for_structure): Issue an error for subclasses for
1749 which the base doesn't have a "desc", since otherwise the
1750 autogenerated routines for the base would silently fail to visit
1751 any subclass fields.
1752 (write_root): Use get_ultimate_base_class, tweaking constness of
1753 tp to match that function's signature.
1755 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1757 * doc/gty.texi (GTY Options): Add note about inheritance to
1758 description of desc and tag.
1759 (Inheritance and GTY): New.
1761 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1763 * gengtype-parse.c (opts_have): Drop "static" so that
1764 we can use this from gengtype.c.
1765 * gengtype.c (set_gc_used_type): Mark any base class as used;
1766 update field traversal to visit inherited fields.
1767 (output_mangled_typename): Convert references to classes within
1768 an inheritance hierarchy to reference the ultimate base class,
1769 since only it will have gt_ functions.
1770 (get_string_option): New.
1771 (walk_subclasses): New.
1772 (walk_type): Treat GTY structs that have a "desc" as being the
1773 root of an inheritance hierarchy. Generate a switch on it
1774 within the marking function which walks all subclasses, adding
1775 cases for them via walk_subclasses. For subclasses, visit all
1776 fields of the type (including inherited ones).
1777 (write_func_for_structure): Don't write fns for subclasses, only
1778 for the ultimate base class within an inheritance hierarchy.
1779 Subclasses-marking will be handled by the base class marking functions.
1780 (write_types): Likewise.
1781 (write_local_func_for_structure): Likewise.
1782 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
1783 a "tag" option (and are thus concrete subclasses).
1784 (write_root): Use the marker function for the ultimate base class.
1785 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
1786 (opts_have): Add declaration.
1788 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
1790 * lra-spills.c (lra_final_code_change): Remove useless move insns
1791 originated from moves of pseudos.
1793 2013-10-28 Jeff Law <law@redhat.com>
1795 * lower-subreg.c (resolve_simple_move): Fix comment typo.
1797 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
1799 * df-scan.c (df_collection_rec): Adjust.
1800 (copy_defs): New constant.
1801 (copy_uses): Likewise.
1802 (copy_eq_uses): Likewise.
1803 (copy_mw): Likewise.
1804 (copy_all): Likewise.
1805 (df_insn_rescan): Adjust.
1806 (df_notes_rescan): Likewise.
1807 (df_swap_refs): Likewise.
1808 (df_sort_and_compress_refs): Likewise.
1809 (df_sort_and_compress_mws): Likewise.
1810 (df_install_refs): Likewise.
1811 (df_install_mws): Likewise.
1812 (df_refs_add_to_chains): Add flags parameter controlling which vectors
1814 (df_bb_refs_record): Adjust.
1815 (df_record_entry_block_defs): Likewise.
1816 (df_record_exit_block_defs): Likewise.
1817 (df_refs_verify): Likewise.
1818 (df_mws_verify): Likewise.
1819 (df_insn_refs_verify): Likewise.
1820 (df_bb_verify): Likewise.
1821 * ipa-pure-const.c (finish_state): Remove.
1822 (propagate): Adjust.
1823 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
1824 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
1825 var-tracking.c: Adjust.
1826 * vec.c (stack_vecs): Remove.
1827 (register_stack_vec): Likewise.
1828 (stack_vec_register_index): Likewise.
1829 (unregister_stack_vec): Likewise.
1830 * vec.h (struct va_stack): Remove.
1831 (struct vec<T, A, vl_ptr>): Specialize as
1832 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
1833 allocation strategy compatable with the vl_ptr layout.
1834 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
1835 specialization anyway.
1836 (class stack_vec): New class.
1837 (vec_stack_alloc): Remove.
1838 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
1840 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1841 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1842 Sergey Lega <sergey.s.lega@intel.com>
1843 Anna Tikhonova <anna.tikhonova@intel.com>
1844 Ilya Tocar <ilya.tocar@intel.com>
1845 Andrey Turetskiy <andrey.turetskiy@intel.com>
1846 Ilya Verbin <ilya.verbin@intel.com>
1847 Kirill Yukhin <kirill.yukhin@intel.com>
1848 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1850 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
1851 (*prefetch_avx512pf_<mode>): New.
1852 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
1853 (avx512f_maskcmp<mode>3): Ditto.
1854 (vashrv16si3): Ditto.
1856 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1857 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1858 Sergey Lega <sergey.s.lega@intel.com>
1859 Anna Tikhonova <anna.tikhonova@intel.com>
1860 Ilya Tocar <ilya.tocar@intel.com>
1861 Andrey Turetskiy <andrey.turetskiy@intel.com>
1862 Ilya Verbin <ilya.verbin@intel.com>
1863 Kirill Yukhin <kirill.yukhin@intel.com>
1864 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1866 * config/i386/i386.md (any_truncate): New.
1867 (trunsuffix): Ditto.
1868 * config/i386/predicates.md (const_8_to_9_operand): New.
1869 (const_10_to_11_operand): Ditto.
1870 (const_12_to_13_operand): Ditto.
1871 (const_14_to_15_operand): Ditto.
1872 (const_16_to_19_operand): Ditto.
1873 (const_20_to_23_operand): Ditto.
1874 (const_24_to_27_operand): Ditto.
1875 (const_28_to_31_operand): Ditto.
1876 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
1877 (cvtusi2<ssescalarmodesuffix>32): New.
1878 (cvtusi2<ssescalarmodesuffix>64): Ditto.
1879 (ufloatv16siv16sf2): Ditto.
1880 (avx512f_fix_notruncv16sfv16si): Ditto.
1881 (avx512f_ufix_notruncv16sfv16si): Ditto.
1882 (avx512f_vcvtss2usi): Ditto.
1883 (avx512f_vcvtss2usiq): Ditto.
1884 (avx512f_vcvttss2usi): Ditto.
1885 (avx512f_vcvttss2usiq): Ditto.
1886 (avx512f_vcvtsd2usi): Ditto.
1887 (avx512f_vcvtsd2usiq): Ditto.
1888 (avx512f_vcvttsd2usi): Ditto.
1889 (avx512f_vcvttsd2usiq): Ditto.
1890 (ufloatv8siv8df): Ditto.
1891 (avx512f_cvtdq2pd512_2): Ditto.
1892 (avx512f_cvtpd2dq512): Ditto.
1893 (avx512f_ufix_notruncv8dfv8si): Ditto.
1894 (avx512f_cvtpd2ps512): Ditto.
1895 (vec_unpacks_lo_v16sf): Ditto.
1896 (vec_unpacks_hi_v16sf): Ditto.
1897 (vec_unpacks_float_hi_v16si): Ditto.
1898 (vec_unpacks_float_lo_v16si): Ditto.
1899 (avx512f_unpckhps512): Ditto.
1900 (avx512f_unpcklps512): Ditto.
1901 (avx512f_movshdup512): Ditto.
1902 (avx512f_movsldup512): Ditto.
1903 (vec_extract_lo_v32hi): Ditto.
1904 (vec_extract_hi_v32hi): Ditto.
1905 (vec_extract_lo_v64qi): Ditto.
1906 (vec_extract_hi_v64qi): Ditto.
1907 (avx512f_unpckhpd512): Ditto.
1908 (avx512f_movddup512): Ditto.
1909 (avx512f_unpcklpd512): Ditto.
1910 (*avx512f_unpcklpd512): Ditto.
1911 (avx512f_shufps512_1): Ditto.
1912 (avx512f_shufpd512_1): Ditto.
1913 (avx512f_interleave_highv8di): Ditto.
1914 (avx512f_interleave_lowv8di): Ditto.
1915 (PMOV_DST_MODE): Ditto.
1916 (pmov_src_mode): Ditto.
1917 (pmov_src_lower): Ditto.
1919 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
1920 (*avx512f_<code>v8div16qi2): Ditto.
1921 (*avx512f_<code>v8div16qi2_store): Ditto.
1922 (vec_widen_umult_even_v16si): Ditto.
1923 (*vec_widen_umult_even_v16si): Ditto.
1924 (vec_widen_smult_even_v16si): Ditto.
1925 (*vec_widen_smult_even_v16si): Ditto.
1926 (avx512f_interleave_highv16si): Ditto.
1927 (avx512f_interleave_lowv16si): Ditto.
1928 (avx512f_<code>v16qiv16si2): Ditto.
1929 (avx512f_<code>v16hiv16si2): Ditto.
1930 (avx512f_<code>v8qiv8di2): Ditto.
1931 (avx512f_<code>v8hiv8di2): Ditto.
1932 (avx512f_<code>v8siv8di2): Ditto.
1933 (avx512cd_maskb_vec_dupv8di): Ditto.
1934 (avx512cd_maskw_vec_dupv16si): Ditto.
1935 (avx512f_vcvtph2ps512): Ditto.
1936 (avx512f_vcvtps2ph512): Ditto.
1937 (VEC_EXTRACT_MODE): Extened with wider modes.
1938 (VEC_PERM_AVX2): Ditto.
1939 (VEC_PERM_CONST): Ditto.
1941 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
1943 * config/arc/arc.c (arc_ccfsm_post_advance):
1944 Add comment about TYPE_RETURN.
1946 2013-10-28 Bin Cheng <bin.cheng@arm.com>
1948 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
1949 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
1950 (strip_offset): Convert offset to unsigned number.
1952 2013-10-27 Tom de Vries <tom@codesourcery.com>
1954 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
1955 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
1957 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
1959 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
1960 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
1961 (TARGET_BIG_ENDIAN): New macro.
1962 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
1963 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
1964 * config/sh/sh.c: Likewise.
1965 * config/sh/sh.h: Likewise.
1967 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
1969 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
1970 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
1972 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1974 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
1976 (cmpstr_t, cmpstrsi): Fix formatting.
1978 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1981 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
1982 addressing, do not use general_operand for memory operands.
1984 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
1987 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1988 * lra-spills.c (lra_final_code_change): Remove useless move insns.
1990 2013-10-26 Jeff Law <law@redhat.com>
1992 * predict.c (PRED_MUDFLAP): Remove.
1993 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
1995 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
1996 (OBJS): Remove tree-nomudflap.o
1997 (GTFILES): Remove tree-mudflap.c
1998 * builtins.c (expand_builtin_alloc): Remove mudflap support.
1999 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2000 (mfwrap_spec, mflib_spec): Likewise.
2001 (cpp_unique_options, cc1_options, static_specs): Likewise.
2002 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
2003 * passes.def: Likewise.
2004 * toplev.c (compile_file, process_options): Likewise.
2005 * tree-inline.c (copy_tree_r): Likewise.
2006 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
2007 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
2008 (build_constant_desc, output_constant_def_contents): Likewise.
2009 (categorize_decl_for_section): Likewise.
2010 * tree-mudflap.c: Removed.
2011 * tree-mudflap.h: Removed.
2012 * tree-nomudflap.c: Removed.
2013 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
2014 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
2015 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2016 * config/sol2.h (MFLIB_SPEC): Likewise.
2017 * doc/install.texi: Remove mudflap references.
2018 * doc/passes.texi: Similarly.
2019 * doc/sourcebuild.texi: Similarly.
2020 * doc/invoke.texi: Remove mudlfap related options.
2022 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2024 PR rtl-optimization/58759
2025 * lra-constraints.c (lra_constraints): Remove wrong condition to
2026 remove insn setting up an equivalent pseudo.
2028 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2030 * config/rs6000/rs6000-protos.h
2031 (rs6000_secondary_memory_needed_mode): New prototype.
2032 * config/rs6000/rs6000.c: Include ira.h.
2033 (TARGET_LRA_P): Redefine.
2034 (rs6000_legitimate_offset_address_p): Call
2035 legitimate_constant_pool_address_p in strict mode for LRA.
2036 (rs6000_legitimate_address_p): Ditto.
2037 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
2038 (rs6000_emit_move): Add LRA version of code to generate load/store
2040 (rs6000_secondary_memory_needed_mode): New.
2041 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
2042 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
2043 constants, memory, and FP registers.
2044 (rs6000_lra_p): New.
2045 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
2046 * config/rs6000/rs6000.opt (mlra): New option.
2047 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2049 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
2051 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
2052 has_single_use () and not do the conversion if has_single_use ()
2053 returns false for the multiplication result.
2055 2013-10-25 David Malcolm <dmalcolm@redhat.com>
2057 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
2058 to "NODE", since this works on a "tree", not an
2060 (CONSTANT_CLASS_P): Likewise.
2063 (INDIRECT_REF_P): Likewise.
2064 (REFERENCE_CLASS_P): Likewise.
2065 (COMPARISON_CLASS_P): Likewise.
2066 (UNARY_CLASS_P): Likewise.
2067 (BINARY_CLASS_P): Likewise.
2068 (STATEMENT_CLASS_P): Likewise.
2069 (VL_EXP_CLASS_P): Likewise.
2070 (EXPRESSION_CLASS_P): Likewise.
2071 (IS_TYPE_OR_DECL_P): Likewise.
2073 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
2075 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
2076 ADDR_EXPR in the defining statement.
2078 2013-10-25 Richard Biener <rguenther@suse.de>
2080 PR tree-optimization/58626
2081 * tree-loop-distribution.c (enum rdg_dep_type): Remove
2082 anti_dd, output_dd and input_dd.
2083 (struct rdg_edge): Remove level and relation members.
2084 (RDGE_LEVEL, RDGE_RELATION): Remove.
2085 (dot_rdg_1): Adjust.
2086 (create_rdg_edge_for_ddr): Remove.
2087 (create_rdg_edges_for_scalar): Adjust.
2088 (create_edge_for_control_dependence): Likewise.
2089 (create_rdg_edges): Split into ...
2090 (create_rdg_flow_edges): ... this
2091 (create_rdg_cd_edges): ... and this.
2093 (build_rdg): Likewise, do not compute data dependences or
2095 (pg_add_dependence_edges): New function.
2097 (distribute_loop): First apply all non-dependence based
2098 partition mergings. Then compute dependences between partitions
2099 and merge and order partitions according to them.
2101 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2103 PR rtl-optimization/58831
2104 * alias.c (init_alias_analysis): At the beginning of each iteration,
2105 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
2107 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2109 * recog.c (search_ofs): New static variable moved from...
2110 (peep2_find_free_register): ...here.
2111 (peephole2_optimize): Initialize it.
2113 2013-10-25 Tobias Burnus <burnus@net-b.de>
2115 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
2117 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
2119 * config/i386/i386.h (TARGET_MPX): New define.
2120 (TARGET_MPX_P): Ditto.
2122 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2124 * config/i386/constraints.md (B): New.
2127 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
2128 * config/i386/i386-modes.def (BND32): New.
2130 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
2131 * config/i386/i386.c (isa_opts): Add mmpx.
2132 (regclass_map): Add bound registers.
2133 (dbx_register_map): Likewise.
2134 (dbx64_register_map): Likewise.
2135 (svr4_dbx_register_map): Likewise.
2137 (ix86_option_override_internal): Support MPX ISA.
2138 (ix86_conditional_register_usage): Support bound registers.
2139 (print_reg): Likewise.
2140 (ix86_code_end): Add MPX bnd prefix.
2141 (output_set_got): Likewise.
2142 (ix86_output_call_insn): Likewise.
2143 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
2144 (ix86_print_operand_punct_valid_p): Likewise.
2145 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
2147 (ix86_class_likely_spilled_p): Add bound regs support.
2148 (ix86_hard_regno_mode_ok): Likewise.
2149 (x86_order_regs_for_local_alloc): Likewise.
2150 (ix86_bnd_prefixed_insn_p): New.
2151 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
2152 (FIXED_REGISTERS): Add bound registers.
2153 (CALL_USED_REGISTERS): Likewise.
2154 (REG_ALLOC_ORDER): Likewise.
2155 (HARD_REGNO_NREGS): Likewise.
2157 (VALID_BND_REG_MODE): New.
2158 (FIRST_BND_REG): New.
2159 (LAST_BND_REG): New.
2160 (reg_class): Add BND_REGS.
2161 (REG_CLASS_NAMES): Likewise.
2162 (REG_CLASS_CONTENTS): Likewise.
2164 (ANY_BND_REG_P): New.
2166 (HI_REGISTER_NAMES): Add bound registers.
2167 * config/i386/i386.md (UNSPEC_BNDMK): New.
2168 (UNSPEC_BNDMK_ADDR): New.
2169 (UNSPEC_BNDSTX): New.
2170 (UNSPEC_BNDLDX): New.
2171 (UNSPEC_BNDLDX_ADDR): New.
2172 (UNSPEC_BNDCL): New.
2173 (UNSPEC_BNDCU): New.
2174 (UNSPEC_BNDCN): New.
2175 (UNSPEC_MPX_FENCE): New.
2178 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
2179 (length_immediate): Likewise.
2180 (prefix_0f): Likewise.
2182 (prefix_rep): Check for bnd prefix.
2183 (length_nobnd): New.
2184 (length): Use length_nobnd if specified.
2189 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
2192 (simple_return_internal): Likewise.
2193 (simple_return_pop_internal): Likewise.
2194 (*indirect_jump): Add MPX bnd prefix.
2195 (*tablejump_1): Likewise.
2196 (simple_return_internal_long): Likewise.
2197 (simple_return_indirect_internal): Likewise.
2201 (*mov<mode>_internal_mpx): New.
2202 (<mode>_<bndcheck>): New.
2203 (*<mode>_<bndcheck>): New.
2208 * config/i386/predicates.md (lea_address_operand): Rename to...
2209 (address_no_seg_operand): ... this.
2210 (address_mpx_no_base_operand): New.
2211 (address_mpx_no_index_operand): New.
2212 (bnd_mem_operator): New.
2213 * config/i386/i386.opt (mmpx): New.
2214 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
2215 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
2217 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2219 * mode-classes.def (MODE_POINTER_BOUNDS): New.
2220 * tree.def (POINTER_BOUNDS_TYPE): New.
2221 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
2222 (POINTER_BOUNDS_MODE): New.
2223 (make_pointer_bounds_mode): New.
2224 * machmode.h (POINTER_BOUNDS_MODE_P): New.
2225 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
2226 (layout_type): Support POINTER_BOUNDS_TYPE.
2227 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
2228 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
2229 (type_contains_placeholder_1): Likewise.
2230 * tree.h (POINTER_BOUNDS_TYPE_P): New.
2231 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
2232 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
2234 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2236 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
2237 mode of address computation when calling memory_address_addr_space.
2239 2013-08-24 Richard Henderson <rth@twiddle.net>
2242 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
2243 instead of create_convert_operand_to.
2244 (maybe_emit_sync_lock_test_and_set): Likewise.
2245 (expand_atomic_compare_and_swap): Likewise.
2246 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
2248 2013-08-24 Sriraman Tallam <tmsriram@google.com>
2250 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
2251 Change param_is to use the struct and not the pointer to the struct.
2253 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
2255 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
2256 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2257 Move to targhooks.c.
2258 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
2259 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2260 Relocate from gimplify.c.
2261 * targhooks.h: Add 2 prototypes.
2262 * tree.h: Delete 2 prototypes.
2264 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2266 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
2267 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
2269 2013-10-24 Cong Hou <congh@google.com>
2271 * convert.c (convert_to_real): Guard those unsafe math function
2272 convertions with flag_unsafe_math_optimizations. Handle sqrt()
2275 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
2278 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
2280 (cgraph_create_edge): Use the new argument.
2281 (cgraph_create_indirect_edge): Likewise.
2283 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
2285 * gcc/config/arc/arc.c (arc_ccfsm_post_advance): Also handle
2287 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
2288 changes statep->state.
2290 2013-10-24 Tobias Burnus <burnus@net-b.de>
2293 * tree-cfg.c (replace_loop_annotate): New function.
2294 (execute_build_cfg): Call it.
2295 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
2296 * internal-fn.c (expand_ANNOTATE): New function.
2297 * internal-fn.def (ANNOTATE): Define as new internal function.
2298 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
2299 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
2300 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
2301 * doc/extend.texi (Pragmas): Document #pragma ivdep.
2302 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
2304 2013-10-17 Ian Bolton <ian.bolton@arm.com>
2305 Marcus Shawcroft <marcus.shawcroft@arm.com>
2307 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
2308 Special case reload SP+C into none GENERAL_REGS.
2310 2013-10-24 Michael Matz <matz@suse.de>
2312 * gengtype.c (is_file_equal): Check that files will be same length.
2314 2013-10-25 Christian Bruel <christian.bruel@st.com>
2316 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
2317 * gcc/config/sh/t-sh (sh-mem.o): New rule.
2318 * gcc/config/sh/sh-mem.cc (expand_block_move): Moved here.
2319 (sh_expand_cmpstr): New function.
2320 * gcc/config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
2321 * gcc/config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
2322 * gcc/config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
2323 (rotlhi3_8): Rename.
2325 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2327 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
2328 * Makefile.in (CXXDEPMODE): Assign and change users.
2329 (CCDEPMODE): Delete.
2330 * configure: Regenerate.
2332 2013-10-23 David Malcolm <dmalcolm@redhat.com>
2334 * gengtype-parse.c (require_without_advance): New.
2335 (type): For GTY-marked types that are not GTY((user)), parse any
2336 base classes, requiring them to be single-inheritance, and not
2337 be templates. For non-GTY-marked types and GTY((user)),
2338 continue to skip over any C++ inheritance specification.
2339 * gengtype-state.c (state_writer::write_state_struct_type):
2340 Write base class of type (if any).
2341 (read_state_struct_type): Read base class of type (if any).
2342 * gengtype.c (new_structure): Add a "base_class" parameter.
2343 (create_optional_field_): Update for new parameter to new_structure.
2344 (adjust_field_rtx_def): Likewise.
2345 (adjust_field_tree_exp): Likewise.
2346 * gengtype.h (struct type): Add "base_class" field to the s
2348 (new_structure): Add "base" parameter.
2350 2013-10-23 Sriraman Tallam <tmsriram@google.com>
2353 * config/i386/i386.c (ix86_option_override_internal):
2354 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
2355 (ix86_valid_target_attribute_tree):
2356 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
2357 Change TARGET_SSE to TARGET_SSE_P (opts->...)
2359 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2361 * tree-ssa-loop.h: Remove include files.
2362 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2363 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
2364 * graphite-clast-to-gimple.c: Likewise.
2365 * graphite-scop-detection.c: Likewise.
2366 * graphite-sese-to-poly.c: Likewise.
2367 * ipa-inline-analysis.c: Likewise.
2368 * ipa-pure-const.c: Likewise.
2369 * loop-init.c: Likewise.
2370 * passes.c: Likewise.
2371 * predict.c: Likewise.
2372 * tree-cfg.c: Likewise.
2373 * tree-cfgcleanup.c: Likewise.
2374 * tree-chrec.c: Likewise.
2375 * tree-data-ref.c: Likewise.
2376 * tree-loop-distribution.c: Likewise.
2377 * tree-parloops.c: Likewise.
2378 * tree-predcom.c: Likewise.
2379 * tree-scalar-evolution.c: Likewise.
2380 * tree-ssa-address.c: Likewise.
2381 * tree-ssa.c: Likewise.
2382 * tree-ssa-dce.c: Likewise.
2383 * tree-ssa-loop.c: Likewise.
2384 * tree-ssa-loop-im.c: Likewise.
2385 * tree-ssa-loop-ivcanon.c: Likewise.
2386 * tree-ssa-loop-ivopts.c: Likewise.
2387 * tree-ssa-loop-manip.c: Likewise.
2388 * tree-ssa-loop-niter.c: Likewise.
2389 * tree-ssa-loop-prefetch.c: Likewise.
2390 * tree-ssa-loop-unswitch.c: Likewise.
2391 * tree-ssa-reassoc.c: Likewise.
2392 * tree-vect-data-refs.c: Likewise.
2393 * tree-vect-loop.c: Likewise.
2394 * tree-vect-loop-manip.c: Likewise.
2395 * tree-vectorizer.c: Likewise.
2396 * tree-vect-stmts.c: Likewise.
2397 * tree-vrp.c: Likewise.
2399 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2401 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
2403 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2405 PR tree-optimization/58775
2406 PR tree-optimization/58791
2407 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
2408 (insert_stmt_after): Rewritten, don't move the stmt, but really
2410 (get_stmt_uid_with_default): Remove.
2411 (build_and_add_sum): Use insert_stmt_after and
2412 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
2413 (update_range_test): Set uid on stmts added by
2414 force_gimple_operand_gsi. Don't immediately modify statements
2415 in inter-bb optimization, just update oe->op values.
2416 (optimize_range_tests): Return bool whether any changed have been made.
2417 (update_ops): New function.
2418 (struct inter_bb_range_test_entry): New type.
2419 (maybe_optimize_range_tests): Perform statement changes here.
2420 (not_dominated_by, appears_later_in_bb, get_def_stmt,
2421 ensure_ops_are_available): Remove.
2422 (find_insert_point): Rewritten.
2423 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
2424 return LHS of the (new resp. old) stmt. Don't call
2425 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
2426 instead of last, move new stmt at the right place.
2427 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
2428 (negate_value): Likewise. Set uids.
2429 (break_up_subtract_bb): Initialize uids.
2430 (reassociate_bb): Adjust rewrite_expr_tree caller.
2431 (do_reassoc): Don't call renumber_gimple_stmt_uids.
2433 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
2436 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
2437 TARGET_32BIT final condition.
2438 (mulsi3_internal2 and splitter): Same.
2440 2013-10-23 Jeff Law <law@redhat.com>
2442 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
2443 through joiner blocks with abnormal outgoing edges.
2445 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
2446 Add parameter JOINERS, to allow/disallow threading through joiner
2448 (thread_block): New. Call thread_block_1.
2449 (mark_threaded_blocks): Remove code to filter out certain cases
2450 of threading through joiner blocks.
2451 (thread_through_all_blocks): Document how we can have a dangling
2452 edge AUX field and clear it.
2454 2013-10-23 Ian Lance Taylor <iant@google.com>
2456 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
2457 (C++ Dialect Options): Likewise.
2458 (Optimize Options): Likewise.
2460 2013-10-23 Tom de Vries <tom@codesourcery.com>
2462 PR tree-optimization/58805
2463 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
2465 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2467 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
2468 sequence based on get_range_info returned range.
2470 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2472 * tree-ssa.h: Remove all #include's
2473 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2474 * alias.c: Move required includes from tree-ssa.h.
2476 * builtins.c: Likewise.
2477 * calls.c: Likewise.
2478 * cfgexpand.c: Likewise.
2479 * cfghooks.c: Likewise.
2480 * cfgloop.c: Likewise.
2481 * cfgloopmanip.c: Likewise.
2482 * cgraph.c: Likewise.
2483 * cgraphbuild.c: Likewise.
2484 * cgraphclones.c: Likewise.
2485 * cgraphunit.c: Likewise.
2487 * except.c: Likewise.
2489 * final.c: Likewise.
2490 * fold-const.c: Likewise.
2491 * ggc-page.c: Likewise.
2492 * gimple-builder.c: Likewise.
2493 * gimple-fold.c: Likewise.
2494 * gimple-iterator.c: Likewise.
2495 * gimple-low.c: Likewise.
2496 * gimple-pretty-print.c: Likewise.
2497 * gimple-ssa-strength-reduction.c: Likewise.
2498 * gimple-streamer-in.c: Likewise.
2499 * gimple-streamer-out.c: Likewise.
2500 * gimplify.c: Likewise.
2501 * graphite-blocking.c: Likewise.
2502 * graphite-clast-to-gimple.c: Likewise.
2503 * graphite-dependences.c: Likewise.
2504 * graphite-interchange.c: Likewise.
2505 * graphite-optimize-isl.c: Likewise.
2506 * graphite-poly.c: Likewise.
2507 * graphite-scop-detection.c: Likewise.
2508 * graphite-sese-to-poly.c: Likewise.
2509 * graphite.c: Likewise.
2510 * ipa-cp.c: Likewise.
2511 * ipa-inline-analysis.c: Likewise.
2512 * ipa-inline-transform.c: Likewise.
2513 * ipa-inline.c: Likewise.
2514 * ipa-prop.c: Likewise.
2515 * ipa-pure-const.c: Likewise.
2516 * ipa-reference.c: Likewise.
2517 * ipa-split.c: Likewise.
2518 * ipa-utils.c: Likewise.
2519 * loop-init.c: Likewise.
2520 * lto-cgraph.c: Likewise.
2521 * lto-section-in.c: Likewise.
2522 * lto-section-out.c: Likewise.
2523 * lto-streamer-in.c: Likewise.
2524 * lto-streamer-out.c: Likewise.
2525 * lto-streamer.c: Likewise.
2526 * omp-low.c: Likewise.
2527 * passes.c: Likewise.
2528 * predict.c: Likewise.
2529 * print-tree.c: Likewise.
2530 * profile.c: Likewise.
2532 * targhooks.c: Likewise.
2533 * tracer.c: Likewise.
2534 * trans-mem.c: Likewise.
2535 * tree-call-cdce.c: Likewise.
2536 * tree-cfg.c: Likewise.
2537 * tree-cfgcleanup.c: Likewise.
2538 * tree-chrec.c: Likewise.
2539 * tree-complex.c: Likewise.
2540 * tree-data-ref.c: Likewise.
2541 * tree-dfa.c: Likewise.
2542 * tree-eh.c: Likewise.
2543 * tree-emutls.c: Likewise.
2544 * tree-if-conv.c: Likewise.
2545 * tree-inline.c: Likewise.
2546 * tree-into-ssa.c: Likewise.
2547 * tree-loop-distribution.c: Likewise.
2548 * tree-mudflap.c: Likewise.
2549 * tree-nested.c: Likewise.
2550 * tree-nrv.c: Likewise.
2551 * tree-object-size.c: Likewise.
2552 * tree-outof-ssa.c: Likewise.
2553 * tree-parloops.c: Likewise.
2554 * tree-phinodes.c: Likewise.
2555 * tree-predcom.c: Likewise.
2556 * tree-pretty-print.c: Likewise.
2557 * tree-profile.c: Likewise.
2558 * tree-scalar-evolution.c: Likewise.
2559 * tree-sra.c: Likewise.
2560 * tree-ssa-address.c: Likewise.
2561 * tree-ssa-alias.c: Likewise.
2562 * tree-ssa-ccp.c: Likewise.
2563 * tree-ssa-coalesce.c: Likewise.
2564 * tree-ssa-copy.c: Likewise.
2565 * tree-ssa-copyrename.c: Likewise.
2566 * tree-ssa-dce.c: Likewise.
2567 * tree-ssa-dom.c: Likewise.
2568 * tree-ssa-dse.c: Likewise.
2569 * tree-ssa-forwprop.c: Likewise.
2570 * tree-ssa-ifcombine.c: Likewise.
2571 * tree-ssa-live.c: Likewise.
2572 * tree-ssa-loop-ch.c: Likewise.
2573 * tree-ssa-loop-im.c: Likewise.
2574 * tree-ssa-loop-ivcanon.c: Likewise.
2575 * tree-ssa-loop-ivopts.c: Likewise.
2576 * tree-ssa-loop-manip.c: Likewise.
2577 * tree-ssa-loop-niter.c: Likewise.
2578 * tree-ssa-loop-prefetch.c: Likewise.
2579 * tree-ssa-loop-unswitch.c: Likewise.
2580 * tree-ssa-loop.c: Likewise.
2581 * tree-ssa-math-opts.c: Likewise.
2582 * tree-ssa-operands.c: Likewise.
2583 * tree-ssa-phiopt.c: Likewise.
2584 * tree-ssa-phiprop.c: Likewise.
2585 * tree-ssa-pre.c: Likewise.
2586 * tree-ssa-propagate.c: Likewise.
2587 * tree-ssa-reassoc.c: Likewise.
2588 * tree-ssa-sccvn.c: Likewise.
2589 * tree-ssa-sink.c: Likewise.
2590 * tree-ssa-strlen.c: Likewise.
2591 * tree-ssa-structalias.c: Likewise.
2592 * tree-ssa-tail-merge.c: Likewise.
2593 * tree-ssa-ter.c: Likewise.
2594 * tree-ssa-threadedge.c: Likewise.
2595 * tree-ssa-threadupdate.c: Likewise.
2596 * tree-ssa-uncprop.c: Likewise.
2597 * tree-ssa-uninit.c: Likewise.
2598 * tree-ssa.c: Likewise.
2599 * tree-ssanames.c: Likewise.
2600 * tree-stdarg.c: Likewise.
2601 * tree-streamer-in.c: Likewise.
2602 * tree-switch-conversion.c: Likewise.
2603 * tree-tailcall.c: Likewise.
2604 * tree-vect-data-refs.c: Likewise.
2605 * tree-vect-generic.c: Likewise.
2606 * tree-vect-loop-manip.c: Likewise.
2607 * tree-vect-loop.c: Likewise.
2608 * tree-vect-patterns.c: Likewise.
2609 * tree-vect-slp.c: Likewise.
2610 * tree-vect-stmts.c: Likewise.
2611 * tree-vectorizer.c: Likewise.
2612 * tree-vrp.c: Likewise.
2615 * value-prof.c: Likewise.
2616 * var-tracking.c: Likewise.
2617 * varpool.c: Likewise.
2618 * vtable-verify.c: Likewise.
2620 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2622 * config/tilegx/tilegx.c: Include "tree.h".
2624 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2626 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
2627 the info, not after it.
2628 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
2629 dump_ssaname_info call.
2630 (pp_gimple_stmt_1): Adjust caller.
2631 (dump_phi_nodes): Likewise. Don't print "# " here.
2633 2013-10-22 Jan Hubicka <jh@suse.cz>
2635 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
2637 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
2638 * i386.c (expand_small_movmem_or_setmem): New function.
2639 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
2641 (alg_usable_p): Add support for value ranges; cleanup.
2642 (ix86_expand_set_or_movmem): Add support for misaligned moves.
2644 2013-10-22 Sterling Augustine <saugustine@google.com>
2646 * doc/invoke.texi: Document -ggnu-pubnames.
2647 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
2649 * dwarf2out.c: Include gdb/gdb-index.h.
2650 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
2651 debug_generate_pub_sections.
2652 (is_java, output_pubtables, output_pubname): New functions.
2653 (include_pubname_in_output): Handle debug_generate_pub_sections at
2655 (size_of_pubnames): Use new local space_for_flags based on
2656 debug_generate_pub_sections.
2657 (output_pubnames): Unify pubnames and pubtypes output logic.
2658 Genericize comments. Call output_pubname.
2659 (dwarf2out_finish): Move logic to output_pubnames and call it.
2661 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
2664 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
2665 Remove CCCmode handling.
2666 <case LTU>: Return 'c' suffix for CCCmode.
2667 <case GEU>: Return 'nc' suffix for CCCmode.
2668 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
2669 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
2670 (*sub<mode>3_cc_overflow): Ditto.
2671 (*subsi3_zext_cc_overflow): Ditto.
2673 2013-10-22 Steve Ellcey <sellcey@mips.com>
2675 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
2676 (AND (NOT OP1) (NOT OP2)).
2678 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
2680 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2681 meaning of merge-high and merge-low masks for little endian; avoid
2682 use of vector-pack masks for little endian for mismatched modes.
2684 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2686 * config/tilepro/tilepro.c: Include "tree.h".
2688 2013-10-22 Andreas Schwab <schwab@suse.de>
2690 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
2693 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
2695 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
2696 [Ed Smith-Rowland]): New entries.
2697 ([Stephen M. Webb]): Update.
2699 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
2701 * tree-ssa-ter.h: Remove duplicate copy of file contents.
2703 2013-10-21 Marek Polacek <polacek@redhat.com>
2706 * fold-const.c (fold_range_test): Return 0 if the type is not
2709 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
2711 * system.h: Move hwint.h include further down.
2712 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
2713 gcc_checking_asserts.
2714 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
2716 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2718 Fix volatile issues in optimize_bit_field_compare.
2719 * fold-const.c (optimize_bit_field_compare): Bail out if
2720 lvolatilep or rvolatilep.
2722 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2724 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
2725 and get_inner_reference returning different pmode for non-volatile
2726 bit-field members dependent on flag_strict_volatile_bitfields.
2727 * stor-layout.c (layout_decl): Remove special handling of
2728 flag_strict_volatile_bitfields.
2729 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
2730 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
2732 2013-10-21 Paulo Matos <pmatos@broadcom.com>
2734 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
2737 2013-10-21 Jeff Law <law@redhat.com>
2739 * tree-ssa-threadedge.c (thread_through_normal_block): New
2740 argument VISITED. Remove VISISTED as a local variable. When we
2741 have a threadable jump, verify the destination of the jump has not
2743 (thread_across_edge): Allocate VISITED bitmap once at function
2744 scope and use it throughout. Make sure to set appropriate bits in
2745 VISITED for E (start of jump thread path).
2746 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
2747 through a joiner if any edge on the path has a recorded jump thread.
2749 2013-10-21 Ian Lance Taylor <iant@google.com>
2751 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
2752 don't imply that attributes can solve all problems.
2753 (Directory Options): Fix typo.
2755 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2757 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
2758 extend and extend_add.
2760 2013-10-21 Richard Biener <rguenther@suse.de>
2762 PR tree-optimization/58794
2763 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
2764 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
2766 2013-10-21 Richard Biener <rguenther@suse.de>
2769 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
2770 to (T) X for sign-changing conversions (or no conversion).
2772 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
2774 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
2776 2013-10-20 Jan Hubicka <jh@suse.cz>
2778 * config/i386/i386-tune.def: Add comment; organize into categories
2780 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2782 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
2783 argument. Update function comment.
2784 (expand_set_or_movmem_via_rep): New function combining
2785 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
2786 (expand_movmem_via_rep_mov): Remove.
2787 expand_setmem_via_rep_stos): Remove.
2788 (expand_movmem_epilogue): Update calls correspondingly.
2789 (expand_setmem_epilogue_via_loop): Likewise.
2791 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
2792 (expand_set_or_movmem_prologue): New function combining
2793 expand_movmem_prologue and expand_setmem_prologue.
2794 (expand_movmem_prologue): Remove.
2795 (expand_setmem_prologue): Remove.
2796 (expand_set_or_movmem_constant_prologue): New function combining
2797 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
2798 (expand_constant_movmem_prologue): Remove.
2799 (expand_constant_setmem_prologue): Remove.
2800 (promote_duplicated_reg): Allow vector-const0 value.
2801 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
2802 and ix86_expand_setmem.
2803 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
2804 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
2806 2013-10-21 Diego Novillo <dnovillo@google.com>
2808 * asan.c: Include tree.h
2809 * bb-reorder.c: Likewise.
2810 * cfgcleanup.c: Likewise.
2811 * cfgloopmanip.c: Likewise.
2812 * data-streamer-in.c: Likewise.
2813 * data-streamer-out.c: Likewise.
2814 * data-streamer.c: Likewise.
2815 * dwarf2cfi.c: Likewise.
2816 * graphite-blocking.c: Likewise.
2817 * graphite-clast-to-gimple.c: Likewise.
2818 * graphite-dependences.c: Likewise.
2819 * graphite-interchange.c: Likewise.
2820 * graphite-optimize-isl.c: Likewise.
2821 * graphite-poly.c: Likewise.
2822 * graphite-scop-detection.c: Likewise.
2823 * graphite-sese-to-poly.c: Likewise.
2824 * graphite.c: Likewise.
2825 * ipa-devirt.c: Likewise.
2826 * ipa-profile.c: Likewise.
2829 * loop-init.c: Likewise.
2830 * loop-unroll.c: Likewise.
2831 * lower-subreg.c: Likewise.
2832 * lto/lto-object.c: Likewise.
2833 * recog.c: Likewise.
2834 * reginfo.c: Likewise.
2835 * tree-loop-distribution.c: Likewise.
2836 * tree-parloops.c: Likewise.
2837 * tree-ssa-strlen.c: Likewise.
2838 * tree-streamer.c: Likewise.
2839 * value-prof.c: Likewise.
2840 * target-globals.c: Likewise.
2841 * expr.h: Include tree-core.h instead of tree.h.
2842 * gimple.h: Likewise.
2843 * ipa-prop.h: Likewise.
2844 * ipa-utils.h: Likewise.
2845 * lto-streamer.h: Likewise.
2846 * streamer-hooks.h: Likewise.
2847 * ipa-reference.h: Include cgraph.h instead of tree.h.
2848 * cgraph.h: Include basic-block.h instead of tree.h.
2849 * tree-streamer.h: Do not include tree.h.
2850 * genattrtab.c (write_header): Generate inclusion of
2852 * genautomata.c (main): Likewise.
2853 * genemit.c: Likewise.
2854 * genopinit.c: Likewise.
2855 * genoutput.c (output_prologue): Likewise.
2856 * genpeep.c: Likewise.
2858 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2860 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
2862 (vec_unpacku_hi_v8hi): Likewise.
2863 (vec_unpacku_lo_v16qi): Likewise.
2864 (vec_unpacku_lo_v8hi): Likewise.
2866 2013-10-20 Jan Hubicka <jh@suse.cz>
2868 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
2869 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
2870 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
2872 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
2874 * config/mips/mips.h (ISA_HAS_WSBH): Define.
2875 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
2877 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
2879 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
2882 * system.h: Undef m_slot.
2884 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2886 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
2887 all elements for both endian flavors.
2889 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2892 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
2893 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
2894 and SI_REG for 64bit SYSV ABI targets.
2896 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2898 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
2899 to multi_reg_return. Clarify that we are skipping USEs of multiple
2900 return registers. Use bool type where appropriate.
2902 2013-10-18 Jan Hubicka <jh@suse.cz>
2904 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
2906 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
2907 (X86_TUNE_PUSH_MEMORY): Likewise.
2908 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
2909 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
2910 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
2912 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
2913 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
2915 (ix86_option_override_internal): Update to use tune features instead
2918 2013-10-18 Cong Hou <congh@google.com>
2920 PR tree-optimization/58508
2921 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
2922 statement that contains data refs with zero-step.
2924 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2926 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
2927 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
2928 * gimple-low.c (gimple_check_call_arg,
2929 gimple_check_call_matching_types): Move to cgraph.c.
2930 * gimple-low.h: Remove prototype.
2931 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
2932 Relocate from gimple-low.c.
2933 * cgraph.h: Add prototype. Don't include basic-block.h.
2934 * gimplify.c: Add gimple-low to include list.
2935 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
2936 * tree-eh.c: Add gimple-low to include list.
2937 * tree-nested.c: Likewise.
2938 * cfgexpand.c: Add tree-ssa-address.h to include list.
2940 * gimple-fold.c: Likewise.
2941 * gimple-ssa-strength-reduction.c: Likewise.
2942 * trans-mem.c: Likewise.
2943 * tree-mudflap.c: Likewise.
2944 * tree-ssa-loop-ivopts.c: Likewise.
2945 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
2946 (degenerate_phi_result): Move to tree-phinodes.c.
2947 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
2948 * tree-ssa-threadedge.c: Likewise.
2949 * tree-vrp.c: Likewise.
2950 * tree-phinodes.c (degenerate_phi_result): Relocate here.
2951 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
2952 * tree-phinodes.h (degenerate_phi_result): Add prototype.
2953 * tree-ssa-copy.c: Include tree-ssa-dom.h.
2954 * tree-ssa-forwprop.c: Likewise.
2955 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
2956 pass_data_cleanup_cfg_post_optimizing,
2957 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
2958 * tree-optimize.c: Delete File.
2959 * graphite.c: Include tree-cfgcleanup.h.
2960 * passes.c: Likewise.
2961 * tree-cfg.c: Likewise.
2962 * tree-profile.c: Likewise.
2963 * tree-ssa-dse.c: Likewise.
2964 * tree-ssa-loop-ivcanon.c: Likewise.
2965 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
2966 * tree-outof-ssa.c: Include sbitmap.h.
2967 * tree-ssa-live.c: Likewise.
2968 * tree-ssa-propagate.c: Likewise.
2969 * tree-ssa-structalias.c: Likewise.
2970 * tree-stdarg.c: Likewise.
2971 * Makefile.in (OBJS): Delete tree-optimize.o.
2972 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
2973 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
2974 * varasm.c: Include basic-block.h.
2975 * cfgloop.h: Include function.h instead of basic-block.h
2976 (bb_loop_depth): Move to cfgloop.c.
2977 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
2979 2013-10-18 Teresa Johnson <tejohnson@google.com>
2981 * predict.c (probably_never_executed): Compare frequency-based
2982 count to number of training runs.
2983 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
2985 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2987 * config/arm/arm.c (cortexa9_extra_costs): New table.
2988 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
2990 2013-10-18 Jeff Law <law@redhat.com>
2992 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
2994 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
2995 (register_jump_thread): Add "registered_jump_thread" debug
2997 * dbgcnt.def (registered_jump_thread): New debug counter.
2999 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3001 * config/rs6000/rs6000.c: Include cgraph.h.
3003 2013-10-18 Teresa Johnson <tejohnson@google.com>
3005 * tree-ssa-tail-merge.c (replace_block_by): Update edge
3006 weights during merging.
3008 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3010 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
3011 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
3012 * tree-ssa-operands.h: Remove prototype.
3013 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
3014 * gimple.c (virtual_operand_p): Relocate from gimple.c.
3015 * gimple.h: Add prototype.
3016 * gimple-ssa.h: Include tree-ssa-operands.h.
3017 * tree-dump.c: Add tree-cfg.h to include list.
3018 * tree-ssa-alias.c: Add ipa-reference.h to include list.
3019 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
3020 * config/i386/i386.c: Don't include tree-flow.h.
3021 * config/rs6000/rs6000.c: Likewise.
3023 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3025 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
3027 2013-10-18 Richard Biener <rguenther@suse.de>
3029 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
3030 or TYPE_UNSIGNED of integral types.
3031 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
3032 NULL_TREE for zero-precision integral types.
3034 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
3036 * config/aarch64/arm_neon.h
3037 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
3039 2013-10-17 Sriraman Tallam <tmsriram@google.com>
3042 * opth-gen.awk: Define target_flags_explicit.
3044 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3046 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
3047 fields to the reg_addr array that describes the valid addressing
3048 mode for any register, general purpose registers, floating point
3049 registers, and Altivec registers.
3050 (FIRST_RELOAD_REG_CLASS): Likewise.
3051 (LAST_RELOAD_REG_CLASS): Likewise.
3052 (struct reload_reg_map_type): Likewise.
3053 (reload_reg_map_type): Likewise.
3054 (RELOAD_REG_VALID): Likewise.
3055 (RELOAD_REG_MULTIPLE): Likewise.
3056 (RELOAD_REG_INDEXED): Likewise.
3057 (RELOAD_REG_OFFSET): Likewise.
3058 (RELOAD_REG_PRE_INCDEC): Likewise.
3059 (RELOAD_REG_PRE_MODIFY): Likewise.
3060 (reg_addr): Likewise.
3061 (mode_supports_pre_incdec_p): New helper functions to say whether
3062 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
3063 (mode_supports_pre_modify_p): Likewise.
3064 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
3065 print the valid address mode bits for each mode.
3066 (rs6000_debug_print_mode): Likewise.
3067 (rs6000_debug_reg_global): Likewise.
3068 (rs6000_setup_reg_addr_masks): New function to set up the address
3069 mask bits for each type.
3070 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
3071 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
3072 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
3073 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
3074 PRE_MODIFY are supported.
3075 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
3076 registers, instead of {src,dest}_av_p.
3077 (rs6000_print_options_internal): Tweak the debug output slightly.
3079 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3081 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
3084 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
3086 * tree-flow.h (struct omp_region): Move to omp-low.c.
3087 Remove omp_ prototypes and variables.
3088 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
3089 (copy_var_decl): Relocate prototype from tree-flow.h.
3090 * gimple.c (copy_var_decl): Relocate from omp-low.c.
3091 * tree.h: Move prototype to omp-low.h.
3092 * omp-low.h: New File. Relocate prototypes here.
3093 * omp-low.c (struct omp_region): Make local here.
3094 (root_omp_region): Make static.
3095 (copy_var_decl) Move to gimple.c.
3096 (new_omp_region): Make static.
3097 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
3098 * tree-cfg.c: Include omp-low.h.
3099 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
3100 * gimplify.c: Include omp-low.h.
3101 * tree-parloops.c: Likewise.
3103 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3105 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
3106 operands are in memory, prefer to force non-matched operand 1 to
3109 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3112 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
3113 restrict TImode addresses to single indirect registers if both
3114 -mquad-memory and -mvsx-timode are used.
3115 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
3116 we should emit load/store quad. Remove using %y for quad memory
3119 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
3120 constraints to allow load/store quad on machines where TImode is
3121 not allowed in VSX registers. Use 'n' instead of 'F' constraint
3122 for TImode to load integer constants.
3124 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3126 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
3128 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
3130 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
3131 handling of STACK_REG.
3133 2013-10-17 Richard Biener <rguenther@suse.de>
3135 PR tree-optimization/58143
3136 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
3138 (rewrite_to_defined_overflow): Likewise.
3139 (move_computations_dom_walker::before_dom): Rewrite stmts
3140 with undefined signed overflow that are not always executed
3141 into unsigned arithmetic.
3143 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3146 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
3147 explicit isa flag to be an options variable, instead of using
3148 global_options_set. Remove define from rs6000.h.
3149 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
3151 * config/rs6000/rs6000.c (rs6000_option_override_internal):
3152 Initialize rs6000_isa_flags_explicit.
3153 (rs6000_function_specific_save): Add gcc_options* parameter, so
3154 that the powerpc builds after the 2013-10-15 changes.
3155 (rs6000_function_specific_restore): Likewise.
3157 2013-10-16 DJ Delorie <dj@redhat.com>
3159 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
3160 op is a REG before checking REGNO.
3161 (rl78_alloc_physical_registers): Verify pattern is a SET before
3164 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3166 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
3168 (vec_unpacks_lo_v4sf): Likewise.
3169 (vec_unpacks_float_hi_v4si): Likewise.
3170 (vec_unpacks_float_lo_v4si): Likewise.
3171 (vec_unpacku_float_hi_v4si): Likewise.
3172 (vec_unpacku_float_lo_v4si): Likewise.
3174 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3176 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
3177 (vsx_concat_v2sf): Likewise.
3179 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
3181 * config/aarch64/aarch64.md
3182 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
3184 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
3186 PR tree-optimization/58697
3187 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
3188 estimated_loop_iterations_int.
3189 (max_stmt_executions_int): Call get_max_loop_iterations_int.
3190 (get_max_loop_iterations_int): New. HWINT version of
3191 get_max_loop_iterations.
3192 * cfgloop.h: Add prototypes.
3193 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
3194 * loop-unroll.c (decide_peel_once_rolling): Call
3195 get_estimated_loop_iterations_int.
3196 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
3197 * tree-ssa-loop-niter.h: Tweak prototypes.
3199 2013-10-16 David Malcolm <dmalcolm@redhat.com>
3201 * gengtype-parse.c (struct_field_seq): Ignore access-control
3202 keywords ("public:" etc).
3204 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
3206 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
3207 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
3209 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
3211 * config/arm/arm.opt (mlra): New option.
3212 * config/arm/arm.c (arm_lra_p): New function.
3213 (TARGET_LRA_P): Define.
3215 2013-10-16 Paulo Matos <pmatos@broadcom.com>
3217 * tree-core.h (tree_code_name): Remove.
3218 * tree.h (get_tree_code_name): New prototype.
3219 * tree.c (tree_code_name): Make static.
3220 (get_tree_code_name): New function.
3221 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
3222 tree_class_check_failed, tree_range_check_failed,
3223 tree_not_class_check_failed, omp_clause_check_failed,
3224 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
3225 wrapper get_tree_code_name instead of calling tree_code_name directly.
3226 * tree-vrp.c (dump_asserts_for): Likewise.
3227 * tree-dump.c (dequeue_and_dump): Likewise.
3228 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
3229 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
3230 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
3231 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
3232 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
3233 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
3234 dump_gimple_omp_for): Likewise.
3235 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
3236 * tree-ssa-pre.c (print_pre_expr): Likewise.
3237 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
3238 * print-tree.c (print_node_brief, print_node): Likewise.
3239 * gimple.c (gimple_check_failed): Likewise.
3240 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
3241 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
3242 * config/mep/mep.c (mep_validate_vliw): Likewise.
3243 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
3244 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
3246 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
3248 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
3251 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
3253 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
3254 (MULTILIB_MATCHES): Add multilib for -march=v8.
3256 2013-10-15 Sriraman Tallam <tmsriram@google.com>
3259 * optc-save-gen.awk: Add extra parameter to the save and restore
3261 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
3262 * tree.c (build_optimization_node): New parameter. Add extra parameter
3263 to call to cl_optimization_save.
3264 (build_target_option_node): New parameter. Add extra parameter
3265 to call to cl_target_option_save.
3266 * tree.h (build_optimization_node): New parameter.
3267 (build_target_option_node): New parameter.
3268 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
3269 build_optimization_node and build_target_option_node.
3270 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
3271 (handle_pragma_push_options): Ditto.
3272 * toplev.c (process_options): Ditto.
3273 * opts.c (init_options_struct): Check for opts_set non-null.
3274 * target.def (target_option.save): New parameter.
3275 (target_option.restore): New parameter.
3276 * tm.texi: Generate.
3277 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
3278 (ix86_pragma_target_parse): Ditto.
3279 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
3281 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
3282 to build_optimization_node and build_target_option_node.
3283 (rs6000_valid_attribute_p): Ditto.
3284 (rs6000_pragma_target_parse): Ditto.
3285 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
3287 * config/i386/i386.h:
3288 TARGET_64BIT_P: New Macro
3289 TARGET_MMX_P: New Macro.
3290 TARGET_3DNOW_P: New Macro.
3291 TARGET_3DNOW_A_P: New Macro.
3292 TARGET_SSE_P: New Macro.
3293 TARGET_SSE2_P: New Macro.
3294 TARGET_SSE3_P: New Macro.
3295 TARGET_SSSE3_P: New Macro.
3296 TARGET_SSE4_1_P: New Macro.
3297 TARGET_SSE4_2_P: New Macro.
3298 TARGET_AVX_P: New Macro.
3299 TARGET_AVX2_P: New Macro.
3300 TARGET_AVX512F_P: New Macro.
3301 TARGET_AVX512PF_P: New Macro.
3302 TARGET_AVX512ER_P: New Macro.
3303 TARGET_AVX512CD_P: New Macro.
3304 TARGET_FMA_P: New Macro.
3305 TARGET_SSE4A_P: New Macro.
3306 TARGET_FMA4_P: New Macro.
3307 TARGET_XOP_P: New Macro.
3308 TARGET_LWP_P: New Macro.
3309 TARGET_ABM_P: New Macro.
3310 TARGET_BMI_P: New Macro.
3311 TARGET_BMI2_P: New Macro.
3312 TARGET_LZCNT_P: New Macro.
3313 TARGET_TBM_P: New Macro.
3314 TARGET_POPCNT_P: New Macro.
3315 TARGET_SAHF_P: New Macro.
3316 TARGET_MOVBE_P: New Macro.
3317 TARGET_CRC32_P: New Macro.
3318 TARGET_AES_P: New Macro.
3319 TARGET_PCLMUL_P: New Macro.
3320 TARGET_CMPXCHG16B_P: New Macro.
3321 TARGET_FSGSBASE_P: New Macro.
3322 TARGET_RDRND_P: New Macro.
3323 TARGET_F16C_P: New Macro.
3324 TARGET_RTM_P: New Macro.
3325 TARGET_HLE_P: New Macro.
3326 TARGET_RDSEED_P: New Macro.
3327 TARGET_PRFCHW_P: New Macro.
3328 TARGET_ADX_P: New Macro.
3329 TARGET_FXSR_P: New Macro.
3330 TARGET_XSAVE_P: New Macro.
3331 TARGET_XSAVEOPT_P: New Macro.
3332 TARGET_LP64_P: New Macro.
3333 TARGET_X32_P: New Macro.
3334 TARGET_FPMATH_DEFAULT_P: New Macro.
3335 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
3336 * config/i386/i386.c (ix86_option_override_internal): New parameters.
3338 Change ix86_tune_string to access opts->x_ix86_tune_string.
3339 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
3340 Change ix86_arch_string to access opts->x_ix86_arch_string.
3341 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
3342 Change ix86_pmode to access opts->x_ix86_pmode.
3343 Change ix86_abi to access opts->x_ix86_abi.
3344 Change ix86_cmodel to access opts->x_ix86_cmodel.
3345 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
3346 Change ix86_isa_flags_explicit to access
3347 opts->x_ix86_isa_flags_explicit.
3348 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
3349 Change ix86_regparm to access opts->x_ix86_regparm.
3350 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
3351 Change ix86_preferred_stack_boundary_arg to access
3352 opts->x_ix86_preferred_stack_boundary_arg.
3353 Change ix86_force_align_arg_pointer to access
3354 opts->x_ix86_force_align_arg_pointer.
3355 Change ix86_incoming_stack_boundar_arg to access
3356 opts->x_ix86_incoming_stack_boundar_arg.
3357 Change ix86_fpmath to access opts->x_ix86_fpmath.
3358 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
3359 Change ix86_recip_name to access opts->x_ix86_recip_name.
3360 Change ix86_stack_protector_guard to access
3361 opts->x_ix86_stack_protector_guard.
3362 Change ix86_tune_memcpy_strategy to access
3363 opts->x_ix86_tune_memcpy_strategy.
3364 Change ix86_tune_memset_strategy to access
3365 opts->x_ix86_tune_memset_strategy.
3366 Change global_options to access opts.
3367 Change global_options_set to access opts_set.
3368 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
3369 Change TARGET_MMX to TARGET_MMX_P (opts->...).
3370 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
3371 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
3372 Change TARGET_SSE to TARGET_SSE_P (opts->...).
3373 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
3374 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
3375 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
3376 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
3377 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
3378 Change TARGET_AVX to TARGET_AVX_P (opts->...).
3379 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
3380 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
3381 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
3382 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
3383 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
3384 Change TARGET_FMA to TARGET_FMA_P (opts->...).
3385 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
3386 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
3387 Change TARGET_XOP to TARGET_XOP_P (opts->...).
3388 Change TARGET_LWP to TARGET_LWP_P (opts->...).
3389 Change TARGET_ABM to TARGET_ABM_P (opts->...).
3390 Change TARGET_BMI to TARGET_BMI_P (opts->...).
3391 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
3392 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
3393 Change TARGET_TBM to TARGET_TBM_P (opts->...).
3394 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
3395 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
3396 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
3397 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
3398 Change TARGET_AES to TARGET_AES_P (opts->...).
3399 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
3400 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
3401 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
3402 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
3403 Change TARGET_F16C to TARGET_F16C_P (opts->...).
3404 Change TARGET_RTM to TARGET_RTM_P (opts->...).
3405 Change TARGET_HLE to TARGET_HLE_P (opts->...).
3406 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
3407 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
3408 Change TARGET_ADX to TARGET_ADX_P (opts->...).
3409 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
3410 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
3411 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
3412 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
3413 Change TARGET_X32 to TARGET_X32_P (opts->...).
3414 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
3415 Change TARGET_FLOAT_RETURNS_IN_80387 to
3416 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
3417 (ix86_function_specific_save): New parameter. Use opts-> fields
3418 to replace global fields.
3419 (ix86_function_specific_restore): Ditto.
3420 (ix86_valid_target_attribute_inner_p): New parameters.
3422 Fix call to ix86_handle_option and set_option.
3423 (ix86_valid_target_attribute_tree): New parameters.
3424 Change global_options to access opts.
3425 Change global_options_set to access opts_set.
3426 Fix call to ix86_valid_target_attribute_inner_p.
3427 Change ix86_tune_string to access opts->x_ix86_tune_string.
3428 Change ix86_arch_string to access opts->x_ix86_arch_string.
3429 Change ix86_fpmath to access opts->x_ix86_fpmath
3430 Fix call to ix86_option_override_internal.
3431 Fix call to ix86_add_new_builtins.
3432 Fix calls to build_optimization_node and build_target_option_node.
3433 (ix86_valid_target_attribute_p): Remove access to global_options.
3434 Use new gcc_options structure func_options.
3435 Fix call to ix86_valid_target_attribute_tree.
3436 Fix call to build_optimization_node.
3437 (get_builtin_code_for_version): Fix call to
3438 ix86_valid_target_attribute_tree.
3440 2013-10-15 David Malcolm <dmalcolm@redhat.com>
3442 * Makefile.in (PICFLAG): New.
3443 (enable_host_shared): New.
3444 (INTERNAL_CFLAGS): Use PICFLAG.
3445 (LIBIBERTY): Use pic build of libiberty.a if configured with
3446 --enable-host-shared.
3447 * configure.ac: Add --enable-host-shared, setting up new
3449 * configure: Regenerate.
3450 * doc/install.texi (--enable-shared): Add note contrasting it with ...
3451 (--enable-host-shared): New option.
3453 2013-10-15 Richard Biener <rguenther@suse.de>
3455 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
3456 for built-in functions.
3458 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
3460 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
3461 (optimize_range_tests_1): New function,
3462 extracted from optimize_range_tests.
3463 (optimize_range_tests_xor): Similarly.
3464 (optimize_range_tests_diff): New function.
3465 (optimize_range_tests): Use optimize_range_tests_1.
3467 2013-10-15 Cong Hou <congh@google.com>
3469 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
3470 requirement of the reduction pattern so that one operand of the
3471 reduction operation can come from outside of the loop.
3473 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3475 * config/arm/neon-schedgen.ml: Remove.
3476 * config/arm/cortex-a9-neon.md: Remove comment regarding
3479 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3481 * config/arm/types: Remove old neon types.
3483 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3485 * config/arm/cortex-a7.md
3486 (cortex_a7_neon_type): New.
3487 (cortex_a7_neon_mul): Update for new types.
3488 (cortex_a7_neon_mla): Likewise.
3489 (cortex_a7_neon): Likewise.
3491 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3493 * config/arm/cortex-a15-neon.md
3494 (cortex_a15_neon_type): New,
3496 (cortex_a15_neon_int_1): Remove.
3497 (cortex_a15_neon_int_2): Likewise.
3498 (cortex_a15_neon_int_3): Likewise.
3499 (cortex_a15_neon_int_4): Likewise.
3500 (cortex_a15_neon_int_5): Likewise.
3501 (cortex_a15_neon_vqneg_vqabs): Likewise.
3502 (cortex_a15_neon_vmov): Likewise.
3503 (cortex_a15_neon_vaba): Likewise.
3504 (cortex_a15_neon_vaba_qqq): Likewise.
3505 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3506 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3507 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3509 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3510 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3511 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
3512 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3513 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3514 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3515 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3516 (cortex_a15_neon_shift_1): Likewise.
3517 (cortex_a15_neon_shift_2): Likewise.
3518 (cortex_a15_neon_shift_3): Likewise.
3519 (cortex_a15_neon_vshl_ddd): Likewise.
3520 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3521 (cortex_a15_neon_vsra_vrsra): Likewise.
3522 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3523 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3524 (cortex_a15_neon_bp_3cycle): Likewise.
3525 (cortex_a15_neon_ldm_2): Likewise.
3526 (cortex_a15_neon_stm_2): Likewise.
3527 (cortex_a15_neon_mcr): Likewise.
3528 (cortex_a15_neon_mrc): Likewise.
3529 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3530 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3531 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3532 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3533 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3534 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3535 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3536 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3537 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3538 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3539 (cortex_a15_neon_bp_simple): Likewise.
3540 (cortex_a15_neon_bp_2cycle): Likewise.
3541 (cortex_a15_neon_bp_3cycle): Likewise.
3542 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3543 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3544 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3545 (cortex_a15_neon_vld2_4_regs): Likewise.
3546 (cortex_a15_neon_vld3_vld4): Likewise.
3547 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3548 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3549 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
3550 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
3551 (cortex_a15_neon_vst3_vst4): Rename to...
3552 (cortex_a15_neon_vst4): This, update for new attributes.
3553 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
3554 (cortex_a15_neon_vld3_vld4_lane): Likewise.
3555 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3556 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3557 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3558 (cortex_a15_neon_ldm_2): Likewise.
3559 (cortex_a15_neon_stm_2): Likewise.
3560 (cortex_a15_neon_mcr): Likewise.
3561 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3562 (cortex_a15_neon_mrc): Likewise.
3563 (cortex_a15_neon_mrrc): Likewise.
3565 (cortex_a15_neon_abd): New.
3566 (cortex_a15_neon_abd_q): Likewise.
3567 (cortex_a15_neon_aba): Likewise.
3568 (cortex_a15_neon_aba_q): Likewise.
3569 (cortex_a15_neon_acc): Likewise.
3570 (cortex_a15_neon_acc_q): Likewise.
3571 (cortex_a15_neon_arith_basic): Likewise.
3572 (cortex_a15_neon_arith_complex): Likewise.
3573 (cortex_a15_neon_multiply): Likewise.
3574 (cortex_a15_neon_multiply_q): Likewise.
3575 (cortex_a15_neon_mla): Likewise.
3576 (cortex_a15_neon_mla_q): Likewise.
3577 (cortex_a15_neon_sat_mla_long): Likewise.
3578 (cortex_a15_neon_shift_acc): Likewise.
3579 (cortex_a15_neon_shift_imm_basic): Likewise.
3580 (cortex_a15_neon_shift_imm_complex): Likewise.
3581 (cortex_a15_neon_shift_reg_basic): Likewise.
3582 (cortex_a15_neon_shift_reg_basic_q): Likewise.
3583 (cortex_a15_neon_shift_reg_complex): Likewise.
3584 (cortex_a15_neon_shift_reg_complex_q): Likewise.
3585 (cortex_a15_neon_fp_negabs): Likewise
3586 (cortex_a15_neon_fp_arith): Likewise
3587 (cortex_a15_neon_fp_arith_q): Likewise
3588 (cortex_a15_neon_fp_cvt_int): Likewise
3589 (cortex_a15_neon_fp_cvt_int_q): Likewise
3590 (cortex_a15_neon_fp_cvt_16): Likewise
3591 (cortex_a15_neon_fp_mul): Likewise
3592 (cortex_a15_neon_fp_mul_q): Likewise
3593 (cortex_a15_neon_fp_mla): Likewise
3594 (cortex_a15_neon_fp_mla_q): Likewise
3595 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
3596 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
3597 (cortex_a15_neon_bitops): Likewise.
3598 (cortex_a15_neon_bitops_q): Likewise.
3599 (cortex_a15_neon_from_gp): Likewise.
3600 (cortex_a15_neon_from_gp_q): Likewise.
3601 (cortex_a15_neon_tbl3_tbl4): Likewise.
3602 (cortex_a15_neon_zip_q): Likewise.
3603 (cortex_a15_neon_to_gp): Likewise.
3604 (cortex_a15_neon_load_a): Likewise.
3605 (cortex_a15_neon_load_b): Likewise.
3606 (cortex_a15_neon_load_c): Likewise.
3607 (cortex_a15_neon_load_d): Likewise.
3608 (cortex_a15_neon_load_e): Likewise.
3609 (cortex_a15_neon_load_f): Likewise.
3610 (cortex_a15_neon_store_a): Likewise.
3611 (cortex_a15_neon_store_b): Likewise.
3612 (cortex_a15_neon_store_c): Likewise.
3613 (cortex_a15_neon_store_d): Likewise.
3614 (cortex_a15_neon_store_e): Likewise.
3615 (cortex_a15_neon_store_f): Likewise.
3616 (cortex_a15_neon_store_g): Likewise.
3617 (cortex_a15_neon_store_h): Likewise.
3618 (cortex_a15_vfp_to_from_gp): Likewise.
3620 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3622 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
3624 (cortex_a9_neon_vshl_ddd): Remove.
3625 (cortex_a9_neon_vst3_vst4): Likewise.
3626 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3628 (cortex_a9_neon_bit_ops_q): New.
3630 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
3631 (cortex_a9_neon_int_2): Likewise.
3632 (cortex_a9_neon_int_3): Likewise.
3633 (cortex_a9_neon_int_4): Likewise.
3634 (cortex_a9_neon_int_5): Likewise.
3635 (cortex_a9_neon_vqneg_vqabs): Likewise.
3636 (cortex_a9_neon_vmov): Likewise.
3637 (cortex_a9_neon_vaba): Likewise.
3638 (cortex_a9_neon_vaba_qqq): Likewise.
3639 (cortex_a9_neon_shift_1): Likewise.
3640 (cortex_a9_neon_shift_2): Likewise.
3641 (cortex_a9_neon_shift_3): Likewise.
3642 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3643 (cortex_a9_neon_vsra_vrsra): Likewise.
3644 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3645 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3646 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3648 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3649 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3650 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3652 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3653 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3654 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3655 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3656 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3657 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3658 (cortex_a9_neon_fp_vsum): Likewise.
3659 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3660 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3661 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3662 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3663 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3664 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3665 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3666 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3667 (cortex_a9_neon_bp_simple): Likewise.
3668 (cortex_a9_neon_bp_2cycle): Likewise.
3669 (cortex_a9_neon_bp_3cycle): Likewise.
3670 (cortex_a9_neon_ldr): Likewise.
3671 (cortex_a9_neon_str): Likewise.
3672 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3673 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3674 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3675 (cortex_a9_neon_vld2_4_regs): Likewise.
3676 (cortex_a9_neon_vld3_vld4): Likewise.
3677 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3678 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3679 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3680 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3681 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3682 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3683 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3684 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3685 (cortex_a9_neon_mcr): Likewise.
3686 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3687 (cortex_a9_neon_mrc): Likewise.
3688 (cortex_a9_neon_mrrc): Likewise.
3690 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3692 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
3694 (cortex_a8_neon_vshl_ddd): Remove.
3695 (cortex_a8_neon_vst3_vst4): Likewise.
3696 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3698 (cortex_a8_neon_bit_ops_q): New.
3700 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
3701 (cortex_a8_neon_int_2): Likewise..
3702 (cortex_a8_neon_int_3): Likewise.
3703 (cortex_a8_neon_int_5): Likewise.
3704 (cortex_a8_neon_vqneg_vqabs): Likewise.
3705 (cortex_a8_neon_int_4): Likewise.
3706 (cortex_a8_neon_vaba): Likewise.
3707 (cortex_a8_neon_vaba_qqq): Likewise.
3708 (cortex_a8_neon_shift_1): Likewise.
3709 (cortex_a8_neon_shift_2): Likewise.
3710 (cortex_a8_neon_shift_3): Likewise.
3711 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3712 (cortex_a8_neon_vsra_vrsra): Likewise.
3713 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3714 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3715 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3717 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3718 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3719 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3721 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3722 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3723 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3724 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3725 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3726 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3727 (cortex_a8_neon_fp_vsum): Likewise.
3728 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3729 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3730 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3731 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3732 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3733 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3734 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3735 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3736 (cortex_a8_neon_bp_simple): Likewise.
3737 (cortex_a8_neon_bp_2cycle): Likewise.
3738 (cortex_a8_neon_bp_3cycle): Likewise.
3739 (cortex_a8_neon_ldr): Likewise.
3740 (cortex_a8_neon_str): Likewise.
3741 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3742 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3743 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3744 (cortex_a8_neon_vld2_4_regs): Likewise.
3745 (cortex_a8_neon_vld3_vld4): Likewise.
3746 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3747 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3748 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3749 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3750 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3751 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3752 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3753 (cortex_a8_neon_mcr): Likewise.
3754 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3755 (cortex_a8_neon_mrc): Likewise.
3756 (cortex_a8_neon_mrrc): Likewise.
3758 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3760 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
3762 * config/aarch64/aarch64-simd.md (neon_type): Remove.
3763 (aarch64_simd_dup<mode>): Add "type" attribute.
3764 (aarch64_dup_lane<mode>): Likewise.
3765 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
3766 (*aarch64_simd_mov<mode>): Likewise.
3767 (aarch64_simd_mov_from_<mode>low): Likewise.
3768 (aarch64_simd_mov_from_<mode>high): Likewise.
3769 (orn<mode>3): Likewise.
3770 (bic<mode>3): Likewise.
3771 (add<mode>3): Likewise.
3772 (sub<mode>3): Likewise.
3773 (mul<mode>3): Likewise.
3774 (*aarch64_mul3_elt<mode>): Likewise.
3775 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
3776 (*aarch64_mul3_elt_to_128df): Likewise.
3777 (*aarch64_mul3_elt_to_64v2df): Likewise.
3778 (neg<mode>2): Likewise.
3779 (abs<mode>2): Likewise.
3780 (abd<mode>_3): Likewise.
3781 (aba<mode>_3): Likewise.
3782 (fabd<mode>_3): Likewise.
3783 (*fabd_scalar<mode>3): Likewise.
3784 (and<mode>3): Likewise.
3785 (ior<mode>3): Likewise.
3786 (xor<mode>3): Likewise.
3787 (one_cmpl<mode>2): Likewise.
3788 (aarch64_simd_vec_set<mode>): Likewise.
3789 (aarch64_simd_lshr<mode>): Likewise.
3790 (aarch64_simd_ashr<mode>): Likewise.
3791 (aarch64_simd_imm_shl<mode>): Likewise.
3792 (aarch64_simd_reg_sshl<mode): Likewise.
3793 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
3794 (aarch64_simd_reg_shl<mode>_signed): Likewise.
3795 (aarch64_simd_vec_setv2di): Likewise.
3796 (aarch64_simd_vec_set<mode>): Likewise.
3797 (aarch64_mla<mode>): Likewise.
3798 (*aarch64_mla_elt<mode>): Likewise.
3799 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
3800 (aarch64_mls<mode>): Likewise.
3801 (*aarch64_mls_elt<mode>): Likewise.
3802 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
3803 (<su><maxmin><mode>3): Likewise.
3804 (move_lo_quad_<mode>): Likewise.
3805 (aarch64_simd_move_hi_quad_<mode>): Likewise.
3806 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
3807 (vec_pack_trunc_<mode>): Likewise.
3808 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
3809 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
3810 (*aarch64_<su>mlal_lo<mode>): Likewise.
3811 (*aarch64_<su>mlal_hi<mode>): Likewise.
3812 (*aarch64_<su>mlsl_lo<mode>): Likewise.
3813 (*aarch64_<su>mlsl_hi<mode>): Likewise.
3814 (*aarch64_<su>mlal<mode>): Likewise.
3815 (*aarch64_<su>mlsl<mode>): Likewise.
3816 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
3817 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
3818 (add<mode>3): Likewise.
3819 (sub<mode>3): Likewise.
3820 (mul<mode>3): Likewise.
3821 (div<mode>3): Likewise.
3822 (neg<mode>2): Likewise.
3823 (abs<mode>2): Likewise.
3824 (fma<mode>4): Likewise.
3825 (*aarch64_fma4_elt<mode>): Likewise.
3826 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
3827 (*aarch64_fma4_elt_to_128df): Likewise.
3828 (*aarch64_fma4_elt_to_64v2df): Likewise.
3829 (fnma<mode>4): Likewise.
3830 (*aarch64_fnma4_elt<mode>): Likewise.
3831 (*aarch64_fnma4_elt_<vswap_width_name><mode>
3832 (*aarch64_fnma4_elt_to_128df): Likewise.
3833 (*aarch64_fnma4_elt_to_64v2df): Likewise.
3834 (<frint_pattern><mode>2): Likewise.
3835 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
3836 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
3837 (vec_unpacks_lo_v4sf): Likewise.
3838 (aarch64_float_extend_lo_v2df): Likewise.
3839 (vec_unpacks_hi_v4sf): Likewise.
3840 (aarch64_float_truncate_lo_v2sf): Likewise.
3841 (aarch64_float_truncate_hi_v4sf): Likewise.
3842 (aarch64_vmls<mode>): Likewise.
3843 (<su><maxmin><mode>3): Likewise.
3844 (<maxmin_uns><mode>3): Likewise.
3845 (reduc_<sur>plus_<mode>): Likewise.
3846 (reduc_<sur>plus_v2di): Likewise.
3847 (reduc_<sur>plus_v2si): Likewise.
3848 (reduc_<sur>plus_<mode>): Likewise.
3849 (aarch64_addpv4sf): Likewise.
3850 (clz<mode>2): Likewise.
3851 (reduc_<maxmin_uns>_<mode>): Likewise.
3852 (reduc_<maxmin_uns>_v2di): Likewise.
3853 (reduc_<maxmin_uns>_v2si): Likewise.
3854 (reduc_<maxmin_uns>_<mode>): Likewise.
3855 (reduc_<maxmin_uns>_v4sf): Likewise.
3856 (aarch64_simd_bsl<mode>_internal): Likewise.
3857 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
3858 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
3859 (aarch64_get_lane<mode>): Likewise.
3860 (*aarch64_combinez<mode>): Likewise.
3861 (aarch64_combine<mode>): Likewise.
3862 (aarch64_simd_combine<mode>): Likewise.
3863 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
3864 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
3865 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
3866 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
3867 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
3868 (aarch64_<sur>h<addsub><mode>): Likewise.
3869 (aarch64_<sur><addsub>hn<mode>): Likewise.
3870 (aarch64_<sur><addsub>hn2<mode>): Likewise.
3871 (aarch64_pmul<mode>): Likewise.
3872 (aarch64_<su_optab><optab><mode>): Likewise.
3873 (aarch64_<sur>qadd<mode>): Likewise.
3874 (aarch64_sqmovun<mode>): Likewise.
3875 (aarch64_<sur>qmovn<mode>): Likewise.
3876 (aarch64_s<optab><mode>): Likewise.
3877 (aarch64_sq<r>dmulh<mode>): Likewise.
3878 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3879 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
3880 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3881 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
3882 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3883 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3884 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
3885 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
3886 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
3887 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
3888 (aarch64_sqdmull<mode>): Likewise.
3889 (aarch64_sqdmull_lane<mode>_internal): Likewise.
3890 (aarch64_sqdmull_n<mode>): Likewise.
3891 (aarch64_sqdmull2<mode>_internal): Likewise.
3892 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
3893 (aarch64_sqdmull2_n<mode>_internal): Likewise.
3894 (aarch64_<sur>shl<mode>): Likewise.
3895 (aarch64_<sur>q<r>shl<mode>
3896 (aarch64_<sur>shll_n<mode>): Likewise.
3897 (aarch64_<sur>shll2_n<mode>): Likewise.
3898 (aarch64_<sur>shr_n<mode>): Likewise.
3899 (aarch64_<sur>sra_n<mode>): Likewise.
3900 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
3901 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
3902 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
3903 (aarch64_cm<optab><mode>): Likewise.
3904 (aarch64_cm<optab>di): Likewise.
3905 (aarch64_cm<optab><mode>): Likewise.
3906 (aarch64_cm<optab>di): Likewise.
3907 (aarch64_cmtst<mode>): Likewise.
3908 (aarch64_cmtstdi): Likewise.
3909 (aarch64_cm<optab><mode>): Likewise.
3910 (*aarch64_fac<optab><mode>): Likewise.
3911 (aarch64_addp<mode>): Likewise.
3912 (aarch64_addpdi): Likewise.
3913 (sqrt<mode>2): Likewise.
3914 (vec_load_lanesoi<mode>): Likewise.
3915 (vec_store_lanesoi<mode>): Likewise.
3916 (vec_load_lanesci<mode>): Likewise.
3917 (vec_store_lanesci<mode>): Likewise.
3918 (vec_load_lanesxi<mode>): Likewise.
3919 (vec_store_lanesxi<mode>): Likewise.
3920 (*aarch64_mov<mode>): Likewise.
3921 (aarch64_ld2<mode>_dreg): Likewise.
3922 (aarch64_ld2<mode>_dreg): Likewise.
3923 (aarch64_ld3<mode>_dreg): Likewise.
3924 (aarch64_ld3<mode>_dreg): Likewise.
3925 (aarch64_ld4<mode>_dreg): Likewise.
3926 (aarch64_ld4<mode>_dreg): Likewise.
3927 (aarch64_tbl1<mode>): Likewise.
3928 (aarch64_tbl2v16qi): Likewise.
3929 (aarch64_combinev16qi): Likewise.
3930 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
3931 (aarch64_st2<mode>_dreg): Likewise.
3932 (aarch64_st2<mode>_dreg): Likewise.
3933 (aarch64_st3<mode>_dreg): Likewise.
3934 (aarch64_st3<mode>_dreg): Likewise.
3935 (aarch64_st4<mode>_dreg): Likewise.
3936 (aarch64_st4<mode>_dreg): Likewise.
3937 (*aarch64_simd_ld1r<mode>): Likewise.
3938 (aarch64_frecpe<mode>): Likewise.
3939 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
3940 (aarch64_frecps<mode>): Likewise.
3942 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3944 * config/arm/iterators.md (V_elem_ch): New.
3946 (VQH_type): Likewise.
3947 * config/arm/arm.md (is_neon_type): New.
3948 (conds): Use is_neon_type.
3949 (anddi3_insn): Update type attribute.
3950 (xordi3_insn): Likewise.
3951 (one_cmpldi2): Likewise.
3952 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
3953 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
3954 (*movmisalign<mode>_neon_store): Likewise.
3955 (*movmisalign<mode>_neon_load): Likewise.
3956 (vec_set<mode>_internal): Likewise.
3957 (vec_set<mode>_internal): Likewise.
3958 (vec_setv2di_internal): Likewise.
3959 (vec_extract<mode>): Likewise.
3960 (vec_extract<mode>): Likewise.
3961 (vec_extractv2di): Likewise.
3962 (*add<mode>3_neon): Likewise.
3963 (adddi3_neon): Likewise.
3964 (*sub<mode>3_neon): Likewise.
3965 (subdi3_neon): Likewise.
3966 (fma<VCVTF:mode>4): Likewise.
3967 (fma<VCVTF:mode>4_intrinsic): Likewise.
3968 (*fmsub<VCVTF:mode>4): Likewise.
3969 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3970 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3971 (ior<mode>3): Likewise.
3972 (and<mode>3): Likewise.
3973 (orn<mode>3_neon): Likewise.
3974 (orndi3_neon): Likewise.
3975 (bic<mode>3_neon): Likewise.
3976 (bicdi3_neon): Likewise.
3977 (xor<mode>3): Likewise.
3978 (one_cmpl<mode>2): Likewise.
3979 (abs<mode>2): Likewise.
3980 (neg<mode>2): Likewise.
3981 (negdi2_neon): Likewise.
3982 (*umin<mode>3_neon): Likewise.
3983 (*umax<mode>3_neon): Likewise.
3984 (*smin<mode>3_neon): Likewise.
3985 (*smax<mode>3_neon): Likewise.
3986 (vashl<mode>3): Likewise.
3987 (vashr<mode>3_imm): Likewise.
3988 (vlshr<mode>3_imm): Likewise.
3989 (ashl<mode>3_signed): Likewise.
3990 (ashl<mode>3_unsigned): Likewise.
3991 (neon_load_count): Likewise.
3992 (ashldi3_neon_noclobber): Likewise.
3993 (ashldi3_neon): Likewise.
3994 (signed_shift_di3_neon): Likewise.
3995 (unsigned_shift_di3_neon): Likewise.
3996 (ashrdi3_neon_imm_noclobber): Likewise.
3997 (lshrdi3_neon_imm_noclobber): Likewise.
3998 (<shift>di3_neon): Likewise.
3999 (widen_ssum<mode>3): Likewise.
4000 (widen_usum<mode>3): Likewise.
4001 (quad_halves_<code>v4si): Likewise.
4002 (quad_halves_<code>v4sf): Likewise.
4003 (quad_halves_<code>v8hi): Likewise.
4004 (quad_halves_<code>v16qi): Likewise.
4005 (reduc_splus_v2di): Likewise.
4006 (neon_vpadd_internal<mode>): Likewise.
4007 (neon_vpsmin<mode>): Likewise.
4008 (neon_vpsmax<mode>): Likewise.
4009 (neon_vpumin<mode>): Likewise.
4010 (neon_vpumax<mode>): Likewise.
4011 (*ss_add<mode>_neon): Likewise.
4012 (*us_add<mode>_neon): Likewise.
4013 (*ss_sub<mode>_neon): Likewise.
4014 (*us_sub<mode>_neon): Likewise.
4015 (neon_vadd<mode>_unspec): Likewise.
4016 (neon_vaddl<mode>): Likewise.
4017 (neon_vaddw<mode>): Likewise.
4018 (neon_vhadd<mode>): Likewise.
4019 (neon_vqadd<mode>): Likewise.
4020 (neon_vaddhn<mode>): Likewise.
4021 (neon_vmul<mode>): Likewise.
4022 (neon_vfms<VCVTF:mode>): Likewise.
4023 (neon_vmlal<mode>): Likewise.
4024 (neon_vmls<mode>): Likewise.
4025 (neon_vmlsl<mode>): Likewise.
4026 (neon_vqdmulh<mode>): Likewise.
4027 (neon_vqdmlal<mode>): Likewise.
4028 (neon_vqdmlsl<mode>): Likewise.
4029 (neon_vmull<mode>): Likewise.
4030 (neon_vqdmull<mode>): Likewise.
4031 (neon_vsub<mode>_unspec): Likewise.
4032 (neon_vsubl<mode>): Likewise.
4033 (neon_vsubw<mode>): Likewise.
4034 (neon_vqsub<mode>): Likewise.
4035 (neon_vhsub<mode>): Likewise.
4036 (neon_vsubhn<mode>): Likewise.
4037 (neon_vceq<mode>): Likewise.
4038 (neon_vcge<mode>): Likewise.
4039 (neon_vcgeu<mode>): Likewise.
4040 (neon_vcgt<mode>): Likewise.
4041 (neon_vcgtu<mode>): Likewise.
4042 (neon_vcle<mode>): Likewise.
4043 (neon_vclt<mode>): Likewise.
4044 (neon_vcage<mode>): Likewise.
4045 (neon_vcagt<mode>): Likewise.
4046 (neon_vtst<mode>): Likewise.
4047 (neon_vabd<mode>): Likewise.
4048 (neon_vabdl<mode>): Likewise.
4049 (neon_vaba<mode>): Likewise.
4050 (neon_vabal<mode>): Likewise.
4051 (neon_vmax<mode>): Likewise.
4052 (neon_vmin<mode>): Likewise.
4053 (neon_vpaddl<mode>): Likewise.
4054 (neon_vpadal<mode>): Likewise.
4055 (neon_vpmax<mode>): Likewise.
4056 (neon_vpmin<mode>): Likewise.
4057 (neon_vrecps<mode>): Likewise.
4058 (neon_vrsqrts<mode>): Likewise.
4059 (neon_vqabs<mode>): Likewise.
4060 (neon_vqneg<mode>): Likewise.
4061 (neon_vcls<mode>): Likewise.
4062 (clz<mode>2): Likewise.
4063 (popcount<mode>2): Likewise.
4064 (neon_vrecpe<mode>): Likewise.
4065 (neon_vrsqrte<mode>): Likewise.
4066 (neon_vget_lane<mode>_sext_internal): Likewise.
4067 (neon_vget_lane<mode>_zext_internal): Likewise.
4068 (neon_vdup_n<mode>): Likewise.
4069 (neon_vdup_n<mode>): Likewise.
4070 (neon_vdup_nv2di): Likewise.
4071 (neon_vdup_lane<mode>_interal): Likewise.
4072 (*neon_vswp<mode>): Likewise.
4073 (neon_vcombine<mode>): Likewise.
4074 (float<mode><V_cvtto>2): Likewise.
4075 (floatuns<mode><V_cvtto>2): Likewise.
4076 (fix_trunc<mode><V_cvtto>2): Likewise.
4077 (fixuns_trunc<mode><V_cvtto>2
4078 (neon_vcvt<mode>): Likewise.
4079 (neon_vcvt<mode>): Likewise.
4080 (neon_vcvtv4sfv4hf): Likewise.
4081 (neon_vcvtv4hfv4sf): Likewise.
4082 (neon_vcvt_n<mode>): Likewise.
4083 (neon_vcvt_n<mode>): Likewise.
4084 (neon_vmovn<mode>): Likewise.
4085 (neon_vqmovn<mode>): Likewise.
4086 (neon_vqmovun<mode>): Likewise.
4087 (neon_vmovl<mode>): Likewise.
4088 (neon_vmul_lane<mode>): Likewise.
4089 (neon_vmul_lane<mode>): Likewise.
4090 (neon_vmull_lane<mode>): Likewise.
4091 (neon_vqdmull_lane<mode>): Likewise.
4092 (neon_vqdmulh_lane<mode>): Likewise.
4093 (neon_vqdmulh_lane<mode>): Likewise.
4094 (neon_vmla_lane<mode>): Likewise.
4095 (neon_vmla_lane<mode>): Likewise.
4096 (neon_vmlal_lane<mode>): Likewise.
4097 (neon_vqdmlal_lane<mode>): Likewise.
4098 (neon_vmls_lane<mode>): Likewise.
4099 (neon_vmls_lane<mode>): Likewise.
4100 (neon_vmlsl_lane<mode>): Likewise.
4101 (neon_vqdmlsl_lane<mode>): Likewise.
4102 (neon_vext<mode>): Likewise.
4103 (neon_vrev64<mode>): Likewise.
4104 (neon_vrev32<mode>): Likewise.
4105 (neon_vrev16<mode>): Likewise.
4106 (neon_vbsl<mode>_internal): Likewise.
4107 (neon_vshl<mode>): Likewise.
4108 (neon_vqshl<mode>): Likewise.
4109 (neon_vshr_n<mode>): Likewise.
4110 (neon_vshrn_n<mode>): Likewise.
4111 (neon_vqshrn_n<mode>): Likewise.
4112 (neon_vqshrun_n<mode>): Likewise.
4113 (neon_vshl_n<mode>): Likewise.
4114 (neon_vqshl_n<mode>): Likewise.
4115 (neon_vqshlu_n<mode>): Likewise.
4116 (neon_vshll_n<mode>): Likewise.
4117 (neon_vsra_n<mode>): Likewise.
4118 (neon_vsri_n<mode>): Likewise.
4119 (neon_vsli_n<mode>): Likewise.
4120 (neon_vtbl1v8qi): Likewise.
4121 (neon_vtbl2v8qi): Likewise.
4122 (neon_vtbl3v8qi): Likewise.
4123 (neon_vtbl4v8qi): Likewise.
4124 (neon_vtbl1v16qi): Likewise.
4125 (neon_vtbl2v16qi): Likewise.
4126 (neon_vcombinev16qi): Likewise.
4127 (neon_vtbx1v8qi): Likewise.
4128 (neon_vtbx2v8qi): Likewise.
4129 (neon_vtbx3v8qi): Likewise.
4130 (neon_vtbx4v8qi): Likewise.
4131 (*neon_vtrn<mode>_insn): Likewise.
4132 (*neon_vzip<mode>_insn): Likewise.
4133 (*neon_vuzp<mode>_insn): Likewise.
4134 (neon_vld1<mode>): Likewise.
4135 (neon_vld1_lane<mode>): Likewise.
4136 (neon_vld1_lane<mode>): Likewise.
4137 (neon_vld1_dup<mode>): Likewise.
4138 (neon_vld1_dup<mode>): Likewise.
4139 (neon_vld1_dupv2di): Likewise.
4140 (neon_vst1<mode>): Likewise.
4141 (neon_vst1_lane<mode>): Likewise.
4142 (neon_vst1_lane<mode>): Likewise.
4143 (neon_vld2<mode>): Likewise.
4144 (neon_vld2<mode>): Likewise.
4145 (neon_vld2_lane<mode>): Likewise.
4146 (neon_vld2_lane<mode>): Likewise.
4147 (neon_vld2_dup<mode>): Likewise.
4148 (neon_vst2<mode>): Likewise.
4149 (neon_vst2<mode>): Likewise.
4150 (neon_vst2_lane<mode>): Likewise.
4151 (neon_vst2_lane<mode>): Likewise.
4152 (neon_vld3<mode>): Likewise.
4153 (neon_vld3qa<mode>): Likewise.
4154 (neon_vld3qb<mode>): Likewise.
4155 (neon_vld3_lane<mode>): Likewise.
4156 (neon_vld3_lane<mode>): Likewise.
4157 (neon_vld3_dup<mode>): Likewise.
4158 (neon_vst3<mode>): Likewise.
4159 (neon_vst3qa<mode>): Likewise.
4160 (neon_vst3qb<mode>): Likewise.
4161 (neon_vst3_lane<mode>): Likewise.
4162 (neon_vst3_lane<mode>): Likewise.
4163 (neon_vld4<mode>): Likewise.
4164 (neon_vld4qa<mode>): Likewise.
4165 (neon_vld4qb<mode>): Likewise.
4166 (neon_vld4_lane<mode>): Likewise.
4167 (neon_vld4_lane<mode>): Likewise.
4168 (neon_vld4_dup<mode>): Likewise.
4169 (neon_vst4<mode>): Likewise.
4170 (neon_vst4qa<mode>): Likewise.
4171 (neon_vst4qb<mode>): Likewise.
4172 (neon_vst4_lane<mode>): Likewise.
4173 (neon_vst4_lane<mode>): Likewise.
4174 (neon_vec_unpack<US>_lo_<mode>): Likewise.
4175 (neon_vec_unpack<US>_hi_<mode>): Likewise.
4176 (neon_vec_<US>mult_lo_<mode>): Likewise.
4177 (neon_vec_<US>mult_hi_<mode>): Likewise.
4178 (neon_vec_<US>shiftl_<mode>): Likewise.
4179 (neon_unpack<US>_<mode>): Likewise.
4180 (neon_vec_<US>mult_<mode>): Likewise.
4181 (vec_pack_trunc_<mode>): Likewise.
4182 (neon_vec_pack_trunc_<mode>): Likewise.
4183 (neon_vabd<mode>_2): Likewise.
4184 (neon_vabd<mode>_3): Likewise.
4186 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4188 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
4189 (load_pair): Update type attribute.
4190 (store_pair): Update type attribute.
4191 * config/aarch64/iterators.md (q): New.
4193 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4195 * config/arm/types.md: Add new types for Neon insns.
4197 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4198 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4199 Sergey Lega <sergey.s.lega@intel.com>
4200 Anna Tikhonova <anna.tikhonova@intel.com>
4201 Ilya Tocar <ilya.tocar@intel.com>
4202 Andrey Turetskiy <andrey.turetskiy@intel.com>
4203 Ilya Verbin <ilya.verbin@intel.com>
4204 Kirill Yukhin <kirill.yukhin@intel.com>
4205 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4207 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
4208 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
4209 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
4211 (srcp14<mode>): Ditto.
4212 (rsqrt14<mode>): Ditto.
4213 (rsqrt14<mode>): Ditto.
4214 (avx512f_vmscalef<mode>): Ditto.
4215 (avx512f_scalef<mode>): Ditto.
4216 (avx512f_getexp<mode>): Ditto.
4217 (avx512f_sgetexp<mode>): Ditto.
4218 (avx512f_fixupimm<mode>): Ditto.
4219 (avx512f_sfixupimm<mode>): Ditto.
4220 (avx512f_rndscale<mode>): Ditto.
4221 (*avx512er_exp2<mode>): Ditto.
4222 (*avx512er_rcp28<mode>): Ditto.
4223 (avx512er_rsqrt28<mode>): Ditto.
4224 (avx512f_getmant<mode>): Ditto.
4225 (avx512f_getmant<mode>): Ditto.
4226 (avx512f_rndscale<mode>): Fix formatting.
4228 2013-10-15 Martin Jambor <mjambor@suse.cz>
4230 * ipa-utils.h (ipa_edge_within_scc): Declare.
4231 * ipa-cp.c (edge_within_scc): Moved...
4232 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
4234 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4235 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4236 Sergey Lega <sergey.s.lega@intel.com>
4237 Anna Tikhonova <anna.tikhonova@intel.com>
4238 Ilya Tocar <ilya.tocar@intel.com>
4239 Andrey Turetskiy <andrey.turetskiy@intel.com>
4240 Ilya Verbin <ilya.verbin@intel.com>
4241 Kirill Yukhin <kirill.yukhin@intel.com>
4242 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4244 * config/i386/predicates.md (const_8_to_15_operand): New.
4245 (const_16_to_31_operand): Ditto.
4246 * config/i386/sse.md (V8FI): New.
4248 (reduc_splus_v8df): Ditto.
4249 (reduc_splus_v16sf): Ditto.
4250 (avx512f_vextract<shuffletype>32x4_1): Ditto.
4251 (vec_extract_hi_<mode>): Ditto.
4252 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
4253 (vec_set_lo_<mode>): Ditto.
4254 (vec_set_hi_<mode>): Ditto.
4255 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
4256 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
4257 (avx512f_pshufd_1): Ditto.
4258 (avx512f_broadcast<mode>): Ditto.
4259 (avx512f_broadcast<mode>): Ditto.
4260 (define_split): Split vec_extract_lo into move.
4261 (ssequartermode): Ditto.
4262 (ssedoublemode): Extened with wider modes.
4263 (vec_extract_lo_<mode>): Ditto.
4265 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4266 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4267 Sergey Lega <sergey.s.lega@intel.com>
4268 Anna Tikhonova <anna.tikhonova@intel.com>
4269 Ilya Tocar <ilya.tocar@intel.com>
4270 Andrey Turetskiy <andrey.turetskiy@intel.com>
4271 Ilya Verbin <ilya.verbin@intel.com>
4272 Kirill Yukhin <kirill.yukhin@intel.com>
4273 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4275 * config/i386/predicates.md (register_or_constm1_operand): New.
4276 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
4277 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
4278 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
4279 UNSPEC_SCATTER_PREFETCH
4281 (avx512f_ucmp<mode>3): Ditto.
4282 (avx512f_vternlog<mode>): Ditto.
4283 (avx512f_align<mode>): Ditto.
4284 (<shift_insn><mode>3): Ditto.
4285 (avx512f_<rotate>v<mode>): Ditto.
4286 (avx512f_<rotate><mode>): Ditto.
4287 (avx512f_eq<mode>3): Ditto.
4288 (avx512f_eq<mode>3_1): Ditto.
4289 (avx512f_gt<mode>3): Ditto.
4290 (avx512f_testm<mode>3): Ditto.
4291 (avx512f_testnm<mode>3): Ditto.
4292 (avx512pf_gatherpf<mode>): Ditto.
4293 (*avx512pf_gatherpf<mode>_mask): Ditto.
4294 (*avx512pf_gatherpf<mode>): Ditto.
4295 (avx512pf_scatterpf<mode>): Ditto.
4296 (*avx512pf_scatterpf<mode>_mask): Ditto.
4297 (*avx512pf_scatterpf<mode>): Ditto.
4298 (avx512f_vec_dup_gpr<mode>): Ditto.
4299 (clz<mode>2): Ditto.
4300 (conflict<mode>): Ditto.
4301 (REDUC_SMINMAX_MODE): Extened with wider modes.
4302 (reduc_<code>_<mode>): Ditto.
4303 (vlshr<mode>3): Ditto.
4304 (vashl<mode>3): Ditto.
4306 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4307 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4308 Sergey Lega <sergey.s.lega@intel.com>
4309 Anna Tikhonova <anna.tikhonova@intel.com>
4310 Ilya Tocar <ilya.tocar@intel.com>
4311 Andrey Turetskiy <andrey.turetskiy@intel.com>
4312 Ilya Verbin <ilya.verbin@intel.com>
4313 Kirill Yukhin <kirill.yukhin@intel.com>
4314 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4316 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
4319 (avx512fmaskmode): Ditto.
4320 (bcstscalarsuff): Ditto.
4321 (avx512f_blendm<mode>): Ditto.
4322 (cmp_imm_predicate): Ditto.
4323 (avx512f_cmp<mode>3): Ditto.
4324 (avx512f_vec_dup<mode>): Ditto.
4325 (avx512f_vec_dup_mem<mode>): Ditto.
4326 (avx512f_vpermi2var<mode>3): Ditto.
4327 (avx512f_vpermt2var<mode>3): Ditto.
4328 (vec_init<mode>): Ditto.
4329 (avx512f_gathersi<mode>): Ditto.
4330 (*avx512f_gathersi<mode>): Ditto.
4331 (*avx512f_gathersi<mode>_2): Ditto.
4332 (avx512f_gatherdi<mode>): Ditto.
4333 (*avx512f_gatherdi<mode>): Ditto.
4334 (*avx512f_gatherdi<mode>_2): Ditto.
4335 (avx512f_scattersi<mode>): Ditto.
4336 (*avx512f_scattersi<mode>): Ditto.
4337 (avx512f_scatterdi<mode>): Ditto.
4338 (*avx512f_scatterdi<mode>): Ditto.
4339 (sseintprefix): Extened with wider modes.
4340 (VEC_GATHER_IDXSI): Ditto.
4341 (VEC_GATHER_IDXDI): Ditto.
4342 (VEC_GATHER_SRCDI): Ditto.
4344 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
4345 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4347 * config/arm/t-aprofile: New file.
4348 * config.gcc: Handle --with-multilib-list option.
4350 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
4352 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
4355 2013-10-15 Richard Biener <rguenther@suse.de>
4357 * tree-loop-distribution.c (build_empty_rdg): Inline into
4359 (rdg_flag_vertex): Inline into single user.
4360 (rdg_flag_vertex_and_dependent): Likewise.
4361 (build_rdg_partition_for_vertex): Remove processed bitmap.
4362 (rdg_build_partitions): Simplify.
4364 2013-10-15 Richard Biener <rguenther@suse.de>
4366 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
4367 Restructure forwarding through conversions and copies to
4368 avoid performing copy-propagation the wrong way. Adjust
4369 recursion invocations.
4370 (forward_propagate_addr_expr): Add argument stating if we
4371 are recursing from a single-use.
4372 (ssa_forward_propagate_and_combine): Adjust.
4374 2013-10-14 David Malcolm <dmalcolm@redhat.com>
4376 * dumpfile.h (gcc::dump_manager): New class, to hold state
4377 relating to dumpfile management.
4378 (get_dump_file_name): Remove in favor of method of dump_manager.
4379 (dump_initialized_p): Likewise.
4380 (dump_start): Likewise.
4381 (dump_finish): Likewise.
4382 (dump_switch_p): Likewise.
4383 (dump_register): Likewise.
4384 (get_dump_file_info): Likewise.
4385 * context.c (gcc::context::context): Construct the dump_manager
4387 * context.h (gcc::context::get_dumps): New.
4388 (gcc::context::m_dumps): New.
4389 * coverage.c (coverage_init): Port to dump_manager API.
4390 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
4391 (extra_dump_files_in_use): Likewise.
4392 (extra_dump_files_alloced): Likewise.
4393 (gcc::dump_manager::dump_manager): New.
4394 (dump_register): Convert to...
4395 (gcc::dump_manager::dump_register): ...method, replacing
4396 function-static next_dump with m_next_dump field.
4397 (get_dump_file_info): Convert to...
4398 (gcc::dump_manager::get_dump_file_info): ...method.
4399 (get_dump_file_name): Convert to...
4400 (gcc::dump_manager::get_dump_file_name): ...method.
4401 (dump_start): Convert to...
4402 (gcc::dump_manager::dump_start): ...method.
4403 (dump_finish): Convert to...
4404 (gcc::dump_manager::dump_finish): ...method.
4405 (dump_begin): Replace body with...
4406 (gcc::dump_manager::dump_begin): ...new method.
4407 (dump_phase_enabled_p): Convert to...
4408 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4409 (dump_phase_enabled_p): Convert to...
4410 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4411 (dump_initialized_p): Convert to...
4412 (gcc::dump_manager::dump_initialized_p): ...method.
4413 (dump_flag_name): Replace body with...
4414 (gcc::dump_manager::dump_flag_name): ...new method.
4415 (dump_enable_all): Convert to...
4416 (gcc::dump_manager::dump_enable_all): ...new method.
4417 (opt_info_enable_passes): Convert to...
4418 (gcc::dump_manager::opt_info_enable_passes): ...new method.
4419 (dump_switch_p_1): Convert to...
4420 (gcc::dump_manager::dump_switch_p_1): ...new method.
4421 (dump_switch_p): Convert to...
4422 (gcc::dump_manager::dump_switch_p): ...new method.
4423 (opt_info_switch_p): Port to dump_manager API.
4424 (enable_rtl_dump_file): Likewise.
4425 * opts-global.c (handle_common_deferred_options): Port to new
4427 * passes.c (pass_manager::finish_optimization_passes): Likewise.
4428 (pass_manager::register_one_dump_file): Likewise.
4429 (pass_manager::register_pass): Likewise.
4430 (pass_init_dump_file): Likewise.
4431 (pass_fini_dump_file): Likewise.
4432 * statistics.c (statistics_early_init): Likewise.
4434 2013-10-14 Richard Biener <rguenther@suse.de>
4436 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
4437 iterative_hash_canonical_type, gimple_canonical_type_hash,
4438 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
4439 gimple_register_canonical_type, print_gimple_types_stats,
4440 free_gimple_type_tables): Move to lto/lto.c
4441 (gt-gimple.h): Do not include.
4442 * gimple.h (gimple_register_canonical_type,
4443 print_gimple_types_stats, free_gimple_type_tables): Remove.
4444 * Makefile.in (GTFILES): Remove gimple.c.
4446 2013-10-14 Travis Snoozy <quandary@remstate.com>
4449 * config/msp430/msp430.c (msp430_option_override): Correct thinko
4450 scanning for msp430x targets.
4452 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
4455 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
4456 (registers_ok_for_ldd_peep): Move around.
4457 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
4458 * config/sparc/sparc.md (widening peepholes): Use it.
4460 2013-10-14 Richard Biener <rguenther@suse.de>
4464 * gimple.c (iterative_hash_canonical_type): Make sure to
4465 record the hash into the correct hashtable slot.
4467 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
4469 PR rtl-optimization/58662
4470 * combine.c (try_combine): Take into account death nodes on I2 when
4471 splitting a PARALLEL of two independent SETs. Fix dump message.
4473 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
4476 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
4477 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
4478 * config/sh/t-sh (sh_treg_combine.o): New entry.
4479 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
4480 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
4481 (register_sh_passes): New function. Register sh_treg_combine pass.
4482 (sh_option_override): Invoke it.
4483 (sh_canonicalize_comparison): Handle op0_preserve_value.
4484 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
4485 opportunities. Canonicalize branch condition.
4486 (nott): Allow only if pseudos can be created for non-SH2A.
4488 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
4491 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
4492 (ix86_expand_movmem): Replace copy_addr_to_reg with
4493 ix86_copy_addr_to_reg.
4494 (ix86_expand_setmem): Likewise.
4496 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
4498 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
4499 provided by ix86_fp_compare_mode instead of CCFPUmode.
4501 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
4503 * config/aarch64/arm_neon.h
4504 (vtbx<1,3>_<psu>8): Fix register constriants.
4506 2013-10-11 Jeff Law <law@redhat.com>
4508 PR tree-optimization/58640
4509 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
4510 threading paths that cross over two loop entry points.
4512 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4514 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
4515 handle vector float as well.
4516 (*vsx_le_perm_load_v4si): Likewise.
4517 (*vsx_le_perm_store_v2di): Likewise.
4518 (*vsx_le_perm_store_v4si): Likewise.
4520 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4522 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
4523 directly to circumvent subtract from splat{31} workaround.
4524 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
4526 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
4527 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
4528 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
4529 (altivec_vperm_<mode>): Convert to define_insn_and_split to
4530 separate big and little endian logic.
4531 (*altivec_vperm_<mode>_internal): New define_insn.
4532 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
4533 separate big and little endian logic.
4534 (*altivec_vperm_<mode>_uns_internal): New define_insn.
4535 (vec_permv16qi): Add little endian logic.
4537 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
4539 * doc/extend.texi (returns_nonnull): Remove arguments.
4541 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4542 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4543 Sergey Lega <sergey.s.lega@intel.com>
4544 Anna Tikhonova <anna.tikhonova@intel.com>
4545 Ilya Tocar <ilya.tocar@intel.com>
4546 Andrey Turetskiy <andrey.turetskiy@intel.com>
4547 Ilya Verbin <ilya.verbin@intel.com>
4548 Kirill Yukhin <kirill.yukhin@intel.com>
4549 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4551 * config/i386/sse.md (VI48F_256_512): New.
4552 (avx2_permvar<mode>): Change to ...
4553 (<avx2_avx512f>_permvar<mode>): This.
4555 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4556 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4557 Sergey Lega <sergey.s.lega@intel.com>
4558 Anna Tikhonova <anna.tikhonova@intel.com>
4559 Ilya Tocar <ilya.tocar@intel.com>
4560 Andrey Turetskiy <andrey.turetskiy@intel.com>
4561 Ilya Verbin <ilya.verbin@intel.com>
4562 Kirill Yukhin <kirill.yukhin@intel.com>
4563 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4565 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
4566 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
4567 * config/i386/sse.md (VI4_AVX): New.
4568 (sf2simodelower): Ditto.
4569 (sse2_cvtps2dq): Change to ...
4570 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
4572 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4573 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4574 Sergey Lega <sergey.s.lega@intel.com>
4575 Anna Tikhonova <anna.tikhonova@intel.com>
4576 Ilya Tocar <ilya.tocar@intel.com>
4577 Andrey Turetskiy <andrey.turetskiy@intel.com>
4578 Ilya Verbin <ilya.verbin@intel.com>
4579 Kirill Yukhin <kirill.yukhin@intel.com>
4580 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4582 * config/i386/sse.md (V_512): New.
4584 (vcond<V_512:mode><VF_512:mode>): Ditto.
4585 (vcond<V_512:mode><VI_512:mode>): Ditto.
4586 (vcondu<V_512:mode><VI_512:mode>): Ditto.
4588 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4589 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4590 Sergey Lega <sergey.s.lega@intel.com>
4591 Anna Tikhonova <anna.tikhonova@intel.com>
4592 Ilya Tocar <ilya.tocar@intel.com>
4593 Andrey Turetskiy <andrey.turetskiy@intel.com>
4594 Ilya Verbin <ilya.verbin@intel.com>
4595 Kirill Yukhin <kirill.yukhin@intel.com>
4596 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4598 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
4599 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
4601 (fma<mode>4): Removed condition.
4602 (fms<mode>4): Ditto.
4603 (fnma<mode>4): Ditto.
4604 (fnms<mode>4): Ditto.
4605 (fma4i_fmadd_<mode>): Ditto.
4606 (*fma_fmadd_<mode>): Ditto.
4607 (*fma_fmsub_<mode>): Ditto.
4608 (*fma_fnmadd_<mode>): Ditto.
4609 (*fma_fnmsub_<mode>): Ditto.
4610 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
4611 (*fma_fmaddsub_<mode>): Ditto.
4612 (*fma_fmsubadd_<mode>): Ditto.
4613 (*fmai_fmadd_<mode>): Ditto.
4614 (*fmai_fmsub_<mode>): Ditto.
4615 (*fmai_fnmadd_<mode>): Ditto.
4616 (*fmai_fnmsub_<mode>): Ditto.
4618 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4619 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4620 Sergey Lega <sergey.s.lega@intel.com>
4621 Anna Tikhonova <anna.tikhonova@intel.com>
4622 Ilya Tocar <ilya.tocar@intel.com>
4623 Andrey Turetskiy <andrey.turetskiy@intel.com>
4624 Ilya Verbin <ilya.verbin@intel.com>
4625 Kirill Yukhin <kirill.yukhin@intel.com>
4626 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4628 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
4629 (VI124_256): Changed to ...
4630 (VI124_256_48_512): This.
4631 (ssepackmode): Extended with wider modes.
4632 (<code><mode>3): Changed iterator.
4633 (*avx2_<code><mode>3): Ditto.
4634 (vec_pack_trunc_<mode>): Ditto.
4636 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4637 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4638 Sergey Lega <sergey.s.lega@intel.com>
4639 Anna Tikhonova <anna.tikhonova@intel.com>
4640 Ilya Tocar <ilya.tocar@intel.com>
4641 Andrey Turetskiy <andrey.turetskiy@intel.com>
4642 Ilya Verbin <ilya.verbin@intel.com>
4643 Kirill Yukhin <kirill.yukhin@intel.com>
4644 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4646 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
4647 (VI8F_256_512): Ditto.
4648 (abs<mode>2): Changed iterator.
4649 (avx2_perm<mode>): Changed to ...
4650 (<avx2_avx512f>_perm<mode>): This.
4651 (avx2_perm<mode>_1): Changed to ...
4652 (<avx2_avx512f>_perm<mode>_1): This.
4654 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4655 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4656 Sergey Lega <sergey.s.lega@intel.com>
4657 Anna Tikhonova <anna.tikhonova@intel.com>
4658 Ilya Tocar <ilya.tocar@intel.com>
4659 Andrey Turetskiy <andrey.turetskiy@intel.com>
4660 Ilya Verbin <ilya.verbin@intel.com>
4661 Kirill Yukhin <kirill.yukhin@intel.com>
4662 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4664 * config/i386/sse.md (VI48_AVX512F): New.
4665 (VI48_AVX2): Changed to ...
4666 (VI48_AVX2_48_AVX512F): This.
4667 (avx2_ashrv<mode>): Changed to ...
4668 (<avx2_avx512f>_ashrv<mode>): This.
4669 (avx2_<shift_insn>v<mode>): Changed to ...
4670 (<avx2_avx512f>_<shift_insn>v<mode>): This.
4672 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4673 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4674 Sergey Lega <sergey.s.lega@intel.com>
4675 Anna Tikhonova <anna.tikhonova@intel.com>
4676 Ilya Tocar <ilya.tocar@intel.com>
4677 Andrey Turetskiy <andrey.turetskiy@intel.com>
4678 Ilya Verbin <ilya.verbin@intel.com>
4679 Kirill Yukhin <kirill.yukhin@intel.com>
4680 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4682 * config/i386/sse.md (VI4_AVX512F): New.
4683 (VI8_AVX2_AVX512F): Ditto.
4684 (mul<mode>3): Extended with wider modes.
4685 (*<sse4_1_avx2>_mul<mode>3): Ditto.
4686 (mul<mode>3): Ditto.
4687 (vec_widen_<s>mult_odd_<mode>): Ditto.
4689 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4690 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4691 Sergey Lega <sergey.s.lega@intel.com>
4692 Anna Tikhonova <anna.tikhonova@intel.com>
4693 Ilya Tocar <ilya.tocar@intel.com>
4694 Andrey Turetskiy <andrey.turetskiy@intel.com>
4695 Ilya Verbin <ilya.verbin@intel.com>
4696 Kirill Yukhin <kirill.yukhin@intel.com>
4697 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4699 * config/i386/sse.md (VI2_AVX512F): New.
4700 (VI124_AVX512F): Ditto.
4701 (sseunpackmode): Extended with wider modes.
4702 (sseunpackfltmode): Ditto.
4703 (vec_unpacks_float_hi_<mode>): Ditto.
4704 (vec_unpacks_float_lo_<mode>): Ditto.
4705 (vec_unpacku_float_hi_<mode>): Ditto.
4706 (vec_unpacku_float_lo_<mode>): Ditto.
4707 (vec_unpacks_lo_<mode>): Ditto.
4708 (vec_unpacks_hi_<mode>): Ditto.
4709 (vec_unpacku_lo_<mode>): Ditto.
4710 (vec_unpacku_hi_<mode>): Ditto.
4712 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4713 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4714 Sergey Lega <sergey.s.lega@intel.com>
4715 Anna Tikhonova <anna.tikhonova@intel.com>
4716 Ilya Tocar <ilya.tocar@intel.com>
4717 Andrey Turetskiy <andrey.turetskiy@intel.com>
4718 Ilya Verbin <ilya.verbin@intel.com>
4719 Kirill Yukhin <kirill.yukhin@intel.com>
4720 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4722 * config/i386/i386.md (multdiv): New.
4723 (multdiv_mnemonic): Ditto.
4724 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
4725 (<sse>_vm<multdiv_mnemonic><mode>3): This.
4726 (<sse>_vmdiv<mode>3): Removed.
4728 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4729 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4730 Sergey Lega <sergey.s.lega@intel.com>
4731 Anna Tikhonova <anna.tikhonova@intel.com>
4732 Ilya Tocar <ilya.tocar@intel.com>
4733 Andrey Turetskiy <andrey.turetskiy@intel.com>
4734 Ilya Verbin <ilya.verbin@intel.com>
4735 Kirill Yukhin <kirill.yukhin@intel.com>
4736 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4738 * config/i386/sse.md (V): Extended with wider modes.
4740 (ssehalfvecmode): Ditto.
4742 (ssepackfltmode): Ditto.
4743 (avx_vec_concat<mode>): Ditto.
4744 (V_256_512): New iterator.
4745 (VF2_512_256): Ditto.
4746 (si2dfmode): New attribute.
4747 (si2dfmodelower): Ditto.
4749 (concat_tg_mode): Ditto.
4750 (floatv4siv4df2): Changed to ...
4751 (float<si2dfmodelower><mode>2): This.
4752 (avx_cvtps2pd256): Changed to ...
4753 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
4754 (vec_pack_trunc_v4df): Changed to ...
4755 (vec_pack_trunc_<mode>): This.
4756 (avx_vpermil<mode>): Changed to ...
4757 (<sse2_avx_avx512f>_vpermil<mode>): This.
4758 (<fixsuffix>fix_truncv8dfv8si2): New.
4759 (vec_pack_sfix_trunc_v8df): Ditto.
4760 (avx512f_rndscale<mode>): Ditto.
4761 (avx512f_roundpd512): Ditto.
4762 (vec_pack_ufix_trunc_<mode>): Updated iterator.
4764 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4765 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4766 Sergey Lega <sergey.s.lega@intel.com>
4767 Anna Tikhonova <anna.tikhonova@intel.com>
4768 Ilya Tocar <ilya.tocar@intel.com>
4769 Andrey Turetskiy <andrey.turetskiy@intel.com>
4770 Ilya Verbin <ilya.verbin@intel.com>
4771 Kirill Yukhin <kirill.yukhin@intel.com>
4772 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4774 * config/i386/i386.md (any_fix): New iterator.
4775 (fixsuffix): New attribute.
4776 * config/i386/sse.md (VF1): Extened with wider modes.
4780 (sseintvecmodelower): Ditto.
4781 (ssescalarmode): Ditto.
4782 (ssescalarnum): Ditto.
4784 (ssexmmmode): Ditto.
4785 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
4786 (<sse>_rcp<mode>2): Change iterator.
4787 (rsqrt<mode>2): Ditto.
4788 (<sse>_rsqrt<mode>2): Ditto.
4789 (avx2_vec_dup<mode>): Ditto.
4790 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
4791 (round<mode>2_sfix): Ditto.
4792 (avx2_pbroadcast<mode>): Ditto.
4793 (*andnot<mode>3): Handle XI mode.
4794 (*<code><mode>3): Ditto.
4795 (AVXTOSSEMODE): Removed.
4796 (avx_vpermil<mode>): Changed to ...
4797 (<sse2_avx_avx512f>_vpermil<mode>): This.
4799 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4800 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4801 Sergey Lega <sergey.s.lega@intel.com>
4802 Anna Tikhonova <anna.tikhonova@intel.com>
4803 Ilya Tocar <ilya.tocar@intel.com>
4804 Andrey Turetskiy <andrey.turetskiy@intel.com>
4805 Ilya Verbin <ilya.verbin@intel.com>
4806 Kirill Yukhin <kirill.yukhin@intel.com>
4807 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4809 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
4810 (<sse>_comi): Ditto.
4811 (<sse>_ucomi): Ditto.
4812 (sse_cvtss2siq_2): Ditto.
4813 (sse2_cvtsd2si): Ditto.
4814 (sse2_cvtsd2siq): Ditto.
4815 (sse2_cvttsd2si): Ditto.
4816 (sse2_cvttsd2siq): Ditto.
4817 (<shift_insn><mode>3): Ditto.
4818 (sse2_cvtsi2sdq): Update constraint and prefix.
4819 (sse_cvtsi2ss): Update prefix.
4820 (sse_cvtsi2ssq): Ditto.
4822 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4824 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
4825 ignore internal calls.
4827 2013-10-11 Richard Biener <rguenther@suse.de>
4829 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
4830 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
4831 TREE_OVERFLOW is set.
4833 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
4835 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
4837 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
4838 * gimple.h: Likewise.
4840 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
4842 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
4845 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
4847 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
4848 * configure: Regenerate.
4850 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4852 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
4853 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
4854 formatting fixes, use pp_colon instead of pp_character (..., ':'),
4855 similarly pp_right_paren.
4856 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
4857 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
4858 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
4859 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
4861 (omp_declare_simd_clauses_equal,
4862 omp_remove_redundant_declare_simd_attrs): New functions.
4863 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
4864 (walk_tree_1): Handle new OpenMP 4.0 clauses.
4865 * tree.h (OMP_LOOP_CHECK): Define.
4866 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
4867 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
4868 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
4869 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
4870 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
4871 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
4872 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
4873 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
4874 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
4875 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
4876 OMP_CLAUSE_SIMDLEN_EXPR): Define.
4877 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
4878 (omp_remove_redundant_declare_simd_attrs): New prototype.
4879 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
4880 GIMPLE_OMP_TEAMS): New codes.
4881 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
4882 * omp-low.c (struct omp_context): Add cancel_label and cancellable
4884 (target_nesting_level): New variable.
4885 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
4886 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
4887 for collapse > 1 static schedule unless ordered.
4888 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
4889 (determine_parallel_type): Adjust get_ws_args_for caller.
4890 (install_var_field): Handle mask & 4 for double indirection.
4891 (scan_sharing_clauses): Ignore shared clause on teams construct.
4892 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
4893 (create_omp_child_function): If inside target or declare target
4894 constructs, set "omp declare target" attribute on the child function.
4895 (find_combined_for): New function.
4896 (scan_omp_parallel): Handle combined loops.
4897 (scan_omp_target, scan_omp_teams): New functions.
4898 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
4899 restrictions and set ctx->cancellable for cancellable constructs.
4900 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
4901 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
4902 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
4903 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
4904 (omp_clause_aligned_alignment): New function.
4905 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
4906 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
4907 on teams constructs. Handle user defined reductions and new
4909 (lower_reduction_clauses): Don't set placeholder to address of ref
4910 if it has already the right type.
4911 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
4912 (expand_parallel_call): Use the new non-_start suffixed builtins,
4913 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
4914 and GOMP_parallel_end after the call.
4915 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
4916 (expand_omp_for_init_counts): Handle combined loops.
4917 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
4919 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
4920 end of cancellable loops.
4921 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
4922 Likewise. Handle collapse > 1 loops.
4923 (expand_omp_simd): Handle combined loops.
4924 (expand_omp_for): Add inner_stmt argument, adjust callers of
4925 expand_omp_for* functions, use expand_omp_for_static*chunk even
4926 for collapse > 1 unless ordered.
4927 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
4928 of cancellable sections.
4929 (expand_omp_single): Remove need_barrier variable, just rely on
4930 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
4931 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
4932 (expand_omp_atomic_load, expand_omp_atomic_store,
4933 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
4934 (expand_omp_target): New function.
4935 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
4936 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
4937 (build_omp_regions_1): Immediately close region for
4938 GF_OMP_TARGET_KIND_UPDATE.
4939 (maybe_add_implicit_barrier_cancel): New function.
4940 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
4942 (lower_omp_single): Likewise. Add clobber after the barrier.
4943 (lower_omp_taskgroup): New function.
4944 (lower_omp_for): Handle combined loops. Adjust
4945 lower_rec_input_clauses caller. Handle cancellation.
4946 (lower_depend_clauses): New function.
4947 (lower_omp_taskreg): Lower depend clauses. Adjust
4948 lower_rec_input_clauses caller. Add clobber after the call. Handle
4950 (lower_omp_target, lower_omp_teams): New functions.
4951 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
4952 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
4953 and GOMP_cancellation_point calls.
4954 (lower_omp): Fold stmts inside of target region.
4955 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
4956 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4957 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
4958 (BT_FN_VOID_OMPFN_PTR_UINT,
4959 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
4960 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
4961 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
4962 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
4963 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
4964 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
4965 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
4966 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
4967 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
4968 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
4969 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
4970 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
4971 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
4972 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
4973 BUILT_IN_GOMP_PARALLEL_END.
4974 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
4975 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4976 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
4977 GF_OMP_FOR_KIND_DISTRIBUTE.
4978 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
4979 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
4980 (dump_gimple_omp_return): Print lhs if it has any.
4981 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
4982 gimple_omp_atomic_seq_cst_p.
4983 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
4984 and GIMPLE_OMP_TEAMS.
4985 * langhooks.c (lhd_omp_mappable_type): New function.
4986 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
4987 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
4989 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
4990 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
4991 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
4992 (struct gimplify_omp_ctx): Add combined_loop field.
4993 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
4994 on stmts inside of target region.
4995 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
4996 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
4997 ORT_TARGET and ORT_TARGET_DATA.
4998 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
5000 (omp_notice_threadprivate_variable): Complain about threadprivate
5001 variables in target region.
5002 (omp_notice_variable): Complain about vars with non-mappable type
5003 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
5004 (omp_check_private): Ignore ORT_TARGET* regions.
5005 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
5006 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
5007 (find_combined_omp_for): New function.
5008 (gimplify_omp_for): Handle gimplification of combined loops.
5009 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
5011 (gimplify_omp_target_update): New function.
5012 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
5013 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
5014 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
5015 (gimplify_body): If fndecl has "omp declare target" attribute, add
5016 implicit ORT_TARGET context around it.
5017 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
5018 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
5019 * tree-nested.c (convert_nonlocal_reference_stmt,
5020 convert_local_reference_stmt, convert_gimple_call): Handle
5021 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5022 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
5023 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
5024 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
5025 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
5026 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
5027 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
5028 BUILT_IN_GOMP_LOOP_END_CANCEL,
5029 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
5030 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
5031 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
5032 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
5033 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
5034 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
5035 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
5036 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
5037 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
5038 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
5039 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
5040 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
5041 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
5042 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
5043 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
5044 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5045 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
5046 gimple_build_omp_teams): New functions.
5047 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
5048 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
5049 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
5050 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5051 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
5052 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
5053 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
5054 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
5055 GF_OMP_ATOMIC_SEQ_CST): New.
5056 (gimple_build_omp_taskgroup, gimple_build_omp_target,
5057 gimple_build_omp_teams): New prototypes.
5058 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
5059 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5060 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
5061 GIMPLE_OMP_SINGLE as end of range.
5062 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
5063 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
5064 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
5065 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
5066 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
5067 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
5068 gimple_omp_target_kind, gimple_omp_target_set_kind,
5069 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
5070 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
5071 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
5072 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
5073 gimple_omp_teams_set_clauses): New inlines.
5074 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
5075 and GIMPLE_OMP_TASKGROUP.
5076 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
5077 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
5078 enum omp_clause_proc_bind_kind): New.
5079 (union omp_clause_subcode): Add depend_kind, map_kind and
5080 proc_bind_kind fields.
5081 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
5082 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5083 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
5084 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
5085 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
5087 2013-10-10 Teresa Johnson <tejohnson@google.com>
5089 * predict.c (tree_estimate_probability): Add new parameter
5090 for estimate_bb_frequencies.
5091 (estimate_bb_frequencies): Add new parameter to force estimation.
5092 (rebuild_frequencies): When max frequency in function is small,
5093 recompute counts from frequencies.
5094 * predict.h (estimate_bb_frequencies): New parameter.
5096 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5098 * ipa-inline.c (ipa_inline): Fix leak of "order" when
5099 optimizations are disabled.
5101 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5103 * coverage.c (coverage_finish): Fix leak of da_file_name.
5105 2013-10-10 Jan Hubicka <jh@suse.cz>
5107 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
5108 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
5111 2013-10-10 Jakub Jelinek <jakub@redhat.com>
5114 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
5115 if any labels are in FALLTHRU_BB, use a special label emitted
5116 immediately after the asm goto insn rather than label_rtx
5118 (expand_asm_stmt): Adjust caller.
5119 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
5120 edge if the last insn in predecessor is a jump with single successor,
5121 but it isn't simplejump_p.
5123 2013-10-10 Richard Biener <rguenther@suse.de>
5125 PR tree-optimization/58656
5126 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
5128 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5130 * gimplify.c: Include expr.h and tm_p.h for targets with special
5131 va-arg padding requirements.
5133 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5135 * tree-flow.h: Move some prototypes to gimple.h.
5136 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
5137 * gimple.h: Relocate some prototypes from tree-flow.h
5138 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
5140 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
5141 (build_va_arg_indirect_ref): Relocate and make static.
5142 (std_gimplify_va_arg_expr): Relocate here.
5143 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
5144 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
5146 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5148 * doc/md.texi: Document the mnemonic attribute.
5150 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5153 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
5154 (set_attr_alternative x ...) when searching for user defined
5157 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5159 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
5161 2013-10-09 Easwaran Raman <eraman@google.com>
5163 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
5164 * cfgexpand.c (defer_stack_allocation): ...use here
5165 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
5167 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
5169 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
5170 (operand_equal_for_value_replacement): New function, extracted from
5171 value_replacement and enhanced to catch more cases.
5172 (value_replacement): Use operand_equal_for_value_replacement.
5174 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5176 * loop-doloop.c (doloop_modify, doloop_optimize): Use
5177 get_max_loop_iterations.
5179 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5181 * config/arm/aarch-common.c (arm_early_load_addr_dep):
5182 Place comment above function.
5184 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5186 * tree-flow.h: Remove all remaining prototypes, enums and structs that
5187 are not related to tree-cfg.c.
5188 * tree-ssa-address.h: New file. Relocate prototypes.
5189 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
5190 (addr_for_mem_ref): New. Combine call to get_address_description and
5191 return addr_for_mem_ref.
5192 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
5193 * tree-ssa-live.h: Adjust prototypes.
5194 * passes.c: Include tree-ssa-live.h.
5195 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
5196 * graphite.c (graphite_transform_loops): Make static.
5197 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5198 make_pass_graphite, pass_data_graphite_transforms,
5199 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
5200 * ipa-pure-const.c (warn_function_noreturn): Make static.
5201 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5202 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5203 Relocate from tree-cfg.c
5204 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
5206 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5207 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5208 Move to ipa-pure-const.c.
5209 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
5210 Relocate from tree-optimize.c.
5211 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
5212 make_pass_fixup_cfg): Move to tree-cfg.c.
5213 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
5214 Relocate some prototypes.
5215 * tree-data-ref.h (tree_check_data_deps) Add prototype.
5216 * tree-dump.c (dump_function_to_file): Remove prototype.
5217 Add tree-flow.h to the include file.
5218 * tree-dump.h: Remove prototype.
5219 * tree-parloops.h: New File. Add prototypes.
5220 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
5221 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
5222 from tree-ssa-loop.c.
5223 * tree-predcom.c (run_tree_predictive_commoning,
5224 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
5225 Relocate here from tree-ssa-loop.c.
5226 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
5227 ssa_name_values.release ().
5228 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
5229 (ssa_name_values): Relocate from tree-flow.h.
5230 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
5231 * tree-ssa-loop.c (run_tree_predictive_commoning,
5232 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
5233 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5234 make_pass_graphite, pass_data_graphite_transforms,
5235 make_pass_graphite_transforms, gate_tree_parallelize_loops,
5236 tree_parallelize_loops, pass_data_parallelize_loops,
5237 make_pass_parallelize_loops): Move to other files.
5238 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
5240 * tree.h: Remove prototypes from tree-address.c.
5242 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5244 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
5245 (struct int_tree_map): Move to tree-hasher.h
5246 (SCALE, LABEL, PERCENT): Move to gimple.h
5247 * tree-flow-inline.h: Delete. Move functions to other files.
5248 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
5249 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
5250 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
5251 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
5252 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
5253 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
5254 * tree-hasher.h: Don't include tree-flow.h.
5255 (struct int_tree_map): Relocate from tree-flow.h.
5256 * tree-sra.c (contains_view_convert_expr_p): Relocate from
5257 tree-flow-inline.h and make static.
5258 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
5260 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
5261 tree-flow-inline.h and make static.
5262 * tree.h (is_global_var, may_be_aliased): Relocate from
5264 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
5265 * value-prof.c: No longer include tree-flow-inline.h.
5266 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
5268 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5270 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
5271 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
5272 (enum move_pos): Move to tree-ssa-loop-im.h
5273 * cfgloop.h: Move some prototypes.
5274 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
5275 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
5276 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
5277 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
5278 (loop_containing_stmt): Relocate from tree-flow-inline.h.
5279 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
5280 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
5281 (enum move_pos): Relocate here.
5282 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
5284 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
5285 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
5286 make_pass_lim): Relocate here from tree-ssa-loop.c.
5287 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
5289 (loop_edge_to_cancel, unloop_loops): Make static.
5290 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5291 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
5292 (tree_complete_unroll, gate_tree_complete_unroll,
5293 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
5294 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5295 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
5296 * tree-ssa-loop-ivopts.c: Remove local prototypes.
5297 (stmt_invariant_in_loop_p): Remove unused function.
5298 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
5299 * tree-ssa-loop-manip.h: New file. Add prototypes.
5300 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
5301 (gcov_type_to_double_int): Move to cfgloop.h.
5302 (double_int_cmp, bound_index,
5303 estimate_numbers_of_iterations_loop): Make static.
5304 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
5305 (max_loop_iterations): Factor out get_max_loop_iterations.
5306 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
5308 * tree-ssa-loop-niter.h: New file. Add prototypes.
5309 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
5310 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
5311 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
5312 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
5313 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
5314 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
5315 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
5316 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
5317 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
5318 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
5319 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5320 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
5321 pass_data_complete_unroll, make_pass_complete_unroll,
5322 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5323 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
5324 tree-ssa-loop-ivcanon.c.
5325 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
5326 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
5327 tree-ssa-loop-prefetch.c.
5328 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
5330 (get_lsm_tmp_name): Relocate and add suffix parameter.
5331 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
5332 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
5333 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
5334 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
5335 (get_estimated_loop_iterations): Factor out accessor from
5336 estimated_loop_iterations in tree-ssa-loop-niter.c.
5337 (get_max_loop_iterations): Factor out accessor from
5338 _max_loop_iterations in tree-ssa-niter.c.
5339 * loop-unroll.c (decide_unroll_constant_iterations,
5340 decide_unroll_runtime_iterations, decide_peel_simple,
5341 decide_unroll_stupid): Use new get_* accessors.
5343 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
5345 PR tree-optimization/20318
5346 * doc/extend.texi (returns_nonnull): New function attribute.
5347 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
5349 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
5350 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
5352 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
5355 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
5356 false if both components are bitfields.
5358 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5360 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
5361 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
5362 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
5363 * config/aarch64/aarch64.h
5364 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
5365 * config/aarch64/aarch64-simd-builtins.def
5366 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
5368 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5370 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
5372 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5374 * config/aarch64/arm_neon.h (vdiv_f64): Added.
5376 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5378 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
5379 (vneg_f64): New intrinsic.
5380 (vneg_s8): Asm replaced with C.
5381 (vneg_s16): Likewise.
5382 (vneg_s32): Likewise.
5383 (vneg_s64): New intrinsic.
5384 (vnegq_f32): Asm replaced with C.
5385 (vnegq_f64): Likewise.
5386 (vnegq_s8): Likewise.
5387 (vnegq_s16): Likewise.
5388 (vnegq_s32): Likewise.
5389 (vnegq_s64): Likewise.
5391 2013-10-09 Renlin Li <Renlin.Li@arm.com>
5393 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
5395 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5397 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
5398 packed stack special handling.
5399 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
5400 back to fixed stack slots for FPRs saved due to stdarg.
5402 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5404 * config/s390/s390.c (s390_frame_info): Restructure function.
5406 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5408 * config/s390/s390.c (struct s390_frame_layout): New field
5410 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
5411 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
5412 regs_ever_clobbered to char*.
5413 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
5414 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
5415 (s390_register_info): Enable FPR save slots. Move/Copy some
5416 functionality into ...
5417 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
5418 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
5420 (s390_frame_info): Do gpr slot allocation here now. stdarg does
5421 not imply a stack frame.
5422 (s390_init_frame_layout): Remove variable clobbered_regs.
5423 (s390_update_register_info): Remove function.
5424 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
5425 cannot be used for register renaming.
5426 (s390_hard_regno_scratch_ok): New function.
5427 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
5428 (s390_initial_elimination_offset): Change offset calculation of
5429 the return address pointer.
5430 (save_gprs): Deal with only r6 being saved from the call-saved regs.
5431 (restore_gprs): Set frame related flag.
5432 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
5433 (s390_emit_prologue): Call s390_register_info instead of
5434 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
5435 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
5436 (s390_optimize_prologue): Call s390_optimize_register_info.
5437 Try to remove also FPR slot save/restore INSNs. Remove frame
5438 related flags from restore INSNs.
5440 2013-10-08 DJ Delorie <dj@redhat.com>
5442 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
5445 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
5446 avoid conflict with the MI use of %c.
5447 * config/rl78/rl78-real.md: change %c to %C throughout.
5448 * config/rl78/rl78-virt.md: Likewise.
5450 2013-10-08 Jan Hubicka <jh@suse.cz>
5452 * config/i386/i386.c (ix86_option_override_internal): Switch
5453 to SSE math for -ffast-math when target ISA supports SSE2.
5455 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
5457 * tree-flow.h: Remove some prototypes.
5458 * tree.h: Remove some protypes, add a couple.
5459 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
5460 using_eh_for_cleanups_p): Add interface routines for front ends.
5461 * tree-eh.h: New file. Add protoptyes.
5462 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
5463 (add_stmt_to_eh_lp_fn): Make static.
5464 (lower_try_finally): Use new using_eh_for_cleanups_p.
5465 * emit-rtl.c: Include tree-eh.h.
5466 * gimple.h: Include tree-eh.h.
5468 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
5470 PR tree-optimization/58480
5471 * tree-vrp.c (infer_nonnull_range): New function.
5472 (infer_value_range): Call infer_nonnull_range.
5474 2013-10-08 Dehao Chen <dehao@google.com>
5476 PR tree-optimization/58619
5477 * tree-inline.c (copy_phis_for_bb): Combine location data
5480 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5483 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
5484 RTX_FRAME_RELATED_P on INSN.
5486 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5488 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
5489 (altivec_expand_vec_perm_const): Call it.
5491 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5493 * config/rs6000/vector.md (mov<mode>): Emit permuted move
5494 sequences for LE VSX loads and stores at expand time.
5495 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
5497 * config/rs6000/rs6000.c (rs6000_const_vec): New.
5498 (rs6000_gen_le_vsx_permute): New.
5499 (rs6000_gen_le_vsx_load): New.
5500 (rs6000_gen_le_vsx_store): New.
5501 (rs6000_gen_le_vsx_move): New.
5502 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
5503 (*vsx_le_perm_load_v4si): New.
5504 (*vsx_le_perm_load_v8hi): New.
5505 (*vsx_le_perm_load_v16qi): New.
5506 (*vsx_le_perm_store_v2di): New.
5507 (*vsx_le_perm_store_v4si): New.
5508 (*vsx_le_perm_store_v8hi): New.
5509 (*vsx_le_perm_store_v16qi): New.
5510 (*vsx_xxpermdi2_le_<mode>): New.
5511 (*vsx_xxpermdi4_le_<mode>): New.
5512 (*vsx_xxpermdi8_le_V8HI): New.
5513 (*vsx_xxpermdi16_le_V16QI): New.
5514 (*vsx_lxvd2x2_le_<mode>): New.
5515 (*vsx_lxvd2x4_le_<mode>): New.
5516 (*vsx_lxvd2x8_le_V8HI): New.
5517 (*vsx_lxvd2x16_le_V16QI): New.
5518 (*vsx_stxvd2x2_le_<mode>): New.
5519 (*vsx_stxvd2x4_le_<mode>): New.
5520 (*vsx_stxvd2x8_le_V8HI): New.
5521 (*vsx_stxvd2x16_le_V16QI): New.
5523 2013-10-07 Renlin Li <Renlin.Li@arm.com>
5525 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
5527 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5529 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
5530 loop to work also for 31bit ABI.
5531 Save the stack pointer for frame_size > 0.
5533 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5535 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
5536 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
5537 constraint letters from expanders.
5538 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
5539 retry count to general_operand.
5540 ("tabort"): Give operand 0 a mode.
5541 ("tabort_1"): Add mode and constraint letter for operand 0.
5542 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
5544 2013-10-04 Jeff Law <law@redhat.com>
5546 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
5548 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
5550 (thread_across_edge): Here. Call it.
5552 2013-10-04 Cary Coutant <ccoutant@google.com>
5554 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
5555 discarding a location list expression (or a piece of one).
5557 2013-10-03 Jan Hubicka <jh@suse.cz>
5559 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
5560 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
5561 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
5563 2013-10-03 Jan Hubicka <jh@suse.cz>
5565 * config/i386/i386.c (ix86_option_override_internal): Do not enable
5566 accumulate-outgoing-args when producing unwind info.
5568 2013-10-03 Wei Mi <wmi@google.com>
5570 * lra-constraints.c (insert_move_for_subreg): New function
5571 extracted from simplify_operand_subreg.
5572 (simplify_operand_subreg): Add reload for paradoxical subreg.
5574 2013-10-03 Rong Xu <xur@google.com>
5576 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
5577 the candidate of builtin_expect such that we should fix the
5578 size/time estimation.
5579 (estimate_function_body_sizes): Do the acutally size/time fix-up
5582 2013-10-03 Rong Xu <xur@google.com>
5584 * predict.c (tree_predict_by_opcode): Get the probability
5585 for builtin_expect from param builtin_expect_probability.
5586 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
5587 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
5588 * doc/invoke.texi: Add documentation for builtin-expect-probability.
5590 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
5593 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
5594 * calls.c (alloca_call_p): Use get_callee_fndecl.
5595 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
5596 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
5598 (vrp_visit_stmt): Remove duplicated code.
5600 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
5602 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
5603 ceildf2, btruncdf2, instead of vsx_* name.
5605 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
5606 iterators to only do V2DF and V4SF here. Move the DF code to
5607 rs6000.md where it is combined with SF mode. Replace <VSv> with
5608 just 'v' since only vector operations are handled with these insns
5609 after moving the DF support to rs6000.md.
5610 (vsx_sub<mode>3): Likewise.
5611 (vsx_mul<mode>3): Likewise.
5612 (vsx_div<mode>3): Likewise.
5613 (vsx_fre<mode>2): Likewise.
5614 (vsx_neg<mode>2): Likewise.
5615 (vsx_abs<mode>2): Likewise.
5616 (vsx_nabs<mode>2): Likewise.
5617 (vsx_smax<mode>3): Likewise.
5618 (vsx_smin<mode>3): Likewise.
5619 (vsx_sqrt<mode>2): Likewise.
5620 (vsx_rsqrte<mode>2): Likewise.
5621 (vsx_fms<mode>4): Likewise.
5622 (vsx_nfma<mode>4): Likewise.
5623 (vsx_copysign<mode>3): Likewise.
5624 (vsx_btrunc<mode>2): Likewise.
5625 (vsx_floor<mode>2): Likewise.
5626 (vsx_ceil<mode>2): Likewise.
5627 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
5628 (vsx_sminsf3): Likewise.
5629 (vsx_fmadf4): Likewise.
5630 (vsx_fmsdf4): Likewise.
5631 (vsx_nfmadf4): Likewise.
5632 (vsx_nfmsdf4): Likewise.
5633 (vsx_cmpdf_internal1): Likewise.
5635 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
5636 simpler to select whether a target has SPE or traditional floating
5637 point support in iterators.
5638 (TARGET_DF_SPE): Likewise.
5639 (TARGET_SF_FPR): Likewise.
5640 (TARGET_DF_FPR): Likewise.
5641 (TARGET_SF_INSN): Macros to say whether floating point support
5642 exists for a given operation for expanders.
5643 (TARGET_DF_INSN): Likewise.
5645 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
5646 combining of SF/DF mode operations, using both traditional and VSX
5654 (abs<mode>2): Combine SF/DF modes using traditional floating point
5655 instructions. Add support for using the upper DF registers with
5656 VSX support, and SF registers with power8-vector support. Update
5657 expanders for operations supported by both the SPE and traditional
5658 floating point units.
5659 (abs<mode>2_fpr): Likewise.
5660 (nabs<mode>2): Likewise.
5661 (nabs<mode>2_fpr): Likewise.
5662 (neg<mode>2): Likewise.
5663 (neg<mode>2_fpr): Likewise.
5664 (add<mode>3): Likewise.
5665 (add<mode>3_fpr): Likewise.
5666 (sub<mode>3): Likewise.
5667 (sub<mode>3_fpr): Likewise.
5668 (mul<mode>3): Likewise.
5669 (mul<mode>3_fpr): Likewise.
5670 (div<mode>3): Likewise.
5671 (div<mode>3_fpr): Likewise.
5672 (sqrt<mode>3): Likewise.
5673 (sqrt<mode>3_fpr): Likewise.
5674 (fre<Fs>): Likewise.
5675 (rsqrt<mode>2): Likewise.
5676 (cmp<mode>_fpr): Likewise.
5677 (smax<mode>3): Likewise.
5678 (smin<mode>3): Likewise.
5679 (smax<mode>3_vsx): Likewise.
5680 (smin<mode>3_vsx): Likewise.
5681 (negsf2): Delete SF operations that are merged with DF.
5688 (fmasf4_fpr): Likewise.
5689 (fmssf4_fpr): Likewise.
5690 (nfmasf4_fpr): Likewise.
5691 (nfmssf4_fpr): Likewise.
5692 (sqrtsf2): Likewise.
5693 (rsqrtsf_internal1): Likewise.
5694 (smaxsf3): Likewise.
5695 (sminsf3): Likewise.
5696 (cmpsf_internal1): Likewise.
5697 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
5698 (negdf2): Delete DF operations that are merged with SF.
5700 (nabsdf2): Likewise.
5706 (rsqrtdf_internal1): Likewise.
5707 (fmadf4_fpr): Likewise.
5708 (fmsdf4_fpr): Likewise.
5709 (nfmadf4_fpr): Likewise.
5710 (nfmsdf4_fpr): Likewise.
5711 (sqrtdf2): Likewise.
5712 (smaxdf3): Likewise.
5713 (smindf3): Likewise.
5714 (cmpdf_internal1): Likewise.
5715 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
5716 (btrunc<mode>2): Delete separate expander, and combine with the
5717 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
5718 (btrunc<mode>2_fpr): Likewise.
5719 (ceil<mode>2): Likewise.
5720 (ceil<mode>2_fpr): Likewise.
5721 (floor<mode>2): Likewise.
5722 (floor<mode>2_fpr): Likewise.
5723 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
5724 Add support for using the upper registers with VSX and
5725 power8-vector. Move insns to be closer to the define_expands. On
5726 VSX systems, prefer the traditional form of FMA over the VSX
5727 version, since the traditional form allows the target not to
5728 overlap with the inputs.
5729 (fms<mode>4_fpr): Likewise.
5730 (nfma<mode>4_fpr): Likewise.
5731 (nfms<mode>4_fpr): Likewise.
5733 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5734 Richard Earnshaw <richard.earnshaw@arm.com>
5736 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
5737 (struct mult_cost_table): Likewise.
5738 (struct mem_cost_table): Likewise.
5739 (struct fp_cost_table): Likewise.
5740 (struct vector_cost_table): Likewise.
5741 (cpu_cost_table): Likewise.
5742 * config/arm/arm.opt (mold-rts-costs): New option.
5743 (mnew-generic-costs): Likewise.
5744 * config/arm/arm.c (generic_extra_costs): New table.
5745 (cortexa15_extra_costs): Likewise.
5746 (arm_slowmul_tune): Use NULL as new costs.
5747 (arm_fastmul_tune): Likewise.
5748 (arm_strongarm_tune): Likewise.
5749 (arm_xscale_tune): Likewise.
5750 (arm_9e_tune): Likewise.
5751 (arm_v6t2_tune): Likewise.
5752 (arm_cortex_a5_tune): Likewise.
5753 (arm_cortex_a9_tune): Likewise.
5754 (arm_v6m_tune): Likewise.
5755 (arm_fa726te_tune): Likewise.
5756 (arm_cortex_a15_tune): Use cortex15_extra_costs.
5757 (arm_cortex_tune): Use generict_extra_costs.
5758 (shifter_op_p): New function.
5759 (arm_unspec_cost): Likewise.
5760 (LIBCALL_COST): Define.
5761 (arm_new_rtx_costs): New function.
5762 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
5763 table is available. Use old costs otherwise unless mnew-generic-costs
5765 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
5766 (cpu_cost_table): Declare.
5768 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
5771 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
5772 (*subs_mul_imm_<mode>)
5773 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
5774 (*sub_<shift>_<mode>)
5775 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
5776 Remove k constraint.
5778 2013-10-03 Ian Bolton <ian.bolton@arm.com>
5780 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
5782 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
5784 2013-10-02 Teresa Johnson <tejohnson@google.com>
5786 * predict.c (probably_never_executed): New function.
5787 (probably_never_executed_bb_p): Invoke probably_never_executed.
5788 (probably_never_executed_edge_p): Ditto.
5789 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
5790 Treat profile insanities conservatively.
5792 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
5794 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
5796 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
5798 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
5799 use it. Use smaller increase for scratch. Don't increase reject
5800 for early clobber scratch.
5801 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
5802 setting eliminated regs except setting fp from hfp.
5803 (lra_eliminate): Check lra_insn_recog_data on NULL.
5805 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5808 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
5809 setting -mvsx-timode by default until the underlying problem is fixed.
5810 (RS6000_CPU, power7 defaults): Likewise.
5812 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5814 * config/x-linux (host-linux.o): Remove header dependencies.
5815 Use $(COMPILE) and $(POSTCOMPILE).
5816 * config/t-linux-android (linux-android.o): Ditto.
5818 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5820 * Makefile.in (expmed.o-warn): Remove.
5822 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5824 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
5825 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
5827 2013-10-02 Teresa Johnson <tejohnson@google.com>
5829 * dojump.c (do_jump_1): Divide probability between
5830 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
5832 2013-10-02 Tom Tromey <tromey@redhat.com>
5834 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
5836 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5838 * tree-flow.h: Remove some prototypes.
5839 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
5840 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
5841 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
5842 mark_virtual_phi_result_for_renaming): Relocate here.
5843 * tree-into-ssa.h: Add prototypes.
5844 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
5845 single_pred_before_succ_order.
5846 (blocks_in_phiopt_order): Rename and move to cfganal.c.
5847 (nonfreeing_call_p) Move to gimple.c.
5848 * cfganal.c (single_pred_before_succ_order): Move and renamed from
5850 * basic-block.h (single_pred_before_succ_order): Add prototype.
5851 * gimple.c (nonfreeing_call_p): Relocate here.
5852 * gimple.h: Add prototype.
5853 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
5854 * tree-ssa-dom.h: New file. Relocate prototypes here.
5855 * tree-ssa.h: Include tree-ssa-dom.h.
5857 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5859 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
5860 Use $(COMPILE) and $(POSTCOMPILE).
5862 * config/alpha/x-alpha (driver-alpha.o): Ditto.
5864 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5866 * tree-flow.h: Remove some prototypes.
5867 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
5868 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5869 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
5870 propagate_tree_value*): Move from here to...
5871 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
5872 propagate_tree_value*): Relocate here.
5873 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5874 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
5875 * gimple-fold.c: Remove gimple-fold.h from include list.
5876 * tree-vrp.c: Remove gimple-fold.h from include list.
5877 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
5878 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
5879 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
5880 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
5881 * sese.c: Add tree-ssa-propagate.h to include list.
5883 2013-10-02 Richard Biener <rguenther@suse.de>
5885 * tree-loop-distribution.c: Include tree-vectorizer.h for
5887 (enum partition_kind): Remove PKIND_REDUCTION.
5888 (struct partition_s): Remove has_writes member, add reduction_p member.
5889 (partition_alloc): Adjust.
5890 (partition_builtin_p): Likewise.
5891 (partition_has_writes): Remove.
5892 (partition_reduction_p): New function.
5893 (partition_merge_into): Likewise.
5894 (generate_code_for_partition): Commonize builtin partition
5896 (rdg_cannot_recompute_vertex_p): Remove.
5897 (already_processed_vertex_p): Likewise.
5898 (rdg_flag_vertex): Do not set has_writes.
5899 (classify_partition): Adjust.
5900 (rdg_build_partitions): Do not set has_writes, treat all
5901 partitions as useful.
5902 (distribute_loop): Record number of library calls generated. Adjust.
5903 (tree_loop_distribution): Report number of loops and library
5904 calls generated as opt-info.
5906 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5908 * tree-flow.h: Include new .h files. Move prototypes.
5909 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
5910 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
5911 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
5912 * tree-pretty-print.h: Add prototypes from tree-flow.h.
5913 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
5914 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
5915 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
5916 * tree.h (get_ref_base_and_extent): Move prototype out.
5917 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
5919 * gimple-low.h: New File. Add prototypes from tree-flow.h.
5920 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
5921 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
5922 * tree-scalar-evolution.c: Include tree.h.
5923 * sese.c: Include tree.h.
5924 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
5925 * dwarf2out.c: Include tree-dfa.h.
5926 * tree-chrec.c: Include tree.h.
5927 * tree-data-ref.c: Include tree.h.
5929 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
5931 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
5934 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5936 * config/t-sol2 (sol2-c.o): Remove header dependencies.
5937 Use $(COMPILE) and $(POSTCOMPILE).
5938 (sol2-cxx.o): Likewise.
5939 (sol2-stubs.o): Likewise.
5941 * config/x-solaris (host-solaris.o): Likewise.
5943 * config/sparc/t-sparc (sparc.o): Remove.
5944 (sparc-c.o): Remove header dependencies.
5945 Use $(COMPILE) and $(POSTCOMPILE).
5946 * config/sparc/x-sparc: Likewise.
5948 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
5950 * config/arc/arc-opts.h: Add 2013 to Copyright years.
5951 * config/arc/arc700.md: Likewise.
5952 * config/arc/arc-modes.def: Likewise.
5953 * config/arc/arc-simd.h: Likewise.
5954 * config/arc/t-arc-uClibc: Likewise.
5955 * config/arc/t-arc-newlib: Likewise.
5957 2013-10-02 Renlin Li <renlin.li@arm.com>
5959 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
5961 (aarch64_expand_epilogue): Likewise.
5963 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5964 Yufeng Zhang <yufeng.zhang@arm.com>
5966 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
5968 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
5969 'base_in' represent a conversion and legal_cast_p_1 holds; set
5970 'base_in' with the returned value from get_unwidened.
5972 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5974 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
5975 plus_constant is not used.
5977 2013-10-01 Wei Mi <wmi@google.com>
5979 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
5980 * config/i386/i386.md: Add define_peephole2 to
5981 break partial reg stall for cvtss2sd/cvtsd2ss.
5983 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5985 * config/arc/arc.c (pass_arc_ifcvt::clone):
5986 Update for ctxt_ -> m_ctxt change.
5988 2013-10-01 Jeff Law <law@redhat.com>
5990 * tree-ssa-threadupdate.c (struct redirection_data): Delete
5991 outgoing_edge and intermediate_edge fields. Instead store the path.
5992 (redirection_data::hash): Hash on the last edge's destination index.
5993 (redirection_data::equal): Check the entire thread path.
5994 (lookup_redirectio_data): Corresponding changes.
5995 (create_edge_and_update_destination_phis): Likewise.
5996 (thread_single_edge): Likewise.
5998 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5999 Diego Novillo <dnovillo@google.com>
6001 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
6002 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
6003 (vld32wh_insn, vld32wl_insn): Delete commented-out old
6004 versions of these patterns.
6006 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
6007 (__builtin_arc_aligned): Likewise.
6009 * config/arc/arc.md: Expand adc_0 comment stating the intended
6010 purpose and why it isn't ready.
6011 Replace commented out call_value_via_label_mixed with a
6012 plain comment about bl_s.
6014 * config/arc/arc.c (stdio.h): Don't include directly.
6015 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
6016 Remove if (1) condition.
6017 (arc_encode_section_info): Fix comment.
6019 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6021 * config/arc/arc.c (arc_conditional_register_usage):
6022 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
6023 Also set reg_alloc_order for DMA config regs.
6025 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6026 Jeremy Bennett <jeremy.bennett@embecosm.com>
6028 * doc/install.texi (--with-cpu): Mention ARC.
6029 (arc-*-elf32): New paragraph.
6030 (arc-linux-uclibc): Likewise.
6031 * doc/md.texi (Machine Constraints): Add ARC part.
6032 * doc/invoke.texi: (menu): Add ARC Options.
6033 (Machine Dependent Options) <ARC Options>: Add synopsis.
6034 (node ARC Options): Add.
6035 * doc/extend.texi (long_call / short_call attribute): Add ARC.
6036 (ARC Built-in Functions): New section defining
6037 generic ARC built-in functions.
6038 (ARC SIMD Built-in Functions): New section defining SIMD specific
6040 (Declaring Attributes of Functions): Extended
6041 description of short_call and long_call attributes for ARC and
6042 added index entries.
6044 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
6045 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
6046 Joern Rennecke <joern.rennecke@embecosm.com>
6047 Muhammad Khurram Riaz <khurram.riaz@arc.com>
6048 Brendan Kehoe <brendan@zen.org>
6049 Michael Eager <eager@eagercon.com>
6050 Simon Cook <simon.cook@embecosm.com>
6051 Jeremy Bennett <jeremy.bennett@embecosm.com>
6053 * config/arc, common/config/arc: New directories.
6055 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6056 Brendan Kehoe <brendan@zen.org>
6057 Simon Cook <simon.cook@embecosm.com>
6059 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
6061 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6063 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
6064 * tree-ssa-coalesce.h: New. Move prototype to here.
6065 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
6066 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
6067 (gimple_can_coalesce_p): Move to...
6068 * gimple.c (gimple_can_coalesce_p): Here.
6070 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6072 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
6073 (dump_decl_set): Move to gimple.c.
6074 * gimple.h: Don't include tree-ssa-operands.h.
6075 (dump_decl_set): Add prototype.
6076 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6077 Move to gimple-ssa.h.
6078 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
6079 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
6080 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
6081 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
6082 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
6084 (dump_decl_set): Relocate here.
6085 * gimple-ssa.h: New file.
6086 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6087 Relocate from gimple.h.
6088 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
6089 * tree-ssa-operands.c (swap_ssa_operands): Rename from
6090 swap_tree_operands and remove non-ssa path.
6091 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
6092 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
6094 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
6095 vect_is_simple_reduction_1): Use swap_ssa_operands.
6096 * tree-flow.h: Move various prototypes to tree-phinodes.h.
6097 (enum need_phi_state): Move to tree-into-ssa.c.
6098 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6099 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
6100 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
6101 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
6102 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6103 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6104 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6105 num_imm_uses): Move to ssa-iterators.h.
6106 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
6107 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
6109 (op_iter_done, op_iter_next_def, op_iter_next_tree,
6110 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
6111 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
6112 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
6113 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
6114 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
6115 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
6116 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
6117 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
6118 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
6119 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
6120 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
6121 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
6122 (set_phi_nodes): Move to tree-phinodes.h.
6123 * tree-ssa-operands.h (enum ssa_op_iter_type,
6124 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
6125 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
6126 (dump_decl_set): Remove prototype.
6127 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
6128 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
6129 (set_phi_nodes): Relocate from tree-flow-inline.h.
6130 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
6132 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
6133 include list. Temporarily add gimple.h to include list.
6134 * ssa-iterators.h: New file.
6135 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6136 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
6137 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
6138 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
6139 Relocate from tree-ssa-operands.h.
6140 (delink_imm_use, link_imm_use_to_list, link_imm_use,
6141 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6142 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6143 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6144 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
6145 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
6146 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
6147 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
6148 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
6149 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
6150 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
6151 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
6152 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
6153 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
6154 Relocate from tree-flow-inline.h.
6155 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
6157 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
6160 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
6161 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
6162 Insert '\t' to output template.
6163 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
6164 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
6165 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
6166 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
6168 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
6170 * doc/install.texi (Host/target specific installation notes for GCC):
6171 Put @anchor before @heading.
6172 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
6173 Use @email for email addresses.
6175 2013-10-01 Jeff Law <law@redhat.com>
6177 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
6178 a vec. Only delete the path if we create one without successfully
6179 registering a jump thread.
6180 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
6182 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
6183 (paths): New vector of jump threading paths.
6184 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
6185 (THREAD_PATH): New accessor macro for the entire thread path.
6186 (lookup_redirection_data): Get intermediate and final outgoing edge
6187 from the thread path.
6188 (create_edge_and_update_destination_phis): Copy the threading path.
6189 (ssa_fix_duplicate_block_edges): Get edges and block types from the
6190 jump threading path.
6191 (ssa_redirect_edges): Get edges and block types from the jump threading
6192 path. Free the path vector.
6193 (thread_block): Get edges from the jump threading path. Look at the
6194 entire path to see if we thread to a loop exit. If we cancel a jump
6195 thread request, then free the path vector.
6196 (thread_single_edge): Get edges and block types from the jump threading
6197 path. Free the path vector.
6198 (thread_through_loop_header): Get edges and block types from the jump
6199 threading path. Free the path vector.
6200 (mark_threaded_blocks): Iterate over the vector of paths and store
6201 the path on the appropriate edge. Get edges and block types from the
6202 jump threading path.
6203 (mark_threaded_blocks): Get edges and block types from the jump
6204 threading path. Free the path vector.
6205 (thread_through_all_blocks): Use the vector of paths rather than
6206 a vector of 3-edge sets.
6207 (register_jump_thread): Accept pointer to a path vector rather
6208 than the path vector itself. Store the path vector for later use.
6211 2013-10-01 Jakub Jelinek <jakub@redhat.com>
6212 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6215 * config/s390/s390.c (s390_split_branches): Modify check for table
6217 (s390_chunkify_start): Rearrange table jump insn check in order to
6218 deal with compare and branch insns correctly.
6220 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
6224 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
6225 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
6226 define_insn_and_split.
6227 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
6228 (shiftsi3_compare): New pattern.
6230 * config/arm/unspecs.md (UNSPEC_RRX): New.
6232 2013-10-01 Alan Modra <amodra@gmail.com>
6234 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
6235 casing inout operands.
6237 2013-10-01 Richard Biener <rguenther@suse.de>
6239 PR tree-optimization/58553
6240 * tree-loop-distribution.c (struct partition_s): Add niter member.
6241 (classify_partition): Populate niter member for the partition
6242 and properly identify whether the relevant store happens before
6243 or after the loop exit.
6244 (generate_memset_builtin): Use niter member from the partition.
6245 (generate_memcpy_builtin): Likewise.
6247 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6249 * vec.h (vec_prefix, vec): Prefix member names with "m_".
6250 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
6252 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6254 * basic-block.h (edge_list): Prefix member names with "m_".
6255 * context.h (context): Likewise.
6256 * domwalk.h (dom_walker): Likewise.
6257 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
6258 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
6259 * hash-table.h (hash_table): Likewise.
6260 * machmode.h (bit_field_mode_iterator): Likewise.
6261 * pass_manager.h (pass_list): Likewise.
6262 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
6263 * tree-pass.h (pass_data): Likewise.
6264 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
6265 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
6266 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
6267 * asan.c (pass_data_asan): Update accordingly.
6268 * cfganal.c (control_dependences::find_control_dependence): Likewise.
6269 (control_dependences::control_dependences): Likewise.
6270 (control_dependences::~control_dependences): Likewise.
6271 (control_dependences::~control_dependences): Likewise.
6272 (control_dependences::get_edges_dependent_on): Likewise.
6273 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
6274 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
6275 * context.c (gcc::context::context): Likewise.
6276 * cprop.c (pass_rtl_cprop::clone): Likewise.
6277 * domwalk.c (dom_walker::walk): Likewise.
6278 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
6279 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
6280 * mode-switching.c (pass_mode_switching::clone): Likewise.
6281 * passes.c (opt_pass::opt_pass): Likewise.
6282 (pass_manager::pass_manager): Likewise.
6283 * predict.c (pass_strip_predict_hints::clone): Likewise.
6284 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
6285 (pass_split_all_insns::clone): Likewise.
6286 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
6288 (bit_field_mode_iterator::next_mode): Likewise.
6289 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
6290 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
6291 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
6292 * tree-complex.c (pass_lower_complex::clone): Likewise.
6293 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
6294 * tree-object-size.c (pass_object_sizes::clone): Likewise.
6295 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
6296 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
6297 (pass_fold_builtins::clone): Likewise.
6298 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
6299 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
6300 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
6301 (pass_cd_dce::clone): Likewise.
6302 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
6303 (pass_phi_only_cprop::clone): Likewise.
6304 * tree-ssa-dse.c (pass_dse::clone): Likewise.
6305 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
6306 * tree-ssa-loop.c (pass_lim::clone): Likewise.
6307 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
6308 * tree-ssa-pre.c (pass_fre::clone): Likewise.
6309 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
6310 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
6311 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
6312 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
6313 * tree-vrp.c (pass_vrp::clone): Likewise.
6314 * tsan.c (pass_tsan::clone): Likewise.
6316 2013-09-30 Jakub Jelinek <jakub@redhat.com>
6319 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
6320 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
6323 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
6324 optimization, punt if sign_bit_p looked through any zero extension.
6326 2013-09-30 Teresa Johnson <tejohnson@google.com>
6328 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
6329 Update redirected out edge count in joiner case.
6330 (ssa_redirect_edges): Common the joiner and non-joiner cases
6331 so that joiner case gets profile updates.
6333 2013-09-30 Richard Biener <rguenther@suse.de>
6335 PR tree-optimization/58554
6336 * tree-loop-distribution.c (classify_partition): Require
6337 unconditionally executed stores for memcpy and memset recognition.
6338 (tree_loop_distribution): Calculate dominance info.
6340 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
6342 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
6343 (NO_PROFILE_COUNTERS): Likewise.
6344 (PROFILE_HOOK): Likewise.
6345 (FUNCTION_PROFILER): Likewise.
6346 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
6348 2013-09-30 Iain Sandoe <iain@codesourcery.com>
6350 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
6351 calls and defines in TARGET_MACHO conditional.
6352 (load_macho_picbase_di): Likewise.
6353 (reload_macho_picbase): Likewise.
6354 (reload_macho_picbase_si): Likewise.
6355 (reload_macho_picbase_di): Likewise.
6356 (nonlocal_goto_receiver): Likewise.
6358 2013-09-30 Nick Clifton <nickc@redhat.com>
6360 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
6361 that use the MSP430X ISA.
6362 (msp430_option_override): Scan -mmcu command line option for any
6363 MCU name that supports the MSP430X ISA.
6364 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
6365 -mmcu options which enable the MSP430X ISA.
6367 2013-09-30 Richard Biener <rguenther@suse.de>
6370 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
6371 before looking for setjmp-like calls.
6373 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6376 * config/darwin-protos.h (machopic_get_function_picbase): New.
6377 * config/darwin.c (machopic_get_function_picbase): New.
6378 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
6379 label for a new func. (load_macho_picbase_di): Likewise.
6380 (reload_macho_picbase): New expand.
6381 (reload_macho_picbase_si): New insn.
6382 (reload_macho_picbase_di): New insn.
6383 (nonlocal_goto_receiver): New define and split.
6384 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
6385 (unspecv enum): Add UNSPECV_NLGR.
6387 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6389 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
6390 that altivec registers are correctly sized on Darwin.
6392 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6394 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
6395 darwin-driver.o): Use COMPILE and POSTCOMPILE.
6396 * config/x-darwin (host-darwin.o): Likewise.
6397 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
6398 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
6399 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
6401 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
6403 * doc/invoke.texi: Fix usage of @tie{} command.
6405 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
6407 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
6409 2013-09-28 Jan Hubicka <jh@suse.cz>
6411 * config/i386/x86-tune.def: Add documentation for each of the options;
6414 2013-09-28 Jan Hubicka <jh@suse.cz>
6416 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
6418 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
6419 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
6420 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
6421 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
6422 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
6424 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
6426 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
6427 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
6428 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
6429 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
6430 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
6431 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
6432 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
6433 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
6434 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
6435 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
6436 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
6437 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
6438 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
6439 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
6440 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
6441 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
6442 gimple.h, godump.c, graphite-clast-to-gimple.c,
6443 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
6444 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
6445 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
6446 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6447 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
6448 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
6449 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
6450 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
6451 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
6452 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
6453 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
6454 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
6455 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
6456 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
6457 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
6458 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
6459 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
6460 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
6461 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
6462 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
6463 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
6464 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
6465 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
6466 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
6467 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6468 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6469 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
6470 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
6471 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
6472 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6473 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6474 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
6475 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
6476 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6477 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6478 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
6479 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
6480 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
6481 whitespace before "(".
6483 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
6485 * expr.h (extract_bit_field): Remove packedp parameter.
6486 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
6487 from forward declaration.
6488 (store_split_bit_field): Remove packedp arg from calls to
6489 extract_fixed_bit_field.
6490 (extract_bit_field_1): Remove packedp parameter and packedp
6491 argument from recursive calls and calls to extract_fixed_bit_field.
6492 (extract_bit_field): Remove packedp parameter and corresponding
6493 arg to extract_bit_field_1.
6494 (extract_fixed_bit_field): Remove packedp parameter. Remove code
6496 (extract_split_bit_field): Remove packedp arg from call to
6497 extract_fixed_bit_field.
6498 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
6499 (copy_blkmode_from_reg): Likewise.
6500 (copy_blkmode_to_reg): Likewise.
6501 (read_complex_part): Likewise.
6502 (store_field): Likewise.
6503 (expand_expr_real_1): Likewise.
6504 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
6505 to extract_bit_field.
6506 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
6507 call to extract_bit_field.
6508 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
6509 call to extract_bit_field.
6510 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
6511 and special packedp behavior from -fstrict-volatile-bitfields
6514 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6516 * lra-eliminations.c (init_elim_table): Guard value_p.
6518 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6520 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
6521 DFmode, DImode, and SFmode in the upper VSX registers based on the
6522 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
6523 if -mpower8-vector. Combine -mvsx-timode handling with the rest
6524 of the VSX register handling.
6526 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
6528 (zero_extendsidi2_lfiwzx): Add support for loading into the
6529 Altivec registers with -mpower8-vector. Use wu/wv constraints to
6530 only do VSX memory options on Altivec registers.
6531 (extendsidi2_lfiwax): Likewise.
6532 (extendsfdf2_fpr): Likewise.
6533 (mov<mode>_hardfloat, SF/SD modes): Likewise.
6534 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
6535 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
6536 (movdi_internal64): Likewise.
6538 2013-09-27 Xinliang David Li <davidxl@google.com>
6540 * opts.c (finish_options): Adjust parameters
6541 according to vect cost model.
6542 (common_handle_option): Set dynamic vect cost
6544 targhooks.c (default_add_stmt_cost): Compute stmt cost
6546 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
6547 Use helper function.
6548 * tree-vectorizer.h (unlimited_cost_model): New function.
6549 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
6550 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
6552 (vect_enhance_data_refs_alignment): Ditto.
6553 * flag-types.h: New enum.
6554 * common/config/i386/i386-common.c (ix86_option_init_struct):
6555 No need to initialize vect_cost_model flag.
6556 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
6559 2013-09-27 Diego Novillo <dnovillo@google.com>
6561 * gimple.h (enum ssa_mode): Remove.
6563 2013-09-27 Paulo Matos <pmatos@broadcom.com>
6565 * cfgloop.h (number_of_loops): Fix typo in check for null.
6567 2013-09-27 Jakub Jelinek <jakub@redhat.com>
6570 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
6571 are children of outermost saved_cfun's loop, and set it up to
6572 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
6573 if loop != loop0 and SESE region contains bbs that belong to loop0.
6575 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6577 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
6578 (binary_scale_code_p, get_base_term, get_index_term): New functions.
6579 (set_address_segment, set_address_base, set_address_index)
6580 (set_address_disp): Accept the argument unconditionally.
6581 (baseness): Remove must_be_base_p and must_be_index_p checks.
6582 (decompose_normal_address): Classify as much as possible in the
6585 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6587 * cse.c (count_reg_usage): Handle INT_LIST.
6588 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
6589 * reginfo.c (reg_scan_mark_refs): Likewise.
6590 * reload1.c (eliminate_regs_1): Likewise.
6592 2013-09-27 Iain Sandoe <iain@codesourcery.com>
6595 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
6598 2013-09-27 Richard Biener <rguenther@suse.de>
6600 PR tree-optimization/58459
6601 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
6602 restriction not propagating into loops.
6604 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
6606 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
6607 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
6608 * doc/tree-ssa.texi (Walking use-def chains): Delete.
6610 2013-09-26 Richard Biener <rguenther@suse.de>
6612 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
6614 2013-09-26 Richard Biener <rguenther@suse.de>
6616 * alias.h (component_uses_parent_alias_set): Rename to ...
6617 (component_uses_parent_alias_set_from): ... this.
6618 * alias.c (component_uses_parent_alias_set): Rename to ...
6619 (component_uses_parent_alias_set_from): ... this and return
6621 (reference_alias_ptr_type_1): Use the result from
6622 component_uses_parent_alias_set_from instead of stripping
6623 components one at a time.
6624 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
6626 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6628 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
6629 Move prototypes to...
6630 * tree-ssa-ter.h: New File. Move prototypes here.
6631 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
6632 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
6633 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
6634 from is_replaceable_p.
6635 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
6636 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
6637 (process_replaceable): Use ter_is_replaceable_p.
6638 (find_replaceable_in_bb): Use ter_is_replaceable_p.
6639 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
6640 newly refactored ssa_is_replaceable_p.
6641 * cfgexpand.c: Include tree-outof-ssa.h.
6642 * ssaexpand.h: Delete.
6644 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6646 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
6647 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
6649 (create_gimple_tmp): Delete.
6650 (get_expr_type, build_assign, build_type_cast): Move to...
6651 * gimple-builder.c: New File.
6652 (get_expr_type): Relocate from gimple.c.
6653 (build_assign, build_type_cast): Change to only create ssanames.
6654 * gimple.h: Move prototypes to...
6655 * gimple-builder.h: New File. Here.
6656 * tree-ssa.h: And here.
6657 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
6658 count_uses_and_derefs): Relocate from gimple.c.
6659 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
6660 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
6661 * tree-ssa-math-opts (execute_cse_reciprocals): Use
6662 gimple_replace_ssa_lhs.
6663 * asan.c: Include gimple-builder.h.
6664 * Makefile.in: Add gimple-builder.o.
6666 2013-09-26 Richard Biener <rguenther@suse.de>
6668 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
6669 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
6670 (loe_visit_block): Use gcc_checking_assert.
6671 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
6672 gimple_assign_ssa_name_copy_p.
6673 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
6676 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
6678 * config/rs6000/t-rs6000 (rs6000.o): Remove.
6679 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
6681 2013-09-26 Richard Biener <rguenther@suse.de>
6683 PR tree-optimization/58539
6684 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
6685 the fact that debug statements are not taking part in loop-closed
6688 2013-09-26 Nick Clifton <nickc@redhat.com>
6690 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
6691 time warning message.
6692 (msp430_print_operand_raw): Delete unused letter parameter.
6693 (TARGET_PRINT_OPERAND_ADDRESS): Define.
6694 (msp430_print_operand_address): New function.
6695 (msp430_print_operand): Move address printing code from here to
6697 * config/msp430/msp430.md (movsipsi2): Add comment in generated
6699 (zero_extendpsisi2): Likewise.
6700 (extendpsisi2): New pattern.
6701 (andneghi3): New pattern.
6703 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
6705 * config/aarch64/aarch64.opt (mlra): New option.
6706 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
6707 (TARGET_LRA_P): Define.
6709 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
6711 * expr.c (expand_assignment): Remove obsolete comment.
6713 2013-09-25 Jeff Law <law@redhat.com>
6715 * tree-flow.h (thread_through_all_blocks): Prototype moved into
6716 tree-ssa-threadupdate.h.
6717 (register_jump_thread): Similarly.
6718 * tree-ssa-threadupdate.h: New header file.
6719 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
6720 * tree-vrp.c: Likewise.
6721 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
6722 (thread_around_empty_blocks): Change type of path vector argument to
6723 an edge,type pair from just an edge. Initialize both elements when
6724 appending to a jump threading path. Tweak references to elements
6726 (thread_across_edge): Similarly. Release memory for the elements
6728 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
6729 (dump_jump_thread_path): New function broken out from
6730 register_jump_thread.
6731 (register_jump_thread): Use dump_jump_thread_path. Change type of
6732 path vector entries. Search the path for NULL edges and dump
6733 the path if one is found. Tweak the conversion of path to 3-edge
6734 form to use the block copy type information embedded in the path.
6736 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6738 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
6740 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6741 Vladimir Makarov <vmakarov@redhat.com>
6743 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
6744 from the least significant bit.
6745 (strip_address_mutations): Add bitfield operations handling.
6746 (must_be_index_p): Add shifting and rotate operations handling.
6747 (set_address_base): Use must_be_base_p predicate.
6748 (set_address_index): Use must_be_index_p predicate.
6750 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
6751 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6752 Sergey Lega <sergey.s.lega@intel.com>
6753 Anna Tikhonova <anna.tikhonova@intel.com>
6754 Ilya Tocar <ilya.tocar@intel.com>
6755 Andrey Turetskiy <andrey.turetskiy@intel.com>
6756 Ilya Verbin <ilya.verbin@intel.com>
6757 Kirill Yukhin <kirill.yukhin@intel.com>
6758 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6760 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
6762 (ix86_expand_vector_move_misalign): Support new unaligned load and
6763 stores and use new names.
6764 (CODE_FOR_sse2_storedqu): Rename to ...
6765 (CODE_FOR_sse2_storedquv16qi): ... this.
6766 (CODE_FOR_sse2_loaddqu): Rename to ...
6767 (CODE_FOR_sse2_loaddquv16qi): ... this.
6768 (CODE_FOR_avx_loaddqu256): Rename to ...
6769 (CODE_FOR_avx_loaddquv32qi): ... this.
6770 (CODE_FOR_avx_storedqu256): Rename to ...
6771 (CODE_FOR_avx_storedquv32qi): ... this.
6772 * config/i386/i386.md (fpint_logic): New.
6773 * config/i386/sse.md (VMOVE): Extend for AVX512.
6777 (VI_UNALIGNED_LOADSTORE): Ditto.
6778 (sse2_avx_avx512f): Ditto.
6779 (sse2_avx2): Extend for AVX512.
6780 (sse4_1_avx2): Ditto.
6781 (avx2_avx512f): New.
6782 (sse): Extend for AVX512.
6785 (avxsizesuffix): Ditto.
6786 (sseintvecmode): Ditto.
6788 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
6789 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
6790 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6791 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
6792 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6793 (<sse2_avx_avx512f>_storedqu<mode): ... this.
6794 (<sse>_movnt<mode>): Replace constraint "x" with "v".
6795 (STORENT_MODE): Extend for AVX512.
6796 (*absneg<mode>2): Replace constraint "x" with "v".
6797 (*mul<mode>3): Ditto.
6798 (*ieee_smin<mode>3): Ditto.
6799 (*ieee_smax<mode>3): Ditto.
6800 (avx_cmp<mode>3): Replace VF with VF_128_256.
6801 (*<sse>_maskcmp<mode>3_comm): Ditto.
6802 (<sse>_maskcmp<mode>3): Ditto.
6803 (<sse>_andnot<mode>3): Extend for AVX512.
6804 (<code><mode>3, anylogic): Replace VF with VF_128_256.
6805 (<code><mode>3, fpint_logic): New.
6806 (*<code><mode>3): Extend for AVX512.
6807 (avx512flogicsuff): New.
6808 (avx512f_<logic><mode>): Ditto.
6809 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
6811 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
6812 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
6813 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
6814 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
6815 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
6816 (xop_vpermil2<mode>3): Ditto.
6817 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
6818 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
6819 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
6820 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
6822 2013-09-25 Tom Tromey <tromey@redhat.com>
6824 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
6825 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
6826 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
6827 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
6828 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
6829 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
6830 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
6831 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
6832 (GRAPHITE_HTAB_H): Remove.
6834 2013-09-25 Tom Tromey <tromey@redhat.com>
6836 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
6838 2013-09-25 Tom Tromey <tromey@redhat.com>
6840 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
6842 2013-09-25 Tom Tromey <tromey@redhat.com>
6844 * config/i386/t-i386 (i386.o): Remove.
6845 (i386-c.o): Use COMPILE and POSTCOMPILE.
6847 2013-09-25 Tom Tromey <tromey@redhat.com>
6849 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
6851 2013-09-25 Tom Tromey <tromey@redhat.com>
6853 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
6854 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
6855 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
6856 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
6857 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
6858 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
6859 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
6860 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
6861 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
6862 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
6863 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
6864 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
6865 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
6866 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
6867 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
6868 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
6869 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
6870 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
6871 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
6872 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
6873 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
6874 (tree-ssa-pre.o, tree-ssa-sccvn.o)
6875 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
6876 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
6877 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
6878 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
6879 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
6880 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
6881 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
6882 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
6883 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
6884 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
6885 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
6886 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
6887 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
6888 (graphite-clast-to-gimple.o, graphite-dependences.o)
6889 (graphite-interchange.o, graphite-poly.o)
6890 (graphite-scop-detection.o, graphite-sese-to-poly.o)
6891 (graphite-optimize-isl.o, tree-vect-loop.o)
6892 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
6893 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
6894 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
6895 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
6896 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
6897 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
6898 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
6899 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
6900 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
6901 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
6902 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
6903 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
6904 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
6905 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
6906 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
6907 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
6908 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
6909 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
6910 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
6911 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
6912 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
6913 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
6914 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
6915 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
6916 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
6917 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
6918 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
6919 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
6920 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
6921 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
6922 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
6923 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
6924 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
6925 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
6926 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
6927 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
6928 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
6929 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
6930 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
6931 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
6932 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
6933 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
6934 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
6935 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
6936 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
6937 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
6938 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
6939 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
6940 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
6941 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
6942 (gcov-dump.o): Remove.
6943 (default-c.o): Use COMPILE and POSTCOMPILE.
6944 (CFLAGS-gcc.o): New variable.
6945 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
6947 2013-09-25 Tom Tromey <tromey@redhat.com>
6949 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
6950 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
6951 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
6952 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
6953 (c-family/c-pragma.o, c-family/c-pretty-print.o)
6954 (c-family/c-semantics.o, c-family/c-ada-spec.o)
6955 (c-family/array-notation-common.o, c-family/stub-objc.o)
6956 (c-family/c-ubsan.o): Remove.
6958 2013-09-25 Tom Tromey <tromey@redhat.com>
6960 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
6962 2013-09-25 Tom Tromey <tromey@redhat.com>
6964 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
6965 to add -DENABLE_SHARED_LIBGCC.
6966 (gcc.o): Don't use subshell.
6968 2013-09-25 Tom Tromey <tromey@redhat.com>
6970 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
6971 * configure.ac: Don't invoke AM_PROG_CC_C_O.
6972 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
6973 * configure, config.in: Rebuild.
6975 2013-09-25 Tom Tromey <tromey@redhat.com>
6977 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
6978 (COMPILE, POSTCOMPILE): New variables.
6979 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
6980 (DEPFILES): New variable.
6981 Include ".Po" files.
6982 * configure.ac: Add checks for dependency checking.
6983 * configure, aclocal.m4: Regenerate.
6985 2013-09-25 Tom Tromey <tromey@redhat.com>
6987 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
6988 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
6990 2013-09-25 Tom Tromey <tromey@redhat.com>
6992 * Makefile.in (generated_files): Add options.h,
6993 target-hooks-def.h, insn-opinit.h,
6994 common/common-target-hooks-def.h, pass-instances.def,
6995 c-family/c-target-hooks-def.h.
6997 2013-09-25 Jeff Law <law@redhat.com>
6999 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
7000 than foo[foo.length () - 1] to access last member in a vec.
7001 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
7003 2013-09-25 Richard Biener <rguenther@suse.de>
7006 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
7008 2013-09-25 Jan Hubicka <jh@suse.cz>
7010 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
7013 2013-09-25 Marek Polacek <polacek@redhat.com>
7016 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
7017 when determining the type name.
7019 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
7021 * config/sh/sh.md: Fix formatting.
7023 2013-09-24 Xinliang David Li <davidxl@google.com>
7025 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
7026 max peel iterations parameter.
7027 * param.def: New parameter.
7028 * doc/invoke.texi: Document New parameter.
7030 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
7032 * gimple-pretty-print.c: Various whitespace tweaks.
7033 * tree-core.h: Likewise.
7034 * tree-pretty-print.c: Likewise.
7035 * tree-ssa-alias.c: Likewise.
7036 * tree-ssa-copy.c: Likewise.
7037 * tree-ssanames.c: Likewise.
7038 * tree-ssanames.h: Likewise.
7039 * tree-vrp.c: Likewise.
7041 2013-09-24 Alan Modra <amodra@gmail.com>
7045 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
7046 for output operands that disallow regs. Don't use EXPAND_WRITE on
7049 2013-09-24 Richard Biener <rguenther@suse.de>
7052 * tree.c (reference_alias_ptr_type): Move ...
7053 * alias.c (reference_alias_ptr_type): ... here and implement
7054 in terms of the new reference_alias_ptr_type_1.
7055 (ref_all_alias_ptr_type_p): New helper.
7056 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
7057 use ref_all_alias_ptr_type_p.
7058 (get_deref_alias_set): Add flag_strict_aliasing check here.
7059 (reference_alias_ptr_type_1): New function, split out from ...
7060 (get_alias_set): ... here.
7061 (alias_ptr_types_compatible_p): New function.
7062 * alias.h (reference_alias_ptr_type): Declare.
7063 (alias_ptr_types_compatible_p): Likewise.
7064 * tree.h (reference_alias_ptr_type): Remove.
7065 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
7066 to compare MEM_REF alias types.
7068 2013-09-24 Richard Biener <rguenther@suse.de>
7070 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
7072 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7074 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
7075 reload helper function arrays into a single array reg_addr.
7076 (reload_fpr_gpr): Likewise.
7077 (reload_gpr_vsx): Likewise.
7078 (reload_vsx_gpr): Likewise.
7079 (struct rs6000_reg_addr): Likewise.
7080 (reg_addr): Likewise.
7081 (rs6000_debug_reg_global): Change rs6000_vector_reload,
7082 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
7083 (rs6000_init_hard_regno_mode_ok): Likewise.
7084 (rs6000_secondary_reload_direct_move): Likewise.
7085 (rs6000_secondary_reload): Likewise.
7087 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
7088 constraints: wu, ww, and wy. Repurpose wv constraint added during
7089 power8 changes. Put wg constraint in alphabetical order.
7091 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
7092 for future work to add ISA 2.07 VSX single precision support.
7093 (-mvsx-scalar-double): Change default from -1 to 1, update
7094 documentation comment.
7095 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
7096 (-mupper-regs-df): New debug switch to control whether DF values
7097 can go in the traditional Altivec registers.
7098 (-mupper-regs-sf): New debug switch to control whether SF values
7099 can go in the traditional Altivec registers.
7101 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
7103 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
7104 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
7105 Add new constraints, wu/ww/wy. Repurpose wv constraint.
7106 (rs6000_debug_legitimate_address_p): Print if we are running
7107 before, during, or after reload.
7108 (rs6000_secondary_reload): Add a comment.
7109 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
7111 * config/rs6000/constraints.md (wa constraint): Sort w<x>
7112 constraints. Update documentation string.
7113 (wd constraint): Likewise.
7114 (wf constraint): Likewise.
7115 (wg constraint): Likewise.
7116 (wn constraint): Likewise.
7117 (ws constraint): Likewise.
7118 (wt constraint): Likewise.
7119 (wx constraint): Likewise.
7120 (wz constraint): Likewise.
7121 (wu constraint): New constraint for ISA 2.07 SFmode scalar
7123 (ww constraint): Likewise.
7124 (wy constraint): Likewise.
7125 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
7126 in the previous submissions.
7127 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
7129 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
7131 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
7132 (REG_BR_PROB): Say that the probability is stored in an int_list.
7133 * reg-notes.def: Update commentary to mention INT_LIST.
7134 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
7135 (INT_LIST): New rtx.
7136 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
7137 * rtlanal.c (int_reg_note_p): New function.
7138 (alloc_reg_note): Assert that the note does not have an int argument.
7139 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
7140 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
7141 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
7142 rather than an INSN_LIST. Handle INT_LIST.
7143 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
7144 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
7145 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
7146 Manipulate them as ints rather than rtxes.
7147 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
7148 * regmove.c (copy_src_to_dest): Likewise.
7149 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
7151 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
7152 into the cases that need it.
7153 * config/arm/arm.c (arm_unwind_emit): Likewise.
7155 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
7156 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
7157 * loop-doloop.c (add_test, doloop_modify): Likewise.
7158 * loop-unswitch.c (compare_and_jump_seq): Likewise.
7159 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
7160 * predict.c (combine_predictions_for_insn): Likewise.
7161 * print-rtl.c (print_rtx): Handle INT_LIST.
7162 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
7163 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
7164 * config/arm/arm.c (emit_unlikely_jump): Likewise.
7165 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
7166 (ix86_split_fp_branch, predict_jump): Likewise.
7167 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
7168 * config/sh/sh.c (expand_cbranchsi4): Likewise.
7169 * config/spu/spu.c (ea_load_store_inline): Likewise.
7171 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
7172 value of a REG_BR_PROB note.
7173 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
7174 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
7175 * emit-rtl.c (try_split): Likewise.
7176 * predict.c (br_prob_note_reliable_p): Likewise.
7177 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
7178 * reorg.c (mostly_true_jump): Likewise.
7179 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
7180 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
7181 * config/i386/i386.c (ix86_print_operand): Likewise.
7182 * config/ia64/ia64.c (ia64_print_operand): Likewise.
7183 * config/mmix/mmix.c (mmix_print_operand): Likewise.
7184 * config/rs6000/rs6000.c (output_cbranch): Likewise.
7185 * config/s390/s390.c (s390_expand_tbegin): Likewise.
7186 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
7187 * config/sparc/sparc.c (output_cbranch): Likewise.
7188 * config/spu/spu.c (get_branch_target): Likewise.
7189 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
7190 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
7192 2013-09-23 Jan Hubicka <jh@suse.cz>
7194 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
7196 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
7197 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
7199 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
7200 (gimple_fold_call): Dump inconsistent devirtualizations; add
7201 sanity check for type based devirtualizations.
7202 * ipa-prop.c: Include ipa-utils.h
7203 (ipa_intraprocedural_devirtualization): Add sanity check.
7204 (try_make_edge_direct_virtual_call): Likewise.
7206 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
7208 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
7209 assignment statements.
7211 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
7213 * gimple-pretty-print.c (dump_ssaname_info): New function.
7214 (dump_gimple_phi): Call it.
7215 (pp_gimple_stmt_1): Likewise.
7216 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
7217 (range_info_def): Declare.
7218 * tree-pretty-print.c (pp_double_int): New function.
7219 (dump_generic_node): Call it.
7220 * tree-pretty-print.h (pp_double_int): Declare.
7221 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
7222 * tree-ssanames.h (range_info_def): New structure.
7223 (value_range_type): Move definition here.
7224 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
7226 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
7228 (set_range_info): New function.
7229 (get_range_info): Likewise.
7230 (duplicate_ssa_name_range_info): Likewise.
7231 (duplicate_ssa_name_fn): Check pointer type and call
7232 duplicate_ssa_name_range_info.
7233 * tree-ssa-copy.c (fini_copy_prop): Likewise.
7234 * tree-vrp.c (value_range_type): Remove definition, now in
7236 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
7237 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
7238 (SSA_NAME_RANGE_INFO): New macro.
7240 2013-09-23 Richard Biener <rguenther@suse.de>
7242 PR tree-optimization/58464
7243 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7244 (phi_trans_add): Change to add conditionally on being not present.
7245 (phi_translate_1): Remove recursion detection here.
7246 (phi_translate): Pre-seed the cache with NULL to catch
7247 recursion here in a more generic way.
7248 (bitmap_find_leader): Adjust comment.
7249 (get_representative_for): Dump value-numbers.
7250 (create_expression_by_pieces): Likewise.
7251 (insert_into_preds_of_block): Likewise.
7253 2013-09-23 Christian Bruel <christian.bruel@st.com>
7256 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
7257 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
7259 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
7262 2013-09-20 Renlin Li <renlin.li@arm.com>
7264 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
7266 (aarch64_expand_epilogue): Likewise.
7267 (aarch64_legitimize_reload_address): Likewise.
7269 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
7271 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
7272 NULL_TREE before pushing them onto the vector. Likewise for labels.
7274 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
7276 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
7277 comparison operators when -fno-trapping-math is in effect.
7278 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
7279 comparison operators in TFmode and assert that unsupported operators
7281 (ia64_print_operand): Likewise.
7283 2013-09-21 Jan Hubicka <jh@suse.cz>
7285 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
7286 (sse_typeless_stores): Enable for core
7287 (sse_load0_by_pxor): Likewise.
7288 (four_jump_limit): Disable for core.
7289 (pad_returns): Likewise.
7290 (avoid_vector_decode): Likewise.
7291 (fuse_cmp_and_branch): Enable for cores.
7292 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
7294 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7297 * config/pa/pa.c (pa_option_override): Disable auto increment and
7298 decrement instructions until reload is completed.
7300 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
7301 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
7302 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
7304 2013-09-20 DJ Delorie <dj@redhat.com>
7305 Nick Clifton <nickc@redhat.com>
7307 * config/rl78/rl78.c: Various whitespace and comment tweaks.
7308 (need_to_save): Save bank 0 on interrupts.
7309 (characterize_address): Strip far address wrappers.
7310 (rl78_as_legitimate_address): Likewise.
7311 (transcode_memory_rtx): Likewise.
7312 (rl78_peep_movhi_p): Disable this peephole after devirt.
7313 (rl78_propogate_register_origins): Forget all origins when a
7315 * config/rl78/rl78-virt.md: Various whitespace tweaks.
7316 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
7318 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
7319 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
7320 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
7322 * config/rl78/constraints.md: Various whitespace and paren tweaks.
7324 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7326 * config/pa/pa.md: In "scc" insn patterns, change output template to
7327 handle const0_rtx in reg_or_0_operand operands.
7329 2013-09-20 Martin Husemann <martin@NetBSD.org>
7332 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
7333 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
7335 2013-09-20 Richard Biener <rguenther@suse.de>
7338 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
7339 remembering SSA name version and block index.
7340 (new_scev_info_str): Adjust.
7341 (hash_scev_info): Likewise. Also hash the block index.
7342 (eq_scev_info): Adjust.
7343 (find_var_scev_info): Likewise.
7344 (struct instantiate_cache_entry): Remove.
7345 (struct instantiate_cache_type): Use a htab to map name, block
7347 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7348 (get_instantiated_value_entry): Likewise.
7349 (hash_idx_scev_info, eq_idx_scev_info): New functions.
7350 (instantiate_scev_name): Adjust.
7352 2013-09-20 Jeff Law <law@redhat.com>
7354 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
7356 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
7358 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
7359 Call aarch64_simd_expand_args to update op[argc].
7361 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7363 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
7366 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7368 * gengtype.c (file_rules): Added rule for *.cc files.
7369 (get_output_file_with_visibility): Give fatal message when no
7372 2013-09-20 Renlin Li <renlin.li@arm.com>
7374 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
7375 (aarch64_expand_epilogue): Likewise.
7376 (aarch64_legitimize_reload_address): Likewise.
7378 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
7381 * expr.c (expand_assignment): Remove misalignp code path.
7383 2013-09-20 Marek Polacek <polacek@redhat.com>
7386 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
7387 TYPE_PRECISION. Add asserts.
7389 2013-09-20 Richard Biener <rguenther@suse.de>
7391 PR tree-optimization/58453
7392 * tree-loop-distribution.c (distribute_loop): Apply the cost
7393 model for -ftree-loop-distribute-patterns, too.
7395 2013-09-20 Richard Biener <rguenther@suse.de>
7398 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
7399 make type comparison less strict.
7401 2013-09-20 Alan Modra <amodra@gmail.com>
7403 * configure: Regenerate.
7404 * aclocal.m4: Regenerate.
7406 2013-09-20 Marek Polacek <polacek@redhat.com>
7409 * doc/extend.texi: Document that attribute used is meant to be used
7410 on variables with static storage duration.
7412 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7414 PR tree-optimization/58472
7415 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
7416 simd_lane_access set inv_p = false.
7417 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
7418 the simduid magic VAR_DECL.
7420 2013-09-19 Jan Hubicka <jh@suse.cz>
7422 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
7424 2013-09-17 Jeff Law <law@redhat.com>
7426 * tree-ssa-dom.c (record_temporary_equivalences): New function
7427 split out of dom_opt_dom_walker::after_dom_children.
7428 (dom_opt_dom_walker::thread_across_edge): Move common code
7429 in here from dom_opt_dom_walker::after_dom_children.
7430 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
7432 2013-09-19 Jan Hubicka <jh@suse.cz>
7434 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
7435 (TARGET_GENERIC): Use PROCESOR_GENERIC
7436 (enum processor_type): Unify generic32 and 64.
7437 * i386.md (cpu): Likewise.
7438 * x86-tune.def (use_leave): Enable for generic32.
7439 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
7440 * athlon.md: Change generic64 to generic in all occurences.
7441 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
7442 (ix86_target_macros_internal): Likewise.
7443 * driver-i386.c (host_detect_local_cpu): Likewise.
7444 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
7446 (generic_memcpy, generic_memset, generic_cost): This one.
7447 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
7448 (m_GENERIC32, m_GENERIC64): Remove.
7449 (m_GENERIC): Turn into one flag.
7450 (processor_target): Unify generic tunnings.
7451 (ix86_option_override_internal): Replace generic32/64 by generic.
7452 (ix86_issue_rate): Likewise.
7453 (ix86_adjust_cost): Likewise.
7455 2013-09-19 Jan Hubicka <jh@suse.cz>
7457 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
7458 of speculative flag.
7460 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7462 * omp-low.c (expand_omp_sections): Always pass len - 1 to
7463 GOMP_sections_start, even if !exit_reachable.
7465 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7467 * lra-constraints.c (need_for_all_save_p): Use macro
7468 HARD_REGNO_CALL_PART_CLOBBERED.
7469 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
7470 set up pseudo conflict hard regs.
7472 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7475 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
7478 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7480 PR rtl-optimization/58438
7481 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
7482 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
7483 from previous subpasses.
7485 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
7487 * arm.c (arm_get_frame_offsets): Validate architecture supports
7488 LDRD/STRD before accepting the tuning preference.
7489 (arm_expand_prologue): Likewise.
7490 (arm_expand_epilogue): Likewise.
7492 2013-09-18 Richard Biener <rguenther@suse.de>
7494 PR tree-optimization/58417
7495 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
7496 have chrecs with symbols defined in the loop as operands.
7497 (chrec_fold_multiply): Likewise.
7498 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
7499 parameters before folding binary operations.
7500 (struct instantiate_cache_entry_hasher): Remove.
7501 (struct instantiate_cache_type): Use a pointer-map.
7502 (instantiate_cache_type::instantiate_cache_type): New function.
7503 (instantiate_cache_type::get): Likewise.
7504 (instantiate_cache_type::set): Likewise.
7505 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7506 (get_instantiated_value_entry): Likewise.
7507 (global_cache): New global.
7508 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
7509 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
7510 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
7511 (instantiate_scev_name): Adjust.
7512 (instantiate_scev): Construct global instead of local cache.
7513 (resolve_mixers): Likewise.
7515 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
7516 Paolo Carlini <paolo.carlini@oracle.com>
7519 * doc/implement-cxx.texi: Fix references to the C++ standards.
7521 2013-09-18 Jakub Jelinek <jakub@redhat.com>
7523 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
7524 * tree-vect-data-refs.c (vect_analyze_data_refs): For
7525 simd_lane_access drs, update also DR_ALIGNED_TO.
7527 2013-09-18 Marek Polacek <polacek@redhat.com>
7530 * doc/extend.texi: Document no_sanitize_undefined attribute.
7531 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
7532 no_sanitize_undefined attribute.
7534 2013-09-18 Nick Clifton <nickc@redhat.com>
7536 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
7537 (ASM_DECLARE_FUNCTION_NAME): Define.
7539 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
7541 * compare-elim.c (find_comparison_dom_walker): New class
7542 (find_comparisons_in_bb): Rename to
7543 find_comparison_dom_walker::before_dom_children
7544 (find_comparisons): Adjust
7545 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
7547 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
7548 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
7549 (init_walk_dominator_tree): Remove declaration.
7550 (fini_walk_dominator_tree): Remove declaration.
7551 * fwprop.c (single_def_use_dom_walker): New class
7552 (single_def_use_enter_block): Convert to
7553 single_def_use_dom_walker::before_dom_children.
7554 (single_def_use_leave_block): Convert to
7555 single_def_use_dom_walker::after_dom_children.
7556 (build_single_def_use_links): Adjust.
7557 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
7559 (find_candidates_in_block): Convert to
7560 find_candidates_dom_walker::before_dom_children.
7561 (execute_strength_reduction): Adjust.
7562 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
7563 (sese_dom_walker): New class.
7564 (sese_dom_walker::sese_dom_walker): New constructor.
7565 (sese_dom_walker::~sese_dom_walker): New destructor.
7566 (build_sese_conditions_before): Convert to
7567 sese_dom_walker::before_dom_children.
7568 (build_sese_conditions_after): Convert to
7569 sese_dom_walker::after_dom_children.
7570 (build_poly_scop): Adjust
7571 * tree-into-ssa.c (rewrite_dom_walker): New class
7572 (rewrite_enter_block): Convert to
7573 rewrite_dom_walker::before_dom_children.
7574 (rewrite_leave_block): Convert to
7575 rewrite_dom_walker::after_dom_children.
7576 (rewrite_update_dom_walker): New class.
7577 (rewrite_update_enter_block): Convert to
7578 rewrite_update_dom_walker::before_dom_children.
7579 (rewrite_update_leave_block): Convert to
7580 rewrite_update_dom_walker::after_dom_children.
7581 (rewrite_blocks, rewrite_into_ssa): Adjust.
7582 (mark_def_dom_walker): New class.
7583 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
7584 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
7585 (mark_def_sites_blocks): Convert to
7586 mark_def_dom_walker::before_dom_children.
7587 (mark_def_site_blocks): Remove.
7588 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
7589 (tree_ssa_dominator_optimize): Adjust.
7590 (dom_thread_across_edge): Convert to method
7591 dom_opt_dom_walker::thread_across_edge.
7592 (dom_opt_enter_block): Convert to member function
7593 dom_opt_dom_walker::before_dom_children.
7594 (dom_opt_leave_block): Convert to member function
7595 dom_opt_dom_walker::after_dom_children.
7596 * tree-ssa-dse.c (dse_dom_walker): New class.
7597 (dse_enter_block): Convert to member function
7598 dse_dom_walker::before_dom_children.
7599 (tree_ssa_dse): Adjust.
7600 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
7601 (determine_invariantness_stmt): Convert to method
7602 invariantness_dom_walker::before_dom_children.
7603 (determine_invariantness): Remove
7604 (move_computations_dom_walker): New class.
7605 (move_computations_stmt): Convert to method
7606 move_computations_dom_walker::before_dom_children.
7607 (move_computations, tree_ssa_lim): Adjust.
7608 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
7609 (nt_init_block): Convert to method
7610 notrappping_dom_walker::before_dom_children.
7611 (nt_fini_block): Convert to method
7612 method nontrapping_dom_walker::after_dom_children.
7613 (get_non_trapping): Adjust.
7614 * tree-ssa-pre.c (eliminate_dom_walker): New class.
7615 (eliminate_bb): Convert to method
7616 eliminate_dom_walker::before_dom_children.
7617 (eliminate_leave_block): Convert to method
7618 eliminate_dom_walker::after_dom_children.
7619 (eliminate): Adjust.
7620 * tree-ssa-strlen.c (strlen_dom_walker): New class.
7621 (strlen_enter_block): Convert to method
7622 strlen_dom_walker::before_dom_children.
7623 (strlen_leave_block): Convert to method
7624 method strlen_dom_walker::after_dom_children.
7625 (tree_ssa_strlen): Adjust.
7626 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
7627 (tree_ssa_uncprop): Adjust.
7628 (uncprop_leave_block): Convert to method
7629 uncprop_dom_walker::after_dom_children.
7630 (uncprop_leave_block): Convert to method
7631 uncprop_dom_walker::before_dom_children.
7633 2013-09-18 Bin Cheng <bin.cheng@arm.com>
7635 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
7636 branch in same basic block. Check both src and dest of the move insn.
7638 2013-09-17 Nick Clifton <nickc@redhat.com>
7640 * config/rl78/rl78-real.md (bf): New pattern.
7642 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
7643 (rl78_print_operand): Do not put a # before a %B.
7644 * config/rl78/rl78.opt: Tweak doc strings.
7646 2013-09-17 DJ Delorie <dj@redhat.com>
7648 * config/rl78/constraints.md (Wcv): Allow up to $r31.
7649 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
7650 (rl78_option_override): Likewise, if -mallregs.
7651 (is_virtual_register): Likewise.
7652 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
7653 (REGNO_OK_FOR_BASE_P): Likewise.
7654 * config/rl78/rl78.opt (-mallregs): New.
7656 2013-09-17 Nick Clifton <nickc@redhat.com>
7658 * config/rl78/rl78.c (need_to_save): Change return type to bool.
7659 For interrupt functions: save all call clobbered registers if the
7660 interrupt handler is not a leaf function.
7661 (rl78_expand_prologue): Always recompute the frame information.
7662 For interrupt functions: only select bank 0 if one of the bank 0
7663 registers is going to be psuhed.
7665 2013-09-17 DJ Delorie <dj@redhat.com>
7667 * config/rl78/constraints.md: For each W* constraint, rename to C*
7668 and create a W* constraint that checks for an optional ES: prefix
7670 * config/rl78/rl78.md (UNS_ES_ADDR): New.
7671 (es_addr): New. Used to wrap far addresses.
7672 * config/rl78/rl78-protos.h (rl78_es_addr): New.
7673 (rl78_es_base): New.
7674 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
7675 wrapped far addresses.
7676 (rl78_print_operand_1): Unwrap far addresses before processing.
7677 (rl78_lo16): Wrap far addresses in unspecs.
7678 (rl78_es_addr): New.
7679 (rl78_es_base): New.
7680 (insn_ok_now): Check for not-yet-wrapped far addresses.
7681 (transcode_memory_rtx): Properly re-wrap far addresses.
7683 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
7685 * config/sparc/t-rtems: Add leon3 multilibs.
7687 2013-09-17 Cong Hou <congh@google.com>
7689 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
7690 when checking the dot production pattern. The type of rhs operand
7691 of multiply is now checked correctly.
7693 2013-09-17 Jeff Law <law@redhat.com>
7695 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
7696 edge implied equivalences into successor phis.
7697 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
7698 here from tree-ssa-threadedge.c.
7699 (mark_threaded_blocks): When threading through a joiner, if both
7700 successors of the joiner's clone reach the same block, verify the
7701 PHI arguments are equal. If not, cancel the jump threading request.
7702 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
7703 tree-ssa-threadupdate.c
7704 (thread_across_edge): Don't check PHI argument equality when
7705 threading through joiner block here.
7707 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
7709 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
7710 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
7711 (warn_uninit, warn_uninitialized_vars,
7712 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7713 Move to tree-ssa-uninit.c.
7714 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
7715 (has_undefined_value_p): New. Pass dependant parts of
7716 ssa_undefined_value_p.
7717 (uninit_undefined_value_p): Use has_undefined_value_p.
7718 (warn_uninit, warn_uninitialized_vars,
7719 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7720 Move from tree-ssa.c.
7721 * tree-ssa.h: Adjust prototypes.
7723 2013-09-17 Jan Hubicka <jh@suse.cz>
7726 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
7727 * ipa-inline.c (can_inline_edge_p): Do not downgrade
7728 FUNCTION_NOT_OPTIMIZED.
7729 * ipa-inline-analysis.c (compute_inline_parameters): Function
7730 not optimized is not inlinable unless it is alwaysinline.
7731 (inline_analyze_function): Force calls in not optimized
7732 function not inlinable.
7734 2013-09-17 Jan Hubicka <jh@suse.cz>
7737 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
7739 * ipa.c (function_and_variable_visibility): Likewise.
7740 * ipa-profile.c (ipa_profile): Likewise.
7742 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
7745 * cgraph.c (cgraph_function_body_availability): Check for ifunc
7746 attribute, and don't inline the resolver in this case.
7748 2013-09-17 Teresa Johnson <tejohnson@google.com>
7750 * coverage.c (get_coverage_counts): Add missing newline.
7752 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7754 PR tree-optimization/58088
7755 * fold-const.c (mask_with_trailing_zeros): New function.
7756 (fold_binary_loc): Make sure we don't recurse infinitely
7757 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
7758 Use mask_with_trailing_zeros where appropriate.
7760 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
7762 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
7763 of 'prev' var to get better distance estimation.
7765 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
7767 * tree-inline.h (struct copy_body_data): Add transform_parameter.
7768 * tree-inline.c (is_parameter_of): New predicate.
7769 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
7770 a parameter has been remapped.
7771 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
7772 (optimize_inline_calls): Initialize transform_parameter.
7773 (copy_gimple_seq_and_replace_locals): Likewise.
7774 (tree_function_versioning): Likewise.
7775 (maybe_inline_call_in_expr): Likewise.
7777 2013-09-17 Nick Clifton <nickc@redhat.com>
7779 * config/msp430/msp430-protos.h: Add prototypes for new functions.
7780 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
7782 (is_attr_func): New function.
7783 (msp430_is_interrupt_func): New function.
7784 (is_naked_func): New function.
7785 (is_reentrant_func): New function.
7786 (is_critical_func): New function.
7787 (msp430_start_function): Add annotations for function attributes.
7788 (msp430_attr): New function.
7789 (msp430_attribute_table): New.
7790 (msp430_function_section): New function.
7791 (TARGET_ASM_FUNCTION_SECTION): Define.
7792 (msp430_builtin): New enum.
7793 (msp430_init_builtins): New function.
7794 (msp430_builtin_devl): New function.
7795 (msp430_expand_builtin): New function.
7796 (TARGET_INIT_BUILTINS): Define.
7797 (TARGET_EXPAND_BUILTINS): Define.
7798 (TARGET_BUILTIN_DECL): Define.
7799 (msp430_expand_prologue): Add support for naked, interrupt,
7800 critical and reentrant functions.
7801 (msp430_expand_epilogue): Likewise.
7802 (msp430_print_operand): Handle 'O' character.
7803 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
7805 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
7806 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
7807 (pushm): Use a 'n' rather than an 'i' constraint.
7808 (msp_return): Add generation of the interrupt return instruction.
7809 (disable_interrupts): New pattern.
7810 (enable_interrupts): New pattern.
7811 (push_intr_state): New pattern.
7812 (pop_intr_state): New pattern.
7813 (bic_SR): New pattern.
7814 (bis_SR): New pattern.
7815 * doc/extend.texi: Document MSP430 function attributes and builtin
7818 2013-09-17 Richard Biener <rguenther@suse.de>
7820 PR tree-optimization/58432
7821 * tree-loop-distribution.c (tree_loop_distribution): Also
7822 scan PHIs for outside loop uses and seed a partition from them.
7824 2013-09-17 Bin Cheng <bin.cheng@arm.com>
7826 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
7827 (restructure_reference): Call backtrace_base_for_ref.
7829 2013-09-17 Alan Modra <amodra@gmail.com>
7832 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
7835 2013-09-16 DJ Delorie <dj@redhat.com>
7837 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
7838 vregs location for RL78/G10.
7839 (rl78_expand_prologue): Avoid SEL on G10.
7840 (rl78_expand_epilogue): Likewise.
7841 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
7842 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
7843 __RL78_G10__ when appropriate.
7844 (ASM_SPEC): Pass -mg10 along to the assembler.
7845 * config/rl78/rl78.md (sel_rb): Disable for G10.
7846 * config/rl78/rl78.opt: Add -mg10 option.
7847 * config/rl78/t-rl78: Add -mg10 multilib.
7849 2013-09-16 Xinliang David Li <davidxl@google.com>
7851 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
7852 * omp-low.c (omp_max_vf): Ditto.
7853 (expand_omp_simd): Ditto.
7854 * tree-vectorizer.c (vectorize_loops): Ditto.
7855 (gate_vect_slp): Ditto.
7856 (gate_increase_alignment): Ditto.
7857 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
7858 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
7859 (gate_tree_vectorize): Name change.
7860 (tree_vectorize): Ditto.
7861 (pass_vectorize::gate): Call new function.
7862 (pass_vectorize::execute): Ditto.
7863 * opts.c: O3 default setting change.
7864 (finish_options): Check new flag.
7865 * doc/invoke.texi: Document new flags.
7866 * common.opt: New flags.
7868 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
7870 * doc/tm.texi.in (Cond Exec Macros): Remove node.
7871 (Condition Code): Don't reference it.
7872 * doc/tm.texi: Regenerate.
7874 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
7877 * lra-constraints.c (undo_optional_reloads): Consider all optional
7878 reload even if it did not get a hard reg.
7880 2013-09-16 Teresa Johnson <tejohnson@google.com>
7882 * dumpfile.c (dump_loc): Remove newline emission.
7883 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
7884 emission to dump_printf_loc calls where missing.
7885 (vect_mark_for_runtime_alias_test): Ditto.
7886 (vect_analyze_data_ref_dependence): Ditto.
7887 (vect_analyze_data_ref_dependences): Ditto.
7888 (vect_slp_analyze_data_ref_dependence): Ditto.
7889 (vect_slp_analyze_data_ref_dependences): Ditto.
7890 (vect_compute_data_ref_alignment): Ditto.
7891 (vect_update_misalignment_for_peel): Ditto.
7892 (vect_verify_datarefs_alignment): Ditto.
7893 (vector_alignment_reachable_p): Ditto.
7894 (vect_get_data_access_cost): Ditto.
7895 (vect_enhance_data_refs_alignment): Ditto.
7896 (vect_find_same_alignment_drs): Ditto.
7897 (vect_analyze_data_refs_alignment): Ditto.
7898 (vect_analyze_group_access): Ditto.
7899 (vect_analyze_data_ref_access): Ditto.
7900 (vect_analyze_data_ref_accesses): Ditto.
7901 (vect_prune_runtime_alias_test_list): Ditto.
7902 (vect_analyze_data_refs): Ditto.
7903 (vect_create_addr_base_for_vector_ref): Ditto.
7904 (vect_create_data_ref_ptr): Ditto.
7905 (vect_grouped_store_supported): Ditto.
7906 (vect_grouped_load_supported): Ditto.
7907 * value-prof.c (check_counter): Ditto.
7908 (check_ic_target): Ditto.
7909 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
7910 (vect_recog_widen_mult_pattern): Ditto.
7911 (vect_recog_widen_sum_pattern): Ditto.
7912 (vect_recog_over_widening_pattern): Ditto.
7913 (vect_recog_widen_shift_pattern): Ditto.
7914 (vect_recog_rotate_pattern): Ditto.
7915 (vect_recog_vector_vector_shift_pattern): Ditto.
7916 (vect_recog_divmod_pattern): Ditto.
7917 (vect_recog_mixed_size_cond_pattern): Ditto.
7918 (vect_recog_bool_pattern): Ditto.
7919 (vect_pattern_recog_1): Ditto.
7920 (vect_pattern_recog): Ditto.
7921 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
7922 (vect_is_simple_iv_evolution): Ditto.
7923 (vect_analyze_scalar_cycles_1): Ditto.
7924 (vect_get_loop_niters): Ditto.
7925 (vect_analyze_loop_1): Ditto.
7926 (vect_analyze_loop_form): Ditto.
7927 (vect_analyze_loop_operations): Ditto.
7928 (vect_analyze_loop_2): Ditto.
7929 (vect_analyze_loop): Ditto.
7930 (report_vect_op): Ditto.
7931 (vect_is_slp_reduction): Ditto.
7932 (vect_is_simple_reduction_1): Ditto.
7933 (vect_get_known_peeling_cost): Ditto.
7934 (vect_estimate_min_profitable_iters): Ditto.
7935 (vect_model_reduction_cost): Ditto.
7936 (vect_model_induction_cost): Ditto.
7937 (get_initial_def_for_induction): Ditto.
7938 (vect_create_epilog_for_reduction): Ditto.
7939 (vectorizable_reduction): Ditto.
7940 (vectorizable_induction): Ditto.
7941 (vectorizable_live_operation): Ditto.
7942 (vect_loop_kill_debug_uses): Ditto.
7943 (vect_transform_loop): Ditto.
7944 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
7945 (vect_stmt_relevant_p): Ditto.
7946 (process_use): Ditto.
7947 (vect_mark_stmts_to_be_vectorized): Ditto.
7948 (vect_model_simple_cost): Ditto.
7949 (vect_model_promotion_demotion_cost): Ditto.
7950 (vect_model_store_cost): Ditto.
7951 (vect_get_store_cost): Ditto.
7952 (vect_model_load_cost): Ditto.
7953 (vect_get_load_cost): Ditto.
7954 (vect_init_vector_1): Ditto.
7955 (vect_get_vec_def_for_operand): Ditto.
7956 (vect_finish_stmt_generation): Ditto.
7957 (vectorizable_call): Ditto.
7958 (vectorizable_conversion): Ditto.
7959 (vectorizable_assignment): Ditto.
7960 (vectorizable_shift): Ditto.
7961 (vectorizable_operation): Ditto.
7962 (vectorizable_store): Ditto.
7963 (vectorizable_load): Ditto.
7964 (vectorizable_condition): Ditto.
7965 (vect_analyze_stmt): Ditto.
7966 (vect_transform_stmt): Ditto.
7967 (vect_is_simple_use): Ditto.
7968 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
7969 (vect_can_advance_ivs_p): Ditto.
7970 (vect_update_ivs_after_vectorizer): Ditto.
7971 (vect_do_peeling_for_loop_bound): Ditto.
7972 (vect_gen_niters_for_prolog_loop): Ditto.
7973 (vect_update_inits_of_drs): Ditto.
7974 (vect_create_cond_for_alias_checks): Ditto.
7975 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
7976 (vect_build_slp_tree_1): Ditto.
7977 (vect_supported_load_permutation_p): Ditto.
7978 (vect_analyze_slp_instance): Ditto.
7979 (vect_analyze_slp): Ditto.
7980 (vect_make_slp_decision): Ditto.
7981 (vect_detect_hybrid_slp): Ditto.
7982 (vect_bb_vectorization_profitable_p): Ditto.
7983 (vect_slp_analyze_bb_1): Ditto.
7984 (vect_update_slp_costs_according_to_vf): Ditto.
7985 (vect_get_mask_element): Ditto.
7986 (vect_transform_slp_perm_load): Ditto.
7987 (vect_schedule_slp_instance): Ditto.
7988 (vect_schedule_slp): Ditto.
7989 (vect_slp_transform_bb): Ditto.
7990 * profile.c (read_profile_edge_counts): Ditto.
7991 (compute_branch_probabilities): Ditto.
7992 * coverage.c (get_coverage_counts): Ditto.
7994 2013-09-16 Diego Novillo <dnovillo@google.com>
7996 * tree-core.h: Add missing comment lines from refactoring of tree.h.
7998 2013-09-16 Jan Hubicka <jh@suse.cz>
8000 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
8001 abstract functions; for static functions check the presence of body.
8003 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8005 * config/aarch64/aarch64-simd-builtins.def (fma): New.
8006 * config/aarch64/aarch64-simd.md
8007 (aarch64_mla_elt<mode>): New.
8008 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8009 (aarch64_mls_elt<mode>): Likewise.
8010 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8011 (aarch64_fma4_elt<mode>): Likewise.
8012 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8013 (aarch64_fma4_elt_to_128v2df): Likewise.
8014 (aarch64_fma4_elt_to_64df): Likewise.
8015 (fnma<mode>4): Likewise.
8016 (aarch64_fnma4_elt<mode>): Likewise.
8017 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8018 (aarch64_fnma4_elt_to_128v2df): Likewise.
8019 (aarch64_fnma4_elt_to_64df): Likewise.
8020 * config/aarch64/iterators.md (VDQSF): New.
8021 * config/aarch64/arm_neon.h
8022 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
8023 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
8025 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8027 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
8028 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
8029 (aarch64_mul3_elt_to_128df): Likewise.
8030 (aarch64_mul3_elt_to_64v2df): Likewise.
8031 * config/aarch64/iterators.md (VEL): Also handle DFmode.
8033 (VMUL_CHANGE_NLANES) Likewise.
8036 * config/aarch64/arm_neon.h
8037 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
8039 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8041 * config/aarch64/arm_neon.h
8042 (vcvtx_high_f32_f64): Fix parameters.
8044 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8045 Uros Bizjak <ubizjak@gmail.com>
8047 * config/alpha.c: Include tree-ssanames.h.
8049 2013-09-16 Richard Biener <rguenther@suse.de>
8051 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
8052 (dot_rdg_1): Handle control_dd.
8053 (create_edge_for_control_dependence): New function.
8054 (create_rdg_edges): Add control dependences if asked for.
8055 (build_rdg): Likewise.
8056 (generate_loops_for_partition): If there are not necessary
8057 control stmts remove all their dependencies.
8058 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
8059 (distribute_loop): Pass on control dependences.
8060 (tree_loop_distribution): Compute control dependences and remove
8061 restriction on number of loop nodes.
8063 2013-09-16 Jakub Jelinek <jakub@redhat.com>
8065 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
8068 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
8070 * cse.c (try_const_anchors): Punt on CC modes.
8072 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8074 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
8076 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
8079 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
8082 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8085 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
8087 (darwin_objc1_section): Likewise.
8088 (darwin_file_end): Emit Image Info section when required.
8090 2013-09-14 Jan Hubicka <jh@suse.cz>
8092 * tree-into-ssa.c (gate_into_ssa): New.
8093 (pass_data_build_ssa): Use it.
8094 * cgraph.h (expand_thunk): Update prototype.
8095 * cgraphunit.c (analyze_function): Expand thunks early.
8096 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
8097 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
8098 set lowered flag; do not add new function.
8099 (assemble_thunks_and_aliases): Update.
8100 * tree-ssa.c (gate_init_datastructures): New gate.
8101 (pass_data_init_datastructures): Use it.
8103 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8106 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
8107 xmm register set described in the psABI.
8109 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
8111 * dwarf2out.c (should_emit_struct_debug): Add check
8112 for type_decl variable is not NULL.
8114 2013-09-13 Jacek Caban <jacek@codeweavers.com>
8116 * config.gcc: Use new winnt-c.c target hooks
8117 * config/t-winnt: New file
8118 * config/winnt-c.c: New file
8119 * doc/tm.texi.in: Document new hook
8120 * doc/tm.texi: Regenerated
8122 2013-09-13 Jan Hubicka <jh@suse.cz>
8125 * ipa-inline.c (check_callers): New function.
8126 (check_caller_edge): Remove.
8127 (want_inline_function_to_all_callers_p): Also permit alises that are
8129 (inline_to_all_callers): Terminate the walk when devirtualization
8130 introduce new calls.
8132 2013-09-13 Jan Hubicka <jh@suse.cz>
8134 * ipa-inline-analysis.c (struct growth_data): Add node.
8135 (do_estimate_growth_1): Fix detection of recursion.
8137 2013-09-13 Jakub Jelinek <jakub@redhat.com>
8139 PR tree-optimization/58392
8140 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
8141 to avoid shadowing of outer loop variable. If
8142 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
8143 replace_by_duplicate_decl simduid of loops that have it set and
8144 set dest_cfun->has_simduid_loops and/or
8145 dest_cfun->has_force_vect_loops.
8146 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
8147 instead of maybe_lookup_decl.
8148 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
8149 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
8150 call. Copy over force_vect and copy and remap simduid. Set
8151 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
8152 (copy_cfg_body): Remove blocks_to_copy argument. Use
8153 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
8154 caller. Don't set cfun->has_simduid_loops and/or
8155 cfun->has_force_vect_loops here.
8156 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
8158 (expand_call_inline, tree_function_versioning): Adjust copy_body
8161 2013-09-13 Martin Jambor <mjambor@suse.cz>
8164 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
8165 the assert if the edge was a speculative one.
8167 2013-09-13 Richard Biener <rguenther@suse.de>
8169 * tree-data-ref.h (known_dependences_p): Move here ...
8170 * tree-loop-distribution.c (known_dependences_p): ... from here.
8171 (dump_rdg_component, debug_rdg_component): Remove.
8173 (generate_loops_for_partition): Use gimple_uid instead of
8174 relying on matching stmt visit order.
8175 (rdg_build_partitions): Take starting stmt vector.
8176 (ldist_gen): Merge into ...
8177 (distribute_loop): ... this function. Do not compute starting
8179 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
8181 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8183 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
8184 Set type attribute correctly. Set predicable_short_it attribute.
8185 (cmpsi_shiftsi): Remove %? from output template.
8187 2013-09-13 Richard Biener <rguenther@suse.de>
8189 * tree-loop-distribution.c (struct rdg_component,
8190 rdg_defs_used_in_other_loops_p, free_rdg_components,
8191 rdg_build_components): Remove.
8192 (stmts_from_loop): Do not record virtual PHIs.
8193 (generate_loops_for_partition): Skip virtual PHIs.
8194 (build_rdg_partition_for_component): Rename to ...
8195 (build_rdg_partition_for_vertex): ... this and adjust.
8196 (rdg_build_partitions): Take a vector of starting vertices
8197 instead of components. Remove unnecessary leftover handling.
8198 (ldist_gen): Do not build components or record other stores.
8199 (distribute_loop): Do not distribute loops containing stmts
8202 2013-09-13 Christian Bruel <christian.bruel@st.com>
8205 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
8207 2013-09-13 Kai Tietz <ktietz@redhat.com>
8209 * config.gcc: Separate cases for mingw and cygwin targets,
8210 and add 64-bit cygwin target case.
8212 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
8213 dll-export inline-functions.
8214 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
8217 2013-09-13 Jeff Law <law@redhat.com>
8221 2013-09-06 Jeff Law <law@redhat.com>
8223 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8224 edge implied equivalences into successor phis.
8226 2013-09-12 DJ Delorie <dj@redhat.com>
8228 * config/rl78/rl78-virt.md: Change from | to \; for asm line
8231 2013-09-12 Brooks Moses <bmoses@google.com>
8234 * Makefile.in: Do not install driver binaries in $(target)/bin.
8236 2013-09-12 DJ Delorie <dj@redhat.com>
8238 * config/rl78/rl78.opt (mrelax): New.
8239 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
8240 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
8242 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
8244 (rl78_expand_epilogue): Likewise.
8246 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
8249 * lra-eliminations.c (remove_reg_equal_offset_note): New.
8250 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
8251 pointer elimination with using remove_reg_equal_offset_note.
8253 2013-09-12 DJ Delorie <dj@redhat.com>
8255 * config/msp430/: New port.
8256 * config.gcc (msp430): Added.
8257 * doc/invoke.texi: Document MSP430 options.
8258 * doc/install.texi: Document msp430-elf
8259 * doc/md.texi: Document msp430-elf
8260 * doc/contrib.texi: Document msp430-elf
8262 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
8265 2013-09-12 Martin Jambor <mjambor@suse.cz>
8268 * ipa-prop.c (remove_described_reference): Give up if the edge in the
8269 reference descriptor is NULL.
8270 (ipa_edge_removal_hook): If owning a reference descriptor, set its
8273 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8275 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
8276 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
8277 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
8278 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
8279 make_temp_ssa_name): move to tree-ssanames.h
8280 * tree-ssa-alias.h: Move prototype.
8281 * tree-ssa.h: Include tree-ssanames.h.
8282 * tree-ssanames.c (FREE_SSANAMES): Move to here.
8283 * tree-ssanames.h: New. Move items from tree-flow*.h
8284 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
8286 2013-09-12 Richard Biener <rguenther@suse.de>
8288 PR tree-optimization/58404
8289 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
8290 propagate non-invariant addresses into dereferences wrapped
8291 in component references.
8293 2013-09-12 Richard Biener <rguenther@suse.de>
8295 PR tree-optimization/58402
8296 * passes.def: Move pass_late_warn_uninitialized later.
8298 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8300 * tree-ssa.h: New. Move content from tree-flow.h and
8302 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
8303 Move prototypes belonging to tree-ssa.c.
8304 * tree-flow-inline.h (redirect_edge_var_map_def,
8305 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
8307 * gimple.h: Adjust prototypes.
8308 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
8310 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
8311 * tree.h: Move prototype to tree-ssa.h.
8312 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
8313 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
8314 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
8315 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
8316 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
8317 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
8318 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
8319 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
8320 graphite-blocking.c, graphite-clast-to-gimple.c,
8321 graphite-dependences.c, graphite-interchange.c,
8322 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
8323 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
8324 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
8325 ipa-reference.c, ipa-split.c, ipa-utils.c,
8326 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
8327 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
8328 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
8329 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
8330 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
8331 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
8332 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
8333 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
8334 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
8335 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
8336 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
8337 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
8338 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
8339 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
8340 value-prof.c, var-tracking.c,
8341 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
8343 2013-09-12 Richard Biener <rguenther@suse.de>
8345 PR tree-optimization/58396
8346 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
8347 (build_rdg): Take a loop-nest parameter, fix memleaks.
8348 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
8350 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8352 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
8355 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
8357 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
8358 The FMA instruction names should have a 'v' prefix.
8360 2013-09-12 Richard Biener <rguenther@suse.de>
8362 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
8363 (dot_rdg): Use popen instead of system in optional code.
8364 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
8365 (already_processed_vertex_p): Adjust.
8366 (has_anti_or_output_dependence, predecessor_has_mem_write,
8367 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
8368 rdg_flag_uses): Remove.
8369 (rdg_flag_vertex): Simplify.
8370 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
8372 (build_rdg_partition_for_component): Process the first vertex
8373 of a component only.
8374 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
8376 2013-09-12 Alan Modra <amodra@gmail.com>
8378 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
8380 2013-09-11 DJ Delorie <dj@redhat.com>
8381 Nick Clifton <nickc@redhat.com>
8383 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
8384 (rl78_stack_based_mem): New.
8385 * config/rl78/constraints.md (Iv08): New.
8394 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
8396 (movsi): Change from expand to insn-and-split.
8397 (ashrsi3): Clobber AX.
8401 * config/rl78/rl78.md (CC_REG): Fix.
8402 (addsi3): Allow memory and immediate operands.
8403 (addsi3_internal): Split into...
8404 (addsi3_internal_virt): ...new, and ...
8405 (addsi3_internal_real): ...new.
8407 (subsi3_internal_virt): New.
8408 (subsi3_internal_real): New.
8409 (mulsi3): Add memory operand.
8410 (mulsi3_rl78): Likewise.
8411 (mulsi3_g13): Likewise.
8412 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
8413 (cbranchqi4_real): Add more constraint options.
8414 (cbranchhi4_real): Expand pattern.
8415 (cbranchhi4_real_signed): New.
8416 (cbranchhi4_real_inverted): New.
8417 (cbranchsi4_real_lt): New.
8418 (cbranchsi4_real_ge): New.
8419 (cbranchsi4_real_signed): New.
8420 (cbranchsi4_real): New.
8422 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
8424 (lshrsi3_virt): Likewise.
8425 (ashlsi3_virt): Likewise.
8426 (cbranchqi4_virt_signed): New.
8427 (cbranchhi4_virt_signed): New.
8428 (cbranchsi4_virt): New.
8429 * config/rl78/rl78.c: Whitespace fixes throughout.
8430 (move_elim_pass): New.
8431 (pass_data_rl78_move_elim): New.
8432 (pass_rl78_move_elim): New.
8433 (make_pass_rl78_move_elim): New.
8434 (rl78_devirt_info): Run devirt earlier.
8435 (rl78_move_elim_info): New.
8436 (rl78_asm_file_start): Register it.
8437 (rl78_split_movsi): New.
8438 (rl78_as_legitimate_address): Allow virtual base registers when
8440 (rl78_addr_space_convert): Remove spurious debug stuff.
8441 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
8442 (rl78_print_operand): More cases for not printing '#'.
8443 (rl78_expand_compare): Remove most of the logic.
8444 (content_memory): New.
8445 (clear_content_memory): New.
8446 (get_content_index): New.
8447 (get_content_name): New.
8448 (display_content_memory): New.
8449 (update_content): New.
8450 (record_content): New.
8451 (already_contains): New.
8452 (insn_ok_now): Re-recog insns with virtual registers.
8453 (add_postponed_content_update): New.
8454 (process_postponed_content_update): New.
8455 (gen_and_emit_move): New.
8456 (transcode_memory_rtx): Record new location content.
8457 Use gen_and_emit_move.
8458 (force_into_acc): New.
8459 (move_to_acc): Use gen_and_emit_move.
8460 (move_from_acc): Likewise.
8461 (move_acc_to_reg): Likewise.
8462 (move_to_x): Likewise.
8463 (move_to_hl): Likewise.
8464 (move_to_de): Likewise.
8465 (rl78_alloc_physical_registers_op1): Record location content.
8466 (has_constraint): New.
8467 (rl78_alloc_physical_registers_op2): Record location content.
8469 (rl78_alloc_physical_registers_ro1): Likewise.
8470 (rl78_alloc_physical_registers_cmp): Likewise.
8471 (rl78_alloc_physical_registers_umul): Likewise.
8472 (rl78_alloc_address_registers_macax): New.
8473 (rl78_alloc_physical_registers): Initialize and set location
8474 content memory as needed.
8475 (rl78_reorg): Make sure split2 is called.
8476 (rl78_rtx_costs): New.
8478 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
8480 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
8481 for (not (neg ...)) and (neg (not ...)) cases.
8483 2013-09-11 Richard Biener <rguenther@suse.de>
8486 * passes.def: Split critical edges before late uninit warning passes.
8487 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
8489 2013-09-11 Jakub Jelinek <jakub@redhat.com>
8491 PR tree-optimization/58385
8492 * fold-const.c (build_range_check): If both low and high are NULL,
8493 use omit_one_operand_loc to preserve exp side-effects.
8495 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8497 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
8499 2013-09-11 Richard Biener <rguenther@suse.de>
8501 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
8502 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
8503 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
8504 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
8505 stmts_from_loop, known_dependences_p, build_empty_rdg,
8506 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
8507 * tree-loop-distribution.c: ... here.
8508 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
8509 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
8510 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
8511 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
8512 * tree-loop-distribution.c: ... here.
8513 * tree-loop-distribution.c: Include gimple-pretty-print.h.
8514 (struct partition_s): Add loops member.
8515 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
8516 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
8517 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
8519 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8520 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8521 Sergey Lega <sergey.s.lega@intel.com>
8522 Anna Tikhonova <anna.tikhonova@intel.com>
8523 Ilya Tocar <ilya.tocar@intel.com>
8524 Andrey Turetskiy <andrey.turetskiy@intel.com>
8525 Ilya Verbin <ilya.verbin@intel.com>
8526 Kirill Yukhin <kirill.yukhin@intel.com>
8527 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8529 * config/i386/constraints.md (k): New.
8531 * config/i386/i386.c (const regclass_map): Add new mask registers.
8532 (dbx_register_map): Ditto.
8533 (dbx64_register_map): Ditto.
8534 (svr4_dbx_register_map): Ditto.
8535 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
8537 (ix86_preferred_reload_class): Disable constants for mask registers.
8538 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
8539 (ix86_hard_regno_mode_ok): Support new mask registers.
8540 (x86_order_regs_for_local_alloc): Ditto.
8541 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
8542 (FIXED_REGISTERS): Add new mask registers.
8543 (CALL_USED_REGISTERS): Ditto.
8544 (REG_ALLOC_ORDER): Ditto.
8545 (VALID_MASK_REG_MODE): New.
8546 (FIRST_MASK_REG): Ditto.
8547 (LAST_MASK_REG): Ditto.
8548 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
8549 (MAYBE_MASK_CLASS_P): New.
8550 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
8551 (REG_CLASS_CONTENTS): Ditto.
8552 (MASK_REGNO_P): New.
8553 (ANY_MASK_REG_P): Ditto.
8554 (HI_REGISTER_NAMES): Add new mask registers.
8555 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
8556 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
8558 (attribute "type"): Add mskmov, msklog.
8559 (attribute "length_immediate"): Support them.
8560 (attribute "memory"): Ditto.
8561 (attribute "prefix_0f"): Ditto.
8562 (*movhi_internal): Support new mask registers.
8563 (*movqi_internal): Ditto.
8564 (define_split): Split out clobber pattern is a logic
8565 insn on mask registers.
8566 (*k<logic><mode>): New.
8567 (*andhi_1): Extend to support mask regs.
8568 (*andqi_1): Extend to support mask regs.
8570 (define_split): Split and-not to and and not if operands
8572 (*<code><mode>_1): Separate HI mode to new pattern...
8573 (*<code>hi_1): This.
8574 (*<code>qi_1): Extend to support mask regs.
8576 (kortestzhi): Ditto.
8577 (kortestchi): Ditto.
8579 (*one_cmpl<mode>2_1): Remove HImode and handle it...
8580 (*one_cmplhi2_1): ...Here, now with mask registers support.
8581 (*one_cmplqi2_1): Support new mask registers.
8582 (HI/QImode arithmetics splitter): Don't split if mask registers
8584 (HI/QImode not splitter): Ditto.
8585 * config/i386/predicated.md (mask_reg_operand): New.
8586 (general_reg_operand): Ditto.
8588 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8590 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
8591 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
8593 2013-09-10 Jeff Law <law@redhat.com>
8595 PR tree-optimization/58380
8596 * tree-ssa-threadupdate.c (thread_block): Recognize another case
8597 of threading through a buried loop header.
8599 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
8600 return value for single successor case.
8602 2013-09-10 Jan Hubicka <jh@suse.cz>
8604 * ipa-devirt.c (ipa_devirt): Enable with LTO.
8606 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
8609 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
8610 support conditional execution.
8611 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
8613 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
8615 * lra.c (lra): Clear lra_optional_reload_pseudos before every
8617 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
8618 Check destination too to check move insn.
8619 (undo_optional_reloads): Add check that the original peudo did not
8620 changed its allocation and the optional reload was inherited on last
8621 inheritance pass. Break loop after deciding to keep optional reload.
8622 (lra_undo_inheritance): Add check that inherited pseudo still in
8625 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
8627 * config/aarch64/aarch64.md (generic_sched): New.
8628 * config/aarch64/aarch64-generic.md (load): Make conditional
8629 on generic_sched attribute.
8630 (nonload): Likewise.
8632 2013-09-10 Jan Hubicka <jh@suse.cz>
8634 * lto-cgraph.c: Include ipa-utils.h.
8635 (compute_ltrans_boundary): Also add possible targets into the boundary.
8637 2013-09-10 Jan Hubicka <jh@suse.cz>
8639 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
8640 VAR_DECL of vtable rather than full expression.
8642 2013-09-10 Jan Hubicka <jh@suse.cz>
8643 Paolo Carlini <paolo.carlini@oracle.com>
8645 * cgraphunit.c (analyze_functions): Save input_location, set it
8646 to UNKNOWN_LOCATION and restore it at the end.
8648 2013-09-10 Martin Jambor <mjambor@suse.cz>
8650 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
8651 represented by a thunk.
8653 2013-09-10 Jeff Law <law@redhat.com>
8655 PR tree-optimization/58343
8656 * tree-ssa-threadupdate.c (thread_block): Identify and disable
8657 jump threading requests through loop headers buried in the middle
8658 of a jump threading path.
8660 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
8661 in return value/type.
8663 2013-09-10 Jakub Jelinek <jakub@redhat.com>
8665 PR rtl-optimization/58365
8666 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
8667 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
8670 2013-09-10 Richard Biener <rguenther@suse.de>
8672 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
8673 * tree-data-ref.c (create_rdg_vertices): Collect all data
8674 references, signal failure to the caller, use data-ref API.
8675 (build_rdg): Compute data references only once. Maintain lifetime
8676 of data references and data dependences from within RDG.
8677 (free_rdg): Free dependence relations.
8678 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
8679 inventing extra dependences.
8680 (distribute_loop): Update for RDG API changes.
8682 2013-09-10 Kai Tietz <ktietz@redhat.com>
8684 * doc/invoke.texi (fms-extensions): Document changed
8685 behavior for ms-abi targets.
8686 * config/i386/i386.c (ix86_option_override_internal):
8687 Set default value of option -fms-extension for ms-abi targets.
8689 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
8691 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
8693 2013-09-10 Alan Modra <amodra@gmail.com>
8696 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
8698 2013-09-10 Alan Modra <amodra@gmail.com>
8700 * config/rs6000/predicates.md (add_cint_operand): New.
8701 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
8702 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
8703 using add_cint_operand.
8704 (largetoc_high_plus_aix): Likewise.
8706 2013-09-09 Jakub Jelinek <jakub@redhat.com>
8708 PR tree-optimization/58364
8709 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
8710 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
8711 the current range can't be an unconditional true or false.
8713 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
8715 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
8717 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
8719 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
8721 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
8724 * doc/invoke.texi (-Wdelete-incomplete): Document it.
8726 2013-09-09 Ian Bolton <ian.bolton@arm.com>
8728 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
8729 NO_REGS for immediate that can't be moved directly into FP_REGS.
8731 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8733 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
8734 comparison with negated operand.
8735 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
8738 2013-09-09 Richard Biener <rguenther@suse.de>
8741 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
8742 of a subloop record all its block as affecting loop-closed SSA form.
8744 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8746 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
8747 of an rtx/bitpos pair.
8748 (store_fixed_bit_field): Update accordingly.
8750 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8752 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
8754 * builtins.c (expand_errno_check): Likewise.
8755 * dwarf2cfi.c (init_return_column_size): Likewise.
8756 * except.c (sjlj_mark_call_sites): Likewise.
8757 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
8758 * lra-constraints.c (emit_inc): Likewise.
8759 * ree.c (combine_set_extension): Likewise.
8760 * regmove.c (fixup_match_2): Likewise.
8761 * reload1.c (inc_for_reload): Likewise.
8763 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8765 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
8766 (force_to_mode, simplify_shift_const_1, simplify_comparison):
8767 Use gen_int_mode with the mode of the associated simplify_* call.
8768 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
8769 * expmed.c (expand_shift_1): Likewise.
8770 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8771 * loop-iv.c (iv_number_of_iterations): Likewise.
8772 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8773 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
8775 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8777 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
8778 of the associated expand_* call.
8779 (asan_emit_stack_protection): Likewise.
8780 * builtins.c (round_trampoline_addr): Likewise.
8781 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
8782 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
8783 (emit_store_flag): Likewise.
8784 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
8785 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
8787 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8788 * ifcvt.c (noce_try_store_flag_constants): Likewise.
8789 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8790 * modulo-sched.c (generate_prolog_epilog): Likewise.
8791 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
8792 (expand_ctz, expand_ffs, expand_unop): Likewise.
8794 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8796 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
8797 of the associated gen_rtx_* call.
8798 * caller-save.c (init_caller_save): Likewise.
8799 * combine.c (find_split_point, make_extraction): Likewise.
8800 (make_compound_operation): Likewise.
8801 * dwarf2out.c (mem_loc_descriptor): Likewise.
8802 * explow.c (plus_constant, probe_stack_range): Likewise.
8803 * expmed.c (expand_mult_const): Likewise.
8804 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
8805 * reload1.c (init_reload): Likewise.
8806 * valtrack.c (cleanup_auto_inc_dec): Likewise.
8807 * var-tracking.c (adjust_mems): Likewise.
8808 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
8809 rather than gen_rtx_fmt_ee.
8811 2013-09-09 Jan Hubicka <jh@suse.cz>
8814 * value-prof.c (gimple_ic): Copy also abnormal edges.
8816 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8818 * asan.c (asan_shadow_cst): Use gen_int_mode.
8820 2013-09-08 Jan Hubicka <jh@suse.cz>
8822 * ipa-profile.c: Add toplevel comment.
8823 (ipa_propagate_frequency_1): Be more conservative when profile is read.
8824 (contains_hot_call_p): New function.
8825 (ipa_propagate_frequency): Set frequencies based on counts when
8827 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
8828 profile; do not tamper with profile after inlining if it is read.
8830 2013-09-08 Jan Hubicka <jh@suse.cz>
8832 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
8835 2013-09-08 Jan Hubicka <jh@suse.cz>
8837 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
8840 2013-09-08 Jeff Law <law@redhat.com>
8843 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
8846 2013-09-08 Andi Kleen <ak@linux.intel.com>
8848 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
8850 2013-09-08 Jan Hubicka <jh@suse.cz>
8852 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
8854 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
8856 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
8857 for non-debug insns.
8858 * lra.c (new_insn_reg): Take the containing insn as a parameter.
8859 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
8860 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
8863 2013-09-08 Jan Hubicka <jh@suse.cz>
8865 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
8866 targets and devirtualize to BUILT_IN_UNREACHABLE.
8867 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
8868 * ipa.c (walk_polymorphic_call_targets): New function.
8869 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
8870 functions; use the new timevar.
8871 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
8872 was removed from the program.
8873 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
8874 not consider it in the walk when its vtable is dead.
8875 (possible_polymorphic_call_targets_1): Pass anonymous flag to
8877 (devirt_variable_node_removal_hook): New function.
8878 (possible_polymorphic_call_targets): Also register
8879 devirt_variable_node_removal_hook.
8880 (ipa_devirt): Do not do non-speculative devirtualization.
8881 (gate_ipa_devirt): One execute if devirtualizing speculatively.
8883 2013-09-08 Jan Hubicka <jh@suse.cz>
8885 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
8886 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
8887 varpool_remove_variable_insertion_hook): Declare.
8888 * varpool.c (varpool_node_hook_list): New structure.
8889 (first_varpool_node_removal_hook,
8890 first_varpool_variable_insertion_hook): New variables.
8891 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
8892 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
8893 varpool_remove_variable_insertion_hook,
8894 varpool_call_variable_insertion_hooks): New functions.
8895 (varpool_remove_node): Use it.
8897 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
8900 * diagnostic.c (diagnostic_build_prefix): When s.file is
8901 "<built-in>" don't output line and column numbers.
8903 2013-09-06 Jan Hubicka <jh@suse.cz>
8905 * cgraphunit.c (expand_thunk): Get body before touching arguments.
8906 * lto-streamer-out.c: Stream thunks, too.
8907 * lto-streamer-in.c (input_function): Pop cfun here
8908 (lto_read_body): Instead of here.
8910 2013-09-06 Caroline Tice <cmtice@google.com>
8912 * doc/install.texi: Add documentation for the --enable-vtable-verify
8913 and the --disable-libvtv configure options.
8915 2013-09-06 Jeff Law <law@redhat.com>
8917 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8918 edge implied equivalences into successor phis.
8920 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
8922 * resource.c (mark_referenced_resources): Handle COND_EXEC.
8924 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
8926 * resource.c (mark_target_live_regs): Compute resources taking
8927 into account if a call is predicated or not.
8929 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
8931 * toplev.c (output_stack_usage): Be prepared for suffixes created by
8932 the compiler in the function names.
8934 2013-09-06 Jan Hubicka <jh@suse.cz>
8937 * ipa-inline.c (has_caller_p): New function.
8938 (want_inline_function_to_all_callers_p): Use it.
8939 (sum_callers, inline_to_all_callers): Break out from ...
8940 (ipa_inline): ... here.
8942 2013-09-06 Jan Hubicka <jh@suse.cz>
8944 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
8945 only when AVX is enabled.
8947 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8949 * config/aarch64/aarch64.md
8950 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
8951 is fpsimd_<load/store>2.
8952 (load_pair<mode>): Likewise.
8953 (store_pair<mode>): Likewise.
8955 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8957 * config/arm/types.md (type): Add "mrs" type.
8958 * config/aarch64/aarch64.md
8959 (aarch64_load_tp_hard): Make type "mrs".
8961 (load_tp_hard): Make type "mrs".
8962 * config/arm/cortex-a15.md: Update with new attributes.
8963 * config/arm/cortex-a5.md: Update with new attributes.
8964 * config/arm/cortex-a53.md: Update with new attributes.
8965 * config/arm/cortex-a7.md: Update with new attributes.
8966 * config/arm/cortex-a8.md: Update with new attributes.
8967 * config/arm/cortex-a9.md: Update with new attributes.
8968 * config/arm/cortex-m4.md: Update with new attributes.
8969 * config/arm/cortex-r4.md: Update with new attributes.
8970 * config/arm/fa526.md: Update with new attributes.
8971 * config/arm/fa606te.md: Update with new attributes.
8972 * config/arm/fa626te.md: Update with new attributes.
8973 * config/arm/fa726te.md: Update with new attributes.
8975 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8977 * config/aarch64/aarch64.md
8978 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
8979 (*movtf_aarch64): Likewise.
8981 (thumb1_movdi_insn): Use "multiple" for type where more than one
8982 instruction is used for a move.
8983 (*arm32_movhf): Likewise.
8984 (*thumb_movdf_insn): Likewise.
8986 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8988 * config/arm/types.md (type): Rename fcpys to fmov.
8990 (*arm_movsi_vfp): Rename type fcpys as fmov.
8991 (*thumb2_movsi_vfp): Likewise
8992 (*movhf_vfp_neon): Likewise
8993 (*movhf_vfp): Likewise
8994 (*movsf_vfp): Likewise
8995 (*thumb2_movsf_vfp): Likewise
8996 (*movsfcc_vfp): Likewise
8997 (*thumb2_movsfcc_vfp): Likewise
8998 * config/aarch64/aarch64-simd.md
8999 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
9000 * config/aarch64/aarch64.md
9001 (*movsi_aarch64): Replace type mov_reg with fmovs.
9002 (*movdi_aarch64): Likewise
9003 (*movsf_aarch64): Likewise
9004 (*movdf_aarch64): Likewise
9006 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
9007 * config/arm/iwmmxt.md
9008 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
9009 * config/arm/arm1020e.md: Update with new attributes.
9010 * config/arm/cortex-a15-neon.md: Update with new attributes.
9011 * config/arm/cortex-a5.md: Update with new attributes.
9012 * config/arm/cortex-a53.md: Update with new attributes.
9013 * config/arm/cortex-a7.md: Update with new attributes.
9014 * config/arm/cortex-a8-neon.md: Update with new attributes.
9015 * config/arm/cortex-a9.md: Update with new attributes.
9016 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9017 * config/arm/cortex-r4f.md: Update with new attributes.
9018 * config/arm/marvell-pj4.md: Update with new attributes.
9019 * config/arm/vfp11.md: Update with new attributes.
9021 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9023 * config/aarch64/aarch64.md
9024 (*madd<mode>): Fix type attribute.
9025 (*maddsi_uxtw): Likewise.
9026 (*msub<mode>): Likewise.
9027 (*msubsi_uxtw): Likewise.
9028 (<su_optab>maddsidi4): Likewise.
9029 (<su_optab>msubsidi4): Likewise.
9031 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9033 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
9034 * config/arm/arm.md (core_cycles): Remove fdiv.
9035 * config/arm/vfp.md:
9036 (*sqrtsf2_vfp): Update for attribute changes.
9037 (*sqrtdf2_vfp): Likewise.
9038 * config/aarch64/aarch64.md:
9039 (sqrt<mode>2): Update for attribute changes.
9040 * config/arm/arm1020e.md: Update with new attributes.
9041 * config/arm/cortex-a15-neon.md: Update with new attributes.
9042 * config/arm/cortex-a5.md: Update with new attributes.
9043 * config/arm/cortex-a53.md: Update with new attributes.
9044 * config/arm/cortex-a7.md: Update with new attributes.
9045 * config/arm/cortex-a8-neon.md: Update with new attributes.
9046 * config/arm/cortex-a9.md: Update with new attributes.
9047 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9048 * config/arm/cortex-r4f.md: Update with new attributes.
9049 * config/arm/marvell-pj4.md: Update with new attributes.
9050 * config/arm/vfp11.md: Update with new attributes.
9052 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9054 * config/arm/types.md
9055 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
9056 * config/aarch64/aarch64.md
9057 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
9059 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
9060 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
9061 (float<GPI:mode><GPF:mode>2): Likewise.
9063 (*truncsisf2_vfp): Update with new attributes.
9064 (*truncsidf2_vfp): Likewise.
9065 (fixuns_truncsfsi2): Likewise.
9066 (fixuns_truncdfsi2): Likewise.
9067 (*floatsisf2_vfp): Likewise.
9068 (*floatsidf2_vfp): Likewise.
9069 (floatunssisf2): Likewise.
9070 (floatunssidf2): Likewise.
9071 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
9072 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
9073 * config/arm/arm1020e.md: Update with new attributes.
9074 * config/arm/cortex-a15-neon.md: Update with new attributes.
9075 * config/arm/cortex-a5.md: Update with new attributes.
9076 * config/arm/cortex-a53.md: Update with new attributes.
9077 * config/arm/cortex-a7.md: Update with new attributes.
9078 * config/arm/cortex-a8-neon.md: Update with new attributes.
9079 * config/arm/cortex-a9.md: Update with new attributes.
9080 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9081 * config/arm/cortex-r4f.md: Update with new attributes.
9082 * config/arm/marvell-pj4.md: Update with new attributes.
9083 * config/arm/vfp11.md: Update with new attributes.
9085 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9087 * config/aarch64/arm_neon.h
9088 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
9089 (vqtbx<1,2,3,4><q>_s8): Likewise.
9091 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9093 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
9094 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
9095 (add<mode>3): Add type attribute.
9096 (add<mode>3): Likewise.
9097 (usadd<mode>3): Likewise.
9098 (ssadd<mode>3): Likewise.
9099 (sub<mode>3): Likewise.
9100 (sub<mode>3): Likewise.
9101 (ussub<mode>3): Likewise.
9102 (sssub<mode>3): Likewise.
9103 (ssmulsa3): Likewise.
9104 (usmulusa3): Likewise.
9105 (arm_usatsihi): Likewise.
9107 (*movdi_vfp): Add types for all instructions.
9108 (*movdi_vfp_cortexa8): Likewise.
9109 (*movhf_vfp_neon): Likewise.
9110 (*movhf_vfp): Likewise.
9111 (*movdf_vfp): Likewise.
9112 (*thumb2_movdf_vfp): Likewise.
9113 (*thumb2_movdfcc_vfp): Likewise.
9114 * config/arm/arm.md: Add type attribute to all insn patterns.
9115 (*thumb1_adddi3): Add type attribute.
9116 (*arm_adddi3): Likewise.
9117 (*adddi_sesidi_di): Likewise.
9118 (*adddi_zesidi_di): Likewise.
9119 (*thumb1_addsi3): Likewise.
9120 (addsi3_compare0): Likewise.
9121 (*addsi3_compare0_scratch): Likewise.
9122 (*compare_negsi_si): Likewise.
9123 (cmpsi2_addneg): Likewise.
9124 (*addsi3_carryin_<optab>): Likewise.
9125 (*addsi3_carryin_alt2_<optab>): Likewise.
9126 (*addsi3_carryin_clobercc_<optab>): Likewise.
9127 (*subsi3_carryin): Likewise.
9128 (*subsi3_carryin_const): Likewise.
9129 (*subsi3_carryin_compare): Likewise.
9130 (*subsi3_carryin_compare_const): Likewise.
9131 (*arm_subdi3): Likewise.
9132 (*thumb_subdi3): Likewise.
9133 (*subdi_di_zesidi): Likewise.
9134 (*subdi_di_sesidi): Likewise.
9135 (*subdi_zesidi_di): Likewise.
9136 (*subdi_sesidi_di): Likewise.
9137 (*subdi_zesidi_ze): Likewise.
9138 (thumb1_subsi3_insn): Likewise.
9139 (*arm_subsi3_insn): Likewise.
9140 (*anddi3_insn): Likewise.
9141 (*anddi_zesidi_di): Likewise.
9142 (*anddi_sesdi_di): Likewise.
9143 (*ne_zeroextracts): Likewise.
9144 (*ne_zeroextracts): Likewise.
9145 (*ite_ne_zeroextr): Likewise.
9146 (*ite_ne_zeroextr): Likewise.
9147 (*anddi_notdi_di): Likewise.
9148 (*anddi_notzesidi): Likewise.
9149 (*anddi_notsesidi): Likewise.
9150 (andsi_notsi_si): Likewise.
9151 (thumb1_bicsi3): Likewise.
9152 (*iordi3_insn): Likewise.
9153 (*iordi_zesidi_di): Likewise.
9154 (*iordi_sesidi_di): Likewise.
9155 (*thumb1_iorsi3_insn): Likewise.
9156 (*xordi3_insn): Likewise.
9157 (*xordi_zesidi_di): Likewise.
9158 (*xordi_sesidi_di): Likewise.
9159 (*arm_xorsi3): Likewise.
9160 (*andsi_iorsi3_no): Likewise.
9161 (*smax_0): Likewise.
9162 (*smax_m1): Likewise.
9163 (*arm_smax_insn): Likewise.
9164 (*smin_0): Likewise.
9165 (*arm_smin_insn): Likewise.
9166 (*arm_umaxsi3): Likewise.
9167 (*arm_uminsi3): Likewise.
9168 (*minmax_arithsi): Likewise.
9169 (*minmax_arithsi_): Likewise.
9170 (*satsi_<SAT:code>): Likewise.
9171 (arm_ashldi3_1bit): Likewise.
9172 (arm_ashrdi3_1bit): Likewise.
9173 (arm_lshrdi3_1bit): Likewise.
9174 (*arm_negdi2): Likewise.
9175 (*thumb1_negdi2): Likewise.
9176 (*arm_negsi2): Likewise.
9177 (*thumb1_negsi2): Likewise.
9178 (*negdi_extendsid): Likewise.
9179 (*negdi_zero_extend): Likewise.
9180 (*arm_abssi2): Likewise.
9181 (*thumb1_abssi2): Likewise.
9182 (*arm_neg_abssi2): Likewise.
9183 (*thumb1_neg_abss): Likewise.
9184 (one_cmpldi2): Likewise.
9185 (extend<mode>di2): Likewise.
9186 (*compareqi_eq0): Likewise.
9187 (*arm_extendhisi2addsi): Likewise.
9188 (*arm_movdi): Likewise.
9189 (*thumb1_movdi_insn): Likewise.
9190 (*arm_movt): Likewise.
9191 (*thumb1_movsi_insn): Likewise.
9192 (pic_add_dot_plus_four): Likewise.
9193 (pic_add_dot_plus_eight): Likewise.
9194 (tls_load_dot_plus_eight): Likewise.
9195 (*thumb1_movhi_insn): Likewise.
9196 (*thumb1_movsf_insn): Likewise.
9197 (*movdf_soft_insn): Likewise.
9198 (*thumb_movdf_insn): Likewise.
9199 (cbranchsi4_insn): Likewise.
9200 (cbranchsi4_scratch): Likewise.
9201 (*negated_cbranchsi4): Likewise.
9202 (*tbit_cbranch): Likewise.
9203 (*tlobits_cbranch): Likewise.
9204 (*tstsi3_cbranch): Likewise.
9205 (*cbranchne_decr1): Likewise.
9206 (*addsi3_cbranch): Likewise.
9207 (*addsi3_cbranch_scratch): Likewise.
9208 (*arm_cmpdi_insn): Likewise.
9209 (*arm_cmpdi_unsig): Likewise.
9210 (*arm_cmpdi_zero): Likewise.
9211 (*thumb_cmpdi_zero): Likewise.
9212 (*deleted_compare): Likewise.
9213 (*mov_scc): Likewise.
9214 (*mov_negscc): Likewise.
9215 (*mov_notscc): Likewise.
9216 (*cstoresi_eq0_thumb1_insn): Likewise.
9217 (cstoresi_nltu_thumb1): Likewise.
9218 (cstoresi_ltu_thu): Likewise.
9219 (thumb1_addsi3_addgeu): Likewise.
9220 (*arm_jump): Likewise.
9221 (*thumb_jump): Likewise.
9222 (*check_arch2): Likewise.
9223 (arm_casesi_internal): Likewise.
9224 (thumb1_casesi_dispatch): Likewise.
9225 (*arm_indirect_jump): Likewise.
9226 (*thumb1_indirect_jump): Likewise.
9228 (*and_scc): Likewise.
9229 (*ior_scc): Likewise.
9230 (*compare_scc): Likewise.
9231 (*cond_move): Likewise.
9232 (*cond_arith): Likewise.
9233 (*cond_sub): Likewise.
9234 (*cmp_ite0): Likewise.
9235 (*cmp_ite1): Likewise.
9236 (*cmp_and): Likewise.
9237 (*cmp_ior): Likewise.
9238 (*ior_scc_scc): Likewise.
9239 (*ior_scc_scc_cmp): Likewise.
9240 (*and_scc_scc): Likewise.
9241 (*and_scc_scc_cmp): Likewise.
9242 (*and_scc_scc_nod): Likewise.
9243 (*negscc): Likewise.
9244 (movcond_addsi): Likewise.
9245 (movcond): Likewise.
9246 (*ifcompare_plus_move): Likewise.
9247 (*if_plus_move): Likewise.
9248 (*ifcompare_move_plus): Likewise.
9249 (*if_move_plus): Likewise.
9250 (*ifcompare_arith_arith): Likewise.
9251 (*if_arith_arith): Likewise.
9252 (*ifcompare_arith_move): Likewise.
9253 (*if_arith_move): Likewise.
9254 (*ifcompare_move_arith): Likewise.
9255 (*if_move_arith): Likewise.
9256 (*ifcompare_move_not): Likewise.
9257 (*if_move_not): Likewise.
9258 (*ifcompare_not_move): Likewise.
9259 (*if_not_move): Likewise.
9260 (*ifcompare_shift_move): Likewise.
9261 (*if_shift_move): Likewise.
9262 (*ifcompare_move_shift): Likewise.
9263 (*if_move_shift): Likewise.
9264 (*ifcompare_shift_shift): Likewise.
9265 (*ifcompare_not_arith): Likewise.
9266 (*ifcompare_arith_not): Likewise.
9267 (*if_arith_not): Likewise.
9268 (*ifcompare_neg_move): Likewise.
9269 (*if_neg_move): Likewise.
9270 (*ifcompare_move_neg): Likewise.
9271 (*if_move_neg): Likewise.
9272 (prologue_thumb1_interwork): Likewise.
9273 (*cond_move_not): Likewise.
9274 (*sign_extract_onebit): Likewise.
9275 (*not_signextract_onebit): Likewise.
9276 (stack_tie): Likewise.
9277 (align_4): Likewise.
9278 (align_8): Likewise.
9279 (consttable_end): Likewise.
9280 (consttable_1): Likewise.
9281 (consttable_2): Likewise.
9282 (consttable_4): Likewise.
9283 (consttable_8): Likewise.
9284 (consttable_16): Likewise.
9285 (*thumb1_tablejump): Likewise.
9286 (prefetch): Likewise.
9287 (force_register_use): Likewise.
9288 (thumb_eh_return): Likewise.
9289 (load_tp_hard): Likewise.
9290 (load_tp_soft): Likewise.
9291 (tlscall): Likewise.
9292 (*arm_movtas_ze): Likewise.
9293 (*arm_rev): Likewise.
9294 (*arm_revsh): Likewise.
9295 (*arm_rev16): Likewise.
9296 * config/arm/thumb2.md
9297 (*thumb2_smaxsi3): Likewise.
9298 (*thumb2_sminsi3): Likewise.
9299 (*thumb32_umaxsi3): Likewise.
9300 (*thumb2_uminsi3): Likewise.
9301 (*thumb2_negdi2): Likewise.
9302 (*thumb2_abssi2): Likewise.
9303 (*thumb2_neg_abss): Likewise.
9304 (*thumb2_movsi_insn): Likewise.
9305 (tls_load_dot_plus_four): Likewise.
9306 (*thumb2_movhi_insn): Likewise.
9307 (*thumb2_mov_scc): Likewise.
9308 (*thumb2_mov_negs): Likewise.
9309 (*thumb2_mov_negs): Likewise.
9310 (*thumb2_mov_nots): Likewise.
9311 (*thumb2_mov_nots): Likewise.
9312 (*thumb2_movsicc_): Likewise.
9313 (*thumb2_movsfcc_soft_insn): Likewise.
9314 (*thumb2_indirect_jump): Likewise.
9315 (*thumb2_and_scc): Likewise.
9316 (*thumb2_ior_scc): Likewise.
9317 (*thumb2_ior_scc_strict_it): Likewise.
9318 (*thumb2_cond_move): Likewise.
9319 (*thumb2_cond_arith): Likewise.
9320 (*thumb2_cond_ari): Likewise.
9321 (*thumb2_cond_sub): Likewise.
9322 (*thumb2_negscc): Likewise.
9323 (*thumb2_movcond): Likewise.
9324 (thumb2_casesi_internal): Likewise.
9325 (thumb2_casesi_internal_pic): Likewise.
9326 (*thumb2_alusi3_short): Likewise.
9327 (*thumb2_mov<mode>_shortim): Likewise.
9328 (*thumb2_addsi_short): Likewise.
9329 (*thumb2_subsi_short): Likewise.
9330 (thumb2_addsi3_compare0): Likewise.
9331 (*thumb2_cbz): Likewise.
9332 (*thumb2_cbnz): Likewise.
9333 (*thumb2_one_cmplsi2_short): Likewise.
9334 (*thumb2_negsi2_short): Likewise.
9335 (*orsi_notsi_si): Likewise.
9336 * config/arm/arm1020e.md: Update with new attributes.
9337 * config/arm/arm1026ejs.md: Update with new attributes.
9338 * config/arm/arm1136jfs.md: Update with new attributes.
9339 * config/arm/arm926ejs.md: Update with new attributes.
9340 * config/arm/cortex-a15.md: Update with new attributes.
9341 * config/arm/cortex-a5.md: Update with new attributes.
9342 * config/arm/cortex-a53.md: Update with new attributes.
9343 * config/arm/cortex-a7.md: Update with new attributes.
9344 * config/arm/cortex-a8.md: Update with new attributes.
9345 * config/arm/cortex-a9.md: Update with new attributes.
9346 * config/arm/cortex-m4.md: Update with new attributes.
9347 * config/arm/cortex-r4.md: Update with new attributes.
9348 * config/arm/fa526.md: Update with new attributes.
9349 * config/arm/fa606te.md: Update with new attributes.
9350 * config/arm/fa626te.md: Update with new attributes.
9351 * config/arm/fa726te.md: Update with new attributes.
9353 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9355 * config/aarch64/aarch64-simd.md
9356 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
9357 <vwx> iterator to ensure correct register choice.
9358 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
9359 (aarch64_sqdmull_n<mode>): Likewise.
9360 (aarch64_sqdmull2_n<mode>_internal): Likewise.
9361 * config/aarch64/arm_neon.h
9362 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
9363 (vml<as><q>_n_<su>16): Likewise.
9364 (vml<as>l_high_lane<q>_<su>16): Likewise.
9365 (vml<as>l_high_n_<su>16): Likewise.
9366 (vml<as>l_lane<q>_<su>16): Likewise.
9367 (vml<as>l_n_<su>16): Likewise.
9368 (vmul<q>_lane<q>_<su>16): Likewise.
9369 (vmul<q>_n_<su>16): Likewise.
9370 (vmull_lane<q>_<su>16): Likewise.
9371 (vmull_n_<su>16): Likewise.
9372 (vmull_high_lane<q>_<su>16): Likewise.
9373 (vmull_high_n_<su>16): Likewise.
9374 (vqrdmulh<q>_n_s16): Likewise.
9376 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
9378 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
9379 have the correct lane parameter.
9381 2013-09-06 Richard Biener <rguenther@suse.de>
9383 * cfganal.c (control_dependences::~control_dependences):
9384 Properly free all of the vector.
9386 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
9389 * config/i386/i386.c (ix86_conditional_register_usage):
9390 Proper initialize extended SSE registers.
9392 2013-09-06 Jan Hubicka <jh@suse.cz>
9394 PR tree-optimization/58311
9395 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
9397 2013-09-06 Jan Hubicka <jh@suse.cz>
9399 * Makefile.in (tree-sra.o): Update dependencies.
9400 * tree-sra.c: Include ipa-utils.h
9401 (scan_function): Use recursive_call_p.
9402 (has_caller_p): New function.
9403 (cgraph_for_node_and_aliases): Count also callers of aliases.
9405 2013-09-06 Jan Hubicka <jh@suse.cz>
9408 * cgraph.h (symtab_semantically_equivalent_p): Declare.
9409 * tree-tailcall.c: Include ipa-utils.h.
9410 (find_tail_calls): Use it.
9411 * ipa-pure-const.c (check_call): Likewise.
9412 * ipa-utils.c (recursive_call_p): New function.
9413 * ipa-utils.h (recursive_call_p): Dclare.
9414 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
9415 (symtab_semantically_equivalent_p): New function.
9416 * Makefile.in (tree-tailcall.o): Update dependencies.
9418 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9420 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
9423 2013-09-06 Richard Biener <rguenther@suse.de>
9425 * lto-streamer.h (lto_global_var_decls): Remove.
9426 * Makefile.in (OBJS): Remove lto-symtab.o.
9427 (lto-symtab.o): Remove.
9428 (GTFILES): Remove lto-symtab.c
9429 * lto-symtab.c: Move to lto/
9431 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9433 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
9434 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
9435 (UNSPEC_FPINT_RINT): New constant definitions.
9436 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
9437 definition with 2 attributes.
9438 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
9439 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
9442 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9444 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
9445 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
9446 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
9447 attribute to "z196_alone".
9448 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
9451 2013-09-06 Richard Biener <rguenther@suse.de>
9453 * basic-block.h (class control_dependences): New.
9454 * tree-ssa-dce.c (control_dependence_map): Remove.
9456 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
9457 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
9458 find_pdom, find_control_dependence, find_all_control_dependences):
9460 (mark_control_dependent_edges_necessary,
9461 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
9462 tree_dce_done, perform_tree_ssa_dce): Adjust.
9463 * cfganal.c (set_control_dependence_map_bit,
9464 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
9465 find_all_control_dependences): Move from tree-ssa-dce.c and
9466 implement as methods of control_dependences class.
9467 (control_dependences::control_dependences): New.
9468 (control_dependences::~control_dependences): Likewise.
9469 (control_dependences::get_edges_dependent_on): Likewise.
9470 (control_dependences::get_edge): Likewise.
9472 2013-09-04 Jan Hubicka <jh@suse.cz>
9474 * tree.c (types_same_for_odr): Drop overactive check.
9475 * ipa-devirt.c (hash_type_name): Likewise.
9477 2013-09-04 Jan Hubicka <jh@suse.cz>
9479 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
9480 (analyze_functions): ... here.
9482 2013-09-04 Jan Hubicka <jh@suse.cz>
9485 * cgraphunit.c (analyze_functions): Clear AUX fields
9486 after processing; initialize assembler name has.
9488 2013-09-05 Jeff Law <law@redhat.com>
9490 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
9491 from thread_around_empty_block. Record threading path into PATH.
9492 Recurse if threading through the initial block is successful.
9493 (thread_across_edge): Corresponding changes to slightly simplify.
9495 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9497 * config/aarch64/aarch64.md
9498 (type): Remove frecpe, frecps, frecpx.
9499 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
9500 fix to be a TARGET_SIMD instruction.
9501 (aarch64_frecps): Remove.
9502 * config/aarch64/aarch64-simd.md
9503 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
9504 (aarch64_frecps<mode>): Handle all float/vector of float modes.
9506 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9507 Sofiane Naci <sofiane.naci@arm.com>
9509 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
9510 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
9511 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
9512 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
9513 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
9514 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
9515 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
9516 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
9517 into "clz, "rbit". Rename "shift" to "shift_imm".
9518 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
9519 changes. Update for attribute changes all occurrences of arlo_* and
9521 * config/arm/arm-fixed.md: Update for attribute changes
9522 all occurrences of arlo_* types.
9523 * config/arm/thumb2.md: Update for attribute changes all occurrences
9525 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
9526 (cortexa7_older_only): Likewise.
9527 (cortexa7_younger): Likewise.
9528 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9529 (1020alu_shift_op): Likewise.
9530 (1020alu_shift_reg_op): Likewise.
9531 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9532 (alu_shift_op): Likewise.
9533 (alu_shift_reg_op): Likewise.
9534 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9535 (11_alu_shift_op): Likewise.
9536 (11_alu_shift_reg_op): Likewise.
9537 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9538 (9_alu_shift_reg_op): Likewise.
9539 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
9541 (cortex_a15_alu_shift): Likewise.
9542 (cortex_a15_alu_shift_reg): Likewise.
9543 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
9545 (cortex_a5_alu_shift): Likewise.
9546 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
9548 (cortex_a53_alu_shift): Likewise.
9549 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
9551 (cortex_a7_alu_reg): Likewise.
9552 (cortex_a7_alu_shift): Likewise.
9553 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
9555 (cortex_a8_alu_shift): Likewise.
9556 (cortex_a8_alu_shift_reg): Likewise.
9557 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9558 (cortex_a9_dp_shift): Likewise.
9559 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
9561 * config/arm/cortex-r4.md
9562 (cortex_r4_alu): Update for attribute changes.
9563 (cortex_r4_mov): Likewise.
9564 (cortex_r4_alu_shift_reg): Likewise.
9565 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9566 (526_alu_shift_op): Likewise.
9567 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9568 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9569 (626te_alu_shift_op): Likewise.
9570 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
9571 (726te_alu_shift_op): Likewise.
9572 (726te_alu_shift_reg_op): Likewise.
9573 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9574 (mp626_alu_shift_op): Likewise.
9575 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
9576 (pj4_alu_conds): Likewise.
9577 (pj4_shift): Likewise.
9578 (pj4_shift_conds): Likewise.
9579 (pj4_alu_shift): Likewise.
9580 (pj4_alu_shift_conds): Likewise.
9581 * config/aarch64/aarch64.md: Update for attribute change
9582 all occurrences of arlo_* and shift* types.
9584 2013-09-05 Mike Stump <mikestump@comcast.net>
9586 * tree.h: Move documentation for tree_function_decl to tree-core.h
9587 with the declaration.
9589 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
9592 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
9593 looking for widest mode.
9595 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
9597 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
9599 2013-09-05 Richard Biener <rguenther@suse.de>
9601 PR tree-optimization/58137
9602 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
9603 Do not create vectors of pointers.
9604 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
9605 types for the components of the vector initializer.
9606 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
9607 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
9609 2013-09-05 Martin Jambor <mjambor@suse.cz>
9611 * ipa-prop.c (remove_described_reference): Accept missing references,
9612 return false if that hppens, otherwise return true.
9613 (cgraph_node_for_jfunc): New function.
9614 (try_decrement_rdesc_refcount): Likewise.
9615 (try_make_edge_direct_simple_call): Use them.
9616 (ipa_edge_removal_hook): Remove references from rdescs.
9617 (ipa_edge_duplication_hook): Clone rdescs and their references
9618 when the new edge has the same caller as the old one.
9619 * cgraph.c (cgraph_resolve_speculation): Remove speculative
9620 reference before removing any edges.
9622 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
9624 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
9626 * thumb2.md (thumb2_storewb_parisi): New pattern.
9628 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
9630 * config/aarch64/aarch64-option-extensions.def: Add
9631 AARCH64_OPT_EXTENSION of 'crc'.
9632 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
9633 (AARCH64_ISA_CRC): Ditto.
9634 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
9635 description of the CRC extension.
9637 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
9639 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
9640 * config/rs6000/linux.h: Ditto.
9641 * alpha/linux.h: Ditto.
9642 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
9643 no_c99_libc_has_function.
9644 * config/c6x/uclinux-elf.h: Ditto.
9645 * config/lm32/uclinux-elf.h: Ditto.
9646 * config/m68k/uclinux.h: Ditto.
9647 * config/moxie/uclinux.h: Ditto.
9648 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
9649 (crisv32-*-linux*, cris-*-linux*): Ditto.
9650 * config/bfin/bfin.c: Include "tm_p.h".
9652 2013-09-05 Richard Biener <rguenther@suse.de>
9654 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
9655 check for a definition without a basic-block.
9657 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9658 Sofiane Naci <sofiane.naci@arm.com>
9660 * config/aarch64/aarch64.md
9661 (*movti_aarch64): Rename r_2_f and f_2_r.
9662 (*movsf_aarch64): Likewise.
9663 (*movdf_aarch64): Likewise.
9664 (*movtf_aarch64): Likewise.
9665 (aarch64_movdi_<mode>low): Likewise.
9666 (aarch64_movdi_<mode>high): Likewise.
9667 (aarch64_mov<mode>high_di): Likewise.
9668 (aarch64_mov<mode>low_di): Likewise.
9669 (aarch64_movtilow_tilow): Likewise.
9670 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
9671 values to config/arm/types.md
9672 (attribute "conds"): Update for attribute change.
9673 (anddi3_insn): Likewise.
9674 (iordi3_insn): Likewise.
9675 (xordi3_insn): Likewise.
9676 (one_cmpldi2): Likewise.
9677 * config/arm/types.md (type): Add Neon types.
9678 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
9679 use "type" attribute.
9680 (movmisalign<mode>_neon_store): Likewise.
9681 (movmisalign<mode>_neon_load): Likewise.
9682 (vec_set<mode>_internal): Likewise.
9683 (vec_setv2di_internal): Likewise.
9684 (vec_extract<mode>): Likewise.
9685 (vec_extractv2di): Likewise.
9686 (add<mode>3_neon): Likewise.
9687 (adddi3_neon): Likewise.
9688 (sub<mode>3_neon): Likewise.
9689 (subdi3_neon): Likewise.
9690 (mul<mode>3_neon): Likewise.
9691 (mul<mode>3add<mode>_neon): Likewise.
9692 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
9693 (fma<VCVTF:mode>4)): Likewise.
9694 (fma<VCVTF:mode>4_intrinsic): Likewise.
9695 (fmsub<VCVTF:mode>4)): Likewise.
9696 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
9697 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
9698 (ior<mode>3): Likewise.
9699 (and<mode>3): Likewise.
9700 (anddi3_neon): Likewise.
9701 (orn<mode>3_neon): Likewise.
9702 (orndi3_neon): Likewise.
9703 (bic<mode>3_neon): Likewise.
9704 (bicdi3_neon): Likewise.
9705 (xor<mode>3): Likewise.
9706 (one_cmpl<mode>2): Likewise.
9707 (abs<mode>2): Likewise.
9708 (neg<mode>2): Likewise.
9709 (umin<mode>3_neon): Likewise.
9710 (umax<mode>3_neon): Likewise.
9711 (smin<mode>3_neon): Likewise.
9712 (smax<mode>3_neon): Likewise.
9713 (vashl<mode>3): Likewise.
9714 (vashr<mode>3_imm): Likewise.
9715 (vlshr<mode>3_imm): Likewise.
9716 (ashl<mode>3_signed): Likewise.
9717 (ashl<mode>3_unsigned): Likewise.
9718 (neon_load_count): Likewise.
9719 (ashldi3_neon_noclobber): Likewise.
9720 (signed_shift_di3_neon): Likewise.
9721 (unsigned_shift_di3_neon): Likewise.
9722 (ashrdi3_neon_imm_noclobber): Likewise.
9723 (lshrdi3_neon_imm_noclobber): Likewise.
9724 (widen_ssum<mode>3): Likewise.
9725 (widen_usum<mode>3): Likewise.
9726 (quad_halves_<code>v4si): Likewise.
9727 (quad_halves_<code>v4sf): Likewise.
9728 (quad_halves_<code>v8hi): Likewise.
9729 (quad_halves_<code>v16qi): Likewise.
9730 (reduc_splus_v2di): Likewise.
9731 (neon_vpadd_internal<mode>): Likewise.
9732 (neon_vpsmin<mode>): Likewise.
9733 (neon_vpsmax<mode>): Likewise.
9734 (neon_vpumin<mode>): Likewise.
9735 (neon_vpumax<mode>): Likewise.
9736 (ss_add<mode>_neon): Likewise.
9737 (us_add<mode>_neon): Likewise.
9738 (ss_sub<mode>_neon): Likewise.
9739 (us_sub<mode>_neon): Likewise.
9740 (neon_vadd<mode>_unspec): Likewise.
9741 (neon_vaddl<mode>): Likewise.
9742 (neon_vaddw<mode>): Likewise.
9743 (neon_vhadd<mode>): Likewise.
9744 (neon_vqadd<mode>): Likewise.
9745 (neon_vaddhn<mode>): Likewise.
9746 (neon_vmul<mode>): Likewise.
9747 (neon_vmla<mode>): Likewise.
9748 (neon_vmlal<mode>): Likewise.
9749 (neon_vmls<mode>): Likewise.
9750 (neon_vmlsl<mode>): Likewise.
9751 (neon_vqdmulh<mode>): Likewise.
9752 (neon_vqdmlal<mode>): Likewise.
9753 (neon_vqdmlsl<mode>): Likewise.
9754 (neon_vmull<mode>): Likewise.
9755 (neon_vqdmull<mode>): Likewise.
9756 (neon_vsub<mode>_unspec): Likewise.
9757 (neon_vsubl<mode>): Likewise.
9758 (neon_vsubw<mode>): Likewise.
9759 (neon_vqsub<mode>): Likewise.
9760 (neon_vhsub<mode>): Likewise.
9761 (neon_vsubhn<mode>): Likewise.
9762 (neon_vceq<mode>): Likewise.
9763 (neon_vcge<mode>): Likewise.
9764 (neon_vcgeu<mode>): Likewise.
9765 (neon_vcgt<mode>): Likewise.
9766 (neon_vcgtu<mode>): Likewise.
9767 (neon_vcle<mode>): Likewise.
9768 (neon_vclt<mode>): Likewise.
9769 (neon_vcage<mode>): Likewise.
9770 (neon_vcagt<mode>): Likewise.
9771 (neon_vtst<mode>): Likewise.
9772 (neon_vabd<mode>): Likewise.
9773 (neon_vabdl<mode>): Likewise.
9774 (neon_vaba<mode>): Likewise.
9775 (neon_vabal<mode>): Likewise.
9776 (neon_vmax<mode>): Likewise.
9777 (neon_vmin<mode>): Likewise.
9778 (neon_vpaddl<mode>): Likewise.
9779 (neon_vpadal<mode>): Likewise.
9780 (neon_vpmax<mode>): Likewise.
9781 (neon_vpmin<mode>): Likewise.
9782 (neon_vrecps<mode>): Likewise.
9783 (neon_vrsqrts<mode>): Likewise.
9784 (neon_vqabs<mode>): Likewise.
9785 (neon_vqneg<mode>): Likewise.
9786 (neon_vcls<mode>): Likewise.
9787 (clz<mode>2): Likewise.
9788 (popcount<mode>2): Likewise.
9789 (neon_vrecpe): Likewise.
9790 (neon_vrsqrte): Likewise.
9791 (neon_vget_lane<mode>_sext_internal): Likewise.
9792 (neon_vget_lane<mode>_zext_internal): Likewise.
9793 (neon_vdup_n<mode>): Likewise.
9794 (neon_vdup_nv2di): Likewise.
9795 (neon_vdpu_lane<mode>_internal): Likewise.
9796 (neon_vswp<mode>): Likewise.
9797 (float<mode><V_cvtto>2): Likewise.
9798 (floatuns<mode><V_cvtto>2): Likewise.
9799 (fix_trunc<mode><V_cvtto>)2): Likewise
9800 (fixuns_trunc<mode><V_cvtto)2): Likewise.
9801 (neon_vcvt<mode>): Likewise.
9802 (neon_vcvtv4sfv4hf): Likewise.
9803 (neon_vcvtv4hfv4sf): Likewise.
9804 (neon_vcvt_n<mode>): Likewise.
9805 (neon_vmovn<mode>): Likewise.
9806 (neon_vqmovn<mode>): Likewise.
9807 (neon_vqmovun<mode>): Likewise.
9808 (neon_vmovl<mode>): Likewise.
9809 (neon_vmul_lane<mode>): Likewise.
9810 (neon_vmull_lane<mode>): Likewise.
9811 (neon_vqdmull_lane<mode>): Likewise.
9812 (neon_vqdmulh_lane<mode>): Likewise.
9813 (neon_vmla_lane<mode>): Likewise.
9814 (neon_vmlal_lane<mode>): Likewise.
9815 (neon_vqdmlal_lane<mode>): Likewise.
9816 (neon_vmls_lane<mode>): Likewise.
9817 (neon_vmlsl_lane<mode>): Likewise.
9818 (neon_vqdmlsl_lane<mode>): Likewise.
9819 (neon_vext<mode>): Likewise.
9820 (neon_vrev64<mode>): Likewise.
9821 (neon_vrev32<mode>): Likewise.
9822 (neon_vrev16<mode>): Likewise.
9823 (neon_vbsl<mode>_internal): Likewise.
9824 (neon_vshl<mode>): Likewise.
9825 (neon_vqshl<mode>): Likewise.
9826 (neon_vshr_n<mode>): Likewise.
9827 (neon_vshrn_n<mode>): Likewise.
9828 (neon_vqshrn_n<mode>): Likewise.
9829 (neon_vqshrun_n<mode>): Likewise.
9830 (neon_vshl_n<mode>): Likewise.
9831 (neon_vqshl_n<mode>): Likewise.
9832 (neon_vqshlu_n<mode>): Likewise.
9833 (neon_vshll_n<mode>): Likewise.
9834 (neon_vsra_n<mode>): Likewise.
9835 (neon_vsri_n<mode>): Likewise.
9836 (neon_vsli_n<mode>): Likewise.
9837 (neon_vtbl1v8qi): Likewise.
9838 (neon_vtbl2v8qi): Likewise.
9839 (neon_vtbl3v8qi): Likewise.
9840 (neon_vtbl4v8qi): Likewise.
9841 (neon_vtbx1v8qi): Likewise.
9842 (neon_vtbx2v8qi): Likewise.
9843 (neon_vtbx3v8qi): Likewise.
9844 (neon_vtbx4v8qi): Likewise.
9845 (neon_vtrn<mode>_internal): Likewise.
9846 (neon_vzip<mode>_internal): Likewise.
9847 (neon_vuzp<mode>_internal): Likewise.
9848 (neon_vld1<mode>): Likewise.
9849 (neon_vld1_lane<mode>): Likewise.
9850 (neon_vld1_dup<mode>): Likewise.
9851 (neon_vld1_dupv2di): Likewise.
9852 (neon_vst1<mode>): Likewise.
9853 (neon_vst1_lane<mode>): Likewise.
9854 (neon_vld2<mode>): Likewise.
9855 (neon_vld2_lane<mode>): Likewise.
9856 (neon_vld2_dup<mode>): Likewise.
9857 (neon_vst2<mode>): Likewise.
9858 (neon_vst2_lane<mode>): Likewise.
9859 (neon_vld3<mode>): Likewise.
9860 (neon_vld3qa<mode>): Likewise.
9861 (neon_vld3qb<mode>): Likewise.
9862 (neon_vld3_lane<mode>): Likewise.
9863 (neon_vld3_dup<mode>): Likewise.
9864 (neon_vst3<mode>): Likewise.
9865 (neon_vst3qa<mode>): Likewise.
9866 (neon_vst3qb<mode>): Likewise.
9867 (neon_vst3_lane<mode>): Likewise.
9868 (neon_vld4<mode>): Likewise.
9869 (neon_vld4qa<mode>): Likewise.
9870 (neon_vld4qb<mode>): Likewise.
9871 (neon_vld4_lane<mode>): Likewise.
9872 (neon_vld4_dup<mode>): Likewise.
9873 (neon_vst4<mode>): Likewise.
9874 (neon_vst4qa<mode>): Likewise.
9875 (neon_vst4qb<mode>): Likewise.
9876 (neon_vst4_lane<mode>): Likewise.
9877 (neon_vec_unpack<US>_lo_<mode>): Likewise.
9878 (neon_vec_unpack<US>_hi_<mode>): Likewise.
9879 (neon_vec_<US>mult_lo_<mode>): Likewise.
9880 (neon_vec_<US>mult_hi_<mode>): Likewise.
9881 (neon_vec_<US>shiftl_<mode>): Likewise.
9882 (neon_unpack<US>_<mode>): Likewise.
9883 (neon_vec_<US>mult_<mode>): Likewise.
9884 (vec_pack_trunc_<mode>): Likewise.
9885 (neon_vec_pack_trunk_<mode>): Likewise.
9886 (neon_vabd<mode>_2): Likewise.
9887 (neon_vabd<mode>_3): Likewise.
9888 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
9889 (thumb2_movsi_vfp): Likewise.
9890 (movdi_vfp): Likewise.
9891 (movdi_vfp_cortexa8): Likewise.
9892 (movhf_vfp_neon): Likewise.
9893 (movhf_vfp): Likewiwse.
9894 (movsf_vfp): Likewiwse.
9895 (thumb2_movsf_vfp): Likewiwse.
9896 (movdf_vfp): Likewise.
9897 (thumb2_movdf_vfp): Likewise.
9898 (movsfcc_vfp): Likewise.
9899 (thumb2_movsfcc_vfp): Likewise.
9900 (movdfcc_vfp): Likewise.
9901 (thumb2_movdfcc_vfp): Likewise.
9902 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
9903 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
9904 (v10_v2c): Likewise.
9905 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
9907 (cortex_a15_neon_int_2): Likewise.
9908 (cortex_a15_neon_int_3): Likewise.
9909 (cortex_a15_neon_int_4): Likewise.
9910 (cortex_a15_neon_int_5): Likewise.
9911 (cortex_a15_neon_vqneg_vqabs): Likewise.
9912 (cortex_a15_neon_vmov): Likewise.
9913 (cortex_a15_neon_vaba): Likewise.
9914 (cortex_a15_neon_vaba_qqq): Likewise.
9915 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9916 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9917 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
9918 scalar_64_32_long_scalar): Likewise.
9919 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9920 (cortex_a15_neon_mla_qqq_8_16): Likewise.
9921 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9922 lotype_qdd_64_32_long): Likewise.
9923 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9924 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9925 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
9926 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9927 (cortex_a15_neon_shift_1): Likewise.
9928 (cortex_a15_neon_shift_2): Likewise.
9929 (cortex_a15_neon_shift_3): Likewise.
9930 (cortex_a15_neon_vshl_ddd): Likewise.
9931 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9932 (cortex_a15_neon_vsra_vrsra): Likewise.
9933 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
9934 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
9935 (cortex_a15_neon_fp_vmul_ddd): Likewise.
9936 (cortex_a15_neon_fp_vmul_qqd): Likewise.
9937 (cortex_a15_neon_fp_vmla_ddd): Likewise.
9938 (cortex_a15_neon_fp_vmla_qqq): Likewise.
9939 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
9940 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
9941 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9942 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9943 (cortex_a15_neon_bp_simple): Likewise.
9944 (cortex_a15_neon_bp_2cycle): Likewise.
9945 (cortex_a15_neon_bp_3cycle): Likewise.
9946 (cortex_a15_neon_vld1_1_2_regs): Likewise.
9947 (cortex_a15_neon_vld1_3_4_regs): Likewise.
9948 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9949 (cortex_a15_neon_vld2_4_regs): Likewise.
9950 (cortex_a15_neon_vld3_vld4): Likewise.
9951 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9952 (cortex_a15_neon_vst1_3_4_regs): Likewise.
9953 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
9954 (cortex_a15_neon_vst3_vst4): Likewise.
9955 (cortex_a15_neon_vld1_vld2_lane): Likewise.
9956 (cortex_a15_neon_vld3_vld4_lane" 10
9957 (cortex_a15_neon_vst1_vst2_lane): Likewise.
9958 (cortex_a15_neon_vst3_vst4_lane): Likewise.
9959 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
9960 (cortex_a15_neon_ldm_2): Likewise.0
9961 (cortex_a15_neon_stm_2): Likewise.
9962 (cortex_a15_neon_mcr): Likewise.
9963 (cortex_a15_neon_mcr_2_mcrr): Likewise.
9964 (cortex_a15_neon_mrc): Likewise.
9965 (cortex_a15_neon_mrrc): Likewise.
9966 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9968 (cortex_a15_alu_shift): Likewise.
9969 (cortex_a15_alu_shift_reg): Likewise.
9970 (cortex_a15_mult32): Likewise.
9971 (cortex_a15_mult64): Likewise.
9972 (cortex_a15_block): Likewise.
9973 (cortex_a15_branch): Likewise.
9974 (cortex_a15_load1): Likewise.
9975 (cortex_a15_load3): Likewise.
9976 (cortex_a15_store1): Likewise.
9977 (cortex_a15_store3): Likewise.
9978 (cortex_a15_call): Likewise.
9979 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
9980 (cortex_a5_f2r): Likewise.
9981 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
9983 (cortex_a53_f2r): Likewise.
9984 * config/arm/cortex-a7.md
9985 (cortex_a7_branch): Update for attribute change.
9986 (cortex_a7_call): Likewise.
9987 (cortex_a7_alu_imm): Likewise.
9988 (cortex_a7_alu_reg): Likewise.
9989 (cortex_a7_alu_shift): Likewise.
9990 (cortex_a7_mul): Likewise.
9991 (cortex_a7_load1): Likewise.
9992 (cortex_a7_store1): Likewise.
9993 (cortex_a7_load2): Likewise.
9994 (cortex_a7_store2): Likewise.
9995 (cortex_a7_load3): Likewise.
9996 (cortex_a7_store3): Likewise.
9997 (cortex_a7_load4): Likewise.
9998 (cortex_a7_store4): Likewise.
9999 (cortex_a7_fpalu): Likewise.
10000 (cortex_a7_fconst): Likewise.
10001 (cortex_a7_fpmuls): Likewise.
10002 (cortex_a7_neon_mul): Likewise.
10003 (cortex_a7_fpmacs): Likewise.
10004 (cortex_a7_neon_mla: Likewise.
10005 (cortex_a7_fpmuld: Likewise.
10006 (cortex_a7_fpmacd: Likewise.
10007 (cortex_a7_fpfmad: Likewise.
10008 (cortex_a7_fdivs: Likewise.
10009 (cortex_a7_fdivd: Likewise.
10010 (cortex_a7_r2f: Likewise.
10011 (cortex_a7_f2r: Likewise.
10012 (cortex_a7_f_flags: Likewise.
10013 (cortex_a7_f_loads: Likewise.
10014 (cortex_a7_f_loadd: Likewise.
10015 (cortex_a7_f_stores: Likewise.
10016 (cortex_a7_f_stored: Likewise.
10017 (cortex_a7_neon): Likewise.
10018 * config/arm/cortex-a8-neon.md
10019 (cortex_a8_neon_mrc): Update for attribute change.
10020 (cortex_a8_neon_mrrc): Likewise.
10021 (cortex_a8_neon_int_1): Likewise.
10022 (cortex_a8_neon_int_2): Likewise.
10023 (cortex_a8_neon_int_3): Likewise.
10024 (cortex_a8_neon_int_4): Likewise.
10025 (cortex_a8_neon_int_5): Likewise.
10026 (cortex_a8_neon_vqneg_vqabs): Likewise.
10027 (cortex_a8_neon_vmov): Likewise.
10028 (cortex_a8_neon_vaba): Likewise.
10029 (cortex_a8_neon_vaba_qqq): Likewise.
10030 (cortex_a8_neon_vsma): Likewise.
10031 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10032 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10033 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10034 long_scalar): Likewise.
10035 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10036 (cortex_a8_neon_mla_qqq_8_16): Likewise.
10037 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10038 long_scalar_qdd_64_32_long): Likewise.
10039 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10040 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10041 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
10042 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10043 (cortex_a8_neon_shift_1): Likewise.
10044 (cortex_a8_neon_shift_2): Likewise.
10045 (cortex_a8_neon_shift_3): Likewise.
10046 (cortex_a8_neon_vshl_ddd): Likewise.
10047 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10048 (cortex_a8_neon_vsra_vrsra): Likewise.
10049 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
10050 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
10051 (cortex_a8_neon_fp_vsum): Likewise.
10052 (cortex_a8_neon_fp_vmul_ddd): Likewise.
10053 (cortex_a8_neon_fp_vmul_qqd): Likewise.
10054 (cortex_a8_neon_fp_vmla_ddd): Likewise.
10055 (cortex_a8_neon_fp_vmla_qqq): Likewise.
10056 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
10057 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
10058 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10059 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10060 (cortex_a8_neon_bp_simple): Likewise.
10061 (cortex_a8_neon_bp_2cycle): Likewise.
10062 (cortex_a8_neon_bp_3cycle): Likewise.
10063 (cortex_a8_neon_ldr): Likewise.
10064 (cortex_a8_neon_str): Likewise.
10065 (cortex_a8_neon_vld1_1_2_regs): Likewise.
10066 (cortex_a8_neon_vld1_3_4_regs): Likewise.
10067 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10068 (cortex_a8_neon_vld2_4_regs): Likewise.
10069 (cortex_a8_neon_vld3_vld4): Likewise.
10070 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10071 (cortex_a8_neon_vst1_3_4_regs): Likewise.
10072 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
10073 (cortex_a8_neon_vst3_vst4): Likewise.
10074 (cortex_a8_neon_vld1_vld2_lane): Likewise.
10075 (cortex_a8_neon_vld3_vld4_lane): Likewise.
10076 (cortex_a8_neon_vst1_vst2_lane): Likewise.
10077 (cortex_a8_neon_vst3_vst4_lane): Likewise.
10078 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
10079 (cortex_a8_neon_mcr): Likewise.
10080 (cortex_a8_neon_mcr_2_mcrr): Likewise.
10081 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10082 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
10084 (ca9_neon_mrrc): Likewise.
10085 (cortex_a9_neon_int_1): Likewise.
10086 (cortex_a9_neon_int_2): Likewise.
10087 (cortex_a9_neon_int_3): Likewise.
10088 (cortex_a9_neon_int_4): Likewise.
10089 (cortex_a9_neon_int_5): Likewise.
10090 (cortex_a9_neon_vqneg_vqabs): Likewise.
10091 (cortex_a9_neon_vmov): Likewise.
10092 (cortex_a9_neon_vaba): Likewise.
10093 (cortex_a9_neon_vaba_qqq): Likewise.
10094 (cortex_a9_neon_vsma): Likewise.
10095 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10096 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10097 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10098 long_scalar): Likewise.
10099 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10100 (cortex_a9_neon_mla_qqq_8_16): Likewise.
10101 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10102 long_scalar_qdd_64_32_long): Likewise.
10103 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10104 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10105 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
10106 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10107 (cortex_a9_neon_shift_1): Likewise.
10108 (cortex_a9_neon_shift_2): Likewise.
10109 (cortex_a9_neon_shift_3): Likewise.
10110 (cortex_a9_neon_vshl_ddd): Likewise.
10111 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10112 (cortex_a9_neon_vsra_vrsra): Likewise.
10113 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
10114 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
10115 (cortex_a9_neon_fp_vsum): Likewise.
10116 (cortex_a9_neon_fp_vmul_ddd): Likewise.
10117 (cortex_a9_neon_fp_vmul_qqd): Likewise.
10118 (cortex_a9_neon_fp_vmla_ddd): Likewise.
10119 (cortex_a9_neon_fp_vmla_qqq): Likewise.
10120 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
10121 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
10122 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10123 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10124 (cortex_a9_neon_bp_simple): Likewise.
10125 (cortex_a9_neon_bp_2cycle): Likewise.
10126 (cortex_a9_neon_bp_3cycle): Likewise.
10127 (cortex_a9_neon_ldr): Likewise.
10128 (cortex_a9_neon_str): Likewise.
10129 (cortex_a9_neon_vld1_1_2_regs): Likewise.
10130 (cortex_a9_neon_vld1_3_4_regs): Likewise.
10131 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10132 (cortex_a9_neon_vld2_4_regs): Likewise.
10133 (cortex_a9_neon_vld3_vld4): Likewise.
10134 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10135 (cortex_a9_neon_vst1_3_4_regs): Likewise.
10136 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
10137 (cortex_a9_neon_vst3_vst4): Likewise.
10138 (cortex_a9_neon_vld1_vld2_lane): Likewise.
10139 (cortex_a9_neon_vld3_vld4_lane): Likewise.
10140 (cortex_a9_neon_vst1_vst2_lane): Likewise.
10141 (cortex_a9_neon_vst3_vst4_lane): Likewise.
10142 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
10143 (cortex_a9_neon_mcr): Likewise.
10144 (cortex_a9_neon_mcr_2_mcrr): Likewise.
10145 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
10146 (cortex_a9_fps): Likewise.
10147 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
10149 (cortex_m4_fmuls): Likewise.
10150 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
10152 (cortex_r4_mrc): Likewise.
10153 * config/arm/iterators.md: Update comment referring to neon_type.
10154 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
10155 (iwmmxt_movsi_insn): Likewise.
10156 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
10158 (pj4_core_to_vfp): Likewise.
10159 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
10161 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
10162 (vfp_fstore): Likewise.
10163 * doc/md.texi: Change references to neon_type to refer to type.
10165 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10167 * tree.h (DECL_BUILT_IN): Fix typo in comment.
10169 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
10171 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
10172 lglobl if not weak.
10174 2013-09-04 Easwaran Raman <eraman@google.com>
10176 PR middle-end/57370
10177 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
10178 (build_and_add_sum): Use it.
10179 (appears_later_in_bb): Simplify code.
10181 2013-09-04 Teresa Johnson <tejohnson@google.com>
10183 * dumpfile.c (dump_finish): Don't close stderr/stdout.
10185 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
10187 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
10189 2013-09-04 Jan Hubicka <jh@suse.cz>
10191 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
10192 * ipa-devirt.c: Include diganostic.h
10193 (odr_type_d): Add types and types_set.
10194 (hash_type_name): Work for types with vtables during LTO.
10195 (odr_hasher::remove): Fix comment; destroy types_set.
10196 (add_type_duplicate): New function,
10197 (get_odr_type): Use it.
10198 (dump_type_inheritance_graph): Dump type duplicates.
10199 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
10201 * tree.c (types_same_for_odr): Give exact answers on types with
10204 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10206 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
10207 explaining their differences.
10209 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
10211 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
10213 2013-09-03 Jeff Law <law@redhat.com>
10215 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
10216 when not threading through a joiner block. Pass joiner/no joiner
10217 state to register_jump_thread.
10218 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
10219 state from argument rather than implying on path length.
10220 Dump the entire jump thread path into debugging dump.
10221 * tree-flow.h (register_jump_thread): Update prototype.
10223 2013-08-29 Xinliang David Li <davidxl@google.com>
10225 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10226 Remove a trivial gcc_assert.
10228 2013-08-29 Xinliang David Li <davidxl@google.com>
10230 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
10231 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
10232 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10233 Delay base decl alignment adjustment.
10234 * tree-vectorizer.c (vect_destroy_datarefs): New function.
10235 * tree-vectorizer.h: New data structure.
10236 (set_dr_misalignment): New function.
10237 (dr_misalignment): Ditto.
10238 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
10239 (vectorizable_load): Ditto.
10240 (ensure_base_align): New function.
10241 (vectorize_loops): Add dbg_cnt support.
10242 (execute_vect_slp): Ditto.
10243 * dbgcnt.def: New debug counter.
10244 * Makefile: New dependency.
10246 2013-09-03 Meador Inge <meadori@codesourcery.com>
10250 2013-08-30 Meador Inge <meadori@codesourcery.com>
10252 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10254 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
10256 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
10257 function descriptor.
10259 2013-09-03 Richard Biener <rguenther@suse.de>
10261 * tree-affine.c (add_elt_to_tree): Fix association issue,
10262 avoid useless converts and make sure to always return a
10263 properly typed result.
10265 2013-09-03 Richard Biener <rguenther@suse.de>
10267 PR middle-end/57656
10268 * fold-const.c (negate_expr_p): Fix division case.
10269 (negate_expr): Likewise.
10271 2013-09-03 Richard Biener <rguenther@suse.de>
10274 * tree-streamer-out.c: Include tm.h.
10275 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
10277 2013-09-03 Jan Hubicka <jh@suse.cz>
10279 * tree-profile.c (tree_profiling): Cleanup CFG when done.
10281 2013-09-03 Alan Modra <amodra@gmail.com>
10283 * config.gcc (powerpc*-*-linux*): Add support for little-endian
10284 multilibs to big-endian target and vice versa.
10285 * config/rs6000/t-linux64: Use := assignment on all vars.
10286 (MULTILIB_EXTRA_OPTS): Remove fPIC.
10287 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
10288 * config/rs6000/t-linux64le: New file.
10289 * config/rs6000/t-linux64bele: New file.
10290 * config/rs6000/t-linux64lebe: New file.
10292 2013-09-02 Jan Hubicka <jh@suse.cz>
10294 * ipa-inline-transform.c (inline_transform): Do not
10295 optimize_inline_calls when not optimizing.
10297 2013-09-02 Jan Hubicka <jh@suse.cz>
10299 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
10300 duplicated nodes for assembler names.
10301 * symtab.c (symtab_unregister_node): Do not attempt to unlink
10302 hard registers from assembler name hash.
10304 2013-09-02 Jan Hubicka <jh@suse.cz>
10306 * ipa-split.c (execute_split_functions): Split externally visible
10307 functions called once.
10309 2013-09-02 Martin Jambor <mjambor@suse.cz>
10312 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
10313 linked list. When finding the correct duplicate, also consider also
10314 the caller in additon to its inlined_to node.
10316 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
10318 * config/aarch64/aarch64-simd-builtins.def
10319 (dup_lane_scalar): Remove.
10320 * config/aarch64/aarch64-simd.md
10321 (aarch64_simd_dup): Add 'w->w' alternative.
10322 (aarch64_dup_lane<mode>): Allow for VALL.
10323 (aarch64_dup_lane_scalar<mode>): Remove.
10324 (aarch64_dup_lane_<vswap_width_name><mode>): New.
10325 (aarch64_get_lane_signed<mode>): Add w->w altenative.
10326 (aarch64_get_lane_unsigned<mode>): Likewise.
10327 (aarch64_get_lane<mode>): Likewise.
10328 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
10329 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
10330 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
10331 (VCON): Change container of V2SF.
10332 (vswap_width_name): Likewise.
10333 * config/aarch64/arm_neon.h
10334 (__aarch64_vdup_lane_any): New.
10335 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10336 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
10337 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10339 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
10341 PR middle-end/56382
10342 * expr.c (emit_move_complex): Do not move complex FP values as parts if
10343 the source or the destination is a single hard register.
10345 2013-09-02 Richard Biener <rguenther@suse.de>
10347 PR middle-end/57511
10348 * tree-scalar-evolution.c (instantiate_scev_name): Allow
10351 2013-09-02 Richard Biener <rguenther@suse.de>
10353 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
10354 arithmetic to sizetype.
10356 2013-09-02 Bin Cheng <bin.cheng@arm.com>
10358 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
10359 Find auto-increment use both before and after candidate.
10361 2013-09-02 Marek Polacek <polacek@redhat.com>
10363 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
10365 2013-09-01 Jan Hubicka <jh@suse.cz>
10367 * Makefile.in: Add ipa-profile.o
10368 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
10369 * cgraph.c (struct cgraph_propagate_frequency_data,
10370 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
10371 ipa-profile.c; replace cgraph_ by ipa_ prefix.
10372 * cgraph.h (cgraph_propagate_frequency): Remove.
10373 * ipa-inline-analysis.c: Include ipa-utils.h;
10374 drop duplicated cfgloop.h.
10375 (inline_update_callee_summaries): Update.
10376 * ipa-profile.c: New file.
10377 * ipa-utils.h (ipa_propagate_frequency): Declare.
10378 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
10379 data-streamer.h, value-prof.h.
10380 (symtab_remove_unreachable_nodes): Update profile.
10381 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
10382 account_time_size, cmp_counts, dump_histogram,
10383 ipa_profile_generate_summary, ipa_profile_write_summary,
10384 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
10385 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
10386 Move to ipa-profile.c.
10388 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
10390 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
10392 2013-09-01 Jan Hubicka <jh@suse.cz>
10394 * common.opt (fdevirtualize-speculatively): New function.
10395 * invoke.texi (fdevirtualize-speculatively): Document.
10396 * ipa-devirt.c: Include ipa-inline.h
10397 (likely_target_p): New function.
10398 (ipa_devirt): New function.
10399 (gate_ipa_devirt): New function.
10400 (pass_data_ipa_devirt): New static var.
10401 (pass_ipa_devirt): Likewise.
10402 (make_pass_ipa_devirt): New function.
10403 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
10404 (common_handle_option): Disable devirtualization when
10405 value range profiling is available.
10406 * passes.def (pass_ipa_devirt): Add.
10407 * timever.def (TV_IPA_DEVIRT): New timevar.
10408 * tree-pass.h (make_pass_ipa_devirt):
10410 2013-09-01 Iain Sandoe <iain@codesourcery.com>
10412 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
10413 include sanitize(undefined).
10415 2013-08-31 Diego Novillo <dnovillo@google.com>
10417 * Makefile.in (TREE_CORE_H): Define.
10419 (GTFILES): Add tree-core.h.
10420 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
10422 * tree-core.h: New file.
10423 Move all data structures, enum, typedefs, global
10424 declarations and constants from ...
10425 * tree.h: ... here.
10427 2013-08-31 Jan Hubicka <jh@suse.cz>
10429 * bulitins.c (expand_builtin): Do not early exit for gcov
10430 instrumented functions.
10432 2013-08-31 Marek Polacek <polacek@redhat.com>
10434 * ubsan.c: Include tm_p.h.
10436 2013-08-31 Jan Hubicka <jh@suse.cz>
10438 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
10441 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
10442 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
10443 * tree-cfg.c (verify_gimple_label): ... here.
10444 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
10445 (ipa_merge_profiles): New function.
10446 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
10447 (lto_input_function_body): Likewise.
10448 * ipa-utils.h (ipa_merge_profiles): Declare.
10449 * lto-streamer.h (lto_input_function_body): Update prototype.
10450 (emit_label_in_global_context_p): Remove.
10451 * lto-symtab.c: Include ipa-utils.h
10452 (lto_cgraph_replace_node): Use ipa_merge_profiles.
10454 2013-08-31 Jan Hubicka <jh@suse.cz>
10456 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
10458 2013-08-31 Jan Hubicka <jh@suse.cz>
10460 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
10462 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
10464 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
10465 "cmp" RTX before signed_comparison_operator check to account
10466 for "code" changes.
10468 2013-08-30 Jan Hubicka <jh@suse.cz>
10470 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
10471 (detect_type_change_1): Rename to ...
10472 (detect_type_change): ... this one; early return on non-polymorphic
10474 (detect_type_change_ssa): Add comp_type parameter; update
10475 use of detect_type_change.
10476 (compute_complex_assign_jump_func): Add param_type parameter;
10477 update use of detect_type_change_ssa.
10478 (compute_complex_ancestor_jump_func): Likewise.
10479 (ipa_get_callee_param_type): New function.
10480 (ipa_compute_jump_functions_for_edge): Compute parameter type;
10481 update calls to the jump function computation functions.
10483 2013-08-30 Teresa Johnson <tejohnson@google.com>
10484 Steven Bosscher <steven@gcc.gnu.org>
10486 * cfgrtl.c (fixup_new_cold_bb): New routine.
10487 (commit_edge_insertions): Invoke fixup_partitions.
10488 (find_partition_fixes): New routine.
10489 (fixup_partitions): Ditto.
10490 (verify_hot_cold_block_grouping): Update comments.
10491 (rtl_verify_edges): Invoke find_partition_fixes.
10492 (rtl_verify_bb_pointers): Update comments.
10493 (rtl_verify_bb_layout): Ditto.
10494 * basic-block.h (probably_never_executed_edge_p): Declare.
10495 (fixup_partitions): Ditto.
10496 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
10497 * bb-reorder.c (sanitize_hot_paths): New function.
10498 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
10499 sanitize_hot_paths.
10500 * predict.c (probably_never_executed_edge_p): New routine.
10501 * cfg.c (check_bb_profile): Add partition insanity warnings.
10503 2013-08-30 Meador Inge <meadori@codesourcery.com>
10505 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10507 2013-08-30 Marek Polacek <polacek@redhat.com>
10509 * Makefile.in (ubsan.o): Add.
10510 (c-family/c-ubsan.o): Add.
10511 (builtins.o): Add ubsan.h dependency.
10512 * ubsan.h: New file.
10513 * ubsan.c: New file.
10514 * common.opt: Add -fsanitize=undefined option.
10515 (flag_sanitize): Add variable.
10516 (fsanitize=): Add option. Add Driver.
10517 (fsanitize=thread): Remove option.
10518 (fsanitize=address): Likewise.
10519 (static-libubsan): New option.
10520 * doc/invoke.texi: Document the new flag and -static-libubsan.
10521 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
10522 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
10523 * builtin-attrs.def (ATTR_COLD): Define.
10524 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10525 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
10526 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
10527 * flag-types.h (sanitize_code): New enum.
10528 * opts.c (common_handle_option): Parse command line arguments
10529 of -fsanitize=. Add -fsanitize=unreachable option.
10530 * varasm.c (get_variable_section): Adjust.
10531 (assemble_noswitch_variable): Likewise.
10532 (assemble_variable): Likewise.
10533 (output_constant_def_contents): Likewise.
10534 (categorize_decl_for_section): Likewise.
10535 (place_block_symbol): Likewise.
10536 (output_object_block): Likewise.
10537 * builtins.def: Likewise.
10538 * toplev.c (compile_file): Likewise.
10539 (process_options): Likewise.
10540 * cppbuiltin.c: Likewise.
10541 * tsan.c (tsan_pass): Likewise.
10542 (tsan_gate): Likewise.
10543 (tsan_gate_O0): Likewise.
10544 * cfgexpand.c (partition_stack_vars): Likewise.
10545 (expand_stack_vars): Likewise.
10546 (defer_stack_allocation): Likewise.
10547 (expand_used_vars): Likewise.
10548 * cfgcleanup.c (old_insns_match_p): Likewise.
10549 * asan.c (asan_finish_file): Likewise.
10550 (asan_instrument): Likewise.
10551 (gate_asan): Likewise.
10552 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
10553 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10554 (asan_global_struct): Use pointer_sized_int_node instead
10555 calling build_nonstandard_integer_type.
10556 (initialize_sanitizer_builtins): Likewise.
10557 (asan_finish_file): Likewise.
10558 * gcc.c: Document %{%:function(args):X}.
10559 (static_spec_functions): Add sanitize.
10560 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
10561 store funcval != NULL there.
10562 (do_spec_1): Adjust handle_spec_function caller.
10563 (handle_braces): Allow %:function(args) as condition.
10564 (sanitize_spec_function): New function.
10565 (ADD_STATIC_LIBUBSAN_LIBS): Define.
10566 (LIBUBSAN_SPEC): Likewise.
10567 (LIBUBSAN_EARLY_SPEC): Likewise.
10568 (SANITIZER_SPEC): Handle libubsan.
10569 (SANITIZER_EARLY_SPEC): Likewise.
10570 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
10571 instead of fsanitize=address.
10572 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
10573 instead of fsanitize=address*.
10574 * builtins.c: Include ubsan.h.
10575 (fold_builtin_0): Instrument __builtin_unreachable.
10576 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
10577 instead of flag_asan.
10578 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
10579 (pointer_sized_int_node): Define.
10580 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
10582 2013-08-30 Mike Stump <mikestump@comcast.net>
10584 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
10587 2013-08-29 Jan Hubicka <jh@suse.cz>
10589 * cgraph.c (cgraph_function_body_availability): Handle weakref
10591 * passes.def: Remove pass_fixup_cfg.
10592 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
10593 track when we need to remove functions.
10594 (gate_ipa_inline): Execute inlining always; add comment why.
10595 (pass_data_ipa_inline): Remove TODO_remove_functions.
10596 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
10597 do not produce summaries.
10598 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
10599 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
10600 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
10601 constant pool and vtable.
10603 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
10605 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
10606 New arm_neon.h's internal macros to specify 64-bit constants.
10607 Avoid using stdint.h's macros.
10609 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
10611 * recog.c (verify_changes): Verify that changes[i].old is non-zero
10612 before applying REG_P.
10614 2013-08-30 Jakub Jelinek <jakub@redhat.com>
10616 PR tree-optimization/58277
10617 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
10618 after seeing too many stmts with vdef in between dombb and current
10619 bb, invalidate everything.
10621 2013-08-30 Richard Biener <rguenther@suse.de>
10623 * fold-const.c (fold_single_bit_test): Fix overflow test.
10625 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
10627 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
10628 and which can live in a register, always retrieve the value on entry.
10629 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
10630 passed by invisible reference specially.
10631 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
10632 (vt_add_function_parameter): Correctly deal with a parameter passed by
10633 invisible reference.
10635 2013-08-30 Jan Hubicka <jh@suse.cz>
10637 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
10639 2013-08-30 Richard Biener <rguenther@suse.de>
10641 PR tree-optimization/58228
10642 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
10643 allow invariant loads in nested loop vectorization.
10645 2013-08-30 Richard Biener <rguenther@suse.de>
10647 PR tree-optimization/58223
10648 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
10649 (has_anti_or_output_dependence): ... this and adjust to also
10650 look for output dependences.
10651 (mark_nodes_having_upstream_mem_writes): Adjust.
10652 (rdg_flag_uses): Likewise.
10654 2013-08-30 Richard Biener <rguenther@suse.de>
10656 PR tree-optimization/58010
10657 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
10658 assert that we have a loop-closed PHI.
10660 2013-08-29 Jan Hubicka <jh@suse.cz>
10662 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
10663 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
10664 * lto-section-in.c (lto_free_function_in_decl_state): New function.
10665 (lto_free_function_in_decl_state_for_node): New function.
10667 2013-08-29 Xinliang David Li <davidxl@google.com>
10669 * loop-unroll.c (report_unroll_peel): Minor message change.
10670 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
10671 Emit alignment peeling message with default -fopt-info.
10672 (vect_loop_versioning): Emit loop version info message.
10673 * tree-vectorizer.c (vectorize_loops): Minor message change.
10674 (execute_vect_slp): Ditto.
10676 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
10678 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
10679 of the clone from the DECL_NAME of the original function.
10681 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
10683 * passes.c (register_pass): Add overload.
10684 * tree-pass.h (register_pass): Forward declare it. Add comment.
10686 2013-08-29 Jan Hubicka <jh@suse.cz>
10688 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
10689 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
10690 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
10691 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10692 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
10693 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
10694 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10695 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
10697 2013-08-29 Teresa Johnson <tejohnson@google.com>
10699 * dumpfile.c (dump_loc): Output column number.
10700 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
10701 * doc/invoke.texi: Document optall -fopt-info flag.
10702 * profile.c (read_profile_edge_counts): Use new dump framework.
10703 (compute_branch_probabilities): Ditto.
10704 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
10705 when pass not in any opt group.
10706 * pass_manager.h (pass_manager::get_pass_profile): New method.
10707 * value-prof.c (check_counter): Use new dump framework.
10708 (check_ic_target): Ditto.
10709 * coverage.c (get_coverage_counts): Ditto.
10710 (coverage_init): Setup new dump framework.
10712 2013-08-29 Richard Biener <rguenther@suse.de>
10714 PR tree-optimization/58246
10715 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
10716 handle the dominance check inside a basic-block.
10718 2013-08-29 Richard Biener <rguenther@suse.de>
10720 PR middle-end/57287
10721 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
10722 of default defs that appear in abnormal PHI nodes.
10724 2013-08-29 Richard Biener <rguenther@suse.de>
10726 PR tree-optimization/57685
10727 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
10728 single-use operands to avoid exponential complexity.
10730 2013-08-28 Dehao Chen <dehao@google.com>
10732 * ipa-inline.c (edge_badness): Fix integer underflow.
10734 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
10736 * gtm-builtins.def (_ITM_free): Declare leaf.
10738 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10741 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
10742 (*tls_local_dynamic_base_64_largepic): Likewise.
10743 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
10744 Remove predicate from call operand.
10745 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
10746 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
10748 2013-08-28 Jeff Law <law@redhat.com>
10750 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
10751 checks for the number of predecessors and successors allowed.
10752 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
10753 which require copying a joiner block if there is a request which
10754 is a subpath that requires no joiner block copying.
10756 2013-08-28 Jan Hubicka <jh@suse.cz>
10758 * lto-streamer-out.c (DFS_write_tree_body): Drop
10759 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
10760 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
10761 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
10762 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
10763 Do not read DECL_ERROR_ISSUED.
10764 (unpack_ts_decl_with_vis_value_fields): Do not read
10766 (lto_input_ts_binfo_tree_pointers): Do not read
10767 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
10768 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
10769 write DECL_ERROR_ISSUED..
10770 (pack_ts_decl_with_vis_value_fields): Do not write
10772 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
10773 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
10774 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
10775 * tree.h (tree_decl_common): Update comment.
10776 (DECL_ERROR_ISSUED): Remove.
10778 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10780 PR middle-end/58257
10781 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
10783 2013-08-28 Jan Hubicka <jh@suse.cz>
10785 * builtins.def (free): Declare leaf.
10787 2013-08-27 David Malcolm <dmalcolm@redhat.com>
10789 * gdbhooks.py: New.
10790 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
10791 * configure: Regenerate.
10793 2013-08-27 Martin Jambor <mjambor@suse.cz>
10795 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
10796 (ipa_ancestor_jf_data): Likewise.
10797 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
10798 (ipa_get_jf_pass_through_type_preserved): New function.
10799 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
10800 (ipa_get_jf_ancestor_type_preserved): New function.
10801 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
10802 (ipa_get_jf_ancestor_result): Likewise.
10803 (propagate_vals_accross_pass_through): Use
10804 ipa_get_jf_pass_through_result to do all the value mappings.
10805 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
10806 type_preserved flag.
10807 (ipa_set_jf_cst_copy): New function.
10808 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
10809 (ipa_set_jf_arith_pass_through): Likewise.
10810 (ipa_set_ancestor_jf): Likewise.
10811 (compute_complex_assign_jump_func): Set type_preserved instead of
10813 (ipa_compute_jump_functions_for_edge): Likewise.
10814 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
10815 (update_jump_functions_after_inlining): Update type_preserved.
10816 Explicitely create jump functions when combining one with pass_through.
10817 (ipa_write_jump_function): Stream the type_preserved flags.
10818 (ipa_read_jump_function): Likewise.
10820 2013-08-27 Jakub Jelinek <jakub@redhat.com>
10821 Aldy Hernandez <aldyh@redhat.com>
10823 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
10824 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
10825 * function.h (struct function): Add has_force_vect_loops and
10827 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
10828 * gimple.c (gimple_build_omp_critical): Add KIND argument and
10830 * gimple.def: Update CLAUSES comments.
10831 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
10832 (gimple_build_omp_for): Add argument to prototype.
10833 (gimple_omp_for_kind): New.
10834 (gimple_omp_for_set_kind): New.
10835 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
10836 GOVD_DATA_SHARE_CLASS.
10837 (enum omp_region_type): Add ORT_SIMD.
10838 (gimple_add_tmp_var): Handle ORT_SIMD.
10839 (gimplify_var_or_parm_decl): Same.
10840 (is_gimple_stmt): Same.
10841 (omp_firstprivatize_variable): Same.
10842 (omp_add_variable): Only use splay_tree_insert if lookup failed.
10843 (omp_notice_variable): Handle ORT_SIMD.
10844 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
10845 (omp_check_private): Handle ORT_SIMD.
10846 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10847 OMP_CLAUSE_SAFELEN.
10848 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
10849 Handle OMP_CLAUSE_LASTPRIVATE.
10850 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10851 OMP_CLAUSE_SAFELEN.
10852 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
10853 (gimplify_expr): Handle OMP_SIMD.
10854 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
10855 (expand_GOMP_SIMD_VF): New.
10856 (expand_GOMP_SIMD_LAST_LANE): New.
10857 * internal-fn.def (GOMP_SIMD_LANE): New.
10858 (GOMP_SIMD_VF): New.
10859 (GOMP_SIMD_LAST_LANE): New.
10860 * omp-low.c: Include target.h.
10861 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
10862 OMP_CLAUSE_SAFELEN.
10863 (check_omp_nesting_restrictions): Same.
10865 (lower_rec_simd_input_clauses): New.
10866 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
10868 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
10869 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
10870 (expand_omp_build_assign): New.
10871 (expand_omp_for_init_counts): New.
10872 (expand_omp_for_init_vars): New.
10873 (extract_omp_for_update_vars): New.
10874 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
10875 and rewrite accordingly.
10876 (expand_omp_simd): New.
10877 (expand_omp_for): Use expand_omp_simd.
10878 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
10879 (lower_omp_for): Do not lower the body.
10880 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
10881 in their own loops.
10882 * tree-flow.h (find_omp_clause): Remove prototype.
10883 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
10884 forcing vectorization of the loop, or if flag_tree_vectorize.
10885 (gate_tree_if_conversion): Similarly.
10886 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
10887 gimple_build_omp_for.
10888 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
10889 * tree-parloops (create_parallel_loop): Pass kind argument to
10890 gimple_build_omp_for.
10891 * tree-pretty-print.c (dump_omp_clause): Add cases for
10892 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
10893 OMP_CLAUSE__SIMDUID_.
10894 (dump_generic_node): Handle OMP_SIMD.
10895 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
10896 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
10897 unroll OMP_SIMD loops here.
10898 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
10899 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
10901 (vect_analyze_data_refs): Handle simd loops.
10902 * tree-vect-loop.c (vectorizable_live_operation): Handle
10904 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
10905 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
10906 (vectorizable_load): Same.
10907 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
10908 (struct simduid_to_vf): New.
10909 (simduid_to_vf::hash): New.
10910 (simduid_to-vf::equal): New.
10911 (struct simd_array_to_simduid): New.
10912 (simd_array_to_simduid::hash): New.
10913 (simd_array_to_simduid::equal): New.
10914 (adjust_simduid_builtins): New.
10915 (struct note_simd_array_uses_struct): New.
10916 (note_simd_array_uses_cb): New.
10917 (note_simd_array_uses): New.
10918 (vectorize_loops): Handle simd hints and adjust simd builtins
10920 * tree-vectorizer.h (struct _stmt_vec_info): Add
10921 simd_lane_access_p field.
10922 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
10923 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
10924 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
10925 (omp_clause_code_name): Same.
10926 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
10927 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
10928 * tree.def (OMP_SIMD): New entry.
10929 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
10930 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
10931 (OMP_CLAUSE_DECL): Adjust range for new clauses.
10932 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
10933 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
10934 (OMP_CLAUSE_LINEAR_STEP): New.
10935 (OMP_CLAUSE_SAFELEN_EXPR): New.
10936 (OMP_CLAUSE__SIMDUID__DECL): New.
10937 (find_omp_clause): New prototype.
10939 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
10941 * config/i386/driver-i386.c (host_detect_local_cpu): Update
10942 Haswell processor detection.
10944 2013-08-27 Christian Widmer <shadow@umbrox.de>
10947 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
10948 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
10949 AVX2 capable processors.
10951 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
10953 * config/aarch64/arm_neon.h: Replace all inline asm implementations
10954 of vget_low_* with implementations in terms of other intrinsics.
10956 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
10958 PR middle-end/57219
10959 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
10960 values to -1, 0 and 1.
10962 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
10964 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
10965 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
10966 (<optab><mode>3_insn): Remove.
10967 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
10968 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
10969 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
10970 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
10971 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
10972 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
10973 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
10974 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
10975 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
10976 (ror<mode>3_insn): Likewise.
10977 * config/aarch64/predicates.md (aarch64_simd_register): New.
10979 2013-08-27 Richard Biener <rguenther@suse.de>
10981 PR tree-optimization/57521
10982 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
10983 one edge is non-critical.
10984 (find_phi_replacement_condition): Make sure to use a non-critical
10985 edge. Cleanup and remove old bug workarounds.
10986 (bb_postdominates_preds): Remove.
10987 (if_convertible_loop_p_1): Do not compute post-dominators.
10988 (combine_blocks): Do not free post-dominators.
10989 (main_tree_if_conversion): Likewise.
10990 (pass_data_if_conversion): Add TODO_verify_ssa.
10992 2013-08-27 DJ Delorie <dj@redhat.com>
10994 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
10996 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
10998 * function.c (assign_parm_find_data_types): Set passed_mode and
10999 nominal_mode to the TYPE_MODE of nominal_type for the built
11000 pointer type in case of the struct-pass-by-reference.
11002 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11004 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
11005 (UINT16_TYPE): Change default to "unsigned int".
11007 * config/avr/avr.opt (mfract-convert-truncate): New option.
11008 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
11009 is set, round negative fractional integers according to n1169
11010 when converting to integer types.
11012 2013-08-26 Jan Hubicka <jh@suse.cz>
11014 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
11015 methods can not be called indirectly when their address is not taken.
11017 2013-08-26 Jan Hubicka <jh@suse.cz>
11019 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
11022 2013-08-26 Jan Hubicka <jh@suse.cz>
11024 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
11027 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11029 * reload.c (find_valid_class): Allow classes that do not include
11030 FIRST_PSEUDO_REGISTER - 1.
11032 2013-08-26 Jan Hubicka <jh@suse.cz>
11034 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
11035 fix edge count/frequency when speculation failed; fix type check
11036 for the direct call.
11038 2013-08-26 Jan Hubicka <jh@suse.cz>
11040 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
11042 2013-08-26 Jan Hubicka <jh@suse.cz>
11044 * ipa-inline-transform.c (inline_transform): Be ready for basic block
11045 to be changed by edge redirection.
11047 2013-08-26 Jan Hubicka <jh@suse.cz>
11049 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
11050 formating; add sanity check.
11051 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
11052 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
11053 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
11054 (resolve_noninline_speculation): Update callee keys, too.
11056 2013-08-26 Jan Hubicka <jh@suse.cz>
11058 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
11059 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
11061 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11063 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
11066 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
11068 * config/i386/i386.c (ix86_debug_options): Remove prototype.
11069 (x86_64_elf_select_section): Ditto.
11070 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
11072 (ix86_pass_by_reference): Ditto.
11073 (output_set_got): Ditto.
11074 (ix86_unary_operator_ok): Ditto.
11075 (ix86_expand_builtin): Ditto.
11077 2013-08-23 Jan Hubicka <jh@suse.cz>
11079 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
11081 2013-08-23 Jan Hubicka <jh@suse.cz>
11083 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
11084 (tree_decl_with_vis): Add FINAL field.
11086 2013-08-23 Jeff Law <law@redhat.com>
11088 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
11089 the debugging dump when the expression is fully redundant.
11091 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11093 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
11094 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
11095 * pretty-print.c (pp_formatted_text_data): Likewise.
11096 (pp_write_text_to_stream): Likewise.
11097 (pp_write_text_as_dot_label_to_stream): Likewise.
11098 (pp_append_r): Likewise.
11099 (pp_format): Likewise.
11100 (pp_flush): Likewise.
11101 (pp_clear_output_area): Likewise.
11102 (pp_append_text): Likewise.
11103 (pp_formatted_text): Likewise.
11104 (pp_remaining_character_count_for_line): Likewise.
11105 (pp_newline): Likewise.
11106 (pp_character): Likewise.
11107 (output_buffer::~output_buffer): Define.
11108 (pretty_printer::~pretty_printer): Destruct output buffer.
11109 * pretty-print.h (output_buffer::~output_buffer): Declare.
11110 (pretty_printer::~pretty_printer): Declare virtual.
11112 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
11115 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
11116 HOST_WIDE_INT_M1U): New macros.
11117 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
11118 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
11119 unsigned -1 for lshift.
11120 * cse.c (cse_insn): Likewise.
11121 * double-int.c (rshift_double, lshift_double): Likewise.
11122 * builtins.c (fold_builtin_bitop): Likewise.
11123 * combine.c (force_to_mode): Likewise.
11124 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
11125 * simplify-rtx.c (simplify_const_unary_operation,
11126 simplify_const_binary_operation): Likewise.
11127 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
11128 check_va_list_escapes): Likewise.
11129 * rtlanal.c (nonzero_bits1): Likewise.
11130 * expmed.c (expand_smod_pow2): Likewise.
11131 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
11133 2013-08-23 Jan Hubicka <jh@suse.cz>
11135 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
11136 as having address taken.
11138 2013-08-23 Jan Hubicka <jh@suse.cz>
11140 * ipa-utils.h (method_class_type): Declare.
11141 * ipa-devirt.c (method_class_type): Export.
11143 * cgraphunit.c (analyze_functions): Do basic devirtualization;
11144 do not walk base classes of anonymous types.
11146 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
11148 PR rtl-optimization/58220
11149 PR regression/58221
11150 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
11151 handle SEQUENCE insns properly.
11153 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11155 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
11157 (pp_newline_and_indent): Likewise.
11158 (pp_separate_with): Likewise.
11159 * pretty-print.c (pp_newline_and_flush): Define.
11160 (pp_newline_and_indent): Likewise.
11161 (pp_separate_with): Likewise.
11163 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11166 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
11167 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
11169 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
11171 * config/i386/predicates.md (ext_sse_reg_operand): New.
11172 * config/i386/i386.md (*movti_internal): Use
11173 predicate to determine if EVEX is needed.
11174 (*movsi_internal): Ditto.
11175 (*movdf_internal): Ditto.
11176 (*movsf_internal): Ditto.
11177 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
11179 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11181 PR tree-optimization/58209
11182 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
11183 (find_tail_calls): Give up for pointer result types if m is non-NULL.
11184 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
11185 emit POINTER_PLUS_EXPR.
11186 (create_tailcall_accumulator): For pointer result type accumulate in
11189 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
11191 * configure.ac: Add backslashes missing from the last change.
11192 * configure: Regenerate.
11194 2013-08-22 Jan Hubicka <jh@susue.cz>
11196 * ipa.c (function_and_variable_visibility): First remember function
11197 was global and then make it local.
11199 2013-08-22 Julian Brown <julian@codesourcery.com>
11201 * configure.ac: Add aarch64 to list of arches which use "nop" in
11203 * configure: Regenerate.
11205 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11207 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
11208 gnu_libc_has_function.
11209 * config/s390/tpf.h: Likewise.
11211 2013-08-22 Jan Hubicka <jh@susue.cz>
11213 * timevar.c (validate_phases): Add cast.
11215 2013-08-22 Jan Hubicka <jh@susue.cz>
11217 * timevar.c (validate_phases): Use size_t for memory.
11218 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
11220 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
11222 * pretty-print.h (output_buffer::output_buffer): Declare.
11223 (pretty_printer::pretty_printer): Likewise.
11224 (pp_construct): Remove.
11225 * pretty-print.c (output_buffer::output_buffer): Define.
11226 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
11227 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
11228 (print_gimple_expr): Likewise.
11229 (print_gimple_seq): Likewise.
11230 (gimple_dump_bb): Likewise.
11231 * sched-vis.c (dump_value_slim): Likewise.
11232 (dump_insn_slim): Likewise.
11233 (dump_rtl_slim): Likewise.
11234 (str_pattern_slim): Likewise.
11235 * tree-mudflap.c (mf_varname_tree): Likewise.
11236 * graph.c (print_graph_cfg): Likewise.
11237 (start_graph_dump): Likewise.
11238 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
11240 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
11241 pretty printer initialization.
11242 * coretypes.h (diagnostic_context): Remove superflous type alias
11244 (pretty_printer): Likewise. Declare directly as a class.
11245 (pretty_print_info): Remove declaration as class.
11246 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
11247 and pp_clear_output_area.
11248 (asan_add_global): Likewise.
11250 2013-08-22 Jan Hubicka <jh@suse.cz>
11252 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
11253 * ipa-utils.h (update_type_inheritance_graph): Declare.
11254 (possible_polymorphic_call_target_p): Declare.
11255 (possible_polymorphic_call_target_p): New.
11256 * ipa-devirt.c: Update toplevel comments.
11257 (cached_polymorphic_call_targets): Move up.
11258 (odr_type_d): Move ID down.
11259 (polymorphic_type_binfo_p): Update comment.
11260 (odr_hasher::remove): Likewise;
11261 (get_odr_type): Set anonymous_namespace.
11262 (dump_odr_type): Dump it.
11263 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
11264 (maybe_record_node): Record node in cached_polymorphic_call_targets.
11265 (record_binfo): Add comment.
11266 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
11268 (devirt_node_removal_hook): Do not iCE when cache is freed.
11269 (possible_polymorphic_call_target_p): New predicate.
11270 (update_type_inheritance_graph): New function.
11272 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
11273 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
11274 Sergey Lega <sergey.s.lega@intel.com>
11275 Anna Tikhonova <anna.tikhonova@intel.com>
11276 Ilya Tocar <ilya.tocar@intel.com>
11277 Andrey Turetskiy <andrey.turetskiy@intel.com>
11278 Ilya Verbin <ilya.verbin@intel.com>
11279 Kirill Yukhin <kirill.yukhin@intel.com>
11280 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11282 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
11283 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
11284 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
11285 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
11286 (OPTION_MASK_ISA_AVX2_UNSET): Update.
11287 (OPTION_MASK_ISA_AVX512F_UNSET): New.
11288 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
11289 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
11290 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
11291 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
11292 OPT_mavx512pf, OPT_mavx512er cases.
11293 * config/i386/constraints.md (v): New constraint.
11294 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
11295 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
11296 (bit_AVX512CD): New.
11297 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11298 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
11299 * config/i386/i386-c.c (ix86_target_macros_internal):
11300 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
11302 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
11303 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
11304 * config/i386/i386.c (regclass_map, dbx_register_map)
11305 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
11306 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
11307 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
11308 -mavx512pf options.
11309 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
11310 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
11311 -mavx512pf options. Fix formatting.
11312 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
11313 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
11314 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
11315 -mavx512cd, -mavx512pf options.
11316 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
11317 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
11318 (ix86_preferred_output_reload_class): Replace SSE_REGS with
11320 (ix86_hard_regno_mode_ok): Support 512-bit registers.
11321 (ix86_set_reg_reg_cost): Ditto.
11322 (x86_order_regs_for_local_alloc): Ditto.
11323 (MAX_VECT_LEN): Extend to 64-byte.
11324 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
11325 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
11326 (TARGET_AVX512ER, TARGET_AVX512CD): New.
11327 (BIGGEST_ALIGNMENT): Extend to 512-bits.
11328 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
11329 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
11330 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
11331 (SSE_REG_MODE_P): Support new modes.
11332 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
11333 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
11334 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
11335 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
11336 (REG_CLASS_CONTENTS): Add new registers.
11337 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
11338 (EXT_REX_SSE_REGNO_P): New.
11339 (HI_REGISTER_NAMES): Add new registers.
11340 * config/i386/i386.md: Define constants for new registers.
11341 (mode): Add new 512-bit modes.
11342 (prefix): Support evex prefix.
11343 (isa): Support avx512f, noavx512f, fma_avx512f.
11344 (ssemodesuffix): Add new 512-bit modes.
11346 (*movxi_internal_avx512f): Ditto.
11347 (*movdi_internal): Replace constraint "x" with the new constraint "v".
11349 (*movsi_internal): Likewise.
11350 (*movdf_internal): Likewise.
11351 (*movsf_internal): Likewise.
11352 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
11353 (<code><mode>3): Likewise.
11354 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
11356 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
11357 with the new constraint "v".
11358 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
11360 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
11361 with the new constraint "v".
11362 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
11363 (<sse2>_storedqu<avxsizesuffix>): Likewise.
11364 (*<plusminus_insn><mode>3): Likewise.
11365 (<sse>_vm<plusminus_insn><mode>3): Likewise.
11366 (*mul<mode>3): Likewise.
11367 (<sse>_vmmul<mode>3): Likewise.
11368 (<sse>_div<mode>3): Likewise.
11369 (<sse>_vmdiv<mode>3): Likewise.
11370 (<sse>_sqrt<mode>2): Likewise.
11371 (<sse>_vmsqrt<mode>2): Likewise.
11372 (*<code><mode>3_finite): Likewise.
11373 (*<code><mode>3) <smaxmin>: Likewise.
11374 (<sse>_vm<code><mode>3): Likewise.
11375 (*<code><mode>3) <any_logic>: Likewise.
11376 (*fma_fmadd_<mode>): Likewise.
11377 (*fma_fmsub_<mode>): Likewise.
11378 (*fma_fnmadd_<mode>): Likewise.
11379 (*fma_fnmsub_<mode>): Likewise.
11380 (*fma_fmaddsub_<mode>): Likewise.
11381 (*fma_fmsubadd_<mode>): Likewise.
11382 (*fmai_fmadd_<mode>): Likewise.
11383 (*fmai_fmsub_<mode>): Likewise.
11384 (*fmai_fnmadd_<mode>): Likewise.
11385 (*fmai_fnmsub_<mode>): Likewise.
11386 (sse_cvtsi2ss): Likewise.
11387 (sse_cvtsi2ssq): Likewise.
11388 (sse_cvtss2si): Likewise.
11389 (sse_cvtss2si_2): Likewise.
11390 (sse_cvtss2siq): Likewise.
11391 (sse_cvtss2siq_2): Likewise.
11392 (sse_cvttss2si): Likewise.
11393 (sse_cvtss2siq_2): Likewise.
11394 (float<sseintvecmodelower><mode>2): Likewise.
11395 (sse2_cvtsd2si_2): Likewise.
11396 (sse2_cvtsd2siq_2): Likewise.
11397 (*<plusminus_insn><mode>3): Likewise.
11398 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
11399 (*<sse4_1_avx2>_mul<mode>3): Likewise.
11400 (ashr<mode>3): Likewise.
11401 (<shift_insn><mode>3): Likewise.
11402 (avx2_<code><mode>3): Likewise.
11403 (*avx2_<code><mode>3): Likewise.
11404 (*andnot<mode>3): Likewise.
11405 (*<code><mode>3) <any_logic>: Likewise.
11406 (abs<mode>2): Likewise.
11407 (avx2_permvar<mode>): Likewise.
11408 (avx2_perm<mode>_1): Likewise.
11409 (*avx_vpermilp<mode>): Likewise.
11410 (avx_vpermilvar<mode>3): Likewise.
11411 (avx2_ashrv<mode>): Likewise.
11412 (avx2_<shift_insn>v<mode>): Likewise.
11413 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
11415 * doc/rtl.texi: Document XImode.
11417 2013-08-21 Jeff Law <law@redhat.com>
11419 * tree-flow.h (register_jump_thread): Pass vector of edges
11420 instead of each important edge.
11421 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
11422 thread path into a vector and pass that to register_jump_thread.
11423 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
11424 passed in edge vector to the current 3-edge form.
11427 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11429 * dce.c (fini_dce): Call df_analyze again just in case
11430 delete_unmarked_insns removed anything.
11432 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
11434 * reload.h (struct reg_equivs): Rename to ..
11435 (struct reg_equivs_s): .. this.
11437 2013-08-20 Martin Liska <marxin.liska@gmail.com>
11439 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
11441 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11443 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
11445 2013-08-21 Jeff Law <law@redhat.com>
11447 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
11448 simplify assignments too. If the RHS collapses to a singleton
11449 range, then return the value for the range.
11451 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
11453 * config/i386/sse.md (V16): Rename to...
11455 (mov<mode>): Update iterator name.
11456 (*mov<mode>_internal): Ditto.
11457 (push<mode>1): Ditto.
11458 (movmisalign<mode>): Ditto.
11460 2013-08-20 Jan Hubicka <jh@suse.cz>
11463 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
11464 entry for direct edges.
11465 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
11467 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11469 Revert my last two changes, r201865 and r201864:
11472 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11474 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11475 instances can own GC refs.
11477 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11478 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11479 (gcc::context::gt_pch_nx): Likewise.
11480 (gcc::context::gt_pch_nx): Likewise.
11481 * ggc.h (gt_ggc_mx <T>): New.
11482 (gt_pch_nx_with_op <T>): New.
11483 (gt_pch_nx <T>): New.
11484 * passes.c (opt_pass::gt_ggc_mx): New.
11485 (opt_pass::gt_pch_nx): New.
11486 (opt_pass::gt_pch_nx_with_op): New.
11487 (pass_manager::gt_ggc_mx): New.
11488 (pass_manager::gt_pch_nx): New.
11489 (pass_manager::gt_pch_nx_with_op): New.
11490 (pass_manager::operator new): Use
11491 ggc_internal_cleared_alloc_stat rather than xcalloc.
11492 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11493 (pass_manager::gt_ggc_mx): New.
11494 (pass_manager::gt_pch_nx): New.
11495 (pass_manager::gt_pch_nx_with_op): New.
11496 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11497 (opt_pass::operator new): New.
11498 (opt_pass::gt_ggc_mx): New.
11499 (opt_pass::gt_pch_nx): New.
11500 (opt_pass::gt_pch_nx_with_op): New.
11503 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11505 * Makefile.in (GTFILES): Add context.h.
11506 * context.c (gcc::context::operator new): New.
11507 (gcc::context::gt_ggc_mx): New.
11508 (gcc::context::gt_pch_nx): New.
11509 (gcc::context::gt_pch_nx): New.
11510 * context.h (gcc::context): Add GTY((user)) marking.
11511 (gcc::context::operator new): New.
11512 (gcc::context::gt_ggc_mx): New.
11513 (gcc::context::gt_pch_nx): New.
11514 (gcc::context::gt_pch_nx): New.
11515 (g): Add GTY marking.
11516 (gt_ggc_mx (gcc::context *)): New.
11517 (gt_pch_nx (gcc::context *)): New.
11518 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11519 void *cookie)): New.
11520 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11522 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11524 * dce.c (fini_dce): Call df_analyze again just in case
11525 delete_unmarked_insns removed anything.
11527 2013-08-20 Teresa Johnson <tejohnson@google.com>
11529 PR rtl-optimizations/57451
11530 * final.c (reemit_insn_block_notes): Prevent lexical blocks
11531 from crossing split section boundaries.
11533 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11535 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
11536 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
11537 with MULTLIB_DEFAULTS.
11539 2013-08-20 Nick Clifton <nickc@redhat.com>
11541 * target.def (narrow_volatile_bitfield): Note that the default
11542 value is false, not !TARGET_STRICT_ALIGN.
11543 * doc/tm.texi: Regenerate.
11545 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
11547 Fix LIB_SPEC for systems without libpthread.
11549 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
11550 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
11552 * config/i386/linux-common.h: Likewise.
11553 * config/mips/linux-common.h: Likewise.
11555 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11557 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
11560 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11562 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11563 instances can own GC refs.
11565 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11566 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11567 (gcc::context::gt_pch_nx): Likewise.
11568 (gcc::context::gt_pch_nx): Likewise.
11569 * ggc.h (gt_ggc_mx <T>): New.
11570 (gt_pch_nx_with_op <T>): New.
11571 (gt_pch_nx <T>): New.
11572 * passes.c (opt_pass::gt_ggc_mx): New.
11573 (opt_pass::gt_pch_nx): New.
11574 (opt_pass::gt_pch_nx_with_op): New.
11575 (pass_manager::gt_ggc_mx): New.
11576 (pass_manager::gt_pch_nx): New.
11577 (pass_manager::gt_pch_nx_with_op): New.
11578 (pass_manager::operator new): Use
11579 ggc_internal_cleared_alloc_stat rather than xcalloc.
11580 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11581 (pass_manager::gt_ggc_mx): New.
11582 (pass_manager::gt_pch_nx): New.
11583 (pass_manager::gt_pch_nx_with_op): New.
11584 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11585 (opt_pass::operator new): New.
11586 (opt_pass::gt_ggc_mx): New.
11587 (opt_pass::gt_pch_nx): New.
11588 (opt_pass::gt_pch_nx_with_op): New.
11590 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11592 * Makefile.in (GTFILES): Add context.h.
11593 * context.c (gcc::context::operator new): New.
11594 (gcc::context::gt_ggc_mx): New.
11595 (gcc::context::gt_pch_nx): New.
11596 (gcc::context::gt_pch_nx): New.
11597 * context.h (gcc::context): Add GTY((user)) marking.
11598 (gcc::context::operator new): New.
11599 (gcc::context::gt_ggc_mx): New.
11600 (gcc::context::gt_pch_nx): New.
11601 (gcc::context::gt_pch_nx): New.
11602 (g): Add GTY marking.
11603 (gt_ggc_mx (gcc::context *)): New.
11604 (gt_pch_nx (gcc::context *)): New.
11605 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11606 void *cookie)): New.
11607 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11609 2013-08-20 Alan Modra <amodra@gmail.com>
11612 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
11613 (rs6000_emit_epilogue): Likewise.
11615 2013-08-19 Dehao Chen <dehao@google.com>
11617 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
11619 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
11620 Jakub Jelinek <jakub@redhat.com>
11622 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
11623 (BUILT_IN_FABSD64): Likewise.
11624 (BUILT_IN_FABSD128): Likewise.
11625 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
11626 (fold_builtin_1): Likewise.
11627 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
11628 destination and source operands.
11629 (*abstd2_fpr): Likewise.
11630 (*nabstd2_fpr): Likewise.
11632 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
11634 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
11637 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
11639 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
11642 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
11644 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
11645 * builtins.c (default_libc_has_function): New.
11646 (gnu_libc_has_function): Ditto.
11647 (no_c99_libc_has_function): Ditto.
11648 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
11649 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
11650 (fold_builtin_sincos): Likewise.
11651 (fold_builtin_cexp): Likewise.
11652 * builtins.def (DEF_C94_BUILTIN): Likewise.
11653 (DEF_C99_BUILTIN): Likewise.
11654 (DEF_C99_C90RES_BUILTIN): Likewise.
11655 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
11656 definitions to using this define.
11657 * config/darwin-protos.h (darwin_libc_has_function): New.
11658 * config/darwin.c (darwin_libc_has_function): Ditto.
11659 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
11660 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
11661 * config/darwin.h: Ditto.
11662 * config/elfos.h: Ditto.
11663 * config/freebsd.h: Ditto.
11664 * config/i386/cygming.h: Ditto.
11665 * config/i386/djgpp.h: Ditto.
11666 * config/i386/i386-interix.h: Ditto.
11667 * config/microblaze/microblaze.h: Ditto.
11668 * config/mmix/mmix.h: Ditto.
11669 * config/gnu-user.h: Ditto.
11670 * config/ia64/hpux.h: Ditto.
11671 * config/pa/pa-hpux.h: Ditto.
11672 * config/pdp11/pdp11.h: Ditto.
11673 * config/picochip/picochip.h: Ditto.
11674 * config/linux.h: Ditto.
11675 * config/netbsd.h: Ditto.
11676 * config/openbsd.h: Ditto.
11677 * config/rs6000/aix43.h: Ditto.
11678 * config/rs6000/aix51.h: Ditto.
11679 * config/rs6000/aix52.h: Ditto.
11680 * config/rs6000/aix53.h: Ditto.
11681 * config/rs6000/aix61.h: Ditto.
11682 * config/rs6000/darwin.h: Ditto.
11683 * config/rs6000/linux.h: Ditto.
11684 * config/rs6000/linux64.h: Ditto.
11685 * config/s390/tpf.h: Ditto.
11686 * config/sol2-10.h: Ditto.
11687 * config/sol2.h: Ditto.
11688 * config/vms/vms.h: Ditto.
11689 * config/vxworks.h: Ditto.
11690 * config/linux-android.c (linux_android_libc_has_function):
11691 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
11692 * config/linux-protos.h (linux_android_libc_has_function):
11694 * config/i386/i386.c (ix86_libc_has_function): New.
11695 * config/i386/i386-protos.h
11696 (ix86_libc_has_function): New declaration.
11697 * config/i386/i386.md
11698 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
11699 ("isinf<mode>2): Likewise.
11700 * convert.c (convert_to_integer): Using new target hook
11701 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
11702 TARGET_C99_FUNCTIONS.
11703 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
11704 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
11705 * coretypes.h (function_class): New enum for different
11706 classes of functions.
11707 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
11708 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
11709 (TARGET_HAS_SINCOS): Likewise.
11710 (TARGET_LIBC_HAS_FUNCTION): New.
11711 * doc/tm.texi: Regenerated.
11712 * targhooks.h (default_libc_has_function): New declaration.
11713 (no_c99_libc_has_function): Ditto.
11714 (gnu_libc_has_function): Ditto.
11715 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
11716 and TARGET_HAS_SINCOS.
11718 2013-08-18 Jan Hubicka <jh@suse.cz>
11720 * Makeifle-in (ipa-devirt.o): New.
11721 (GTFILES): Add ipa-utils.h and ipa-devirt.c
11722 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
11723 (analyze_functions): Look into possible targets of polymorphic call.
11724 * dumpfile.c (dump_files): Add type-inheritance dump.
11725 * dumpfile.h (TDI_inheritance): New.
11726 * ipa-devirt.c: New file.
11727 * ipa-utils.h (odr_type_d): Forward declare.
11728 (odr_type): New type.
11729 (build_type_inheritance_graph): Declare.
11730 (possible_polymorphic_call_targets): Declare and introduce inline
11731 variant when only edge is pased.
11732 (dump_possible_polymorphic_call_targets): Likewise.
11733 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
11734 * tree.c (type_in_anonymous_namespace_p): Break out from ...
11735 (types_same_for_odr): ... here.
11736 * tree.h (type_in_anonymous_namespace_p): Declare.
11738 2013-08-18 Jakub Jelinek <jakub@redhat.com>
11740 PR tree-optimization/58006
11741 * tree-parloops.c (take_address_of): Don't ICE if get_name
11743 (eliminate_local_variables_stmt): Remove clobber stmts.
11745 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
11747 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
11748 error message is issued for an alias to undefined symbol.
11750 2013-08-18 Jan Hubicka <jh@suse.cz>
11752 * cgraph.c (cgraph_create_indirect_edge): Discover
11753 polymorphic calls and record basic info into indirect_info.
11754 * gimple-fold.c (gimple_fold_call): When doing BINFO based
11755 devirtualization, ignore objc function calls.
11756 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
11757 call with no parm index info.
11758 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
11759 * tree.c (virtual_method_call_p): New function.
11760 * tree.h (virtual_method_call_p): Declare.
11762 2013-08-16 Jan Hubicka <jh@suse.cz>
11764 PR middle-end/58179
11765 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
11767 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
11769 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
11772 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11774 * gengtype.c (type_for_name): Add special-case support for
11775 locating types within the "gcc::" namespace.
11776 (open_base_files): Emit a "using namespace gcc" directive.
11778 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11781 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
11782 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
11784 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
11785 array instead of each individual operand as a separate argument.
11786 (emit_fusion_gpr_load): Likewise.
11787 (expand_fusion_gpr_load): Add new function declaration.
11789 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
11790 signature to have the operands passed as an array, instead of as
11791 separate arguments. Allow ZERO_EXTEND to be in the memory
11792 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
11793 depend on the register live/dead flags when peepholes are run.
11794 (expand_fusion_gpr_load): New function to be called from the
11795 peephole2 pass, to change the register that addis sets to be the
11797 (emit_fusion_gpr_load): Change the calling signature to have the
11798 operands passed as an array, instead of as separate arguments.
11799 Allow ZERO_EXTEND to be in the memory address, and also
11800 SIGN_EXTEND if -mpower8-fusion-sign.
11802 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
11803 unspec enumeration.
11804 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
11805 adjust the register addis loads up in the peephole2 pass. Do not
11806 depend on the register live/dead state when the peephole pass is done.
11808 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11810 * gengtype.c (create_user_defined_type): Ensure that the kind
11811 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
11812 declaration is seen before the GTY((user)) marking.
11814 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
11817 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
11819 2013-08-16 Jan Hubicka <jh@suse.cz>
11821 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
11822 arugment expected_type.
11823 (gimple_fold_call): Use it.
11824 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
11825 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
11826 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
11827 (try_make_edge_direct_virtual_call): Likewise.
11828 * tree.c (obj_type_ref_class): New.
11829 * tree.h (obj_type_ref_class): Use it.
11831 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
11833 * sched-vis.c (rtl_slim_pp_initialized): Remove.
11834 (rtl_slim_pp): Likewise.
11835 (init_rtl_slim_pretty_print): Likewise.
11836 (dump_value_slim): Don't call it. Use local pretty printer.
11837 (dump_insn_slim): Likewise.
11838 (dump_rtl_slim): Likewise.
11839 (str_pattern_slim): Likewise.
11840 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
11843 2013-08-16 Jakub Jelinek <jakub@redhat.com>
11845 PR tree-optimization/58164
11846 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
11847 walk gimple_goto_dest of GIMPLE_GOTO.
11849 PR tree-optimization/58165
11850 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
11851 bi_call must be the last stmt in a bb, don't split_block, instead
11852 use fallthru edge from it and give up if there is none.
11853 Release conds vector when returning early.
11855 2013-08-14 Xinliang David Li <davidxl@google.com>
11857 * config/i386/i386.c (ix86_option_override_internal):
11858 Remove unused variable and field.
11860 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11863 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
11864 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
11865 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
11866 and Linux, correct BLKmode alignment when 128-bit alignment is
11867 required and compatibility flag is not set.
11868 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
11869 for zero-size arguments when compatibility flag is not set.
11871 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11873 PR tree-optimization/58145
11874 * tree-sra.c (build_ref_for_offset): If prev_base has
11875 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
11877 2013-08-14 Xinliang David Li <davidxl@google.com>
11879 * config/i386/i386.c (ix86_option_override_internal):
11880 Fix uninitialized variable error.
11882 2013-08-14 Xinliang David Li <davidxl@google.com>
11884 * config/i386/i386.opt: Define two new options.
11885 * config/i386/x86-tune.def: Add arch selector field in macros.
11886 * config/i386/i386.h: Adjust macro definition.
11887 * config/i386/i386.c (ix86_option_override_internal):
11889 (parse_mtune_ctrl_str): New function.
11890 (set_ix86_tune_features): New function.
11891 (ix86_function_specific_restore): Call the new helper function.
11893 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
11895 PR rtl-optimization/57662
11896 * sel-sched.c (code_motion_process_successors): When the current insn
11897 is removed after the recursive traversal, break from the loop.
11898 Add comments and debug printouts.
11900 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11901 Alexandre Oliva <aoliva@redhat.com>
11904 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
11905 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
11906 there also UNSPEC_PLTOFF.
11908 2013-08-14 Marek Polacek <polacek@redhat.com>
11910 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
11911 NUM_CONDITIONS bit positions.
11913 2013-08-13 Cary Coutant <ccoutant@google.com>
11915 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
11916 (attr_checksum): Hash vector contents instead of pointer.
11917 (attr_checksum_ordered): Likewise.
11919 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
11921 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
11922 when Pmode != word_mode. Add length_address attribute.
11923 (sse3_monitor_<mode>): Merge from sse3_monitor and
11924 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
11925 Pmode != word_mode. Update insn length attribute.
11926 * config/i386/i386.c (ix86_option_override_internal): Update
11927 ix86_gen_monitor selection for merged sse3_monitor insn.
11929 2013-08-13 Julian Brown <julian@codesourcery.com>
11931 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
11932 perform invalid legitimization on greater-than-word-size modes for
11933 TARGET_E500_DOUBLE.
11935 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
11937 * ira.c (setup_class_translate_array): Use aclass instead of cl
11938 for classes not fully covered by allocno classes.
11940 2013-08-13 Jakub Jelinek <jakub@redhat.com>
11942 PR tree-optimization/57661
11943 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
11944 * tree-inline.c (tree_function_versioning): Initialize it.
11945 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
11946 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
11947 that is not being copied.
11950 * asan.c (instrument_strlen_call): Fix typo in comment.
11951 Use char * type even for the lhs of POINTER_PLUS_EXPR.
11953 2013-08-13 Steve Ellcey <sellcey@mips.com>
11955 * config/mips/mips.md (prefetch): Use lw instead of ld on
11956 loongson in 32bit mode.
11958 2013-08-13 Nick Clifton <nickc@redhat.com>
11960 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
11962 2013-08-13 Jan Hubicka <jh@suse.cz>
11964 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
11965 introduced edge; fix typo in sanity check.
11966 (cgraph_resolve_speculation): Export; improve diagnostic.
11967 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
11968 speculation at type mismatch.
11969 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
11970 (cgraph_resolve_speculation): Declare.
11971 (symtab_can_be_discarded): New function.
11972 * value-prof.c (gimple_ic_transform): Remove actual transform code.
11973 * ipa-inline-transform.c (speculation_removed): New global var.
11974 (clone_inlined_nodes): See if speculation can be removed.
11975 (inline_call): If speculations was removed, we growths may not match.
11976 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
11977 (speculation_useful_p): New function.
11978 (resolve_noninline_speculation): New function.
11979 (inline_small_functions): Resolve useless speculations.
11980 * ipa-inline.h (speculation_useful_p): Declare
11981 * ipa.c (can_replace_by_local_alias): Simplify.
11982 (ipa_profile): Produce speculative calls in non-lto, too;
11983 add simple cost model; produce local aliases.
11985 2013-08-13 David Malcolm <dmalcolm@redhat.com>
11987 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
11990 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
11992 * config/i386/i386.c (ix86_function_versions): Use error + inform.
11994 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
11996 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
11997 iterator instead of X87MODEF.
11999 2013-08-12 Perez Read <netfirewall@gmail.com>
12002 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
12003 operand 0 for intel asm alternative.
12004 (*movabs<mode>_2): Ditto for operand 1.
12006 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
12008 * config/aarch64/arm_none.h
12009 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
12011 2013-08-12 Nick Clifton <nickc@redhat.com>
12013 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
12015 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
12017 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
12018 expand for QI/HImode operand to produce more effictive code for
12019 unsigned char(short) --> float(double) conversion.
12021 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
12023 * doc/invoke.texi: Mention that -ftls-model does not force the final
12026 2013-08-12 Marek Polacek <polacek@redhat.com>
12027 Marc Glisse <marc.glisse@inria.fr>
12029 PR tree-optimization/57980
12030 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
12031 when creating -1 constant.
12033 2013-08-10 Jan Hubicka <jh@suse.cz>
12035 Workaround binutils PR14342.
12036 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
12037 (gimple_init_edge_profiler): Likewise.
12038 (gimple_gen_ic_func_profiler): Likewise.
12040 2013-08-09 Jan Hubicka <jh@suse.cz>
12042 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
12044 2013-08-09 Xinliang David Li <davidxl@google.com>
12046 * config/i386/stringop.def: New file.
12047 * config/i386/stringop.opt: New file.
12048 * config/i386/i386-opts.h: Include stringopt.def.
12049 * config/i386/i386.opt: Include stringopt.opt.
12050 * config/i386/i386.c (ix86_option_override_internal):
12051 Override default size based stringop inline strategies with options.
12052 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
12055 2013-08-09 Jan Hubicka <jh@suse.cz>
12057 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
12059 2013-08-09 Jan Hubicka <jh@suse.cz>
12061 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
12063 (dump_cgraph_node): Dump profile-id.
12064 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
12065 and common_target_probability.
12066 * lto-cgraph.c (lto_output_edge): Stream common targets.
12067 (lto_output_node): Stream profile ids.
12068 (input_node): Stream profile ids.
12069 (input_edge): Stream common targets.
12070 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
12071 * ipa.c: Include value-prof.h
12072 (ipa_profile_generate_summary): Turn indirect call statement histograms
12073 into common targets.
12074 (ipa_profile): Turn common targets into speculative edges.
12076 2013-08-09 Jan Hubicka <jh@suse.cz>
12078 * cgraph.h (cgraph_node): Add profile_id.
12079 * value-prof.c (cgraph_node_map): Turn into pointer_map.
12080 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
12081 (del_node_map): Update.
12082 (find_func_by_funcdef_no): Replace by ...
12083 (find_func_by_profile_id): ... this one.
12084 (gimple_ic_transform): Do not remove useful histograms when
12085 speculation is not done; dump info when indirect call removal
12087 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
12088 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
12089 (__gcov_indirect_call_profiler_v2): .. this one.
12090 * profile.h (init_node_map): Update.
12091 * coverage.c (coverage_compute_profile_id): New function.
12092 * coverage.h (coverage_compute_profile_id): Declare.
12093 * tree-profile.c (init_ic_make_global_vars): Make
12094 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
12095 (gimple_init_edge_profiler): Update prototype of
12096 __gcov_indirect_call_profiler.
12097 (gimple_gen_ic_func_profiler): Simplify.
12098 (tree_profiling): Use init_node_map
12100 2013-08-09 Jan Hubicka <jh@suse.cz>
12102 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
12103 non-speculative refs.
12104 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
12105 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
12106 (cgraph_set_call_stmt): Likewise.
12107 (cgraph_create_edge_1): Fix release checking compilatoin;
12108 clear lto_stmt_uid.
12109 (cgraph_free_edge): Free indirect info.
12110 (cgraph_turn_edge_to_speculative): New function.
12111 (cgraph_speculative_call_info): New function.
12112 (cgraph_make_edge_direct): Return direct edge; handle speculation.
12113 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
12114 (dump_cgraph_node): Dump speculation.
12115 (verify_edge_count_and_frequency): Accept speculative edges.
12116 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
12117 (verify_cgraph_node): Handle speculation.
12118 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
12119 (cgraph_set_call_stmt): Update prototype.
12120 (cgraph_make_edge_direct): Update prototype.
12121 (cgraph_speculative_call_info): Declare.
12122 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
12123 to change; update call of ipa_find_references.
12124 * ipa-ref.c (ipa_record_reference): Fix return value; clear
12125 lto_stmt_uid and speculative flags.
12126 (ipa_dump_references): Dump speculation.
12127 (ipa_clone_references): Clone speculative flag.
12128 (ipa_clone_referring): Likewise.
12129 (ipa_clone_ref): New function.
12130 (ipa_find_reference): Look into lto_stmt_uids
12131 (ipa_clear_stmts_in_references): Do not clear speculative calls.
12132 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
12133 (ipa_find_reference): Update declaration.
12134 (ipa_clone_ref): Declare.
12135 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
12136 stream speculative flag.
12137 (lto_output_ref): Stream statements uids and speculation.
12138 (input_ref): Likewise.
12139 (input_edge): Stream speuclation.
12140 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
12141 (cgraph_set_call_stmt_including_clones): Handle speculation.
12142 * ipa-inline.c (heap_edge_removal_hook): New function.
12143 (inline_small_functions): Register it.
12144 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
12145 also initialize refs.
12146 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
12148 (try_make_edge_direct_simple_call): Likewise.
12149 (try_make_edge_direct_simple_call): Likewise.
12150 (update_indirect_edges_after_inlining): Likewise.
12151 (remove_described_reference): Look proper lto_stmt_uid.
12152 (propagate_controlled_uses): Likewise.
12153 (propagate_controlled_uses): Liekwise.
12154 * tree-inline.c (copy_bb): Copy speculative edges.
12155 (redirect_all_calls): New function.
12156 (copy_cfg_body): Do redirection after loop info is updated.
12157 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
12159 2013-08-09 Jan Hubicka <jh@suse.cz>
12161 * lto-streamer-out.c (output_function): Renumber PHIs.
12162 * lto-streamer-in.c (input_function): Likewise.
12164 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
12166 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
12167 (get_lane_unsigned): Likewise.
12168 (dup_lane_scalar): Likewise.
12169 (get_lane): enable for VALL.
12170 * config/aarch64/aarch64-simd.md
12171 (aarch64_dup_lane_scalar<mode>): Remove.
12172 (aarch64_get_lane_signed<mode>): Likewise.
12173 (aarch64_get_lane_unsigned<mode>): Likewise.
12174 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
12175 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
12176 (aarch64_get_lane<mode>): Enable for all vector modes.
12177 (aarch64_get_lanedi): Remove misleading constraints.
12178 * config/aarch64/arm_neon.h
12179 (__aarch64_vget_lane_any): Define.
12180 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
12181 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
12182 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
12183 * config/aarch64/iterators.md (VDQQH): New.
12184 (VDQQHS): Likewise.
12185 (vwcore): Likewise.
12187 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
12189 * configure.ac: Add GAS check for LEON instructions on SPARC.
12190 * configure: Regenerate.
12191 * config.in: Likewise.
12192 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
12194 * config/sparc/sparc.opt (LEON, LEON3): New masks.
12195 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
12197 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
12198 (AS_LEON_FLAG): New macro.
12199 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
12200 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
12201 Deal with LEON and LEON3 for the memory model.
12202 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
12203 (atomic_compare_and_swap<mode>_1): Likewise.
12204 (*atomic_compare_and_swap<mode>_1): Likewise.
12206 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12208 * config/arm/neon.md (vcond): Fix floating-point vector
12209 comparisons against 0.
12211 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
12213 * lra-constraints.c (emit_spill_move): Remove assert.
12214 (process_alt_operands): Add more debugging
12215 output. Increase reject for spilling into memory. Decrease
12216 reject for reloading scratch.
12217 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
12219 2013-08-08 Steve Ellcey <sellcey@mips.com>
12221 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
12222 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
12223 micromips incompatible. Add nan2008.
12224 (MULTILIB_DIRNAMES): Add nan2008.
12225 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12226 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
12227 and micromips incompatible. Add nan2008.
12228 (MULTILIB_DIRNAMES): Add nan2008.
12229 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12231 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
12233 PR rtl-optimization/58079
12234 * combine.c (combine_simplify_rtx): Avoid using SUBST if
12235 simplify_comparison has widened a comparison with an integer.
12237 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12239 * config/arm/neon.md (movmisalign<mode>): Disable when we
12240 don't allow unaligned accesses.
12241 (*movmisalign<mode>_neon_store): Likewise.
12242 (*movmisalign<mode>_neon_load): Likewise.
12243 (*movmisalign<mode>_neon_store): Likewise.
12244 (*movmisalign<mode>_neon_load): Likewise.
12246 2013-08-08 Jan Hubicka <jh@suse.cz>
12248 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
12249 (make_pass_rebuild_cgraph_edges): Also clear references.
12250 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
12251 * ipa-inline-transform.c (inline_transform): Remove all references
12253 * cgraphunit.c (expand_function): Remove all references after
12255 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
12256 (ipa_find_reference): Rewrite to iterator.
12257 (remove_stmt_references): Likewise.
12258 (ipa_clear_stmts_in_references): New function.
12259 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
12260 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
12262 * ipa-split.c (split_function): Remove references in split function.
12264 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12267 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
12268 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
12270 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12273 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
12274 suggested mode for the assignment isn't compatible with the
12275 registers required.
12277 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12280 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
12282 2013-08-07 Xinliang David Li <davidxl@google.com>
12284 * config/i386/i386.opt: New option -mtune-ctrl=.
12285 * config/i386/x86-tune.def: New file.
12286 * config/i386/i386.h: include x86-tune.def.
12287 * config/i386/i386.c (ix86_option_override_internal):
12288 Parsing -mtune-ctrl= option and set tune features.
12290 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
12293 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
12294 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
12297 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12299 * rtl.h (update_alignments): Declare.
12300 * final.c (grow_label_align): New function extracted from...
12301 (shorten_branches): ...here. Call it.
12302 (update_alignments): New function.
12303 * reorg.c (sibling_labels): New variable.
12304 (get_label_before): Add SIBLING parameter. If it is non-zero, push
12305 the new label along with it onto the sibling_labels vector.
12306 (fill_simple_delay_slots): Adjust call to get_label_before.
12307 (fill_slots_from_thread): Likewise.
12308 (relax_delay_slots): Likewise.
12309 (make_return_insns): Likewise.
12310 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
12312 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12314 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
12315 document its semantics.
12316 (diagnostic_report_diagnostic): Adjust accordingly.
12318 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12320 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
12321 (sparc_option_override): ...and port to new C++ pass API.
12322 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
12324 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
12326 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
12328 2013-08-06 Caroline Tice <cmtice@google.com>
12330 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
12331 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
12332 * tree-pass.h: Add pass_vtable_verify.
12333 * varasm.c (assemble_variable): Add code to properly set the comdat
12334 section and name for the .vtable_map_vars section.
12335 (assemble_vtyv_preinit_initializer): New function.
12336 (default_sectin_type_flags): Make sure .vtable_map_vars section has
12338 * output.h: Add function decl for assemble_vtv_preinit_initializer.
12339 * vtable-verify.c: New file.
12340 * vtable-verify.h: New file.
12341 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
12342 initialiation levels.
12343 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
12344 * passes.def: Insert pass_vtable_verify.
12345 * aclocal.m4: Reorder includes.
12346 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
12347 -fvtv-counts options.
12348 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
12349 as appropriate, if -fvtable-verify=... is used.
12350 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
12351 -fvtable-verify=... is used.
12352 * Makefile.in (OBJS): Add vtable-verify.o to list.
12353 (vtable-verify.o): Add new build rule.
12354 (GTFILES): Add vtable-verify.c to list.
12355 * common.opt (fvtable-verify=): New flag.
12356 (vtv_priority): Values for fvtable-verify= flag.
12357 (fvtv-counts): New flag.
12358 (fvtv-debug): New flag.
12359 * tree.h (save_vtable_map_decl): New extern function decl.
12361 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12363 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
12364 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
12365 (pass_data_rl78_devirt): ...new pass_data instance and...
12366 (make_pass_rl78_devirt): ...new function.
12367 (rl78_asm_file_start): Port pass registration to new C++ API.
12369 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12371 * coretypes.h (rtl_opt_pass): Add.
12372 (gcc::context): Add.
12373 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
12374 (epiphany_init): Port to new C++ pass API.
12375 (epiphany_optimize_mode_switching): Likewise.
12376 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
12377 (pass_manager::get_pass_mode_switching): New.
12378 (pass_manager::get_pass_peephole2): New.
12379 * mode-switching.c (pass_mode_switching): Add clone method.
12380 * recog.c (pass_peephole2): Add clone method.
12381 (pass_split_all_insns): Add clone method.
12383 2013-08-06 David Malcolm <dmalcolm@redhat.com>
12385 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
12386 (mips_option_override): ...here, porting to new C++ API for passes.
12388 2013-08-06 Jan Hubicka <jh@suse.cz>
12390 * cgraph.c (cgraph_get_body): New function based on lto.c
12392 * cgraph.h (cgraph_get_body): Declare.
12393 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
12395 * cgraphunit.c (expand_function): Get body prior expanding.
12396 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
12397 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
12399 * passes.c (do_per_function_toporder): Get body.
12400 * tree-inline.c (expand_call_inline): Get body prior inlining it.
12401 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
12403 2013-08-06 Martin Jambor <mjambor@suse.cz>
12406 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
12407 is not re-finalized. Rename second parameter to no_collect.
12409 2013-08-06 Martin Jambor <mjambor@suse.cz>
12411 PR middle-end/58041
12412 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
12413 MEM_REF has proper alignment information.
12415 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
12418 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
12420 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
12421 rtx (*) (rtx, ...) with insn_gen_fn.
12422 * genoutput.c (output_insn_data): Cast gen_? function pointers to
12423 insn_gen_fn::stored_funcptr. Add initializer braces.
12425 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12427 Rewrite how instances of passes are cloned to remove assumptions
12428 about their sizes (thus allowing pass subclasses to have
12429 additional data fields, albeit non-GC-managed ones at this point).
12431 * passes.c (make_pass_instance): Now that passes have clone
12432 methods, rewrite this function to eliminate XNEW and memcpy
12433 calls that used hardcoded sizes. Since this function no longer
12434 creates pass instances, rename it to...
12435 (add_pass_instance): ...this. Document the old way that passes were
12436 numbered and flagged, and rework this function to continue using it.
12437 (next_pass_1): Add an initial_pass argument for use by
12439 (position_pass): When adding multiple instances of a pass, use
12440 the pass's clone method, rather than relying on the XNEW/memcpy
12441 within the former make_pass_instance (now add_pass_instance).
12442 (pass_manager::pass_manager): When invoking next_pass_1, also supply
12443 the initial instance of the current pass within the pass manager.
12445 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12447 This is the automated part of the conversion of passes from C
12448 structs to C++ classes.
12450 Patch autogenerated by refactor_passes.py from
12451 https://github.com/davidmalcolm/gcc-refactoring-scripts
12452 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
12454 * asan.c (pass_asan): Convert from a global struct to a subclass of
12455 gimple_opt_pass along with...
12456 (pass_data_asan): ...new pass_data instance and...
12457 (make_pass_asan): ...new function.
12458 (pass_asan_O0): Convert from a global struct to a subclass of
12459 gimple_opt_pass along with...
12460 (pass_data_asan_O0): ...new pass_data instance and...
12461 (make_pass_asan_O0): ...new function.
12462 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
12463 subclass of rtl_opt_pass along with...
12464 (pass_data_inc_dec): ...new pass_data instance and...
12465 (make_pass_inc_dec): ...new function.
12466 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
12467 a subclass of rtl_opt_pass along with...
12468 (pass_data_reorder_blocks): ...new pass_data instance and...
12469 (make_pass_reorder_blocks): ...new function.
12470 (pass_duplicate_computed_gotos): Convert from a global struct to a
12471 subclass of rtl_opt_pass along with...
12472 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
12473 (make_pass_duplicate_computed_gotos): ...new function.
12474 (pass_partition_blocks): Convert from a global struct to a subclass of
12475 rtl_opt_pass along with...
12476 (pass_data_partition_blocks): ...new pass_data instance and...
12477 (make_pass_partition_blocks): ...new function.
12478 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
12479 struct to a subclass of rtl_opt_pass along with...
12480 (pass_data_branch_target_load_optimize1): ...new pass_data instance
12482 (make_pass_branch_target_load_optimize1): ...new function.
12483 (pass_branch_target_load_optimize2): Convert from a global struct to a
12484 subclass of rtl_opt_pass along with...
12485 (pass_data_branch_target_load_optimize2): ...new pass_data instance
12487 (make_pass_branch_target_load_optimize2): ...new function.
12488 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
12489 of rtl_opt_pass along with...
12490 (pass_data_jump): ...new pass_data instance and...
12491 (make_pass_jump): ...new function.
12492 (pass_jump2): Convert from a global struct to a subclass of
12493 rtl_opt_pass along with...
12494 (pass_data_jump2): ...new pass_data instance and...
12495 (make_pass_jump2): ...new function.
12496 * cfgexpand.c (pass_expand): Convert from a global struct to a
12497 subclass of rtl_opt_pass along with...
12498 (pass_data_expand): ...new pass_data instance and...
12499 (make_pass_expand): ...new function.
12500 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
12501 of rtl_opt_pass along with...
12502 (pass_data_free_cfg): ...new pass_data instance and...
12503 (make_pass_free_cfg): ...new function.
12504 (pass_into_cfg_layout_mode): Convert from a global struct to a
12505 subclass of rtl_opt_pass along with...
12506 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
12507 (make_pass_into_cfg_layout_mode): ...new function.
12508 (pass_outof_cfg_layout_mode): Convert from a global struct to a
12509 subclass of rtl_opt_pass along with...
12510 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
12511 (make_pass_outof_cfg_layout_mode): ...new function.
12512 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
12513 struct to a subclass of gimple_opt_pass along with...
12514 (pass_data_build_cgraph_edges): ...new pass_data instance and...
12515 (make_pass_build_cgraph_edges): ...new function.
12516 (pass_rebuild_cgraph_edges): Convert from a global struct to a
12517 subclass of gimple_opt_pass along with...
12518 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
12519 (make_pass_rebuild_cgraph_edges): ...new function.
12520 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
12521 subclass of gimple_opt_pass along with...
12522 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
12524 (make_pass_remove_cgraph_callee_edges): ...new function.
12525 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
12526 struct to a subclass of rtl_opt_pass along with...
12527 (pass_data_stack_adjustments): ...new pass_data instance and...
12528 (make_pass_stack_adjustments): ...new function.
12529 * combine.c (pass_combine): Convert from a global struct to a subclass
12530 of rtl_opt_pass along with...
12531 (pass_data_combine): ...new pass_data instance and...
12532 (make_pass_combine): ...new function.
12533 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
12534 global struct to a subclass of rtl_opt_pass along with...
12535 (pass_data_compare_elim_after_reload): ...new pass_data instance
12537 (make_pass_compare_elim_after_reload): ...new function.
12538 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
12539 of rtl_opt_pass along with...
12540 (pass_data_rtl_cprop): ...new pass_data instance and...
12541 (make_pass_rtl_cprop): ...new function.
12542 * cse.c (pass_cse): Convert from a global struct to a subclass of
12543 rtl_opt_pass along with...
12544 (pass_data_cse): ...new pass_data instance and...
12545 (make_pass_cse): ...new function.
12546 (pass_cse2): Convert from a global struct to a subclass of
12547 rtl_opt_pass along with...
12548 (pass_data_cse2): ...new pass_data instance and...
12549 (make_pass_cse2): ...new function.
12550 (pass_cse_after_global_opts): Convert from a global struct to a
12551 subclass of rtl_opt_pass along with...
12552 (pass_data_cse_after_global_opts): ...new pass_data instance and...
12553 (make_pass_cse_after_global_opts): ...new function.
12554 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
12555 of rtl_opt_pass along with...
12556 (pass_data_ud_rtl_dce): ...new pass_data instance and...
12557 (make_pass_ud_rtl_dce): ...new function.
12558 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
12559 rtl_opt_pass along with...
12560 (pass_data_fast_rtl_dce): ...new pass_data instance and...
12561 (make_pass_fast_rtl_dce): ...new function.
12562 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
12563 a subclass of rtl_opt_pass along with...
12564 (pass_data_df_initialize_opt): ...new pass_data instance and...
12565 (make_pass_df_initialize_opt): ...new function.
12566 (pass_df_initialize_no_opt): Convert from a global struct to a
12567 subclass of rtl_opt_pass along with...
12568 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
12569 (make_pass_df_initialize_no_opt): ...new function.
12570 (pass_df_finish): Convert from a global struct to a subclass of
12571 rtl_opt_pass along with...
12572 (pass_data_df_finish): ...new pass_data instance and...
12573 (make_pass_df_finish): ...new function.
12574 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
12575 rtl_opt_pass along with...
12576 (pass_data_rtl_dse1): ...new pass_data instance and...
12577 (make_pass_rtl_dse1): ...new function.
12578 (pass_rtl_dse2): Convert from a global struct to a subclass of
12579 rtl_opt_pass along with...
12580 (pass_data_rtl_dse2): ...new pass_data instance and...
12581 (make_pass_rtl_dse2): ...new function.
12582 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
12583 subclass of rtl_opt_pass along with...
12584 (pass_data_dwarf2_frame): ...new pass_data instance and...
12585 (make_pass_dwarf2_frame): ...new function.
12586 * except.c (pass_set_nothrow_function_flags): Convert from a global
12587 struct to a subclass of rtl_opt_pass along with...
12588 (pass_data_set_nothrow_function_flags): ...new pass_data instance
12590 (make_pass_set_nothrow_function_flags): ...new function.
12591 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
12592 subclass of rtl_opt_pass along with...
12593 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
12595 (make_pass_convert_to_eh_region_ranges): ...new function.
12596 * final.c (pass_compute_alignments): Convert from a global struct to a
12597 subclass of rtl_opt_pass along with...
12598 (pass_data_compute_alignments): ...new pass_data instance and...
12599 (make_pass_compute_alignments): ...new function.
12600 (pass_final): Convert from a global struct to a subclass of
12601 rtl_opt_pass along with...
12602 (pass_data_final): ...new pass_data instance and...
12603 (make_pass_final): ...new function.
12604 (pass_shorten_branches): Convert from a global struct to a subclass of
12605 rtl_opt_pass along with...
12606 (pass_data_shorten_branches): ...new pass_data instance and...
12607 (make_pass_shorten_branches): ...new function.
12608 (pass_clean_state): Convert from a global struct to a subclass of
12609 rtl_opt_pass along with...
12610 (pass_data_clean_state): ...new pass_data instance and...
12611 (make_pass_clean_state): ...new function.
12612 * function.c (pass_instantiate_virtual_regs): Convert from a global
12613 struct to a subclass of rtl_opt_pass along with...
12614 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
12615 (make_pass_instantiate_virtual_regs): ...new function.
12616 (pass_leaf_regs): Convert from a global struct to a subclass of
12617 rtl_opt_pass along with...
12618 (pass_data_leaf_regs): ...new pass_data instance and...
12619 (make_pass_leaf_regs): ...new function.
12620 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
12621 subclass of rtl_opt_pass along with...
12622 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
12624 (make_pass_thread_prologue_and_epilogue): ...new function.
12625 (pass_match_asm_constraints): Convert from a global struct to a
12626 subclass of rtl_opt_pass along with...
12627 (pass_data_match_asm_constraints): ...new pass_data instance and...
12628 (make_pass_match_asm_constraints): ...new function.
12629 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
12630 subclass of rtl_opt_pass along with...
12631 (pass_data_rtl_fwprop): ...new pass_data instance and...
12632 (make_pass_rtl_fwprop): ...new function.
12633 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
12634 rtl_opt_pass along with...
12635 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
12636 (make_pass_rtl_fwprop_addr): ...new function.
12637 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
12638 rtl_opt_pass along with...
12639 (pass_data_rtl_pre): ...new pass_data instance and...
12640 (make_pass_rtl_pre): ...new function.
12641 (pass_rtl_hoist): Convert from a global struct to a subclass of
12642 rtl_opt_pass along with...
12643 (pass_data_rtl_hoist): ...new pass_data instance and...
12644 (make_pass_rtl_hoist): ...new function.
12645 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
12646 subclass of gimple_opt_pass along with...
12647 (pass_data_lower_cf): ...new pass_data instance and...
12648 (make_pass_lower_cf): ...new function.
12649 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
12650 from a global struct to a subclass of gimple_opt_pass along with...
12651 (pass_data_strength_reduction): ...new pass_data instance and...
12652 (make_pass_strength_reduction): ...new function.
12653 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
12654 of rtl_opt_pass along with...
12655 (pass_data_rtl_ifcvt): ...new pass_data instance and...
12656 (make_pass_rtl_ifcvt): ...new function.
12657 (pass_if_after_combine): Convert from a global struct to a subclass of
12658 rtl_opt_pass along with...
12659 (pass_data_if_after_combine): ...new pass_data instance and...
12660 (make_pass_if_after_combine): ...new function.
12661 (pass_if_after_reload): Convert from a global struct to a subclass of
12662 rtl_opt_pass along with...
12663 (pass_data_if_after_reload): ...new pass_data instance and...
12664 (make_pass_if_after_reload): ...new function.
12665 * init-regs.c (pass_initialize_regs): Convert from a global struct to
12666 a subclass of rtl_opt_pass along with...
12667 (pass_data_initialize_regs): ...new pass_data instance and...
12668 (make_pass_initialize_regs): ...new function.
12669 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
12670 of ipa_opt_pass_d along with...
12671 (pass_data_ipa_cp): ...new pass_data instance and...
12672 (make_pass_ipa_cp): ...new function.
12673 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
12674 global struct to a subclass of gimple_opt_pass along with...
12675 (pass_data_inline_parameters): ...new pass_data instance and...
12676 (make_pass_inline_parameters): ...new function.
12677 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
12678 subclass of gimple_opt_pass along with...
12679 (pass_data_early_inline): ...new pass_data instance and...
12680 (make_pass_early_inline): ...new function.
12681 (pass_ipa_inline): Convert from a global struct to a subclass of
12682 ipa_opt_pass_d along with...
12683 (pass_data_ipa_inline): ...new pass_data instance and...
12684 (make_pass_ipa_inline): ...new function.
12685 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
12686 struct to a subclass of gimple_opt_pass along with...
12687 (pass_data_local_pure_const): ...new pass_data instance and...
12688 (make_pass_local_pure_const): ...new function.
12689 (pass_ipa_pure_const): Convert from a global struct to a subclass of
12690 ipa_opt_pass_d along with...
12691 (pass_data_ipa_pure_const): ...new pass_data instance and...
12692 (make_pass_ipa_pure_const): ...new function.
12693 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
12694 to a subclass of ipa_opt_pass_d along with...
12695 (pass_data_ipa_reference): ...new pass_data instance and...
12696 (make_pass_ipa_reference): ...new function.
12697 * ipa-split.c (pass_split_functions): Convert from a global struct to
12698 a subclass of gimple_opt_pass along with...
12699 (pass_data_split_functions): ...new pass_data instance and...
12700 (make_pass_split_functions): ...new function.
12701 (pass_feedback_split_functions): Convert from a global struct to a
12702 subclass of gimple_opt_pass along with...
12703 (pass_data_feedback_split_functions): ...new pass_data instance and...
12704 (make_pass_feedback_split_functions): ...new function.
12705 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
12706 global struct to a subclass of simple_ipa_opt_pass along with...
12707 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
12709 (make_pass_ipa_function_and_variable_visibility): ...new function.
12710 (pass_ipa_free_inline_summary): Convert from a global struct to a
12711 subclass of simple_ipa_opt_pass along with...
12712 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
12713 (make_pass_ipa_free_inline_summary): ...new function.
12714 (pass_ipa_whole_program_visibility): Convert from a global struct to a
12715 subclass of ipa_opt_pass_d along with...
12716 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
12718 (make_pass_ipa_whole_program_visibility): ...new function.
12719 (pass_ipa_profile): Convert from a global struct to a subclass of
12720 ipa_opt_pass_d along with...
12721 (pass_data_ipa_profile): ...new pass_data instance and...
12722 (make_pass_ipa_profile): ...new function.
12723 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
12724 ipa_opt_pass_d along with...
12725 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
12726 (make_pass_ipa_cdtor_merge): ...new function.
12727 * ira.c (pass_ira): Convert from a global struct to a subclass of
12728 rtl_opt_pass along with...
12729 (pass_data_ira): ...new pass_data instance and...
12730 (make_pass_ira): ...new function.
12731 (pass_reload): Convert from a global struct to a subclass of
12732 rtl_opt_pass along with...
12733 (pass_data_reload): ...new pass_data instance and...
12734 (make_pass_reload): ...new function.
12735 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
12736 subclass of rtl_opt_pass along with...
12737 (pass_data_cleanup_barriers): ...new pass_data instance and...
12738 (make_pass_cleanup_barriers): ...new function.
12739 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
12740 of rtl_opt_pass along with...
12741 (pass_data_loop2): ...new pass_data instance and...
12742 (make_pass_loop2): ...new function.
12743 (pass_rtl_loop_init): Convert from a global struct to a subclass of
12744 rtl_opt_pass along with...
12745 (pass_data_rtl_loop_init): ...new pass_data instance and...
12746 (make_pass_rtl_loop_init): ...new function.
12747 (pass_rtl_loop_done): Convert from a global struct to a subclass of
12748 rtl_opt_pass along with...
12749 (pass_data_rtl_loop_done): ...new pass_data instance and...
12750 (make_pass_rtl_loop_done): ...new function.
12751 (pass_rtl_move_loop_invariants): Convert from a global struct to a
12752 subclass of rtl_opt_pass along with...
12753 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
12754 (make_pass_rtl_move_loop_invariants): ...new function.
12755 (pass_rtl_unswitch): Convert from a global struct to a subclass of
12756 rtl_opt_pass along with...
12757 (pass_data_rtl_unswitch): ...new pass_data instance and...
12758 (make_pass_rtl_unswitch): ...new function.
12759 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
12760 subclass of rtl_opt_pass along with...
12761 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
12763 (make_pass_rtl_unroll_and_peel_loops): ...new function.
12764 (pass_rtl_doloop): Convert from a global struct to a subclass of
12765 rtl_opt_pass along with...
12766 (pass_data_rtl_doloop): ...new pass_data instance and...
12767 (make_pass_rtl_doloop): ...new function.
12768 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
12769 a subclass of rtl_opt_pass along with...
12770 (pass_data_lower_subreg): ...new pass_data instance and...
12771 (make_pass_lower_subreg): ...new function.
12772 (pass_lower_subreg2): Convert from a global struct to a subclass of
12773 rtl_opt_pass along with...
12774 (pass_data_lower_subreg2): ...new pass_data instance and...
12775 (make_pass_lower_subreg2): ...new function.
12776 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
12777 struct to a subclass of ipa_opt_pass_d along with...
12778 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
12779 (make_pass_ipa_lto_gimple_out): ...new function.
12780 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
12781 of ipa_opt_pass_d along with...
12782 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
12783 (make_pass_ipa_lto_finish_out): ...new function.
12784 * mode-switching.c (pass_mode_switching): Convert from a global struct
12785 to a subclass of rtl_opt_pass along with...
12786 (pass_data_mode_switching): ...new pass_data instance and...
12787 (make_pass_mode_switching): ...new function.
12788 * modulo-sched.c (pass_sms): Convert from a global struct to a
12789 subclass of rtl_opt_pass along with...
12790 (pass_data_sms): ...new pass_data instance and...
12791 (make_pass_sms): ...new function.
12792 * omp-low.c (pass_expand_omp): Convert from a global struct to a
12793 subclass of gimple_opt_pass along with...
12794 (pass_data_expand_omp): ...new pass_data instance and...
12795 (make_pass_expand_omp): ...new function.
12796 (pass_lower_omp): Convert from a global struct to a subclass of
12797 gimple_opt_pass along with...
12798 (pass_data_lower_omp): ...new pass_data instance and...
12799 (make_pass_lower_omp): ...new function.
12800 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
12801 of gimple_opt_pass along with...
12802 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
12803 (make_pass_diagnose_omp_blocks): ...new function.
12804 * passes.c (pass_early_local_passes): Convert from a global struct to
12805 a subclass of simple_ipa_opt_pass along with...
12806 (pass_data_early_local_passes): ...new pass_data instance and...
12807 (make_pass_early_local_passes): ...new function.
12808 (pass_all_early_optimizations): Convert from a global struct to a
12809 subclass of gimple_opt_pass along with...
12810 (pass_data_all_early_optimizations): ...new pass_data instance and...
12811 (make_pass_all_early_optimizations): ...new function.
12812 (pass_all_optimizations): Convert from a global struct to a subclass
12813 of gimple_opt_pass along with...
12814 (pass_data_all_optimizations): ...new pass_data instance and...
12815 (make_pass_all_optimizations): ...new function.
12816 (pass_all_optimizations_g): Convert from a global struct to a subclass
12817 of gimple_opt_pass along with...
12818 (pass_data_all_optimizations_g): ...new pass_data instance and...
12819 (make_pass_all_optimizations_g): ...new function.
12820 (pass_rest_of_compilation): Convert from a global struct to a subclass
12821 of rtl_opt_pass along with...
12822 (pass_data_rest_of_compilation): ...new pass_data instance and...
12823 (make_pass_rest_of_compilation): ...new function.
12824 (pass_postreload): Convert from a global struct to a subclass of
12825 rtl_opt_pass along with...
12826 (pass_data_postreload): ...new pass_data instance and...
12827 (make_pass_postreload): ...new function.
12828 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
12829 subclass of rtl_opt_pass along with...
12830 (pass_data_gcse2): ...new pass_data instance and...
12831 (make_pass_gcse2): ...new function.
12832 * postreload.c (pass_postreload_cse): Convert from a global struct to
12833 a subclass of rtl_opt_pass along with...
12834 (pass_data_postreload_cse): ...new pass_data instance and...
12835 (make_pass_postreload_cse): ...new function.
12836 * predict.c (pass_profile): Convert from a global struct to a subclass
12837 of gimple_opt_pass along with...
12838 (pass_data_profile): ...new pass_data instance and...
12839 (make_pass_profile): ...new function.
12840 (pass_strip_predict_hints): Convert from a global struct to a subclass
12841 of gimple_opt_pass along with...
12842 (pass_data_strip_predict_hints): ...new pass_data instance and...
12843 (make_pass_strip_predict_hints): ...new function.
12844 * recog.c (pass_peephole2): Convert from a global struct to a subclass
12845 of rtl_opt_pass along with...
12846 (pass_data_peephole2): ...new pass_data instance and...
12847 (make_pass_peephole2): ...new function.
12848 (pass_split_all_insns): Convert from a global struct to a subclass of
12849 rtl_opt_pass along with...
12850 (pass_data_split_all_insns): ...new pass_data instance and...
12851 (make_pass_split_all_insns): ...new function.
12852 (pass_split_after_reload): Convert from a global struct to a subclass
12853 of rtl_opt_pass along with...
12854 (pass_data_split_after_reload): ...new pass_data instance and...
12855 (make_pass_split_after_reload): ...new function.
12856 (pass_split_before_regstack): Convert from a global struct to a
12857 subclass of rtl_opt_pass along with...
12858 (pass_data_split_before_regstack): ...new pass_data instance and...
12859 (make_pass_split_before_regstack): ...new function.
12860 (pass_split_before_sched2): Convert from a global struct to a subclass
12861 of rtl_opt_pass along with...
12862 (pass_data_split_before_sched2): ...new pass_data instance and...
12863 (make_pass_split_before_sched2): ...new function.
12864 (pass_split_for_shorten_branches): Convert from a global struct to a
12865 subclass of rtl_opt_pass along with...
12866 (pass_data_split_for_shorten_branches): ...new pass_data instance
12868 (make_pass_split_for_shorten_branches): ...new function.
12869 * ree.c (pass_ree): Convert from a global struct to a subclass of
12870 rtl_opt_pass along with...
12871 (pass_data_ree): ...new pass_data instance and...
12872 (make_pass_ree): ...new function.
12873 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
12874 subclass of rtl_opt_pass along with...
12875 (pass_data_stack_regs): ...new pass_data instance and...
12876 (make_pass_stack_regs): ...new function.
12877 (pass_stack_regs_run): Convert from a global struct to a subclass of
12878 rtl_opt_pass along with...
12879 (pass_data_stack_regs_run): ...new pass_data instance and...
12880 (make_pass_stack_regs_run): ...new function.
12881 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
12882 subclass of rtl_opt_pass along with...
12883 (pass_data_cprop_hardreg): ...new pass_data instance and...
12884 (make_pass_cprop_hardreg): ...new function.
12885 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
12886 subclass of rtl_opt_pass along with...
12887 (pass_data_reginfo_init): ...new pass_data instance and...
12888 (make_pass_reginfo_init): ...new function.
12889 * regmove.c (pass_regmove): Convert from a global struct to a subclass
12890 of rtl_opt_pass along with...
12891 (pass_data_regmove): ...new pass_data instance and...
12892 (make_pass_regmove): ...new function.
12893 * regrename.c (pass_regrename): Convert from a global struct to a
12894 subclass of rtl_opt_pass along with...
12895 (pass_data_regrename): ...new pass_data instance and...
12896 (make_pass_regrename): ...new function.
12897 * reorg.c (pass_delay_slots): Convert from a global struct to a
12898 subclass of rtl_opt_pass along with...
12899 (pass_data_delay_slots): ...new pass_data instance and...
12900 (make_pass_delay_slots): ...new function.
12901 (pass_machine_reorg): Convert from a global struct to a subclass of
12902 rtl_opt_pass along with...
12903 (pass_data_machine_reorg): ...new pass_data instance and...
12904 (make_pass_machine_reorg): ...new function.
12905 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
12906 of rtl_opt_pass along with...
12907 (pass_data_sched): ...new pass_data instance and...
12908 (make_pass_sched): ...new function.
12909 (pass_sched2): Convert from a global struct to a subclass of
12910 rtl_opt_pass along with...
12911 (pass_data_sched2): ...new pass_data instance and...
12912 (make_pass_sched2): ...new function.
12913 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
12914 to a subclass of rtl_opt_pass along with...
12915 (pass_data_stack_ptr_mod): ...new pass_data instance and...
12916 (make_pass_stack_ptr_mod): ...new function.
12917 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
12918 to a subclass of rtl_opt_pass along with...
12919 (pass_data_rtl_store_motion): ...new pass_data instance and...
12920 (make_pass_rtl_store_motion): ...new function.
12921 * tracer.c (pass_tracer): Convert from a global struct to a subclass
12922 of gimple_opt_pass along with...
12923 (pass_data_tracer): ...new pass_data instance and...
12924 (make_pass_tracer): ...new function.
12925 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
12926 to a subclass of gimple_opt_pass along with...
12927 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
12928 (make_pass_diagnose_tm_blocks): ...new function.
12929 (pass_lower_tm): Convert from a global struct to a subclass of
12930 gimple_opt_pass along with...
12931 (pass_data_lower_tm): ...new pass_data instance and...
12932 (make_pass_lower_tm): ...new function.
12933 (pass_tm_init): Convert from a global struct to a subclass of
12934 gimple_opt_pass along with...
12935 (pass_data_tm_init): ...new pass_data instance and...
12936 (make_pass_tm_init): ...new function.
12937 (pass_tm_mark): Convert from a global struct to a subclass of
12938 gimple_opt_pass along with...
12939 (pass_data_tm_mark): ...new pass_data instance and...
12940 (make_pass_tm_mark): ...new function.
12941 (pass_tm_edges): Convert from a global struct to a subclass of
12942 gimple_opt_pass along with...
12943 (pass_data_tm_edges): ...new pass_data instance and...
12944 (make_pass_tm_edges): ...new function.
12945 (pass_tm_memopt): Convert from a global struct to a subclass of
12946 gimple_opt_pass along with...
12947 (pass_data_tm_memopt): ...new pass_data instance and...
12948 (make_pass_tm_memopt): ...new function.
12949 (pass_ipa_tm): Convert from a global struct to a subclass of
12950 simple_ipa_opt_pass along with...
12951 (pass_data_ipa_tm): ...new pass_data instance and...
12952 (make_pass_ipa_tm): ...new function.
12953 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
12954 subclass of gimple_opt_pass along with...
12955 (pass_data_call_cdce): ...new pass_data instance and...
12956 (make_pass_call_cdce): ...new function.
12957 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
12958 subclass of gimple_opt_pass along with...
12959 (pass_data_build_cfg): ...new pass_data instance and...
12960 (make_pass_build_cfg): ...new function.
12961 (pass_split_crit_edges): Convert from a global struct to a subclass of
12962 gimple_opt_pass along with...
12963 (pass_data_split_crit_edges): ...new pass_data instance and...
12964 (make_pass_split_crit_edges): ...new function.
12965 (pass_warn_function_return): Convert from a global struct to a
12966 subclass of gimple_opt_pass along with...
12967 (pass_data_warn_function_return): ...new pass_data instance and...
12968 (make_pass_warn_function_return): ...new function.
12969 (pass_warn_function_noreturn): Convert from a global struct to a
12970 subclass of gimple_opt_pass along with...
12971 (pass_data_warn_function_noreturn): ...new pass_data instance and...
12972 (make_pass_warn_function_noreturn): ...new function.
12973 (pass_warn_unused_result): Convert from a global struct to a subclass
12974 of gimple_opt_pass along with...
12975 (pass_data_warn_unused_result): ...new pass_data instance and...
12976 (make_pass_warn_unused_result): ...new function.
12977 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
12978 a subclass of gimple_opt_pass along with...
12979 (pass_data_merge_phi): ...new pass_data instance and...
12980 (make_pass_merge_phi): ...new function.
12981 * tree-complex.c (pass_lower_complex): Convert from a global struct to
12982 a subclass of gimple_opt_pass along with...
12983 (pass_data_lower_complex): ...new pass_data instance and...
12984 (make_pass_lower_complex): ...new function.
12985 (pass_lower_complex_O0): Convert from a global struct to a subclass of
12986 gimple_opt_pass along with...
12987 (pass_data_lower_complex_O0): ...new pass_data instance and...
12988 (make_pass_lower_complex_O0): ...new function.
12989 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
12990 subclass of gimple_opt_pass along with...
12991 (pass_data_lower_eh): ...new pass_data instance and...
12992 (make_pass_lower_eh): ...new function.
12993 (pass_refactor_eh): Convert from a global struct to a subclass of
12994 gimple_opt_pass along with...
12995 (pass_data_refactor_eh): ...new pass_data instance and...
12996 (make_pass_refactor_eh): ...new function.
12997 (pass_lower_resx): Convert from a global struct to a subclass of
12998 gimple_opt_pass along with...
12999 (pass_data_lower_resx): ...new pass_data instance and...
13000 (make_pass_lower_resx): ...new function.
13001 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
13002 of gimple_opt_pass along with...
13003 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
13004 (make_pass_lower_eh_dispatch): ...new function.
13005 (pass_cleanup_eh): Convert from a global struct to a subclass of
13006 gimple_opt_pass along with...
13007 (pass_data_cleanup_eh): ...new pass_data instance and...
13008 (make_pass_cleanup_eh): ...new function.
13009 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
13010 to a subclass of simple_ipa_opt_pass along with...
13011 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
13012 (make_pass_ipa_lower_emutls): ...new function.
13013 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
13014 a subclass of gimple_opt_pass along with...
13015 (pass_data_if_conversion): ...new pass_data instance and...
13016 (make_pass_if_conversion): ...new function.
13017 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
13018 subclass of gimple_opt_pass along with...
13019 (pass_data_build_ssa): ...new pass_data instance and...
13020 (make_pass_build_ssa): ...new function.
13021 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
13022 global struct to a subclass of gimple_opt_pass along with...
13023 (pass_data_loop_distribution): ...new pass_data instance and...
13024 (make_pass_loop_distribution): ...new function.
13025 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
13026 subclass of gimple_opt_pass along with...
13027 (pass_data_mudflap_1): ...new pass_data instance and...
13028 (make_pass_mudflap_1): ...new function.
13029 (pass_mudflap_2): Convert from a global struct to a subclass of
13030 gimple_opt_pass along with...
13031 (pass_data_mudflap_2): ...new pass_data instance and...
13032 (make_pass_mudflap_2): ...new function.
13033 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
13034 subclass of gimple_opt_pass along with...
13035 (pass_data_mudflap_1): ...new pass_data instance and...
13036 (make_pass_mudflap_1): ...new function.
13037 (pass_mudflap_2): Convert from a global struct to a subclass of
13038 gimple_opt_pass along with...
13039 (pass_data_mudflap_2): ...new pass_data instance and...
13040 (make_pass_mudflap_2): ...new function.
13041 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
13042 gimple_opt_pass along with...
13043 (pass_data_nrv): ...new pass_data instance and...
13044 (make_pass_nrv): ...new function.
13045 (pass_return_slot): Convert from a global struct to a subclass of
13046 gimple_opt_pass along with...
13047 (pass_data_return_slot): ...new pass_data instance and...
13048 (make_pass_return_slot): ...new function.
13049 * tree-object-size.c (pass_object_sizes): Convert from a global struct
13050 to a subclass of gimple_opt_pass along with...
13051 (pass_data_object_sizes): ...new pass_data instance and...
13052 (make_pass_object_sizes): ...new function.
13053 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
13054 global struct to a subclass of gimple_opt_pass along with...
13055 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
13057 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13058 (pass_fixup_cfg): Convert from a global struct to a subclass of
13059 gimple_opt_pass along with...
13060 (pass_data_fixup_cfg): ...new pass_data instance and...
13061 (make_pass_fixup_cfg): ...new function.
13062 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
13063 (make_pass_mudflap_1): ...new function.
13064 (pass_mudflap_2): Replace declaration with that of...
13065 (make_pass_mudflap_2): ...new function.
13066 (pass_asan): Replace declaration with that of...
13067 (make_pass_asan): ...new function.
13068 (pass_asan_O0): Replace declaration with that of...
13069 (make_pass_asan_O0): ...new function.
13070 (pass_tsan): Replace declaration with that of...
13071 (make_pass_tsan): ...new function.
13072 (pass_tsan_O0): Replace declaration with that of...
13073 (make_pass_tsan_O0): ...new function.
13074 (pass_lower_cf): Replace declaration with that of...
13075 (make_pass_lower_cf): ...new function.
13076 (pass_refactor_eh): Replace declaration with that of...
13077 (make_pass_refactor_eh): ...new function.
13078 (pass_lower_eh): Replace declaration with that of...
13079 (make_pass_lower_eh): ...new function.
13080 (pass_lower_eh_dispatch): Replace declaration with that of...
13081 (make_pass_lower_eh_dispatch): ...new function.
13082 (pass_lower_resx): Replace declaration with that of...
13083 (make_pass_lower_resx): ...new function.
13084 (pass_build_cfg): Replace declaration with that of...
13085 (make_pass_build_cfg): ...new function.
13086 (pass_early_tree_profile): Replace declaration with that of...
13087 (make_pass_early_tree_profile): ...new function.
13088 (pass_cleanup_eh): Replace declaration with that of...
13089 (make_pass_cleanup_eh): ...new function.
13090 (pass_sra): Replace declaration with that of...
13091 (make_pass_sra): ...new function.
13092 (pass_sra_early): Replace declaration with that of...
13093 (make_pass_sra_early): ...new function.
13094 (pass_early_ipa_sra): Replace declaration with that of...
13095 (make_pass_early_ipa_sra): ...new function.
13096 (pass_tail_recursion): Replace declaration with that of...
13097 (make_pass_tail_recursion): ...new function.
13098 (pass_tail_calls): Replace declaration with that of...
13099 (make_pass_tail_calls): ...new function.
13100 (pass_tree_loop): Replace declaration with that of...
13101 (make_pass_tree_loop): ...new function.
13102 (pass_tree_loop_init): Replace declaration with that of...
13103 (make_pass_tree_loop_init): ...new function.
13104 (pass_lim): Replace declaration with that of...
13105 (make_pass_lim): ...new function.
13106 (pass_tree_unswitch): Replace declaration with that of...
13107 (make_pass_tree_unswitch): ...new function.
13108 (pass_predcom): Replace declaration with that of...
13109 (make_pass_predcom): ...new function.
13110 (pass_iv_canon): Replace declaration with that of...
13111 (make_pass_iv_canon): ...new function.
13112 (pass_scev_cprop): Replace declaration with that of...
13113 (make_pass_scev_cprop): ...new function.
13114 (pass_empty_loop): Replace declaration with that of...
13115 (make_pass_empty_loop): ...new function.
13116 (pass_record_bounds): Replace declaration with that of...
13117 (make_pass_record_bounds): ...new function.
13118 (pass_graphite): Replace declaration with that of...
13119 (make_pass_graphite): ...new function.
13120 (pass_graphite_transforms): Replace declaration with that of...
13121 (make_pass_graphite_transforms): ...new function.
13122 (pass_if_conversion): Replace declaration with that of...
13123 (make_pass_if_conversion): ...new function.
13124 (pass_loop_distribution): Replace declaration with that of...
13125 (make_pass_loop_distribution): ...new function.
13126 (pass_vectorize): Replace declaration with that of...
13127 (make_pass_vectorize): ...new function.
13128 (pass_slp_vectorize): Replace declaration with that of...
13129 (make_pass_slp_vectorize): ...new function.
13130 (pass_complete_unroll): Replace declaration with that of...
13131 (make_pass_complete_unroll): ...new function.
13132 (pass_complete_unrolli): Replace declaration with that of...
13133 (make_pass_complete_unrolli): ...new function.
13134 (pass_parallelize_loops): Replace declaration with that of...
13135 (make_pass_parallelize_loops): ...new function.
13136 (pass_loop_prefetch): Replace declaration with that of...
13137 (make_pass_loop_prefetch): ...new function.
13138 (pass_iv_optimize): Replace declaration with that of...
13139 (make_pass_iv_optimize): ...new function.
13140 (pass_tree_loop_done): Replace declaration with that of...
13141 (make_pass_tree_loop_done): ...new function.
13142 (pass_ch): Replace declaration with that of...
13143 (make_pass_ch): ...new function.
13144 (pass_ccp): Replace declaration with that of...
13145 (make_pass_ccp): ...new function.
13146 (pass_phi_only_cprop): Replace declaration with that of...
13147 (make_pass_phi_only_cprop): ...new function.
13148 (pass_build_ssa): Replace declaration with that of...
13149 (make_pass_build_ssa): ...new function.
13150 (pass_build_alias): Replace declaration with that of...
13151 (make_pass_build_alias): ...new function.
13152 (pass_build_ealias): Replace declaration with that of...
13153 (make_pass_build_ealias): ...new function.
13154 (pass_dominator): Replace declaration with that of...
13155 (make_pass_dominator): ...new function.
13156 (pass_dce): Replace declaration with that of...
13157 (make_pass_dce): ...new function.
13158 (pass_dce_loop): Replace declaration with that of...
13159 (make_pass_dce_loop): ...new function.
13160 (pass_cd_dce): Replace declaration with that of...
13161 (make_pass_cd_dce): ...new function.
13162 (pass_call_cdce): Replace declaration with that of...
13163 (make_pass_call_cdce): ...new function.
13164 (pass_merge_phi): Replace declaration with that of...
13165 (make_pass_merge_phi): ...new function.
13166 (pass_split_crit_edges): Replace declaration with that of...
13167 (make_pass_split_crit_edges): ...new function.
13168 (pass_pre): Replace declaration with that of...
13169 (make_pass_pre): ...new function.
13170 (pass_profile): Replace declaration with that of...
13171 (make_pass_profile): ...new function.
13172 (pass_strip_predict_hints): Replace declaration with that of...
13173 (make_pass_strip_predict_hints): ...new function.
13174 (pass_lower_complex_O0): Replace declaration with that of...
13175 (make_pass_lower_complex_O0): ...new function.
13176 (pass_lower_complex): Replace declaration with that of...
13177 (make_pass_lower_complex): ...new function.
13178 (pass_lower_vector): Replace declaration with that of...
13179 (make_pass_lower_vector): ...new function.
13180 (pass_lower_vector_ssa): Replace declaration with that of...
13181 (make_pass_lower_vector_ssa): ...new function.
13182 (pass_lower_omp): Replace declaration with that of...
13183 (make_pass_lower_omp): ...new function.
13184 (pass_diagnose_omp_blocks): Replace declaration with that of...
13185 (make_pass_diagnose_omp_blocks): ...new function.
13186 (pass_expand_omp): Replace declaration with that of...
13187 (make_pass_expand_omp): ...new function.
13188 (pass_expand_omp_ssa): Replace declaration with that of...
13189 (make_pass_expand_omp_ssa): ...new function.
13190 (pass_object_sizes): Replace declaration with that of...
13191 (make_pass_object_sizes): ...new function.
13192 (pass_strlen): Replace declaration with that of...
13193 (make_pass_strlen): ...new function.
13194 (pass_fold_builtins): Replace declaration with that of...
13195 (make_pass_fold_builtins): ...new function.
13196 (pass_stdarg): Replace declaration with that of...
13197 (make_pass_stdarg): ...new function.
13198 (pass_early_warn_uninitialized): Replace declaration with that of...
13199 (make_pass_early_warn_uninitialized): ...new function.
13200 (pass_late_warn_uninitialized): Replace declaration with that of...
13201 (make_pass_late_warn_uninitialized): ...new function.
13202 (pass_cse_reciprocals): Replace declaration with that of...
13203 (make_pass_cse_reciprocals): ...new function.
13204 (pass_cse_sincos): Replace declaration with that of...
13205 (make_pass_cse_sincos): ...new function.
13206 (pass_optimize_bswap): Replace declaration with that of...
13207 (make_pass_optimize_bswap): ...new function.
13208 (pass_optimize_widening_mul): Replace declaration with that of...
13209 (make_pass_optimize_widening_mul): ...new function.
13210 (pass_warn_function_return): Replace declaration with that of...
13211 (make_pass_warn_function_return): ...new function.
13212 (pass_warn_function_noreturn): Replace declaration with that of...
13213 (make_pass_warn_function_noreturn): ...new function.
13214 (pass_cselim): Replace declaration with that of...
13215 (make_pass_cselim): ...new function.
13216 (pass_phiopt): Replace declaration with that of...
13217 (make_pass_phiopt): ...new function.
13218 (pass_forwprop): Replace declaration with that of...
13219 (make_pass_forwprop): ...new function.
13220 (pass_phiprop): Replace declaration with that of...
13221 (make_pass_phiprop): ...new function.
13222 (pass_tree_ifcombine): Replace declaration with that of...
13223 (make_pass_tree_ifcombine): ...new function.
13224 (pass_dse): Replace declaration with that of...
13225 (make_pass_dse): ...new function.
13226 (pass_nrv): Replace declaration with that of...
13227 (make_pass_nrv): ...new function.
13228 (pass_rename_ssa_copies): Replace declaration with that of...
13229 (make_pass_rename_ssa_copies): ...new function.
13230 (pass_sink_code): Replace declaration with that of...
13231 (make_pass_sink_code): ...new function.
13232 (pass_fre): Replace declaration with that of...
13233 (make_pass_fre): ...new function.
13234 (pass_check_data_deps): Replace declaration with that of...
13235 (make_pass_check_data_deps): ...new function.
13236 (pass_copy_prop): Replace declaration with that of...
13237 (make_pass_copy_prop): ...new function.
13238 (pass_vrp): Replace declaration with that of...
13239 (make_pass_vrp): ...new function.
13240 (pass_uncprop): Replace declaration with that of...
13241 (make_pass_uncprop): ...new function.
13242 (pass_return_slot): Replace declaration with that of...
13243 (make_pass_return_slot): ...new function.
13244 (pass_reassoc): Replace declaration with that of...
13245 (make_pass_reassoc): ...new function.
13246 (pass_rebuild_cgraph_edges): Replace declaration with that of...
13247 (make_pass_rebuild_cgraph_edges): ...new function.
13248 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
13249 (make_pass_remove_cgraph_callee_edges): ...new function.
13250 (pass_build_cgraph_edges): Replace declaration with that of...
13251 (make_pass_build_cgraph_edges): ...new function.
13252 (pass_local_pure_const): Replace declaration with that of...
13253 (make_pass_local_pure_const): ...new function.
13254 (pass_tracer): Replace declaration with that of...
13255 (make_pass_tracer): ...new function.
13256 (pass_warn_unused_result): Replace declaration with that of...
13257 (make_pass_warn_unused_result): ...new function.
13258 (pass_diagnose_tm_blocks): Replace declaration with that of...
13259 (make_pass_diagnose_tm_blocks): ...new function.
13260 (pass_lower_tm): Replace declaration with that of...
13261 (make_pass_lower_tm): ...new function.
13262 (pass_tm_init): Replace declaration with that of...
13263 (make_pass_tm_init): ...new function.
13264 (pass_tm_mark): Replace declaration with that of...
13265 (make_pass_tm_mark): ...new function.
13266 (pass_tm_memopt): Replace declaration with that of...
13267 (make_pass_tm_memopt): ...new function.
13268 (pass_tm_edges): Replace declaration with that of...
13269 (make_pass_tm_edges): ...new function.
13270 (pass_split_functions): Replace declaration with that of...
13271 (make_pass_split_functions): ...new function.
13272 (pass_feedback_split_functions): Replace declaration with that of...
13273 (make_pass_feedback_split_functions): ...new function.
13274 (pass_strength_reduction): Replace declaration with that of...
13275 (make_pass_strength_reduction): ...new function.
13276 (pass_ipa_lower_emutls): Replace declaration with that of...
13277 (make_pass_ipa_lower_emutls): ...new function.
13278 (pass_ipa_function_and_variable_visibility): Replace declaration with
13280 (make_pass_ipa_function_and_variable_visibility): ...new function.
13281 (pass_ipa_tree_profile): Replace declaration with that of...
13282 (make_pass_ipa_tree_profile): ...new function.
13283 (pass_early_local_passes): Replace declaration with that of...
13284 (make_pass_early_local_passes): ...new function.
13285 (pass_ipa_whole_program_visibility): Replace declaration with that
13287 (make_pass_ipa_whole_program_visibility): ...new function.
13288 (pass_ipa_lto_gimple_out): Replace declaration with that of...
13289 (make_pass_ipa_lto_gimple_out): ...new function.
13290 (pass_ipa_increase_alignment): Replace declaration with that of...
13291 (make_pass_ipa_increase_alignment): ...new function.
13292 (pass_ipa_inline): Replace declaration with that of...
13293 (make_pass_ipa_inline): ...new function.
13294 (pass_ipa_free_lang_data): Replace declaration with that of...
13295 (make_pass_ipa_free_lang_data): ...new function.
13296 (pass_ipa_free_inline_summary): Replace declaration with that of...
13297 (make_pass_ipa_free_inline_summary): ...new function.
13298 (pass_ipa_cp): Replace declaration with that of...
13299 (make_pass_ipa_cp): ...new function.
13300 (pass_ipa_reference): Replace declaration with that of...
13301 (make_pass_ipa_reference): ...new function.
13302 (pass_ipa_pure_const): Replace declaration with that of...
13303 (make_pass_ipa_pure_const): ...new function.
13304 (pass_ipa_pta): Replace declaration with that of...
13305 (make_pass_ipa_pta): ...new function.
13306 (pass_ipa_lto_finish_out): Replace declaration with that of...
13307 (make_pass_ipa_lto_finish_out): ...new function.
13308 (pass_ipa_tm): Replace declaration with that of...
13309 (make_pass_ipa_tm): ...new function.
13310 (pass_ipa_profile): Replace declaration with that of...
13311 (make_pass_ipa_profile): ...new function.
13312 (pass_ipa_cdtor_merge): Replace declaration with that of...
13313 (make_pass_ipa_cdtor_merge): ...new function.
13314 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
13316 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13317 (pass_init_datastructures): Replace declaration with that of...
13318 (make_pass_init_datastructures): ...new function.
13319 (pass_fixup_cfg): Replace declaration with that of...
13320 (make_pass_fixup_cfg): ...new function.
13321 (pass_expand): Replace declaration with that of...
13322 (make_pass_expand): ...new function.
13323 (pass_instantiate_virtual_regs): Replace declaration with that of...
13324 (make_pass_instantiate_virtual_regs): ...new function.
13325 (pass_rtl_fwprop): Replace declaration with that of...
13326 (make_pass_rtl_fwprop): ...new function.
13327 (pass_rtl_fwprop_addr): Replace declaration with that of...
13328 (make_pass_rtl_fwprop_addr): ...new function.
13329 (pass_jump): Replace declaration with that of...
13330 (make_pass_jump): ...new function.
13331 (pass_jump2): Replace declaration with that of...
13332 (make_pass_jump2): ...new function.
13333 (pass_lower_subreg): Replace declaration with that of...
13334 (make_pass_lower_subreg): ...new function.
13335 (pass_cse): Replace declaration with that of...
13336 (make_pass_cse): ...new function.
13337 (pass_fast_rtl_dce): Replace declaration with that of...
13338 (make_pass_fast_rtl_dce): ...new function.
13339 (pass_ud_rtl_dce): Replace declaration with that of...
13340 (make_pass_ud_rtl_dce): ...new function.
13341 (pass_rtl_dce): Replace declaration with that of...
13342 (make_pass_rtl_dce): ...new function.
13343 (pass_rtl_dse1): Replace declaration with that of...
13344 (make_pass_rtl_dse1): ...new function.
13345 (pass_rtl_dse2): Replace declaration with that of...
13346 (make_pass_rtl_dse2): ...new function.
13347 (pass_rtl_dse3): Replace declaration with that of...
13348 (make_pass_rtl_dse3): ...new function.
13349 (pass_rtl_cprop): Replace declaration with that of...
13350 (make_pass_rtl_cprop): ...new function.
13351 (pass_rtl_pre): Replace declaration with that of...
13352 (make_pass_rtl_pre): ...new function.
13353 (pass_rtl_hoist): Replace declaration with that of...
13354 (make_pass_rtl_hoist): ...new function.
13355 (pass_rtl_store_motion): Replace declaration with that of...
13356 (make_pass_rtl_store_motion): ...new function.
13357 (pass_cse_after_global_opts): Replace declaration with that of...
13358 (make_pass_cse_after_global_opts): ...new function.
13359 (pass_rtl_ifcvt): Replace declaration with that of...
13360 (make_pass_rtl_ifcvt): ...new function.
13361 (pass_into_cfg_layout_mode): Replace declaration with that of...
13362 (make_pass_into_cfg_layout_mode): ...new function.
13363 (pass_outof_cfg_layout_mode): Replace declaration with that of...
13364 (make_pass_outof_cfg_layout_mode): ...new function.
13365 (pass_loop2): Replace declaration with that of...
13366 (make_pass_loop2): ...new function.
13367 (pass_rtl_loop_init): Replace declaration with that of...
13368 (make_pass_rtl_loop_init): ...new function.
13369 (pass_rtl_move_loop_invariants): Replace declaration with that of...
13370 (make_pass_rtl_move_loop_invariants): ...new function.
13371 (pass_rtl_unswitch): Replace declaration with that of...
13372 (make_pass_rtl_unswitch): ...new function.
13373 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
13374 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13375 (pass_rtl_doloop): Replace declaration with that of...
13376 (make_pass_rtl_doloop): ...new function.
13377 (pass_rtl_loop_done): Replace declaration with that of...
13378 (make_pass_rtl_loop_done): ...new function.
13379 (pass_web): Replace declaration with that of...
13380 (make_pass_web): ...new function.
13381 (pass_cse2): Replace declaration with that of...
13382 (make_pass_cse2): ...new function.
13383 (pass_df_initialize_opt): Replace declaration with that of...
13384 (make_pass_df_initialize_opt): ...new function.
13385 (pass_df_initialize_no_opt): Replace declaration with that of...
13386 (make_pass_df_initialize_no_opt): ...new function.
13387 (pass_reginfo_init): Replace declaration with that of...
13388 (make_pass_reginfo_init): ...new function.
13389 (pass_inc_dec): Replace declaration with that of...
13390 (make_pass_inc_dec): ...new function.
13391 (pass_stack_ptr_mod): Replace declaration with that of...
13392 (make_pass_stack_ptr_mod): ...new function.
13393 (pass_initialize_regs): Replace declaration with that of...
13394 (make_pass_initialize_regs): ...new function.
13395 (pass_combine): Replace declaration with that of...
13396 (make_pass_combine): ...new function.
13397 (pass_if_after_combine): Replace declaration with that of...
13398 (make_pass_if_after_combine): ...new function.
13399 (pass_ree): Replace declaration with that of...
13400 (make_pass_ree): ...new function.
13401 (pass_partition_blocks): Replace declaration with that of...
13402 (make_pass_partition_blocks): ...new function.
13403 (pass_match_asm_constraints): Replace declaration with that of...
13404 (make_pass_match_asm_constraints): ...new function.
13405 (pass_regmove): Replace declaration with that of...
13406 (make_pass_regmove): ...new function.
13407 (pass_split_all_insns): Replace declaration with that of...
13408 (make_pass_split_all_insns): ...new function.
13409 (pass_fast_rtl_byte_dce): Replace declaration with that of...
13410 (make_pass_fast_rtl_byte_dce): ...new function.
13411 (pass_lower_subreg2): Replace declaration with that of...
13412 (make_pass_lower_subreg2): ...new function.
13413 (pass_mode_switching): Replace declaration with that of...
13414 (make_pass_mode_switching): ...new function.
13415 (pass_sms): Replace declaration with that of...
13416 (make_pass_sms): ...new function.
13417 (pass_sched): Replace declaration with that of...
13418 (make_pass_sched): ...new function.
13419 (pass_ira): Replace declaration with that of...
13420 (make_pass_ira): ...new function.
13421 (pass_reload): Replace declaration with that of...
13422 (make_pass_reload): ...new function.
13423 (pass_clean_state): Replace declaration with that of...
13424 (make_pass_clean_state): ...new function.
13425 (pass_branch_prob): Replace declaration with that of...
13426 (make_pass_branch_prob): ...new function.
13427 (pass_value_profile_transformations): Replace declaration with that
13429 (make_pass_value_profile_transformations): ...new function.
13430 (pass_postreload_cse): Replace declaration with that of...
13431 (make_pass_postreload_cse): ...new function.
13432 (pass_gcse2): Replace declaration with that of...
13433 (make_pass_gcse2): ...new function.
13434 (pass_split_after_reload): Replace declaration with that of...
13435 (make_pass_split_after_reload): ...new function.
13436 (pass_branch_target_load_optimize1): Replace declaration with that
13438 (make_pass_branch_target_load_optimize1): ...new function.
13439 (pass_thread_prologue_and_epilogue): Replace declaration with that
13441 (make_pass_thread_prologue_and_epilogue): ...new function.
13442 (pass_stack_adjustments): Replace declaration with that of...
13443 (make_pass_stack_adjustments): ...new function.
13444 (pass_peephole2): Replace declaration with that of...
13445 (make_pass_peephole2): ...new function.
13446 (pass_if_after_reload): Replace declaration with that of...
13447 (make_pass_if_after_reload): ...new function.
13448 (pass_regrename): Replace declaration with that of...
13449 (make_pass_regrename): ...new function.
13450 (pass_cprop_hardreg): Replace declaration with that of...
13451 (make_pass_cprop_hardreg): ...new function.
13452 (pass_reorder_blocks): Replace declaration with that of...
13453 (make_pass_reorder_blocks): ...new function.
13454 (pass_branch_target_load_optimize2): Replace declaration with that
13456 (make_pass_branch_target_load_optimize2): ...new function.
13457 (pass_leaf_regs): Replace declaration with that of...
13458 (make_pass_leaf_regs): ...new function.
13459 (pass_split_before_sched2): Replace declaration with that of...
13460 (make_pass_split_before_sched2): ...new function.
13461 (pass_compare_elim_after_reload): Replace declaration with that of...
13462 (make_pass_compare_elim_after_reload): ...new function.
13463 (pass_sched2): Replace declaration with that of...
13464 (make_pass_sched2): ...new function.
13465 (pass_stack_regs): Replace declaration with that of...
13466 (make_pass_stack_regs): ...new function.
13467 (pass_stack_regs_run): Replace declaration with that of...
13468 (make_pass_stack_regs_run): ...new function.
13469 (pass_df_finish): Replace declaration with that of...
13470 (make_pass_df_finish): ...new function.
13471 (pass_compute_alignments): Replace declaration with that of...
13472 (make_pass_compute_alignments): ...new function.
13473 (pass_duplicate_computed_gotos): Replace declaration with that of...
13474 (make_pass_duplicate_computed_gotos): ...new function.
13475 (pass_variable_tracking): Replace declaration with that of...
13476 (make_pass_variable_tracking): ...new function.
13477 (pass_free_cfg): Replace declaration with that of...
13478 (make_pass_free_cfg): ...new function.
13479 (pass_machine_reorg): Replace declaration with that of...
13480 (make_pass_machine_reorg): ...new function.
13481 (pass_cleanup_barriers): Replace declaration with that of...
13482 (make_pass_cleanup_barriers): ...new function.
13483 (pass_delay_slots): Replace declaration with that of...
13484 (make_pass_delay_slots): ...new function.
13485 (pass_split_for_shorten_branches): Replace declaration with that of...
13486 (make_pass_split_for_shorten_branches): ...new function.
13487 (pass_split_before_regstack): Replace declaration with that of...
13488 (make_pass_split_before_regstack): ...new function.
13489 (pass_convert_to_eh_region_ranges): Replace declaration with that
13491 (make_pass_convert_to_eh_region_ranges): ...new function.
13492 (pass_shorten_branches): Replace declaration with that of...
13493 (make_pass_shorten_branches): ...new function.
13494 (pass_set_nothrow_function_flags): Replace declaration with that of...
13495 (make_pass_set_nothrow_function_flags): ...new function.
13496 (pass_dwarf2_frame): Replace declaration with that of...
13497 (make_pass_dwarf2_frame): ...new function.
13498 (pass_final): Replace declaration with that of...
13499 (make_pass_final): ...new function.
13500 (pass_rtl_seqabstr): Replace declaration with that of...
13501 (make_pass_rtl_seqabstr): ...new function.
13502 (pass_release_ssa_names): Replace declaration with that of...
13503 (make_pass_release_ssa_names): ...new function.
13504 (pass_early_inline): Replace declaration with that of...
13505 (make_pass_early_inline): ...new function.
13506 (pass_inline_parameters): Replace declaration with that of...
13507 (make_pass_inline_parameters): ...new function.
13508 (pass_update_address_taken): Replace declaration with that of...
13509 (make_pass_update_address_taken): ...new function.
13510 (pass_convert_switch): Replace declaration with that of...
13511 (make_pass_convert_switch): ...new function.
13512 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
13513 to a subclass of simple_ipa_opt_pass along with...
13514 (pass_data_ipa_tree_profile): ...new pass_data instance and...
13515 (make_pass_ipa_tree_profile): ...new function.
13516 * tree-sra.c (pass_sra_early): Convert from a global struct to a
13517 subclass of gimple_opt_pass along with...
13518 (pass_data_sra_early): ...new pass_data instance and...
13519 (make_pass_sra_early): ...new function.
13520 (pass_sra): Convert from a global struct to a subclass of
13521 gimple_opt_pass along with...
13522 (pass_data_sra): ...new pass_data instance and...
13523 (make_pass_sra): ...new function.
13524 (pass_early_ipa_sra): Convert from a global struct to a subclass of
13525 gimple_opt_pass along with...
13526 (pass_data_early_ipa_sra): ...new pass_data instance and...
13527 (make_pass_early_ipa_sra): ...new function.
13528 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
13529 subclass of gimple_opt_pass along with...
13530 (pass_data_ccp): ...new pass_data instance and...
13531 (make_pass_ccp): ...new function.
13532 (pass_fold_builtins): Convert from a global struct to a subclass of
13533 gimple_opt_pass along with...
13534 (pass_data_fold_builtins): ...new pass_data instance and...
13535 (make_pass_fold_builtins): ...new function.
13536 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
13537 subclass of gimple_opt_pass along with...
13538 (pass_data_copy_prop): ...new pass_data instance and...
13539 (make_pass_copy_prop): ...new function.
13540 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
13541 global struct to a subclass of gimple_opt_pass along with...
13542 (pass_data_rename_ssa_copies): ...new pass_data instance and...
13543 (make_pass_rename_ssa_copies): ...new function.
13544 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
13545 subclass of gimple_opt_pass along with...
13546 (pass_data_dce): ...new pass_data instance and...
13547 (make_pass_dce): ...new function.
13548 (pass_dce_loop): Convert from a global struct to a subclass of
13549 gimple_opt_pass along with...
13550 (pass_data_dce_loop): ...new pass_data instance and...
13551 (make_pass_dce_loop): ...new function.
13552 (pass_cd_dce): Convert from a global struct to a subclass of
13553 gimple_opt_pass along with...
13554 (pass_data_cd_dce): ...new pass_data instance and...
13555 (make_pass_cd_dce): ...new function.
13556 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
13557 subclass of gimple_opt_pass along with...
13558 (pass_data_dominator): ...new pass_data instance and...
13559 (make_pass_dominator): ...new function.
13560 (pass_phi_only_cprop): Convert from a global struct to a subclass of
13561 gimple_opt_pass along with...
13562 (pass_data_phi_only_cprop): ...new pass_data instance and...
13563 (make_pass_phi_only_cprop): ...new function.
13564 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
13565 subclass of gimple_opt_pass along with...
13566 (pass_data_dse): ...new pass_data instance and...
13567 (make_pass_dse): ...new function.
13568 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
13569 a subclass of gimple_opt_pass along with...
13570 (pass_data_forwprop): ...new pass_data instance and...
13571 (make_pass_forwprop): ...new function.
13572 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
13573 struct to a subclass of gimple_opt_pass along with...
13574 (pass_data_tree_ifcombine): ...new pass_data instance and...
13575 (make_pass_tree_ifcombine): ...new function.
13576 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
13577 subclass of gimple_opt_pass along with...
13578 (pass_data_ch): ...new pass_data instance and...
13579 (make_pass_ch): ...new function.
13580 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
13581 subclass of gimple_opt_pass along with...
13582 (pass_data_tree_loop): ...new pass_data instance and...
13583 (make_pass_tree_loop): ...new function.
13584 (pass_tree_loop_init): Convert from a global struct to a subclass of
13585 gimple_opt_pass along with...
13586 (pass_data_tree_loop_init): ...new pass_data instance and...
13587 (make_pass_tree_loop_init): ...new function.
13588 (pass_lim): Convert from a global struct to a subclass of
13589 gimple_opt_pass along with...
13590 (pass_data_lim): ...new pass_data instance and...
13591 (make_pass_lim): ...new function.
13592 (pass_tree_unswitch): Convert from a global struct to a subclass of
13593 gimple_opt_pass along with...
13594 (pass_data_tree_unswitch): ...new pass_data instance and...
13595 (make_pass_tree_unswitch): ...new function.
13596 (pass_predcom): Convert from a global struct to a subclass of
13597 gimple_opt_pass along with...
13598 (pass_data_predcom): ...new pass_data instance and...
13599 (make_pass_predcom): ...new function.
13600 (pass_vectorize): Convert from a global struct to a subclass of
13601 gimple_opt_pass along with...
13602 (pass_data_vectorize): ...new pass_data instance and...
13603 (make_pass_vectorize): ...new function.
13604 (pass_graphite): Convert from a global struct to a subclass of
13605 gimple_opt_pass along with...
13606 (pass_data_graphite): ...new pass_data instance and...
13607 (make_pass_graphite): ...new function.
13608 (pass_graphite_transforms): Convert from a global struct to a subclass
13609 of gimple_opt_pass along with...
13610 (pass_data_graphite_transforms): ...new pass_data instance and...
13611 (make_pass_graphite_transforms): ...new function.
13612 (pass_check_data_deps): Convert from a global struct to a subclass of
13613 gimple_opt_pass along with...
13614 (pass_data_check_data_deps): ...new pass_data instance and...
13615 (make_pass_check_data_deps): ...new function.
13616 (pass_iv_canon): Convert from a global struct to a subclass of
13617 gimple_opt_pass along with...
13618 (pass_data_iv_canon): ...new pass_data instance and...
13619 (make_pass_iv_canon): ...new function.
13620 (pass_scev_cprop): Convert from a global struct to a subclass of
13621 gimple_opt_pass along with...
13622 (pass_data_scev_cprop): ...new pass_data instance and...
13623 (make_pass_scev_cprop): ...new function.
13624 (pass_record_bounds): Convert from a global struct to a subclass of
13625 gimple_opt_pass along with...
13626 (pass_data_record_bounds): ...new pass_data instance and...
13627 (make_pass_record_bounds): ...new function.
13628 (pass_complete_unroll): Convert from a global struct to a subclass of
13629 gimple_opt_pass along with...
13630 (pass_data_complete_unroll): ...new pass_data instance and...
13631 (make_pass_complete_unroll): ...new function.
13632 (pass_complete_unrolli): Convert from a global struct to a subclass of
13633 gimple_opt_pass along with...
13634 (pass_data_complete_unrolli): ...new pass_data instance and...
13635 (make_pass_complete_unrolli): ...new function.
13636 (pass_parallelize_loops): Convert from a global struct to a subclass
13637 of gimple_opt_pass along with...
13638 (pass_data_parallelize_loops): ...new pass_data instance and...
13639 (make_pass_parallelize_loops): ...new function.
13640 (pass_loop_prefetch): Convert from a global struct to a subclass of
13641 gimple_opt_pass along with...
13642 (pass_data_loop_prefetch): ...new pass_data instance and...
13643 (make_pass_loop_prefetch): ...new function.
13644 (pass_iv_optimize): Convert from a global struct to a subclass of
13645 gimple_opt_pass along with...
13646 (pass_data_iv_optimize): ...new pass_data instance and...
13647 (make_pass_iv_optimize): ...new function.
13648 (pass_tree_loop_done): Convert from a global struct to a subclass of
13649 gimple_opt_pass along with...
13650 (pass_data_tree_loop_done): ...new pass_data instance and...
13651 (make_pass_tree_loop_done): ...new function.
13652 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
13653 struct to a subclass of gimple_opt_pass along with...
13654 (pass_data_cse_reciprocals): ...new pass_data instance and...
13655 (make_pass_cse_reciprocals): ...new function.
13656 (pass_cse_sincos): Convert from a global struct to a subclass of
13657 gimple_opt_pass along with...
13658 (pass_data_cse_sincos): ...new pass_data instance and...
13659 (make_pass_cse_sincos): ...new function.
13660 (pass_optimize_bswap): Convert from a global struct to a subclass of
13661 gimple_opt_pass along with...
13662 (pass_data_optimize_bswap): ...new pass_data instance and...
13663 (make_pass_optimize_bswap): ...new function.
13664 (pass_optimize_widening_mul): Convert from a global struct to a
13665 subclass of gimple_opt_pass along with...
13666 (pass_data_optimize_widening_mul): ...new pass_data instance and...
13667 (make_pass_optimize_widening_mul): ...new function.
13668 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
13669 subclass of gimple_opt_pass along with...
13670 (pass_data_phiopt): ...new pass_data instance and...
13671 (make_pass_phiopt): ...new function.
13672 (pass_cselim): Convert from a global struct to a subclass of
13673 gimple_opt_pass along with...
13674 (pass_data_cselim): ...new pass_data instance and...
13675 (make_pass_cselim): ...new function.
13676 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
13677 subclass of gimple_opt_pass along with...
13678 (pass_data_phiprop): ...new pass_data instance and...
13679 (make_pass_phiprop): ...new function.
13680 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
13681 subclass of gimple_opt_pass along with...
13682 (pass_data_pre): ...new pass_data instance and...
13683 (make_pass_pre): ...new function.
13684 (pass_fre): Convert from a global struct to a subclass of
13685 gimple_opt_pass along with...
13686 (pass_data_fre): ...new pass_data instance and...
13687 (make_pass_fre): ...new function.
13688 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
13689 subclass of gimple_opt_pass along with...
13690 (pass_data_reassoc): ...new pass_data instance and...
13691 (make_pass_reassoc): ...new function.
13692 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
13693 subclass of gimple_opt_pass along with...
13694 (pass_data_sink_code): ...new pass_data instance and...
13695 (make_pass_sink_code): ...new function.
13696 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
13697 subclass of gimple_opt_pass along with...
13698 (pass_data_strlen): ...new pass_data instance and...
13699 (make_pass_strlen): ...new function.
13700 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
13701 struct to a subclass of gimple_opt_pass along with...
13702 (pass_data_build_alias): ...new pass_data instance and...
13703 (make_pass_build_alias): ...new function.
13704 (pass_build_ealias): Convert from a global struct to a subclass of
13705 gimple_opt_pass along with...
13706 (pass_data_build_ealias): ...new pass_data instance and...
13707 (make_pass_build_ealias): ...new function.
13708 (pass_ipa_pta): Convert from a global struct to a subclass of
13709 simple_ipa_opt_pass along with...
13710 (pass_data_ipa_pta): ...new pass_data instance and...
13711 (make_pass_ipa_pta): ...new function.
13712 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
13713 subclass of gimple_opt_pass along with...
13714 (pass_data_uncprop): ...new pass_data instance and...
13715 (make_pass_uncprop): ...new function.
13716 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
13717 global struct to a subclass of gimple_opt_pass along with...
13718 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
13719 (make_pass_late_warn_uninitialized): ...new function.
13720 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
13721 to a subclass of gimple_opt_pass along with...
13722 (pass_data_init_datastructures): ...new pass_data instance and...
13723 (make_pass_init_datastructures): ...new function.
13724 (pass_early_warn_uninitialized): Convert from a global struct to a
13725 subclass of gimple_opt_pass along with...
13726 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
13727 (make_pass_early_warn_uninitialized): ...new function.
13728 (pass_update_address_taken): Convert from a global struct to a
13729 subclass of gimple_opt_pass along with...
13730 (pass_data_update_address_taken): ...new pass_data instance and...
13731 (make_pass_update_address_taken): ...new function.
13732 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
13733 struct to a subclass of gimple_opt_pass along with...
13734 (pass_data_release_ssa_names): ...new pass_data instance and...
13735 (make_pass_release_ssa_names): ...new function.
13736 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
13737 subclass of gimple_opt_pass along with...
13738 (pass_data_stdarg): ...new pass_data instance and...
13739 (make_pass_stdarg): ...new function.
13740 * tree-switch-conversion.c (pass_convert_switch): Convert from a
13741 global struct to a subclass of gimple_opt_pass along with...
13742 (pass_data_convert_switch): ...new pass_data instance and...
13743 (make_pass_convert_switch): ...new function.
13744 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
13745 to a subclass of gimple_opt_pass along with...
13746 (pass_data_tail_recursion): ...new pass_data instance and...
13747 (make_pass_tail_recursion): ...new function.
13748 (pass_tail_calls): Convert from a global struct to a subclass of
13749 gimple_opt_pass along with...
13750 (pass_data_tail_calls): ...new pass_data instance and...
13751 (make_pass_tail_calls): ...new function.
13752 * tree-vect-generic.c (pass_lower_vector): Convert from a global
13753 struct to a subclass of gimple_opt_pass along with...
13754 (pass_data_lower_vector): ...new pass_data instance and...
13755 (make_pass_lower_vector): ...new function.
13756 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
13757 gimple_opt_pass along with...
13758 (pass_data_lower_vector_ssa): ...new pass_data instance and...
13759 (make_pass_lower_vector_ssa): ...new function.
13760 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
13761 to a subclass of gimple_opt_pass along with...
13762 (pass_data_slp_vectorize): ...new pass_data instance and...
13763 (make_pass_slp_vectorize): ...new function.
13764 (pass_ipa_increase_alignment): Convert from a global struct to a
13765 subclass of simple_ipa_opt_pass along with...
13766 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
13767 (make_pass_ipa_increase_alignment): ...new function.
13768 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
13769 gimple_opt_pass along with...
13770 (pass_data_vrp): ...new pass_data instance and...
13771 (make_pass_vrp): ...new function.
13772 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
13773 subclass of simple_ipa_opt_pass along with...
13774 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
13775 (make_pass_ipa_free_lang_data): ...new function.
13776 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
13777 gimple_opt_pass along with...
13778 (pass_data_tsan): ...new pass_data instance and...
13779 (make_pass_tsan): ...new function.
13780 (pass_tsan_O0): Convert from a global struct to a subclass of
13781 gimple_opt_pass along with...
13782 (pass_data_tsan_O0): ...new pass_data instance and...
13783 (make_pass_tsan_O0): ...new function.
13784 * var-tracking.c (pass_variable_tracking): Convert from a global
13785 struct to a subclass of rtl_opt_pass along with...
13786 (pass_data_variable_tracking): ...new pass_data instance and...
13787 (make_pass_variable_tracking): ...new function.
13788 * web.c (pass_web): Convert from a global struct to a subclass of
13789 rtl_opt_pass along with...
13790 (pass_data_web): ...new pass_data instance and...
13791 (make_pass_web): ...new function.
13792 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
13793 declaration with that of...
13794 (make_pass_mode_switch_use): ...new function.
13795 (pass_resolve_sw_modes): Replace declaration with that of...
13796 (make_pass_resolve_sw_modes): ...new function.
13797 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
13798 from a global struct to a subclass of rtl_opt_pass along with...
13799 (pass_data_mode_switch_use): ...new pass_data instance and...
13800 (make_pass_mode_switch_use): ...new function.
13801 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
13802 from a global struct to a subclass of rtl_opt_pass along with...
13803 (pass_data_resolve_sw_modes): ...new pass_data instance and...
13804 (make_pass_resolve_sw_modes): ...new function.
13805 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
13806 struct to a subclass of rtl_opt_pass along with...
13807 (pass_data_insert_vzeroupper): ...new pass_data instance and...
13808 (make_pass_insert_vzeroupper): ...new function.
13809 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
13810 global struct to a subclass of rtl_opt_pass along with...
13811 (pass_data_work_around_errata): ...new pass_data instance and...
13812 (make_pass_work_around_errata): ...new function.
13813 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
13814 struct to a subclass of rtl_opt_pass along with...
13815 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
13816 (make_pass_mips_machine_reorg2): ...new function.
13818 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13820 * passes.c (pass_manager::operator new): New.
13822 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13824 Handwritten part of conversion of passes to C++ classes.
13826 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
13827 (toplev.o): Add dep on PASS_MANAGER_H.
13828 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
13829 of early local pases to reflect this moving from a global to a
13830 member of gcc::pass_manager.
13831 (cgraph_add_new_function): Likewise.
13832 * lto-cgraph.c (lto_output_node): Update for conversion of
13833 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
13834 * passes.c (opt_pass::clone): New.
13835 (opt_pass::gate): New.
13836 (opt_pass::execute): New.
13837 (opt_pass::opt_pass): New.
13838 (pass_manager::execute_early_local_passes): New.
13839 (pass_manager::execute_pass_mode_switching): new.
13840 (finish_optimization_passes): Convert to...
13841 (pass_manager::finish_optimization_passes): ...this.
13842 (finish_optimization_passes): Update for conversion of passes to
13844 (register_dump_files_1): Use has_gate since we cannot portably
13845 check a vtable entry against NULL.
13846 (dump_one_pass): Likewise.
13847 (ipa_write_summaries_2): Likewise.
13848 (ipa_write_optimization_summaries_1): Likewise.
13849 (ipa_read_summaries_1): Likewise.
13850 (ipa_read_optimization_summaries_1): Likewise.
13851 (execute_ipa_stmt_fixups): Likewise.
13852 (pass_manager::pass_manager): Rewrite pass-creation, invoking
13853 pass-creation functions rather than wiring up globals, and
13854 storing the results in fields of pass_manager generated using
13855 pass-instances.def.
13856 (pass_manager::dump_profile_report): Update for conversion of
13857 passes to C++ classes.
13858 (pass_manager::execute_ipa_summary_passes): Likewise.
13859 (execute_one_ipa_transform_pass): Likewise.
13860 (execute_one_pass): Use has_gate and has_execute since we cannot
13861 portably check a vtable entry against NULL.
13862 * pass_manager.h (pass_manager::finish_optimization_passes): New.
13863 (pass_manager): Use pass-instances.def to add fields for the
13864 various pass instances.
13865 * toplev.c (finalize): Update for move of
13866 finish_optimization_passes to a method of gcc::pass_manager.
13867 * toplev.h (finish_optimization_passes): Move to method of class
13869 * tree-pass.h (struct pass_data): New.
13870 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
13871 (opt_pass::gate): Convert to virtual function.
13872 (opt_pass::~opt_pass): New.
13873 (opt_pass::clone): New.
13874 (opt_pass::execute): Convert to virtual function.
13875 (opt_pass::opt_pass): New.
13876 (opt_pass::ctxt_): new.
13877 (gimple_opt_pass): Convert to subclass of opt_pass.
13878 (gimple_opt_pass::gimple_opt_pass): New.
13879 (rtl_opt_pass): Convert to subclass of opt_pass.
13880 (rtl_opt_pass::rtl_opt_pass): New.
13881 (ipa_opt_pass_d): Convert to subclass of opt_pass.
13882 (ipa_opt_pass_d::ipa_opt_pass_d): New.
13883 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
13884 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
13885 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
13886 invocation of pass_mode_switching to reflect this moving from a
13887 global to a member of gcc::pass_manager.
13888 (ix86_option_override): Rework how pass_insert_vzeroupper is
13889 added to the pass_manager to reflect autogenerated changes.
13890 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
13892 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
13894 PR rtl-optimization/57708
13895 * recog.c (peep2_find_free_register): Validate all regs in a
13898 2013-08-05 Jan Hubicka <jh@suse.cz>
13901 * cgraph.c (verify_cgraph_node): Accept local flags from other
13903 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
13904 (function_and_variable_visibility): Likewise.
13905 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
13907 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13909 * graph.c (init_graph_slim_pretty_print): Remove.
13910 (print_graph_cfg): Do not call it. Use local pretty printer.
13911 (start_graph_dump): Likewise.
13913 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13915 * gimple-pretty-print.c (buffer): Remove.
13916 (initialized): Likewise.
13917 (maybe_init_pretty_print): Likewise.
13918 (print_gimple_stmt): Do not call it. Use non-static local
13919 pretty_printer variable.
13920 (print_gimple_expr): Likewise.
13921 (print_gimple_seq): Likewise.
13922 (gimple_dump_bb): Likewise.
13924 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13926 * asan.c (asan_pp): Remove.
13927 (asan_pp_initialized): Likewise.
13928 (asan_pp_initialize): Likewise.
13929 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
13930 (asan_emit_stack_protection): Tidy. Use local pretty printer.
13931 (asan_add_global): Likewise.
13933 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
13935 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
13936 * diagnostic.h (diagnostic_flush_buffer): Adjust.
13937 * pretty-print.c (pp_formatted_text_data): Likewise.
13938 (pp_indent): Rename from pp_base_indent.
13939 (pp_format): Rename from pp_base_format.
13940 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
13941 (pp_format_verbatim): Rename from pp_base_format_verbatim.
13942 (pp_flush): Rename from pp_base_flush.
13943 (pp_set_line_maximum_length): Rename from
13944 pp_base_set_line_maximum_length.
13945 (pp_clear_output_area): Rename from pp_base_clear_output_area.
13946 (pp_set_prefix): Rename from pp_base_set_prefix.
13947 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
13948 (pp_emit_prefix): Rename from pp_base_emit_prefix.
13949 (pp_append_text): Rename from pp_base_append_text.
13950 (pp_formatted_text): Rename from pp_base_formatted_text.
13951 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
13952 (pp_remaining_character_count_for_line): Rename from
13953 pp_base_remaining_character_count_for_line.
13954 (pp_newline): Rename from pp_base_newline.
13955 (pp_character): Rename from pp_base_character.
13956 (pp_string): Rename from pp_base_string.
13957 (pp_maybe_space): Rename from pp_base_maybe_space.
13958 * asan.c (asan_pp_string): Adjust.
13959 (asan_emit_stack_protection): Likewise.
13960 (asan_add_global): Likewise.
13961 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
13962 * tree-mudflap.c (mf_varname_tree): Likewise.
13963 * tree-pretty-print.c (pp_tree_identifier): Rename from
13964 pp_base_tree_identifier.
13965 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
13966 Declare as function.
13968 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13970 * pretty-print.h (pp_bar_bar): New.
13971 (pp_ampersand_ampersand): Likewise.
13972 (pp_less_equal): Likewise.
13973 (pp_greater_equal): Likewise.
13974 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
13975 printer functions instead of pp_string or operators and punctuators.
13976 (dump_gimple_call): Likewise.
13977 (dump_gimple_omp_for): Likewise.
13978 (dump_gimple_transaction): Likewise.
13979 (dump_gimple_phi): Likewise.
13980 (pp_gimple_stmt_1): Likewise.
13981 * sched-vis.c (print_insn): Likewise.
13982 * tree-mudflap.c (mf_varname_tree): Likewise.
13983 * tree-pretty-print.c (dump_block_node): Likewise.
13984 (dump_generic_node): Likewise.
13986 2013-08-02 Jan Hubicka <jh@suse.cz>
13988 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
13990 * lto-streamer-out.c (tree_is_indexable): Results decls and
13991 parm decls are not indexable.
13992 (DFS_write_tree_body): Do not follow args and results.
13993 (hash_tree): Likewise.
13994 (output_functions): Rearrange so struct function is needed
13995 only when real body is output; be able to also ouptut abstract
13996 functions; output DECL_ARGUMENTS and DECL_RESULT.
13997 (lto_output): When not in WPA, ale store abstract functions.
13998 (write_symbol): Do not care about RESULT_DECL.
13999 (output_symbol_p): Handle correctly sbtract decls.
14000 * lto-streamer-in.c (input_function): Rearrange so struct
14001 function can be NULL at entry; allow streaming of
14002 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
14003 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
14004 sanity check during LTO.
14005 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
14006 RESULT_DECl and DECL_ARGUMENTS.
14007 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
14010 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14012 * pretty-print.h (pp_underscore): New.
14014 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
14015 printer functions instead of pp_character.
14016 (dump_binary_rhs): Likewise.
14017 (dump_ternary_rhs): Likewise.
14018 (dump_gimple_call_args): Likewise.
14019 (pp_points_to_solution): Likewise.
14020 (dump_gimple_call): Likewise.
14021 (dump_gimple_switch): Likewise.
14022 (dump_gimple_cond): Likewise.
14023 (dump_gimple_bind): Likewise.
14024 (dump_gimple_try): Likewise.
14025 (dump_gimple_omp_for): Likewise.
14026 (dump_gimple_omp_continue): Likewise.
14027 (dump_gimple_omp_single): Likewise.
14028 (dump_gimple_omp_sections): Likewise.
14029 (dump_gimple_omp_block): Likewise.
14030 (dump_gimple_omp_critical): Likewise.
14031 (dump_gimple_transaction): Likewise.
14032 (dump_gimple_asm): Likewise.
14033 (dump_gimple_phi): Likewise.
14034 (dump_gimple_omp_parallel): Likewise.
14035 (dump_gimple_omp_task): Likewise.
14036 (dump_gimple_omp_atomic_load): Likewise.
14037 (dump_gimple_omp_atomic_store): Likewise.
14038 (dump_gimple_mem_ops): Likewise.
14039 (pp_gimple_stmt_1): Likewise.
14040 (pp_cfg_jump): Likewise.
14041 (dump_implicit_edges): Likewise.
14042 (gimple_dump_bb_for_graph): Likewise.
14043 * graph.c (draw_cfg_node): Likewise.
14044 * langhooks.c (lhd_print_error_function): Likewise.
14045 * sched-vis.c (print_exp): Likewise.
14046 (print_value): Likewise.
14047 (print_pattern): Likewise.
14048 (print_insn): Likewise.
14049 (rtl_dump_bb_for_graph): Likewise.
14050 * tree-pretty-print.c (dump_function_declaration): Likewise.
14051 (dump_array_domain): Likewise.
14052 (dump_omp_clause): Likewise.
14053 (dump_location): Likewise.
14054 (dump_generic_node): Likewise.
14055 (print_struct_decl): Likewise.
14056 * diagnostic.c (diagnostic_show_locus): Use pp_space.
14058 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14060 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
14061 candidate table when replacing a candidate statement.
14062 (replace_rhs_if_not_dup): Likewise.
14063 (replace_one_candidate): Likewise.
14065 2013-08-02 Jan Hubicka <jh@suse.cz>
14066 Martin Liska <marxin.liska@gmail.com>
14068 * cgraphunit.c (add_new_function): Fix logic when adding from
14070 (assemble_thunk): Rename to ...
14071 (expand_thunk); .. this one; export; get it working with
14072 general functions; make produced gimple valid.
14073 * cgraph.h (expand_thunk): Declare.
14075 2013-08-02 Jan Hubicka <jh@suse.cz>
14077 * ipa-cp.c (gather_context_independent_values): Use
14078 ipa_get_param_move_cost.
14079 (get_replacement_map): Remove PARAM; move parameter folding
14081 (create_specialized_node): Update.
14082 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
14083 assert that we have gimple body; update move_cost.
14084 (count_formal_params): Assert that we have gimple body.
14085 (ipa_dump_param): New function.
14086 (ipa_alloc_node_params): Break out from ...
14087 (ipa_initialize_node_params): ... here.
14088 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
14089 (ipa_write_node_info): Stream move costs.
14090 (ipa_read_node_info): Read move costs.
14091 (ipa_update_after_lto_read): Do not recompute node params.
14092 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
14093 (ipa_get_param): Check we are not in WPA.
14094 (ipa_get_param_move_cost): New.
14095 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
14096 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
14097 parm numbers to be present.
14099 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14101 PR rtl-optimization/58048
14102 * lra-constraints.c (process_alt_operands): Don't check asm
14103 operand on register.
14105 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
14107 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
14108 the implied StoreLoad barrier for atomic operations if before.
14110 2013-08-02 Jan Hubicka <jh@suse.cz>
14111 Martin Liska <marxin.liska@gmail.com>
14113 * cgraph.c (cgraph_function_body_availability): Do not check
14115 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
14116 symtab_node_availability): Declare.
14117 * ipa.c (can_replace_by_local_alias): New.
14118 (function_and_variable_visibility): Use it.
14119 * symtab.c (symtab_for_node_and_aliases,
14120 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
14122 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14124 PR rtl-optimization/57963
14125 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
14126 (lra_constraints): Use them.
14128 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
14130 * config/arm/types.md (define_attr "type"): Add "load_acq"
14132 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
14134 (cortex_a53_store1): Likewise.
14136 2013-08-01 Jan Hubicka <jh@suse.cz>
14138 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
14139 partitions are not needed.
14141 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
14143 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
14144 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
14145 MAYBE_NON_Q_CLASS_P where appropriate.
14147 2013-08-01 Jan Hubicka <jh@suse.cz>
14149 * cgraph.h (release_function_body): Declare.
14150 * tree.c (free_lang_data_in_decl): Free, parameters and return values
14151 of unused delcarations.
14153 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14155 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
14156 RTL form when subtracting a constant.
14158 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14160 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
14161 Generate canonical plus rtx with negated immediate instead of minus
14163 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
14165 2013-08-01 Jan Hubicka <jh@suse.cz>
14167 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
14168 (cgraph_release_function_body): Likewise.
14169 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
14170 * cgraph.h (cgrpah_node): Rename abstract_and_needed
14171 to used_as_abstract_origin.
14172 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
14173 symbols used as abstract origins.
14174 * cgraphunit.c (analyze_functions): Update.
14175 * ipa.c (symtab_remove_unreachable_nodes): Recompute
14176 used_as_abstract_origin.
14177 * tree-inline.c (tree_function_versioning): Update
14178 used_as_abstract_origin; be ready for DECL_RESULT and
14179 DECL_ARGUMENTS to be NULL.
14181 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
14182 for abstract functions.
14183 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
14186 2013-08-01 Jan Hubicka <jh@suse.cz>
14188 * profile.c (compute_value_histograms): Fix thinko.
14190 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14192 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
14193 aarch-common-protos.h to extra_headers.
14194 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14195 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
14196 * config/aarch64/t-aarch64 (aarch-common.o): Define.
14198 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14200 * config/aarch64/aarch64.md (define_attr "type"): Delete.
14201 Include "../arm/types.md". Define "type" attribute for all patterns.
14202 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
14205 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
14207 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
14208 to support power8 load fusion.
14209 (fusion_gpr_mem_load): Likewise.
14211 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
14213 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
14214 declarations for power8 load fusion.
14215 (emit_fusion_gpr_load): Likewise.
14217 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
14218 tuning for power8, turn on fusion mode by default. Turn on sign
14219 extending fusion mode if normal fusion mode is on, and we are at
14221 (fusion_gpr_load_p): New function, return true if we can fuse an
14222 addis instruction with a dependent load to a GPR.
14223 (emit_fusion_gpr_load): Emit the instructions for power8 load
14226 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
14227 (VSX load fusion peepholes): New peepholes to fuse together an
14228 addi instruction with a VSX load instruction.
14230 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
14231 peepholes to fuse an addis instruction with a load to a GPR base
14232 register. If we are supporting sign extending fusions, convert
14233 sign extending loads to zero extending loads and add an explicit
14236 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14238 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
14239 aarch-common-protos.h to extra_headers.
14240 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14241 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
14242 (arm_early_store_addr_dep): Likewise.
14243 (arm_no_early_alu_shift_dep): Likewise.
14244 (arm_no_early_alu_shift_value_dep): Likewise.
14245 (arm_no_early_mul_dep): Likewise.
14246 (arm_no_early_store_addr_dep): Likewise.
14247 (arm_mac_accumulator_is_mul_result): Likewise.
14248 (arm_mac_accumulator_is_result): Likewise.
14249 * config/arm/aarch-common.c: ... here. New file.
14250 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
14252 (arm_early_store_addr_dep): Likewise.
14253 (arm_no_early_alu_shift_dep): Likewise.
14254 (arm_no_early_alu_shift_value_dep): Likewise.
14255 (arm_no_early_mul_dep): Likewise.
14256 (arm_no_early_store_addr_dep): Likewise.
14257 (arm_mac_accumulator_is_mul_result): Likewise.
14258 (arm_mac_accumulator_is_result): Likewise.
14259 * config/arm/aarch-common-protos.h: ... here. New file.
14260 * config/arm/t-arm (aarch-common.o): Define.
14262 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14264 * config/arm/arm.md: Include new file "types.md".
14265 (define_attr "type"): Move from here to ...
14266 (define_attr "mul32"): Likewise.
14267 (define_attr "mul64"): Likewise.
14268 * config/arm/types.md: ... here. New file.
14270 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
14272 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
14273 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
14275 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
14277 * gen-pass-instances.awk: Fix offset of substr().
14279 2013-07-31 David Malcolm <dmalcolm@redhat.com>
14281 * Makefile.in (pass-instances.def): New.
14282 (passes.o): Replace dependency on passes.def with one on
14285 * gen-pass-instances.awk: New.
14287 * passes.c (pass_manager::pass_manager): Use pass-instances.def
14288 rather than passes.def, updating local definition of NEXT_PASS
14289 macro to add an extra NUM parameter (currently unused).
14291 2013-07-30 David Malcolm <dmalcolm@redhat.com>
14293 * Makefile.in (PASS_MANAGER_H): New.
14294 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
14295 (passes.o): Likewise.
14296 (statistics.o): Likewise.
14297 (cgraphunit.o): Likewise.
14298 (context.o): Depend on PASS_MANAGER_H.
14300 * pass_manager.h: New.
14302 * cgraphunit.c (cgraph_add_new_function): Update for moves
14303 of globals to fields of pass_manager.
14304 (analyze_function): Likewise.
14305 (expand_function): Likewise.
14306 (ipa_passes): Likewise.
14307 (compile): Likewise.
14309 * context.c (context::context): New.
14310 * context.h (context::context): New.
14311 (context::get_passes): New.
14312 (context::passes_): New.
14314 * lto-cgraph.c (input_node): Update for moves of globals to
14315 fields of pass_manager.
14317 * passes.c (all_passes): Remove, in favor of a field of the
14318 same name within the new class pass_manager.
14319 (all_small_ipa_passes): Likewise.
14320 (all_lowering_passes): Likewise.
14321 (all_regular_ipa_passes): Likewise.
14322 (all_late_ipa_passes): Likewise.
14323 (all_lto_gen_passes): Likewise.
14324 (passes_by_id): Likewise.
14325 (passes_by_id_size): Likewise.
14326 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14327 the new class pass_manager.
14328 (set_pass_for_id): Convert to...
14329 (pass_manager::set_pass_for_id): ...method.
14330 (get_pass_for_id): Convert to...
14331 (pass_manager::get_pass_for_id): ...method.
14332 (register_one_dump_file): Move body of implementation into...
14333 (pass_manager::register_one_dump_file): ...here.
14334 (register_dump_files_1): Convert to...
14335 (pass_manager::register_dump_files_1): ...method.
14336 (register_dump_files): Convert to...
14337 (pass_manager::register_dump_files): ...method.
14338 (create_pass_tab): Update for moves of globals to fields of
14340 (dump_passes): Move body of implementation into...
14341 (pass_manager::dump_passes): ...here.
14342 (register_pass): Move body of implementation into...
14343 (pass_manager::register_pass): ...here.
14344 (init_optimization_passes): Convert into...
14345 (pass_manager::pass_manager): ...constructor for new
14346 pass_manager class, and initialize the pass_lists array.
14347 (check_profile_consistency): Update for moves of globals to
14348 fields of pass_manager.
14349 (dump_profile_report): Move body of implementation into...
14350 (pass_manager::dump_profile_report): ...here.
14351 (ipa_write_summaries_1): Update for moves of pass lists from
14352 being globals to fields of pass_manager.
14353 (ipa_write_optimization_summaries): Likewise.
14354 (ipa_read_summaries): Likewise.
14355 (ipa_read_optimization_summaries): Likewise.
14356 (execute_all_ipa_stmt_fixups): Likewise.
14358 * statistics.c (statistics_fini): Update for moves of globals to
14359 fields of pass_manager.
14361 * toplev.c (general_init): Replace call to
14362 init_optimization_passes with construction of the pass_manager
14365 * tree-pass.h (all_passes): Remove, in favor of a field of the
14366 same name within the new class pass_manager.
14367 (all_small_ipa_passes): Likewise.
14368 (all_lowering_passes): Likewise.
14369 (all_regular_ipa_passes): Likewise.
14370 (all_lto_gen_passes): Likewise.
14371 (all_late_ipa_passes): Likewise.
14372 (passes_by_id): Likewise.
14373 (passes_by_id_size): Likewise.
14374 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14375 the new class pass_manager.
14376 (get_pass_for_id): Remove.
14378 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14380 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
14383 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14385 * arm.md (mulhi3): New expand pattern.
14387 2013-07-30 Jan Hubicka <jh@suse.cz>
14388 Martin Liska <marxin.liska@gmail.com>
14390 * profile.c (compute_value_histograms): Do not ICE when
14391 there is mismatch only on some counters.
14393 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14395 PR rtl-optimization/57637
14396 * function.c (move_insn_for_shrink_wrap): Also check the
14397 GEN set of the LIVE problem for the liveness analysis
14398 if it exists, otherwise give up.
14400 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14402 PR tree-optimization/57993
14403 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
14404 replaced statement in the candidate table.
14405 (phi_add_costs): Return infinite cost when the hidden basis does
14406 not dominate all phis on which the candidate is dependent.
14407 (replace_one_candidate): Record replaced statement in the
14410 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
14412 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
14413 (ashlv2si3): New expander.
14414 (*ashlv2si3_i): New define_insn_and_split.
14415 * predicates.md (float_operation): Allow patterns with three
14416 basic sub-patterns.
14418 PR rtl-optimization/58021
14419 * mode-switching.c (create_pre_exit): Always split off preceding
14420 insns if we are not at the basic block head.
14422 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
14424 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
14425 (UCLIBC_DYNAMIC_LINKER): New macro.
14426 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
14428 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
14429 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
14431 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
14432 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
14433 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
14434 for SF and DF modes. Use ieee_quad_format for TF mode.
14435 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
14436 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
14437 (mips_option_override): Handle `-mnan=legacy'.
14438 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
14439 `-mabs=2008' and `-mnan=2008'.
14440 (OPTION_DEFAULT_SPECS): Add "nan" default.
14441 (ASM_SPEC): Handle `-mnan='.
14442 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
14443 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
14444 comment accordingly.
14445 (neg<mode>2): Likewise.
14446 * config/mips/mips.opt (mabs, mnan): New options.
14447 * doc/install.texi (Configuration): Document `--with-nan=' option.
14448 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
14450 (MIPS Options): Document them.
14451 * config.gcc <mips*-*-*>: Handle `--with-nan='.
14452 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
14453 * configure: Regenerate.
14454 * config.in: Regenerate.
14456 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14458 * config/i386/i386.md (float post-reload splitters): Do not check
14459 for subregs of SSE registers.
14461 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14462 H.J. Lu <hongjiu.lu@intel.com>
14466 * config/i386/i386.md (post-reload splitter
14467 to avoid partial SSE reg dependency stalls): New pattern.
14469 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
14471 * config/s390/s390.md ("movcc"): Swap load and store instructions.
14473 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
14475 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14476 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
14478 2013-07-26 Cary Coutant <ccoutant@google.com>
14480 * dwarf2out.c (die_checksum_ordered): Don't include template
14481 instantiations in signature.
14482 (is_template_parameter): New function.
14483 (is_template_instantiation): New function.
14484 (generate_skeleton_bottom_up): Don't include template instantiations
14486 (generate_skeleton): Likewise.
14487 (break_out_comdat_types): Move recursive call to break out nested
14489 (prune_unused_types_mark_generic_parms_dies): Call
14490 is_template_parameter.
14492 2013-07-26 Ian Bolton <ian.bolton@arm.com>
14494 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
14495 uses vector registers.
14496 * config/aarch64/iterators.md: Add attributes rtn and vas.
14498 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14499 Richard Earnshaw <richard.earnshaw@arm.com>
14501 * combine.c (simplify_comparison): Re-canonicalize operands
14503 * config/arm/arm.md (movcond_addsi): New splitter.
14505 2013-07-25 Sterling Augustine <saugustine@google.com>
14507 * dwarf2out.c (size_of_pubnames): Move code to...
14508 (include_pubname_in_output): ...here. New.
14509 (want_pubnames): Rearrange.
14510 (output_pubnames): Call include_pubname_in_output. Move assertion.
14512 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14514 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
14516 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14519 * doc/extend.texi: Remove obsolete builtins. Fix
14520 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
14522 2013-07-25 Jan Hubicka <jh@suse.cz>
14524 * cgraph.c (release_function_body): Break out from ...
14525 (cgraph_release_function_body): ... this one; also release DECL_RESULT
14526 and DECL_ARGUMENTS.
14527 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
14528 old_tree in the map.
14529 (create_specialized_node): Update.
14530 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
14532 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
14533 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
14534 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
14535 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
14538 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14540 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
14541 addsi3_carryin_alt2_<optab>): Correct output template.
14543 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14545 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
14546 Adjust for arm_restrict_it.
14547 Remove trailing whitespace.
14549 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
14551 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
14552 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
14554 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
14556 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
14558 PR rtl-optimization/57960
14559 * lra-constraints.c (process_alt_operands): Use the right mode
14560 when checking strict_low.
14562 2013-07-25 Jan Hubicka <jh@suse.cz>
14564 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
14565 * cgraph.c (cgraph_remove_node): Do not release function body
14566 when in cgraph streaming.
14567 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
14568 in other partitions are not considered reachable; fix handling of
14571 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14573 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
14575 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14580 * config/arm/arm.md ("*sibcall_insn): Replace use of
14581 Ss with US. Adjust output for v5 and v4t.
14582 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
14584 * config/arm/constraints.md ("Ss"): Rename to US.
14586 2013-07-25 Terry Guo <terry.guo@arm.com>
14588 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
14589 shift_add/shift_sub0/shift_sub1 RTXs.
14591 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
14592 Anton Blanchard <anton@au1.ibm.com>
14594 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
14595 (altivec_vpks<VI_char>ss): Likewise.
14596 (altivec_vpks<VI_char>us): Likewise.
14597 (altivec_vpku<VI_char>us): Likewise.
14598 (altivec_vpku<VI_char>um): Likewise.
14600 2013-07-24 David Malcolm <dmalcolm@redhat.com>
14602 Introduce context class.
14604 * Makefile.in (CONTEXT_H): New.
14605 (OBJS): Add context.o.
14606 (toplev.o): Add CONTEXT_H to dependencies.
14609 * toplev.c (general_init): Create the singleton gcc::context instance.
14615 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
14617 PR rtl-optimization/57968
14618 * mode-switching.c (create_pre_exit): Allow instructions that
14619 don't set a return register to need a non-exit mode.
14621 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14622 Anton Blanchard <anton@au1.ibm.com>
14624 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
14625 operands to vperm for little endian.
14626 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
14627 of lvsl to create the control mask for a vperm for little endian.
14629 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14630 Anton Blanchard <anton@au1.ibm.com>
14632 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
14633 two operands for little-endian.
14635 2013-07-23 Steve Ellcey <sellcey@mips.com>
14637 * config/mips/mips.c (mips_case_values_threshold): New.
14638 (TARGET_CASE_VALUES_THRESHOLD): Define.
14640 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14641 Anton Blanchard <anton@au1.ibm.com>
14643 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
14644 selection of field for vector splat in little endian mode.
14646 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
14648 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
14649 expanders to rs6000.md.
14650 (ior<mode>3): Likewise.
14651 (and<mode>3): Likewise.
14652 (one_cmpl<mode>2): Likewise.
14653 (nor<mode>3): Likewise.
14654 (andc<mode>3): Likewise.
14655 (eqv<mode>3): Likewise.
14656 (nand<mode>3): Likewise.
14657 (orc<mode>3): Likewise.
14659 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
14662 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
14663 to split multi-word logical operations.
14664 (rs6000_split_logical_di): Likewise.
14665 (rs6000_split_logical): Likewise.
14667 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
14668 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
14669 and allow TImode operations in 32-bit.
14670 (vsx_and<mode>3_64bit): Likewise.
14671 (vsx_ior<mode>3_32bit): Likewise.
14672 (vsx_ior<mode>3_64bit): Likewise.
14673 (vsx_xor<mode>3_32bit): Likewise.
14674 (vsx_xor<mode>3_64bit): Likewise.
14675 (vsx_one_cmpl<mode>2_32bit): Likewise.
14676 (vsx_one_cmpl<mode>2_64bit): Likewise.
14677 (vsx_nor<mode>3_32bit): Likewise.
14678 (vsx_nor<mode>3_64bit): Likewise.
14679 (vsx_andc<mode>3_32bit): Likewise.
14680 (vsx_andc<mode>3_64bit): Likewise.
14681 (vsx_eqv<mode>3_32bit): Likewise.
14682 (vsx_eqv<mode>3_64bit): Likewise.
14683 (vsx_nand<mode>3_32bit): Likewise.
14684 (vsx_nand<mode>3_64bit): Likewise.
14685 (vsx_orc<mode>3_32bit): Likewise.
14686 (vsx_orc<mode>3_64bit): Likewise.
14688 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
14689 logical types in GPRs.
14691 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
14692 logical insns to rs6000.md, and allow TImode operations in
14694 (altivec_ior<mode>3): Likewise.
14695 (altivec_xor<mode>3): Likewise.
14696 (altivec_one_cmpl<mode>2): Likewise.
14697 (altivec_nor<mode>3): Likewise.
14698 (altivec_andc<mode>3): Likewise.
14700 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
14701 attributes for moving the 128-bit logical operations into
14703 (BOOL_REGS_OUTPUT): Likewise.
14704 (BOOL_REGS_OP1): Likewise.
14705 (BOOL_REGS_OP2): Likewise.
14706 (BOOL_REGS_UNARY): Likewise.
14707 (BOOL_REGS_AND_CR0): Likewise.
14708 (one_cmpl<mode>2): Add support for DI logical operations on
14709 32-bit, splitting the operations to 32-bit.
14710 (anddi3): Likewise.
14711 (iordi3): Likewise.
14712 (xordi3): Likewise.
14713 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
14714 changes to combine the 32/64-bit code, allow logical operations on
14715 TI mode in 32-bit, and to use similar match_operator patterns like
14716 scalar mode uses. Combine the Altivec and VSX code for logical
14717 operations, and move it here.
14718 (ior<mode>3, 128-bit types): Likewise.
14719 (xor<mode>3, 128-bit types): Likewise.
14720 (one_cmpl<mode>3, 128-bit types): Likewise.
14721 (nor<mode>3, 128-bit types): Likewise.
14722 (andc<mode>3, 128-bit types): Likewise.
14723 (eqv<mode>3, 128-bit types): Likewise.
14724 (nand<mode>3, 128-bit types): Likewise.
14725 (orc<mode>3, 128-bit types): Likewise.
14726 (and<mode>3_internal): Likewise.
14727 (bool<mode>3_internal): Likewise.
14728 (boolc<mode>3_internal1): Likewise.
14729 (boolc<mode>3_internal2): Likewise.
14730 (boolcc<mode>3_internal1): Likewise.
14731 (boolcc<mode>3_internal2): Likewise.
14732 (eqv<mode>3_internal1): Likewise.
14733 (eqv<mode>3_internal2): Likewise.
14734 (one_cmpl1<mode>3_internal): Likewise.
14736 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14738 * config/microblaze/microblaze.c (microblaze_expand_prologue):
14739 Rename flag_stack_usage to flag_stack_usage_info.
14741 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14743 * config/microblaze/sync.md: New file.
14744 * config/microblaze/microblaze.md: Include sync.md
14745 * config/microblaze/microblaze.c: Add print_operand 'y'.
14746 * config/microblaze/constraints.md: Add memory_contraint
14747 'Q' which is a single register.
14749 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
14751 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
14753 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14755 * reload.c (find_reloads): Exit loop once we find this operand
14756 cannot be reloaded somehow for this alternative.
14758 * reload.c (find_reloads): Exit loop once we find a hard register.
14760 * rtlanal.c (computed_jump_p): Exit loop once we find label
14763 * i386.c (ix86_pad_returns): Exit loop after setting replace.
14765 * cfgloopmanip.c (remove_path): Exit loop after setting
14768 * gensupport.c (subst_dup): Avoid loop if code is not
14769 MATCH_DUP nor MATCH_OP_DUP.
14771 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
14773 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
14775 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14777 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
14778 true for SP_REGNUM if mode == ptr_mode.
14779 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
14780 with value R0_REGNUM + 31.
14782 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14784 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
14785 pad pointer-typed argument downward.
14787 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14789 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
14790 and __ILP32__ when the ILP32 model is in use.
14792 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14794 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
14795 (aarch64_load_symref_appropriately): In the case of
14796 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
14797 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
14798 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
14799 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
14800 if 'mode' doesn't equal to 'ptr_mode'.
14801 (aarch64_output_mi_thunk): Add an assertion on the alignment of
14802 'vcall_offset'; change to call aarch64_emit_move differently depending
14803 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
14804 to calculate the upper bound of 'vcall_offset'.
14805 (aarch64_cannot_force_const_mem): Change to also return true if
14807 (aarch64_legitimize_reload_address): In the case of large
14808 displacements, add new local variable 'xmode' and an assertion
14809 based on it; change to use 'xmode' to generate the new rtx and
14811 (aarch64_asm_trampoline_template): Change to generate the template
14812 differently depending on TARGET_ILP32 or not; change to use
14813 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
14814 (aarch64_trampoline_size): Removed.
14815 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
14816 and replace immediate literals with it. Change to use 'ptr_mode'
14817 instead of 'DImode' and call convert_memory_address if the mode
14818 of 'fnaddr' doesn't equal to 'ptr_mode'.
14819 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
14821 (aarch64_elf_asm_destructor): Likewise.
14822 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
14823 on TARGET_ILP32 instead of aarch64_trampoline_size.
14824 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
14825 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
14826 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
14827 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
14828 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
14829 (storewb_pair<GPI:mode>_<P:mode>): ... this.
14830 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
14831 depending on the value of 'mode'.
14832 (add_losym_<mode>): New.
14833 (ldr_got_small_<mode>): New, based on ldr_got_small.
14834 (ldr_got_small): Remove.
14835 (ldr_got_small_sidi): New.
14836 * config/aarch64/iterators.md (P): New.
14837 (PTR): Change to 'ptr_mode' in the condition.
14839 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14841 * config.gcc (aarch64*-*-*): Support --with-abi.
14842 (aarch64*-*-elf): Support --with-multilib-list.
14843 (aarch64*-*-linux*): Likewise.
14844 (supported_defaults): Add abi to aarch64*-*-*.
14845 * configure.ac: Mention AArch64 for --with-multilib-list.
14846 * configure: Re-generated.
14847 * config/aarch64/biarchilp32.h: New file.
14848 * config/aarch64/biarchlp64.h: New file.
14849 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
14851 (MULTILIB_DEFAULTS): Ditto.
14852 (DRIVER_SELF_SPECS): Ditto.
14853 (ASM_SPEC): Update to also substitute -mabi.
14854 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
14855 file whose name depends on -mabi= and -mbig-endian.
14856 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
14858 (POINTER_SIZE): New define.
14859 (POINTERS_EXTEND_UNSIGNED): Ditto.
14860 (enum aarch64_abi_type): New enumeration tag.
14861 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
14862 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
14863 (TARGET_ILP32): New define.
14864 * config/aarch64/aarch64.opt (mabi): New.
14865 (aarch64_abi): New.
14866 (ilp32, lp64): New values for -mabi.
14867 * config/aarch64/t-aarch64 (comma): New define.
14868 (MULTILIB_OPTIONS): Ditto.
14869 (MULTILIB_DIRNAMES): Ditto.
14870 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
14871 * doc/invoke.texi: Document -mabi for AArch64.
14873 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
14875 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
14877 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14878 Anton Blanchard <anton@au1.ibm.com>
14880 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
14881 endianness when selecting field to splat.
14883 2013-07-22 Eric Christopher <echristo@gmail.com>
14885 * dwarf2out.c (die_odr_checksum): New function to use
14886 CHECKSUM_ macros and ULEB128 for DIE tag.
14887 (generate_type_signature): Use.
14889 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
14891 * config.gcc (sparc*-*-*): Accept leon3 processor.
14892 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
14893 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
14894 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
14895 * config/sparc/sparc.opt (enum processor_type): Add leon3.
14896 (mfix-ut699): Adjust comment.
14897 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
14898 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
14899 (CPP_CPU_SPEC): Likewise.
14900 (ASM_CPU_SPEC): Likewise.
14901 * config/sparc/sparc.c (leon3_cost): New constant.
14902 (sparc_option_override): Add leon3 support.
14903 (mem_ref): New function.
14904 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
14905 (sparc_do_work_around_errata): Look into the instruction in the delay
14906 slot and adjust accordingly. Add fix for the data cache nullify issues
14907 of the UT699. Change insertion position for the NOP.
14908 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
14909 (leon3_load): New reservation.
14910 (leon_store): Bump latency to 2.
14911 (grfpu): New automaton.
14912 (grfpu_alu): New unit.
14913 (grfpu_ds): Likewise.
14914 (leon_fp_alu): Adjust.
14915 (leon_fp_mult): Delete.
14916 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
14917 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
14918 * config/sparc/sparc.md (cpu): Add leon3.
14919 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
14920 (swapsi): Likewise.
14921 (atomic_test_and_set): Likewise.
14922 (ldstub): Likewise.
14924 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
14926 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
14927 default for R5900 targets.
14928 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
14929 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
14930 * config/mips/mips.c (mips_option_override): Report an error for
14931 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
14932 for -march=r5900 -mhard-float.
14934 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14936 * df-problems.c (can_move_insns_across): Exit loop once we
14937 find a non-fixed, non-global register.
14939 * ipa-pure-const.c (propagate_nothrow): Exit loop after
14942 * omega.c (omega_eliminate_red): Break after setting red_found.
14943 (omega_problem_has_red_equations): Similarly after setting found.
14944 (omega_query_variable): Similarly after setting coupled.
14946 2013-07-22 Marek Polacek <polacek@redhat.com>
14948 * gimplify.c: Don't include gimple.h twice.
14950 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14952 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
14953 instead of TARGET_THUMB1.
14954 (Pz): New constraint.
14955 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
14957 (compare_negsi_si): Likewise.
14958 (compare_addsi2_op0): Likewise.
14959 (compare_addsi2_op1): Likewise.
14960 (addsi3_carryin_<optab>): Likewise.
14961 (addsi3_carryin_alt2_<optab>): Likewise.
14962 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
14963 for arm_restrict_it.
14964 (subsi3_carryin): Likewise.
14965 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
14966 (minmax_arithsi): Disable for arm_restrict_it.
14967 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
14968 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
14969 (satsi_<SAT:code>_shift): Likewise.
14970 (arm_shiftsi3): Add alternative for 16-bit encoding.
14971 (arm32_movhf): Disable for arm_restrict_it.
14972 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
14973 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
14975 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
14977 * config/arm/arm.md (attribute "insn"): Delete.
14978 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
14979 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
14980 (not_shiftsi): Update for attribute change.
14981 (not_shiftsi_compare0): Likewise.
14982 (not_shiftsi_compare0_scratch): Likewise.
14983 (arm_one_cmplsi2): Likewise.
14984 (thumb1_one_cmplsi2): Likewise.
14985 (notsi_compare0): Likewise.
14986 (notsi_compare0_scratch): Likewise.
14987 (thumb1_movdi_insn): Likewise.
14988 (arm_movsi_insn): Likewise.
14989 (movhi_insn_arch4): Likewise.
14990 (movhi_bytes): Likewise.
14991 (arm_movqi_insn): Likewise.
14992 (thumb1_movqi_insn): Likewise.
14993 (arm32_movhf): Likewise.
14994 (thumb1_movhf): Likewise.
14995 (arm_movsf_soft_insn): Likewise.
14996 (thumb1_movsf_insn): Likewise.
14997 (thumb_movdf_insn): Likewise.
14998 (movsicc_insn): Likewise.
14999 (movsfcc_soft_insn): Likewise.
15000 (and_scc): Likewise.
15001 (cond_move): Likewise.
15002 (if_move_not): Likewise.
15003 (if_not_move): Likewise.
15004 (if_shift_move): Likewise.
15005 (if_move_shift): Likewise.
15006 (if_shift_shift): Likewise.
15007 (if_not_arith): Likewise.
15008 (if_arith_not): Likewise.
15009 (cond_move_not): Likewise.
15010 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
15011 (neon_mov<mode>): Likewise.
15012 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
15013 (thumb2_movsi_vfp): Likewise.
15014 (movsf_vfp): Likewise.
15015 (thumb2_movsf_vfp): Likewise.
15016 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15018 (cortexa7_older_only): Likewise.
15019 (cortexa7_younger): Likewise.
15020 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
15021 (1020alu_shift_op): Likewise.
15022 (1020alu_shift_reg_op): Likewise.
15023 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
15024 (alu_shift_op): Likewise.
15025 (alu_shift_reg_op): Likewise.
15026 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
15027 (11_alu_shift_op): Likewise.
15028 (11_alu_shift_reg_op): Likewise.
15029 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
15030 (9_alu_shift_reg_op): Likewise.
15031 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15033 (cortex_a15_alu_shift): Likewise.
15034 (cortex_a15_alu_shift_reg): Likewise.
15035 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
15036 (cortex_a5_alu_shift): Likewise.
15037 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
15039 (cortex_a53_alu_shift): Likewise.
15040 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15042 (cortex_a7_alu_reg): Likewise.
15043 (cortex_a7_alu_shift): Likewise.
15044 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15045 (cortex_a8_alu_shift): Likewise.
15046 (cortex_a8_alu_shift_reg): Likewise.
15047 (cortex_a8_mov): Likewise.
15048 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
15049 (cortex_a9_dp_shift): Likewise.
15050 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
15051 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
15052 (cortex_r4_mov): Likewise.
15053 (cortex_r4_alu_shift): Likewise.
15054 (cortex_r4_alu_shift_reg): Likewise.
15055 * config/arm/fa526.md (526_alu_op): Update for attribute change.
15056 (526_alu_shift_op): Likewise.
15057 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
15058 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
15059 (626te_alu_shift_op): Likewise.
15060 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
15061 (726te_alu_op): Likewise.
15062 (726te_alu_shift_op): Likewise.
15063 (726te_alu_shift_reg_op): Likewise.
15064 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
15065 (mp626_alu_shift_op): Likewise.
15066 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
15067 (pj4_alu_e1_conds): Likewise.
15068 (pj4_alu): Likewise.
15069 (pj4_alu_conds): Likewise.
15070 (pj4_shift): Likewise.
15071 (pj4_shift_conds): Likewise.
15072 (pj4_alu_shift): Likewise.
15073 (pj4_alu_shift_conds): Likewise.
15075 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15077 * config/arm/predicates.md (shiftable_operator_strict_it):
15079 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
15080 Disable cond_exec version for arm_restrict_it.
15081 (thumb2_smaxsi3): Convert to generate cond_exec.
15082 (thumb2_sminsi3): Likewise.
15083 (thumb32_umaxsi3): Likewise.
15084 (thumb2_uminsi3): Likewise.
15085 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
15086 (thumb2_neg_abssi2): Likewise.
15087 (thumb2_mov_scc): Add alternative for 16-bit encoding.
15088 (thumb2_movsicc_insn): Adjust alternatives.
15089 (thumb2_mov_negscc): Disable for arm_restrict_it.
15090 (thumb2_mov_negscc_strict_it): New pattern.
15091 (thumb2_mov_notscc_strict_it): New pattern.
15092 (thumb2_mov_notscc): Disable for arm_restrict_it.
15093 (thumb2_ior_scc): Likewise.
15094 (thumb2_ior_scc_strict_it): New pattern.
15095 (thumb2_cond_move): Adjust for arm_restrict_it.
15096 (thumb2_cond_arith): Disable for arm_restrict_it.
15097 (thumb2_cond_arith_strict_it): New pattern.
15098 (thumb2_cond_sub): Adjust for arm_restrict_it.
15099 (thumb2_movcond): Likewise.
15100 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
15101 (thumb2_zero_extendhisi2_v6): Likewise.
15102 (thumb2_zero_extendqisi2_v6): Likewise.
15103 (orsi_notsi_si): Likewise.
15104 (orsi_not_shiftsi_si): Likewise.
15106 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
15108 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
15109 instruction sequence is 1 byte shorter.
15111 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
15113 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
15114 it is not needed after split.
15116 2013-07-20 Iain Sandoe <iain@codesourcery.com>
15119 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
15120 second label for nonlocal goto receivers. Don't output pic base labels
15121 unless we're producing PIC; mark that action unreachable().
15122 (ix86_save_reg): If the function contains a nonlocal label, save the
15124 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
15125 * config/darwin.c (emitted_pic_label_num): New GTY.
15126 (update_pic_label_number_if_needed): New.
15127 (machopic_output_function_base_name): Adjust for nonlocal receiver
15129 (machopic_should_output_picbase_label): New.
15130 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
15131 (nonlocal_goto_receiver): New insn and split.
15133 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
15135 * config/aarch64/aarch64-builtins.c
15136 (aarch64_fold_builtin): Fold abs in all modes.
15137 * config/aarch64/aarch64-simd-builtins.def
15138 (abs): Enable for all modes.
15139 * config/aarch64/arm_neon.h
15140 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
15141 (vabs_f64): Add missing intrinsic.
15143 2013-07-19 Ian Bolton <ian.bolton@arm.com>
15145 * config/aarch64/arm_neon.h (vabs_s64): New function
15147 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
15150 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
15151 * config/avr/avr.md (adjust_len): Add `round'.
15152 * config/avr/avr-protos.h (avr_out_round): New prototype.
15153 (avr_out_plus): Add `out_label' argument.
15154 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
15155 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
15156 Handle the case where `insn' is just a pattern.
15157 (avr_out_bitop): Handle the case where `insn' is just a pattern.
15158 (avr_out_round): New function.
15159 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
15161 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
15163 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15164 Add check for flag_stack_usage to handle -fstack-usage support
15166 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
15168 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
15169 interaction for new Power8 flags and VSX.
15171 2013-07-18 Sriraman Tallam <tmsriram@google.com>
15173 PR middle-end/57698
15174 * tree-inline.c (expand_call_inline): Emit errors during
15175 early_inlining only if optimization is not turned on.
15177 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15181 * passes.c (init_optimization_passes): Move the construction of
15182 the pass hierarchy into a new passes.def file.
15184 * Makefile.in (passes.o): Add dependency on passes.def.
15186 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15188 * passes.c (init_optimization_passes): Introduce macros for
15189 constructing the tree of passes (INSERT_PASSES_AFTER,
15190 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
15191 TERMINATE_PASS_LIST).
15193 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
15194 Wei Mi <wmi@google.com>
15196 PR rtl-optimization/57878
15197 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
15199 (reload_pseudo_compare_func): Check nregs first for reload
15202 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15204 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
15206 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
15208 * read-rtl.c (validate_const_int): Once an invalid character is
15209 seen, quit the loop.
15211 * gengtype.c (write_roots): Similarly once we find the "deletable"
15212 or "if_marked" option.
15214 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15216 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
15217 "xtab" and "sat". Move value "clz" from here to ...
15218 (attriubte "type"): ... here.
15219 (satsi_<SAT:code>): Delete "insn" attribute.
15220 (satsi_<SAT:code>_shift): Likewise.
15221 (arm_zero_extendqisi2addsi): Likewise.
15222 (arm_extendqisi2addsi): Likewise.
15223 (clzsi2): Update for attribute changes.
15224 (rbitsi2): Likewise.
15225 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
15227 (arm_usatsihi): Likewise.
15228 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15230 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15232 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
15233 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
15234 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
15235 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
15236 in alphabetical order.
15237 (attribute "core_cycles"): Update for attribute changes.
15238 (arm_addsi3): Likewise.
15239 (addsi3_compare0): Likewise.
15240 (addsi3_compare0_scratch): Likewise.
15241 (addsi3_compare_op1): Likewise.
15242 (addsi3_compare_op2): Likewise.
15243 (compare_addsi2_op0): Likewise.
15244 (compare_addsi2_op1): Likewise.
15245 (addsi3_carryin_shift_<optab>): Likewise.
15246 (subsi3_carryin_shift): Likewise.
15247 (rsbsi3_carryin_shift): Likewise.
15248 (arm_subsi3_insn): Likewise.
15249 (subsi3_compare0): Likewise.
15250 (subsi3_compare): Likewise.
15251 (arm_andsi3_insn): Likewise.
15252 (thumb1_andsi3_insn): Likewise.
15253 (andsi3_compare0): Likewise.
15254 (andsi3_compare0_scratch): Likewise.
15255 (zeroextractsi_compare0_scratch
15256 (andsi_not_shiftsi_si): Likewise.
15257 (iorsi3_insn): Likewise.
15258 (iorsi3_compare0): Likewise.
15259 (iorsi3_compare0_scratch): Likewise.
15260 (arm_xorsi3): Likewise.
15261 (thumb1_xorsi3_insn): Likewise.
15262 (xorsi3_compare0): Likewise.
15263 (xorsi3_compare0_scratch): Likewise.
15264 (satsi_<SAT:code>_shift): Likewise.
15266 (arm_shiftsi3): Likewise.
15267 (shiftsi3_compare0): Likewise.
15268 (not_shiftsi): Likewise.
15269 (not_shiftsi_compare0): Likewise.
15270 (not_shiftsi_compare0_scratch): Likewise.
15271 (arm_one_cmplsi2): Likewise.
15272 (thumb_one_complsi2): Likewise.
15273 (notsi_compare0): Likewise.
15274 (notsi_compare0_scratch): Likewise.
15275 (thumb1_zero_extendhisi2): Likewise.
15276 (arm_zero_extendhisi2): Likewise.
15277 (arm_zero_extendhisi2_v6): Likewise.
15278 (arm_zero_extendhisi2addsi): Likewise.
15279 (thumb1_zero_extendqisi2): Likewise.
15280 (thumb1_zero_extendqisi2_v6): Likewise.
15281 (arm_zero_extendqisi2): Likewise.
15282 (arm_zero_extendqisi2_v6): Likewise.
15283 (arm_zero_extendqisi2addsi): Likewise.
15284 (thumb1_extendhisi2): Likewise.
15285 (arm_extendhisi2): Likewise.
15286 (arm_extendhisi2_v6): Likewise.
15287 (arm_extendqisi): Likewise.
15288 (arm_extendqisi_v6): Likewise.
15289 (arm_extendqisi2addsi): Likewise.
15290 (thumb1_extendqisi2): Likewise.
15291 (thumb1_movdi_insn): Likewise.
15292 (arm_movsi_insn): Likewise.
15293 (movsi_compare0): Likewise.
15294 (movhi_insn_arch4): Likewise.
15295 (movhi_bytes): Likewise.
15296 (arm_movqi_insn): Likewise.
15297 (thumb1_movqi_insn): Likewise.
15298 (arm32_movhf): Likewise.
15299 (thumb1_movhf): Likewise.
15300 (arm_movsf_soft_insn): Likewise.
15301 (thumb1_movsf_insn): Likewise.
15302 (movdf_soft_insn): Likewise.
15303 (thumb_movdf_insn): Likewise.
15304 (arm_cmpsi_insn): Likewise.
15305 (cmpsi_shiftsi): Likewise.
15306 (cmpsi_shiftsi_swp): Likewise.
15307 (arm_cmpsi_negshiftsi_si): Likewise.
15308 (movsicc_insn): Likewise.
15309 (movsfcc_soft_insn): Likewise.
15310 (arith_shiftsi): Likewise.
15311 (arith_shiftsi_compare0
15312 (arith_shiftsi_compare0_scratch
15313 (sub_shiftsi): Likewise.
15314 (sub_shiftsi_compare0
15315 (sub_shiftsi_compare0_scratch
15316 (and_scc): Likewise.
15317 (cond_move): Likewise.
15318 (if_plus_move): Likewise.
15319 (if_move_plus): Likewise.
15320 (if_move_not): Likewise.
15321 (if_not_move): Likewise.
15322 (if_shift_move): Likewise.
15323 (if_move_shift): Likewise.
15324 (if_shift_shift): Likewise.
15325 (if_not_arith): Likewise.
15326 (if_arith_not): Likewise.
15327 (cond_move_not): Likewise.
15328 (thumb1_ashlsi3): Set type attribute.
15329 (thumb1_ashrsi3): Likewise.
15330 (thumb1_lshrsi3): Likewise.
15331 (thumb1_rotrsi3): Likewise.
15332 (shiftsi3_compare0_scratch): Likewise.
15333 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
15334 (neon_mov<mode>): Likewise.
15335 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
15337 (thumb2_movsi_insn): Likewise.
15338 (thumb2_cmpsi_neg_shiftsi): Likewise.
15339 (thumb2_extendqisi_v6): Likewise.
15340 (thumb2_zero_extendhisi2_v6): Likewise.
15341 (thumb2_zero_extendqisi2_v6): Likewise.
15342 (thumb2_shiftsi3_short): Likewise.
15343 (thumb2_addsi3_compare0_scratch): Likewise.
15344 (orsi_not_shiftsi_si): Likewise.
15345 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
15346 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
15348 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
15349 (1020alu_shift_op): Likewise.
15350 (1020alu_shift_reg_op): Likewise.
15351 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
15352 (alu_shift_op): Likewise.
15353 (alu_shift_reg_op): Likewise.
15354 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
15355 (11_alu_shift_op): Likewise.
15356 (11_alu_shift_reg_op): Likewise.
15357 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
15358 (9_alu_shift_reg_op): Likewise.
15359 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15361 (cortex_a15_alu_shift): Likewise.
15362 (cortex_a15_alu_shift_reg): Likewise.
15363 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
15365 (cortex_a5_alu_shift): Likewise.
15366 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
15368 (cortex_a53_alu_shift): Likewise.
15369 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15371 (cortex_a7_alu_reg): Likewise.
15372 (cortex_a7_alu_shift): Likewise.
15373 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
15375 (cortex_a8_alu_shift): Likewise.
15376 (cortex_a8_alu_shift_reg): Likewise.
15377 (cortex_a8_mov): Likewise.
15378 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
15379 (cortex_a9_dp_shift): Likewise.
15380 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
15382 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
15384 (cortex_r4_mov): Likewise.
15385 (cortex_r4_alu_shift): Likewise.
15386 (cortex_r4_alu_shift_reg): Likewise.
15387 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
15388 (526_alu_shift_op): Likewise.
15389 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
15390 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
15391 (626te_alu_shift_op): Likewise.
15392 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
15393 (726te_alu_op): Likewise.
15394 (726te_alu_shift_op): Likewise.
15395 (726te_alu_shift_reg_op): Likewise.
15396 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
15397 (mp626_alu_shift_op): Likewise.
15398 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
15399 (pj4_alu_e1_conds): Likewise.
15400 (pj4_alu): Likewise.
15401 (pj4_alu_conds): Likewise.
15402 (pj4_shift): Likewise.
15403 (pj4_shift_conds): Likewise.
15404 (pj4_alu_shift): Likewise.
15405 (pj4_alu_shift_conds): Likewise.
15406 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15408 (cortexa7_older_only): Likewise.
15409 (cortexa7_younger): Likewise.
15411 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15413 * ipa-pure-const.c (generate_summary): Rename to...
15414 (pure_const_generate_summary): ... this.
15416 2013-07-17 Iain Sandoe <iain@codesourcery.com>
15418 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
15420 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
15423 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
15426 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15428 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
15429 enabled without -march=zEC12.
15430 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
15433 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15435 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
15436 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
15437 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
15438 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
15439 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
15440 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
15441 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
15443 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
15444 and ISA_HAS_NMADD3_NMSUB3.
15445 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
15446 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
15447 (nmsub4<mode>, nmsub3<mode>): Likewise.
15448 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
15450 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15452 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
15453 TARGET_MIPS5400 checking.
15455 2013-07-16 Jakub Jelinek <jakub@redhat.com>
15456 Peter Bergner <bergner@vnet.ibm.com>
15458 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
15459 registers in the comment.
15460 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
15461 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
15462 rather than FIRST_PSEUDO_REGISTERS.
15464 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
15466 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
15467 enable extra ISA flags with TARGET_HTM.
15469 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15471 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
15474 2013-07-15 Cong Hou <congh@google.com>
15476 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
15477 in compare function for sorting.
15479 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
15481 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
15482 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
15483 * config/rs6000/rs6000.opt: Add -mhtm option.
15484 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
15485 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
15486 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
15487 __HTM__ if the HTM instructions are available.
15488 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
15489 htm_spr_reg_operand): New define_predicates.
15490 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
15491 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
15493 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
15494 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
15495 HTM builtin functions.
15496 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
15497 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
15498 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
15499 (rs6000_builtin_mask_calculate): Likewise.
15500 (rs6000_option_override_internal): Likewise.
15501 (bdesc_htm): Add new HTM builtin support.
15502 (htm_spr_num): New function.
15503 (htm_spr_regno): Likewise.
15504 (rs6000_htm_spr_icode): Likewise.
15505 (htm_expand_builtin): Likewise.
15506 (htm_init_builtins): Likewise.
15507 (rs6000_expand_builtin): Add support for HTM builtin functions.
15508 (rs6000_init_builtins): Likewise.
15509 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
15511 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
15512 (TARGET_HTM, MASK_HTM): Define macros.
15513 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
15514 (FIXED_REGISTERS): Likewise.
15515 (CALL_USED_REGISTERS): Likewise.
15516 (CALL_REALLY_USED_REGISTERS): Likewise.
15517 (REG_ALLOC_ORDER): Likewise.
15518 (enum reg_class): Likewise.
15519 (REG_CLASS_NAMES): Likewise.
15520 (REG_CLASS_CONTENTS): Likewise.
15521 (REGISTER_NAMES): Likewise.
15522 (ADDITIONAL_REGISTER_NAMES): Likewise.
15523 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
15524 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
15525 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
15526 * config/rs6000/htm.md: New file.
15527 * config/rs6000/htmintrin.h: New file.
15528 * config/rs6000/htmxlintrin.h: New file.
15530 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
15532 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15533 Define SYMBOL_TINY_GOT, update comment.
15534 * config/aarch64/aarch64.c
15535 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
15536 (aarch64_expand_mov_immediate): Likewise.
15537 (aarch64_print_operand): Likewise.
15538 (aarch64_classify_symbol): Likewise.
15539 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
15540 (ldr_got_tiny): Define.
15542 2013-07-13 Tobias Grosser <tobias@grosser.es>
15544 PR tree-optimization/54094
15545 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
15546 scheduling dimension for the parallelism check from the polyhedral
15547 information in the AST.
15548 * graphite-dependences.c (carries_deps): Do not assume the schedule is
15551 2013-07-13 Jason Merrill <jason@redhat.com>
15553 * print-tree.c (debug_vec_tree): Use debug_raw.
15554 (debug_raw (vec<tree, va_gc> &)): New.
15555 (debug_raw (vec<tree, va_gc> *)): New.
15556 * tree.h: Declare them.
15558 2013-07-13 Bin Cheng <bin.cheng@arm.com>
15560 * ifcvt.c (ifcvt_after_combine): New static variable.
15561 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
15563 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
15564 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
15565 rest_of_handle_if_after_reload): Pass new argument for if_convert.
15567 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
15569 * config/mips/mips.c (mips_expand_call): Remove empty statement.
15571 2013-07-12 Michael Matz <matz@suse.de>
15573 PR middle-end/55771
15574 * convert.c (convert_to_real): Reject non-float inner types.
15576 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
15578 * config/aarch64/aarch64-protos.h
15579 (aarch64_simd_immediate_valid_for_move): Remove.
15580 * config/aarch64/aarch64.c (simd_immediate_info): New member.
15581 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
15583 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
15585 2013-07-11 Steve Ellcey <sellcey@mips.com>
15587 * config/mips/mips.c (mips_conditional_register_usage): Do not
15588 use t[0-7] registers in MIPS16 mode when optimizing for size.
15590 2013-07-11 Sriraman Tallam <tmsriram@google.com>
15592 * config/i386/i386.c (dispatch_function_versions): Fix array
15593 indexing of function_version_info to match actual_versions.
15595 2013-07-11 Teresa Johnson <tejohnson@google.com>
15597 * vec.h (struct va_gc): Move release out-of-line.
15598 (va_gc::release): Call ggc_free on released vec.
15600 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15602 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
15603 Require GOT register as additional operand in UNSPEC.
15604 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
15605 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15606 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15607 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
15608 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
15609 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
15610 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
15612 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
15615 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
15616 name seen by assembler/linker rather if available.
15618 2013-07-11 Andreas Schwab <schwab@suse.de>
15620 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
15622 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
15624 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
15626 2013-07-10 Joseph Myers <joseph@codesourcery.com>
15628 * doc/tm.texi.in: Move hook documentation to ....
15629 * target.def: ... here.
15631 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
15632 text on @hook line.
15633 * doc/tm.texi: Regenerate.
15635 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
15638 * doc/invoke.texi: Document -Wconditionally-supported.
15640 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15643 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
15646 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15649 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
15650 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
15651 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
15652 Remove duplicate devices.
15653 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
15654 * config/avr/t-multilib: Regenerate.
15655 * config/avr/avr-tables.opt: Regenerate.
15656 * doc/avr-mmcu.texi: Regenerate.
15658 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15661 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
15663 2013-07-10 Graham Stott <graham.stott@btinternet.com>
15665 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
15666 the cost of MULT when optimizing for size.
15668 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
15670 * config/cr16/cr16-protos.h: Don't include target.h.
15672 2013-07-09 Joseph Myers <joseph@codesourcery.com>
15674 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
15675 adjust register size for TDmode and TFmode for VSX registers.
15677 2013-07-08 Kai Tietz <ktietz@redhat.com>
15680 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
15681 hook_bool_const_tree_true.
15683 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15685 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
15686 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
15687 * config/s390/s390.md: Define FPR*_REGNUM constants.
15688 Fix FPR2_REGNUM constant (18 -> 17).
15689 ("*trunc<BFP:mode><DFP_ALL:mode>2")
15690 ("*trunc<DFP_ALL:mode><BFP:mode>2")
15691 ("trunc<BFP:mode><DFP_ALL:mode>2")
15692 ("trunc<DFP_ALL:mode><BFP:mode>2")
15693 ("*extend<BFP:mode><DFP_ALL:mode>2")
15694 ("*extend<DFP_ALL:mode><BFP:mode>2")
15695 ("extend<BFP:mode><DFP_ALL:mode>2")
15696 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
15699 2013-07-08 Graham Stott <graham.stott@btinternet.com>
15701 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
15703 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15705 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
15706 and cfun_fpr_bit_p to cfun_fpr_save_p.
15707 (s390_frame_area, s390_register_info, s390_frame_info)
15708 (s390_emit_prologue, s390_emit_epilogue)
15709 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
15711 * config/s390/s390.h: Define *_REGNUM macros for floating point
15714 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
15716 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
15718 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
15720 PR rtl-optimization/57786
15721 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
15722 and break out of the loop when it is set to false.
15724 2013-07-08 Jakub Jelinek <jakub@redhat.com>
15727 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
15728 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
15729 (const_int 63)) 0)).
15730 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
15731 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
15732 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
15734 PR rtl-optimization/57829
15735 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
15736 mask bits outside of mode are just sign-extension from mode to HWI.
15738 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
15740 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
15741 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
15742 adjust_address instead of change_address to keep info about alignment.
15743 (emit_strmov): Remove.
15744 (emit_memmov): New function.
15745 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
15746 (expand_movmem_epilogue): Likewise and return updated rtx for
15748 (expand_constant_movmem_prologue): Likewise and return updated rtx for
15749 destination and source.
15750 (decide_alignment): Refactor, handle vector_loop.
15751 (ix86_expand_movmem): Likewise.
15752 (ix86_expand_setmem): Likewise.
15753 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
15755 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
15757 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
15758 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
15760 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15762 * config/i386/sse.md (sse_movlhps): Change alternative 3
15763 of operand 2 to "m".
15765 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15768 * config/i386/sse.md (iptr): New mode attribute.
15769 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
15770 (<sse>_vm<plusminus_insn><mode>3): Ditto.
15771 (<sse>_vmmul<mode>3): Ditto.
15772 (<sse>_vmdiv<mode>3): Ditto.
15773 (sse_vmrcpv4sf2): Ditto.
15774 (<sse>_vmsqrt<mode>2): Ditto.
15775 (sse_vmrsqrtv4sf2): Ditto.
15776 (<sse>_vm<code><mode>3): Ditto.
15777 (avx_vmcmp<mode>3): Ditto.
15778 (<sse>_vmmaskcmp<mode>3): Ditto.
15779 (<sse>_comi): Ditto.
15780 (<sse>_ucomi): Ditto.
15781 (*xop_vmfrcz_<mode>): Ditto.
15782 (*fmai_fmadd_<mode>): Ditto.
15783 (*fmai_fmsub_<mode>): Ditto.
15784 (*fmai_fnmadd_<mode>): Ditto.
15785 (*fmai_fnmsub_<mode>): Ditto.
15786 (*fma4i_vmfmadd_<mode>): Ditto.
15787 (*fma4i_vmfmsub_<mode>): Ditto.
15788 (*fma4i_vmfnmadd_<mode>): Ditto.
15789 (*fma4i_vmfnmsub_<mode>): Ditto.
15790 (*xop_vmfrcz_<mode>): Ditto.
15791 (sse_cvtps2pi): Ditto.
15792 (sse_cvttps2pi): Ditto.
15793 (sse_cvtss2si): Ditto.
15794 (sse_cvtss2si_2): Ditto.
15795 (sse_cvtss2siq_2): Ditto.
15796 (sse_cvttss2si): Ditto.
15797 (sse_cvttss2siq): Ditto.
15798 (sse_cvtsd2si): Ditto.
15799 (sse_cvtsd2si_2): Ditto.
15800 (sse_cvtsd2siq_2): Ditto.
15801 (sse_cvttsd2si): Ditto.
15802 (sse_cvttsd2siq): Ditto.
15803 (sse_cvtsd2ss): Ditto.
15804 (sse_cvtss2sd): Ditto.
15805 (avx2_pbroadcast<mode>): Ditto.
15806 (avx2_pbroadcast<mode>_1): Ditto.
15807 (*avx_vperm_broadcast_v4sf): Ditto.
15809 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
15810 (sse_movlhps): Ditto.
15811 (sse_storehps): Ditto.
15812 (sse_loadhps): Ditto.
15813 (sse_storelps): Ditto.
15814 (sse_loadlps): Ditto.
15815 (*vec_concatv4sf): Ditto.
15816 (*vec_interleave_highv2df): Ditto.
15817 (*vec_interleave_lowv2df): Ditto.
15818 (*vec_extractv2df_1_sse): Ditto.
15819 (*vec_extractv2df_0_sse): Ditto.
15820 (sse2_storelpd): Ditto.
15821 (sse2_loadlpd): Ditto.
15822 (sse2_movsd): Ditto.
15823 (*vec_concatv4si): Ditto.
15824 (vec_concatv2di): Ditto.
15826 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
15827 for Intel asm dialect.
15828 (mmx_punpcklwd): Ditto.
15829 (mmx_punpckldq): Ditto.
15831 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
15832 for intel assembler dialect.
15834 2013-07-06 Jakub Jelinek <jakub@redhat.com>
15837 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
15838 for BUILT_IN_C{LZ,LRSB}*.
15839 * tree.h (CASE_INT_FN): Add FN##IMAX case.
15840 * tree-vrp.c (extract_range_basic): Handle
15841 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
15842 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
15843 fall thru to code calling set_value*.
15844 * builtins.c (expand_builtin): Remove *IMAX cases.
15845 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
15846 if width is bigger than 2*HWI.
15848 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
15850 PR rtl-optimization/55342
15851 * lra-int.h (lra_subreg_reload_pseudos): New.
15852 * lra.c: Add undoing optional reloads to the block diagram.
15853 (lra_subreg_reload_pseudos): New.
15854 (lra_optional_reload_pseudos): Change comments.
15855 (lra): Init and clear lra_subreg_reload_pseudos. Clear
15856 lra_optional_reload_pseudos after undo transformations.
15857 * lra-assigns.c (pseudo_prefix_title): New.
15858 (lra_setup_reg_renumber): Use it.
15859 (spill_for): Ditto. Check subreg reload pseudos too.
15860 (assign_by_spills): Consider subreg reload pseudos too.
15861 * lra-constraints.c (simplify_operand_subreg): Use
15862 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
15863 (curr_insn_transform): Recognize and do optional reloads.
15864 (undo_optional_reloads): New.
15865 (lra_undo_inheritance): Call undo_optional_reloads.
15867 2013-07-05 Thomas Quinot <quinot@adacore.com>
15869 * tree-complex.c (expand_complex_operations_1): Fix typo.
15871 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
15873 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
15874 (tune_params): New member 'const vec_costs'.
15875 * config/aarch64/aarch64.c (generic_vector_cost): New.
15876 (generic_tunings): New member 'generic_vector_cost'.
15877 (aarch64_builtin_vectorization_cost): New.
15878 (aarch64_add_stmt_cost): New.
15879 (TARGET_VECTORIZE_ADD_STMT_COST): New.
15880 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
15882 2013-07-03 Jakub Jelinek <jakub@redhat.com>
15885 * config/i386/predicates.md (vsib_address_operand): Disallow
15886 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
15888 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
15890 PR middle-end/55030
15891 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
15892 expand_builtin_setjmp_receiver.
15893 (expand_label): Adjust, call expand_builtin_setjmp_receiver
15894 with NULL for the label parameter.
15895 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
15896 the frame-pointer. Adjust comments.
15897 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
15898 only if LABEL is non-NULL.
15900 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
15902 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
15903 (ARM_ABI_AAPCS64): Ditto.
15905 (ARM_DEFAULT_ABI): Ditto.
15907 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
15909 * config/aarch64/aarch64-builtins.c
15910 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
15911 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
15913 * config/aarch64/aarch64-simd.md
15914 (aarch64_ld1<VALL:mode>): New.
15915 (aarch64_st1<VALL:mode>): Likewise.
15916 * config/aarch64/arm_neon.h
15917 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
15919 2013-07-02 Sriraman Tallam <tmsriram@google.com>
15921 * config/i386/i386.c (gate_insert_vzeroupper): Check if
15923 (ix86_option_override_internal):Turn on all -mavx target flags by
15924 default as they are dependent on AVX anyway.
15926 2013-07-02 Cary Coutant <ccoutant@google.com>
15928 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
15929 deterministic hash.
15930 (loc_checksum_ordered): Likewise.
15931 (hash_loc_operands): Remove inline keyword.
15933 2013-07-02 Jakub Jelinek <jakub@redhat.com>
15935 PR tree-optimization/57741
15936 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
15937 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
15938 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
15939 Allow REAL_CST step_exprs if flag_associative_math.
15940 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
15942 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15944 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
15946 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15948 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
15950 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15952 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
15954 (iorsi3_insn): Likewise.
15955 (arm_xorsi3): Likewise.
15957 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
15959 * arm.md (attribute "wtype"): Delete. Move attribute values from here
15961 (attribute "type"): ... here, and prefix with "wmmx_".
15962 (attribute "core_cycles"): Update for attribute changes.
15963 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
15964 (tbcstv4hi): Likewise.
15965 (tbcstv2si): Likewise.
15966 (iwmmxt_iordi3): Likewise.
15967 (iwmmxt_xordi3): Likewise.
15968 (iwmmxt_anddi3): Likewise.
15969 (iwmmxt_nanddi3): Likewise.
15970 (iwmmxt_arm_movdi): Likewise.
15971 (iwmmxt_movsi_insn): Likewise.
15972 (mov<mode>_internal): Likewise.
15973 (and<mode>3_iwmmxt): Likewise.
15974 (ior<mode>3_iwmmxt): Likewise.
15975 (xor<mode>3_iwmmxt): Likewise.
15976 (add<mode>3_iwmmxt): Likewise.
15977 (ssaddv8qi3): Likewise.
15978 (ssaddv4hi3): Likewise.
15979 (ssaddv2si3): Likewise.
15980 (usaddv8qi3): Likewise.
15981 (usaddv4hi3): Likewise.
15982 (usaddv2si3): Likewise.
15983 (sub<mode>3_iwmmxt): Likewise.
15984 (sssubv8qi3): Likewise.
15985 (sssubv4hi3): Likewise.
15986 (sssubv2si3): Likewise.
15987 (ussubv8qi3): Likewise.
15988 (ussubv4hi3): Likewise.
15989 (ussubv2si3): Likewise.
15990 (mulv4hi3_iwmmxt): Likewise.
15991 (smulv4hi3_highpart): Likewise.
15992 (umulv4hi3_highpart): Likewise.
15993 (iwmmxt_wmacs): Likewise.
15994 (iwmmxt_wmacsz): Likewise.
15995 (iwmmxt_wmacu): Likewise.
15996 (iwmmxt_wmacuz): Likewise.
15997 (iwmmxt_clrdi): Likewise.
15998 (iwmmxt_clrv8qi): Likewise.
15999 (iwmmxt_clr4hi): Likewise.
16000 (iwmmxt_clr2si): Likewise.
16001 (iwmmxt_uavgrndv8qi3): Likewise.
16002 (iwmmxt_uavgrndv4hi3): Likewise.
16003 (iwmmxt_uavgv8qi3): Likewise.
16004 (iwmmxt_uavgv4hi3): Likewise.
16005 (iwmmxt_tinsrb): Likewise.
16006 (iwmmxt_tinsrh): Likewise.
16007 (iwmmxt_tinsrw): Likewise.
16008 (iwmmxt_textrmub): Likewise.
16009 (iwmmxt_textrmsb): Likewise.
16010 (iwmmxt_textrmuh): Likewise.
16011 (iwmmxt_textrmsh): Likewise.
16012 (iwmmxt_textrmw): Likewise.
16013 (iwmxxt_wshufh): Likewise.
16014 (eqv8qi3): Likewise.
16015 (eqv4hi3): Likewise.
16016 (eqv2si3): Likewise.
16017 (gtuv8qi3): Likewise.
16018 (gtuv4hi3): Likewise.
16019 (gtuv2si3): Likewise.
16020 (gtv8qi3): Likewise.
16021 (gtv4hi3): Likewise.
16022 (gtv2si3): Likewise.
16023 (smax<mode>3_iwmmxt): Likewise.
16024 (umax<mode>3_iwmmxt): Likewise.
16025 (smin<mode>3_iwmmxt): Likewise.
16026 (umin<mode>3_iwmmxt): Likewise.
16027 (iwmmxt_wpackhss): Likewise.
16028 (iwmmxt_wpackwss): Likewise.
16029 (iwmmxt_wpackdss): Likewise.
16030 (iwmmxt_wpackhus): Likewise.
16031 (iwmmxt_wpackwus): Likewise.
16032 (iwmmxt_wpackdus): Likewise.
16033 (iwmmxt_wunpckihb): Likewise.
16034 (iwmmxt_wunpckihh): Likewise.
16035 (iwmmxt_wunpckihw): Likewise.
16036 (iwmmxt_wunpckilb): Likewise.
16037 (iwmmxt_wunpckilh): Likewise.
16038 (iwmmxt_wunpckilw): Likewise.
16039 (iwmmxt_wunpckehub): Likewise.
16040 (iwmmxt_wunpckehuh): Likewise.
16041 (iwmmxt_wunpckehuw): Likewise.
16042 (iwmmxt_wunpckehsb): Likewise.
16043 (iwmmxt_wunpckehsh): Likewise.
16044 (iwmmxt_wunpckehsw): Likewise.
16045 (iwmmxt_wunpckelub): Likewise.
16046 (iwmmxt_wunpckeluh): Likewise.
16047 (iwmmxt_wunpckeluw): Likewise.
16048 (iwmmxt_wunpckelsb): Likewise.
16049 (iwmmxt_wunpckelsh): Likewise.
16050 (iwmmxt_wunpckelsw): Likewise.
16051 (ror<mode>3): Likewise.
16052 (ashr<mode>3_iwmmxt): Likewise.
16053 (lshr<mode>3_iwmmxt): Likewise.
16054 (ashl<mode>3_iwmmxt): Likewise.
16055 (ror<mode>3_di): Likewise.
16056 (ashr<mode>3_di): Likewise.
16057 (lshr<mode>3_di): Likewise.
16058 (ashl<mode>3_di): Likewise.
16059 (iwmmxt_wmadds): Likewise.
16060 (iwmmxt_wmaddu): Likewise.
16061 (iwmmxt_tmia): Likewise.
16062 (iwmmxt_tmiaph): Likewise.
16063 (iwmmxt_tmiabb): Likewise.
16064 (iwmmxt_tmiatb): Likewise.
16065 (iwmmxt_tmiabt): Likewise.
16066 (iwmmxt_tmiatt): Likewise.
16067 (iwmmxt_tmovmskb): Likewise.
16068 (iwmmxt_tmovmskh): Likewise.
16069 (iwmmxt_tmovmskw): Likewise.
16070 (iwmmxt_waccb): Likewise.
16071 (iwmmxt_wacch): Likewise.
16072 (iwmmxt_waccw): Likewise.
16073 (iwmmxt_waligni): Likewise.
16074 (iwmmxt_walignr): Likewise.
16075 (iwmmxt_walignr0): Likewise.
16076 (iwmmxt_walignr1): Likewise.
16077 (iwmmxt_walignr2): Likewise.
16078 (iwmmxt_walignr3): Likewise.
16079 (iwmmxt_wsadb): Likewise.
16080 (iwmmxt_wsadh): Likewise.
16081 (iwmmxt_wsadbz): Likewise.
16082 (iwmmxt_wsadhz): Likewise.
16083 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
16084 (iwmmxt_wabsdiffb): Likewise.
16085 (iwmmxt_wabsdiffh): Likewise.
16086 (iwmmxt_wabsdiffw): Likewise.
16087 (iwmmxt_waddsubhx): Likewise
16088 (iwmmxt_wsubaddhx): Likewise.
16089 (addc<mode>3): Likewise.
16090 (iwmmxt_avg4): Likewise.
16091 (iwmmxt_avg4r): Likewise.
16092 (iwmmxt_wmaddsx): Likewise.
16093 (iwmmxt_wmaddux): Likewise.
16094 (iwmmxt_wmaddsn): Likewise.
16095 (iwmmxt_wmaddun): Likewise.
16096 (iwmmxt_wmulwsm): Likewise.
16097 (iwmmxt_wmulwum): Likewise.
16098 (iwmmxt_wmulsmr): Likewise.
16099 (iwmmxt_wmulumr): Likewise.
16100 (iwmmxt_wmulwsmr): Likewise.
16101 (iwmmxt_wmulwumr): Likewise.
16102 (iwmmxt_wmulwl): Likewise.
16103 (iwmmxt_wqmulm): Likewise.
16104 (iwmmxt_wqmulwm): Likewise.
16105 (iwmmxt_wqmulmr): Likewise.
16106 (iwmmxt_wqmulwmr): Likewise.
16107 (iwmmxt_waddbhusm): Likewise.
16108 (iwmmxt_waddbhusl): Likewise.
16109 (iwmmxt_wqmiabb): Likewise.
16110 (iwmmxt_wqmiabt): Likewise.
16111 (iwmmxt_wqmiatb): Likewise.
16112 (iwmmxt_wqmiatt): Likewise.
16113 (iwmmxt_wqmiabbn): Likewise.
16114 (iwmmxt_wqmiabtn): Likewise.
16115 (iwmmxt_wqmiatbn): Likewise.
16116 (iwmmxt_wqmiattn): Likewise.
16117 (iwmmxt_wmiabb): Likewise.
16118 (iwmmxt_wmiabt): Likewise.
16119 (iwmmxt_wmiatb): Likewise.
16120 (iwmmxt_wmiatt): Likewise.
16121 (iwmmxt_wmiabbn): Likewise.
16122 (iwmmxt_wmiabtn): Likewise.
16123 (iwmmxt_wmiatbn): Likewise.
16124 (iwmmxt_wmiattn): Likewise.
16125 (iwmmxt_wmiawbb): Likewise.
16126 (iwmmxt_wmiawbt): Likewise.
16127 (iwmmxt_wmiawtb): Likewise.
16128 (iwmmxt_wmiawtt): Likewise.
16129 (iwmmxt_wmiawbbn): Likewise.
16130 (iwmmxt_wmiawbtn): Likewise.
16131 (iwmmxt_wmiawtbn): Likewise.
16132 (iwmmxt_wmiawttn): Likewise.
16133 (iwmmxt_wmerge): Likewise.
16134 (iwmmxt_tandc<mode>3): Likewise.
16135 (iwmmxt_torc<mode>3): Likewise.
16136 (iwmmxt_torvsc<mode>3): Likewise.
16137 (iwmmxt_textrc<mode>3): Likewise.
16138 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
16139 (wmmxt_pack): Likewise.
16140 (wmmxt_mult_c1): Likewise.
16141 (wmmxt_mult_c2): Likewise.
16142 (wmmxt_alu_c1): Likewise.
16143 (wmmxt_alu_c2): Likewise.
16144 (wmmxt_alu_c3): Likewise.
16145 (wmmxt_transfer_c1): Likewise.
16146 (wmmxt_transfer_c2): Likewise.
16147 (wmmxt_transfer_c3): Likewise.
16148 (marvell_f_iwmmxt_wstr): Likewise.
16149 (marvell_f_iwmmxt_wldr): Likewise.
16151 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
16153 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
16155 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16158 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16159 * lra-constraints.c (need_for_split_p): Check call used hard regs
16160 living through calls.
16162 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
16163 call used regs for call insn.
16165 2013-06-28 Jakub Jelinek <jakub@redhat.com>
16168 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
16169 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
16171 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16173 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
16175 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16177 * lra-constraints.c (need_for_split_p): Check call used hard regs
16178 living through calls.
16180 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
16183 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
16184 to tie with any other modes. Eliminate Altivec vector mode tests,
16185 since these are a subset of ALTIVEC or VSX vector modes. Simplify
16186 code, to return 0 if testing MODE2 for a condition, if we've
16187 already tested MODE1 for the same condition.
16189 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16191 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
16194 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16196 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
16197 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
16199 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16201 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
16203 (aarch64_symbolic_constant_p): Remove.
16204 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
16205 static. Fix line length and white space.
16206 (aarch64_symbolic_constant_p): Remove.
16207 * config/aarch64/predicates.md (aarch64_valid_symref):
16208 Use aarch64_classify_symbol_expression.
16210 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16212 * config/arm/constraints.md (Ts): New constraint.
16213 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
16215 (compare_scc): Use "Ts" constraint for operand 0.
16216 (ior_scc_scc): Likewise.
16217 (and_scc_scc): Likewise.
16218 (and_scc_scc_nodom): Likewise.
16219 (ior_scc_scc_cmp): Likewise for operand 7.
16220 (and_scc_scc_cmp): Likewise.
16221 * config/arm/thumb2.md (thumb2_movsi_insn):
16222 Add alternatives for 16-bit encodings.
16223 (thumb2_movhi_insn): Likewise.
16224 (thumb2_movsicc_insn): Likewise.
16225 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
16226 (thumb2_negscc): Use "Ts" constraint.
16227 Move mvn instruction outside cond_exec block.
16228 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
16229 for 16-bit encodings.
16231 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16233 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
16235 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
16236 (mulsi3subsi): Likewise.
16237 (mulsidi3adddi): Likewise.
16238 (mulsidi3_v6): Likewise.
16239 (umulsidi3_v6): Likewise.
16240 (umulsidi3adddi_v6): Likewise.
16241 (smulsi3_highpart_v6): Likewise.
16242 (umulsi3_highpart_v6): Likewise.
16243 (mulhisi3tb): Likewise.
16244 (mulhisi3bt): Likewise.
16245 (mulhisi3tt): Likewise.
16246 (maddhisi4): Likewise.
16247 (maddhisi4tb): Likewise.
16248 (maddhisi4tt): Likewise.
16249 (maddhidi4): Likewise.
16250 (maddhidi4tb): Likewise.
16251 (maddhidi4tt): Likewise.
16252 (zeroextractsi_compare0_scratch): Likewise.
16253 (insv_zero): Likewise.
16254 (insv_t2): Likewise.
16255 (anddi_notzesidi_di): Likewise.
16256 (anddi_notsesidi_di): Likewise.
16257 (andsi_notsi_si): Likewise.
16258 (iordi_zesidi_di): Likewise.
16259 (xordi_zesidi_di): Likewise.
16260 (andsi_iorsi3_notsi): Likewise.
16261 (smax_0): Likewise.
16262 (smax_m1): Likewise.
16263 (smin_0): Likewise.
16264 (not_shiftsi): Likewise.
16265 (unaligned_loadsi): Likewise.
16266 (unaligned_loadhis): Likewise.
16267 (unaligned_loadhiu): Likewise.
16268 (unaligned_storesi): Likewise.
16269 (unaligned_storehi): Likewise.
16270 (extv_reg): Likewise.
16271 (extzv_t2): Likewise.
16272 (divsi3): Likewise.
16273 (udivsi3): Likewise.
16274 (arm_zero_extendhisi2addsi): Likewise.
16275 (arm_zero_extendqisi2addsi): Likewise.
16276 (compareqi_eq0): Likewise.
16277 (arm_extendhisi2_v6): Likewise.
16278 (arm_extendqisi2addsi): Likewise.
16279 (arm_movt): Likewise.
16280 (thumb2_ldrd): Likewise.
16281 (thumb2_ldrd_base): Likewise.
16282 (thumb2_ldrd_base_neg): Likewise.
16283 (thumb2_strd): Likewise.
16284 (thumb2_strd_base): Likewise.
16285 (thumb2_strd_base_neg): Likewise.
16286 (arm_negsi2): Add alternative for 16-bit encoding.
16287 (arm_one_cmplsi2): Likewise.
16289 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16291 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
16292 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
16293 (movdfcc): Likewise.
16294 * config/arm/vfp.md (*thumb2_movsf_vfp):
16295 Disable predication for arm_restrict_it.
16296 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
16297 (*thumb2_movdfcc_vfp): Likewise.
16298 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
16299 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
16300 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
16301 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
16302 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
16303 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
16304 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
16305 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
16306 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
16307 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
16308 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
16309 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
16310 Disable predication for arm_restrict_it.
16312 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
16314 * config/i386/bmiintrin.h (_bextr_u32): New.
16315 (_bextr_u64): Ditto.
16317 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
16319 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
16320 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
16321 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
16322 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
16323 * config/mips/n32-elf.h: ...this new file.
16325 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
16328 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
16329 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
16331 2013-06-27 Catherine Moore <clm@codesourcery.com>
16333 * config/mips/mips-tables.opt: Regenerate.
16334 * config/mips/mips-cpus.def: Add m14ke and m14kec.
16335 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
16336 * doc/invoke.texi: Add -m14kc.
16338 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16341 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
16342 constraints of operand 1 and 2.
16345 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
16346 to match RTL canonicalization. Swap predicates and
16347 constraints of operand 1 and 2.
16349 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
16351 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
16352 Process OP_INOUT regs for splitting too.
16354 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16356 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
16357 decl before the loop, initialize to NULL.
16358 (vectorizable_load): Initialize ptr_incr to NULL.
16360 2013-06-27 Martin Jambor <mjambor@suse.cz>
16363 * ipa-ref.h (ipa_maybe_record_reference): Declare.
16364 * ipa-ref.c (ipa_maybe_record_reference): New function.
16365 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
16366 * ipa-cp.c (create_specialized_node): Record potential references from
16368 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
16370 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16372 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
16373 parameter 'mode' of type 'enum machine_mode mode'; change to pass
16374 'mode' to force_reg.
16375 (aarch64_add_offset): Update calls to aarch64_force_temporary.
16376 (aarch64_expand_mov_immediate): Likewise.
16378 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16380 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
16381 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
16383 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16385 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
16386 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16387 (struct machine_function): Add tbegin_p.
16388 (s390_canonicalize_comparison): Fold CC mode compares to
16389 conditional jump if possible.
16390 (s390_emit_jump): Return the emitted jump.
16391 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
16392 Handle CCRAWmode compares.
16393 (s390_option_override): Default to -mhtm if available.
16394 (s390_reg_clobbered_rtx): Handle floating point regs as well.
16395 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
16396 FPRs instead of df_regs_ever_live_p.
16397 (s390_optimize_nonescaping_tx): New function.
16398 (s390_init_frame_layout): Extend clobbered_regs array to cover
16400 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
16401 (s390_expand_tbegin): New function.
16402 (enum s390_builtin): New enum definition.
16403 (code_for_builtin): New array definition.
16404 (s390_init_builtins): New function.
16405 (s390_expand_builtin): New function.
16406 (TARGET_INIT_BUILTINS): Define.
16407 (TARGET_EXPAND_BUILTIN): Define.
16408 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
16409 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
16410 (s390_alc_comparison): Likewise.
16411 * config/s390/s390-modes.def: Add CCRAWmode.
16412 * config/s390/s390.h (processor_flags): Add PF_TX.
16413 (TARGET_CPU_HTM): Define macro.
16414 (TARGET_HTM): Define macro.
16415 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
16416 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
16417 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16418 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
16419 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
16420 (TBEGIN_MASK, TBEGINC_MASK): New constants.
16421 ("*cc_to_int"): Move up.
16422 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
16423 constants other than 0.
16424 ("*ccraw_to_int"): New insn and splitter definition.
16425 ("tbegin", "tbegin_nofloat", "tbegin_retry")
16426 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
16427 ("tx_assist"): New expander.
16428 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
16429 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
16430 * config/s390/s390.opt: Add -mhtm option.
16431 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
16432 * config/s390/htmxlintrin.h: New file.
16433 * config/s390/htmintrin.h: New file.
16434 * config/s390/s390intrin.h: New file.
16435 * doc/extend.texi: Document htm builtins.
16436 * config.gcc: Add the new header files to extra_headers.
16438 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
16440 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
16441 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
16443 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
16444 Pat Haugen <pthaugen@us.ibm.com>
16445 Peter Bergner <bergner@vnet.ibm.com>
16447 * config/rs6000/power8.md: New.
16448 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
16449 setting for power8 entry.
16450 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
16451 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
16452 test for Power4/Power5 only.
16453 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
16455 (force_new_group): Adjust comment.
16456 * config/rs6000/rs6000.md: Include power8.md.
16458 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
16460 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
16461 * config/arm/arm-protos.h (arm_max_conditional_execute): New
16463 (tune_params): Update comment.
16464 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
16465 (arm_max_conditional_execute): New function.
16466 (thumb2_final_prescan_insn): Use max_insn_skipped and
16467 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
16469 2013-06-25 Jakub Jelinek <jakub@redhat.com>
16471 PR tree-optimization/57705
16472 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
16473 SSA_NAME step, provided that it is not defined inside the loop.
16474 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
16475 (get_initial_def_for_induction): Handle SSA_NAME IV step.
16477 2013-06-25 Martin Jambor <mjambor@suse.cz>
16479 PR middle-end/57670
16480 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
16481 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
16483 (ipa_note_param_call): Initialize member_ptr flag.
16484 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
16485 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
16486 (ipa_write_indirect_edge_info): Stream member_ptr flag.
16487 (ipa_read_indirect_edge_info): Likewise.
16489 2013-06-25 Richard Biener <rguenther@suse.de>
16491 PR middle-end/56977
16492 * passes.c (init_optimization_passes): Move pass_fold_builtins
16493 and pass_dce earlier with -Og.
16495 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
16497 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
16498 <BIT_FIELD_REF>: Remove trailing TAB.
16499 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
16502 2013-06-24 Martin Jambor <mjambor@suse.cz>
16504 PR tree-optimization/57358
16505 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
16506 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
16507 (ipa_analyze_params_uses): Generate pessimistic info when true.
16509 2013-06-24 Martin Jambor <mjambor@suse.cz>
16511 PR tree-optimization/57539
16512 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
16513 global.inlined_to of the new node to it. All callers changed.
16514 * ipa-inline-transform.c (clone_inlined_nodes): New variable
16515 inlining_into, pass it to cgraph_clone_node.
16516 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
16517 ipa_free_edge_args_substructures.
16518 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
16519 rdesc linked list. Do not assert rdesc edges have inlined caller.
16520 Assert we have found an rdesc in the rdesc list.
16522 2013-06-24 Richard Biener <rguenther@suse.de>
16524 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
16525 (pointer_set_lookup): Declare.
16526 (class pointer_map): New template class implementing a
16527 generic pointer to T map.
16528 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
16529 pointer_map<T>::contains, pointer_map<T>::insert,
16530 pointer_map<T>::traverse): New functions.
16531 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
16532 (pointer_set_lookup): New function.
16533 (pointer_set_contains): Use pointer_set_lookup.
16534 (pointer_set_insert): Likewise.
16535 (insert_aux): Remove.
16536 (struct pointer_map_t): Embed a pointer_set_t.
16537 (pointer_map_create): Adjust.
16538 (pointer_map_destroy): Likewise.
16539 (pointer_map_contains): Likewise.
16540 (pointer_map_insert): Likewise.
16541 (pointer_map_traverse): Likewise.
16542 * tree-streamer.h (struct streamer_tree_cache_d): Use a
16543 pointer_map<unsigned> instead of a pointer_map_t.
16544 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
16545 (streamer_tree_cache_lookup): Likewise.
16546 (streamer_tree_cache_create): Likewise.
16547 (streamer_tree_cache_delete): Likewise.
16548 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
16549 pointer_map<unsigned> instead of a pointer_map_t.
16550 (lto_init_tree_ref_encoder): Adjust.
16551 (lto_destroy_tree_ref_encoder): Likewise.
16552 * lto-section-out.c (lto_output_decl_index): Likewise.
16553 (lto_record_function_out_decl_state): Likewise.
16554 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
16556 2013-06-24 Richard Biener <rguenther@suse.de>
16558 PR tree-optimization/57488
16559 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
16561 2013-06-24 Alan Modra <amodra@gmail.com>
16563 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
16564 (gen_easy_altivec_constant): Likewise.
16565 * config/rs6000/predicates.md (easy_vector_constant_add_self,
16566 easy_vector_constant_msb): Likewise.
16568 2013-06-23 Jakub Jelinek <jakub@redhat.com>
16571 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
16572 add missing return true.
16574 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
16577 * config/sh/predicates.md (general_extend_operand): Invoke
16578 general_movsrc_operand for memory operands.
16579 (general_movsrc_operand): Allow reg+reg addressing, do not use
16580 general_operand for memory operands.
16582 2013-06-23 Sriraman Tallam <tmsriram@google.com>
16584 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
16585 when current target options does not apply.
16586 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
16587 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
16588 * config/i386/bmiintrin.h: Pass appropriate target
16589 attributes to header.
16590 * config/i386/mmintrin.h: Ditto.
16591 * config/i386/nmmintrin.h: Ditto.
16592 * config/i386/avx2intrin.h: Ditto.
16593 * config/i386/fxsrintrin.h: Ditto.
16594 * config/i386/tbmintrin.h: Ditto.
16595 * config/i386/xsaveintrin.h: Ditto.
16596 * config/i386/f16cintrin.h: Ditto.
16597 * config/i386/xtestintrin.h: Ditto.
16598 * config/i386/xsaveoptintrin.h: Ditto.
16599 * config/i386/bmi2intrin.h: Ditto.
16600 * config/i386/lzcntintrin.h: Ditto.
16601 * config/i386/smmintrin.h: Ditto.
16602 * config/i386/wmmintrin.h: Ditto.
16603 * config/i386/x86intrin.h: Remove all header include guards.
16604 * config/i386/prfchwintrin.h: Ditto.
16605 * config/i386/pmmintrin.h: Ditto.
16606 * config/i386/tmmintrin.h: Ditto.
16607 * config/i386/xmmintrin.h: Ditto.
16608 * config/i386/popcntintrin.h: Ditto.
16609 * config/i386/rdseedintrin.h: Ditto.
16610 * config/i386/ammintrin.h: Ditto.
16611 * config/i386/emmintrin.h: Ditto.
16612 * config/i386/immintrin.h: Remove all header include guards.
16613 * config/i386/fma4intrin.h: Ditto.
16614 * config/i386/lwpintrin.h: Ditto.
16615 * config/i386/xopintrin.h: Ditto.
16616 * config/i386/ia32intrin.h: Ditto.
16617 * config/i386/avxintrin.h: Ditto.
16618 * config/i386/rtmintrin.h: Ditto.
16619 * config/i386/fmaintrin.h: Ditto.
16620 * config/i386/mm3dnow.h: Ditto.
16622 2013-06-22 Sriraman Tallam <tmsriram@google.com>
16624 * common/config/i386/i386-common.c: Handle LZCNT.
16626 2013-06-22 Andi Kleen <ak@linux.intel.com>
16628 * doc/extend.texi: Use __atomic_store_n instead of
16629 __atomic_store in HLE example.
16631 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
16633 * config/sh/sh.c: Remove <cstdlib> workaround.
16635 2013-06-21 Andi Kleen <ak@linux.intel.com>
16637 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
16639 2013-06-21 Andi Kleen <ak@linux.intel.com>
16641 * doc/extend.texi: Document that __atomic_clear and
16642 __atomic_test_and_set should only be used with bool.
16644 2013-06-20 Jan Hubicka <jh@suse.cz>
16646 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
16647 types_same_for_odr.
16648 * tree.c (decls_same_for_odr): New function.
16649 (same_for_edr): New function.
16650 (types_same_for_odr): New function.
16651 (get_binfo_at_offset): Use it.
16652 * tree.h (types_same_for_odr): Declare.
16654 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
16655 Jason Merrill <jason@redhat.com>
16657 * system.h: Include <cstdlib> as well as <stdlib.h>.
16659 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
16662 * config/i386/i386.c (construct_container): Report error if
16663 long double is used with disabled x87 float returns.
16665 2013-06-20 Jan Hubicka <jh@suse.cz>
16667 * lto-cgraph.c (input_symtab): Do not set cgraph state.
16669 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
16671 PR rtl-optimization/57425
16672 PR rtl-optimization/57569
16673 * alias.c (write_dependence_p): Remove parameters mem_mode and
16674 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
16675 Changed all callers.
16676 (canon_anti_dependence): Get comments and semantics in sync.
16677 Add parameter mem_canonicalized. Changed all callers.
16678 * rtl.h (canon_anti_dependence): Update prototype.
16680 2013-06-20 Richard Biener <rguenther@suse.de>
16682 * data-streamer-in.c (streamer_read_uhwi): Optimize single
16683 byte case, inline streamer_read_uchar and defer section
16686 2013-06-20 Richard Biener <rguenther@suse.de>
16688 PR tree-optimization/57584
16689 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
16690 SSA names into the expanded expression that take part in
16691 abnormal coalescing.
16693 2013-06-19 Sharad Singhai <singhai@google.com>
16695 * gcov.c (print_usage): Handle new option.
16696 (process_args): Ditto.
16697 (get_gcov_intermediate_filename): New function.
16698 (output_intermediate_file): New function.
16699 (output_gcov_file): New function
16700 (generate_results): Handle new option.
16701 (release_function): Relase demangled name.
16702 (read_graph_file): Handle demangled name.
16703 (output_lines): Ditto.
16704 * doc/gcov.texi: Document gcov intermediate format.
16706 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
16709 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
16710 (lra_emit_add): Use the functions. Add comment about Y as an
16713 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
16716 * collect2.c (collect_atexit): New.
16717 (collect_exit): Delete.
16718 (main): Register collect_atexit with atexit.
16719 (collect_wait): Change collect_exit to exit.
16721 * collect2.h (collect_exit): Delete.
16722 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
16724 2013-06-19 Wei Mi <wmi@google.com>
16726 PR rtl-optimization/57518
16727 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
16728 if regno is used in paradoxical subreg.
16729 (update_equiv_regs): Check pdx_subregs[regno] before
16730 set a reg to be equivalent with a mem.
16732 2013-06-19 Matthias Klose <doko@ubuntu.com>
16735 * file-find.h (find_a_file): Add a mode parameter.
16736 * file-find.c (find_a_file): Likewise.
16737 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
16738 with X_OK for the executables.
16739 * collect2.c (main): Call find_a_file with X_OK.
16741 2013-06-19 Steve Ellcey <sellcey@mips.com>
16744 * config/mips/mips.md (casesi_internal_mips16_<mode>):
16745 Use NEXT_INSN instead of next_real_insn.
16747 2013-06-19 Jan Hubicka <jh@suse.cz>
16749 * cgraph.h (const_value_known_p): Replace by ...
16750 (ctor_for_folding): .. this one.
16751 * cgraphunit.c (process_function_and_variable_attributes): Use it.
16752 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
16753 * expr.c (expand_expr_real_1): Likewise.
16754 (string_constant): Likewise.
16755 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
16756 * ipa.c (process_references): Likewise.
16757 (symtab_remove_unreachable_nodes): Likewise.
16758 * ipa-inline-analysis.c (param_change_prob): Likewise.
16759 * gimple-fold.c (canonicalize_constructor_val): Likewise.
16760 (get_base_constructor): Likwise.
16761 * varpool.c (varpool_remove_node): Likewise.
16762 (varpool_remove_initializer): LIkewise.
16763 (dump_varpool_node): LIkwise.
16764 (const_value_known_p): Rewrite to ...
16765 (ctor_for_folding): ... this one.
16767 2013-06-19 Jakub Jelinek <jakub@redhat.com>
16770 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
16771 PERSONALITY in $PATH derived prefixes.
16773 2013-06-19 Jeff Law <law@redhat.com>
16775 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
16778 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
16779 (simplify_bitwise_binary): Use it to simpify certain binary ops on
16782 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
16784 * config/arm/vfp.md: Move VFP instruction classification documentation
16786 * config/arm/arm.md: ... here. Update instruction classification
16789 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
16791 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
16792 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
16793 pattern. Use more efficient sequences on ARMv5 and Thumb2.
16795 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
16798 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16799 with NEXT_INSN. Use tablejump_p to check for jump table data
16802 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
16805 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
16806 that now in C++ the value is correct per the C++ standards.
16808 2013-06-19 Richard Biener <rguenther@suse.de>
16810 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
16811 for global context.
16813 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16816 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16819 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16820 with next_active_insn.
16822 2013-06-18 Sriraman Tallam <tmsriram@google.com>
16824 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
16825 functions are inlined during failures to flag an error.
16826 * tree-inline.c (expand_call_inline): Allow the error to be flagged
16827 in early inline pass.
16829 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
16831 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
16834 2013-06-18 Julian Brown <julian@codesourcery.com>
16836 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
16837 Permit virtual register pre-reload if !strict.
16838 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
16840 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
16842 * config/arm/neon.md (movmisalign<mode>): Use
16843 neon_perm_struct_or_reg_operand instead of
16844 neon_struct_or_register_operand.
16845 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
16846 neon_permissive_struct_operand instead of neon_struct_operand.
16847 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
16848 neon_vector_mem_operand.
16849 * config/arm/predicates.md (neon_struct_operand): Adjust call to
16850 neon_vector_mem_operand.
16851 (neon_permissive_struct_operand): New.
16852 (neon_struct_or_register_operand): Rename to...
16853 (neon_perm_struct_or_reg_operand): This. Adjust call to
16854 neon_vector_mem_operand.
16856 2013-06-18 Richard Biener <rguenther@suse.de>
16858 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
16859 * lto-streamer.h: Include pointer-set.h.
16860 (struct lto_decl_slot): Remove.
16861 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
16862 Remove next_index entry.
16863 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16864 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16865 (lto_init_tree_ref_encoder): Adjust.
16866 (lto_destroy_tree_ref_encoder): Likewise.
16867 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16868 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16869 (lto_output_decl_index): Adjust.
16870 (lto_new_out_decl_state): Likewise.
16871 (lto_record_function_out_decl_state): Likewise.
16872 * lto-streamer-out.c (copy_function): Likewise.
16874 2013-06-18 Richard Biener <rguenther@suse.de>
16876 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
16877 * cgraphunit.c: Include cfgloop.h.
16878 (init_lowered_empty_function): Initialize the loop tree.
16879 (assemble_thunk): Insert new BBs into loops.
16881 2013-06-18 Richard Biener <rguenther@suse.de>
16883 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
16884 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
16885 the map from cache entry to cache index optional.
16886 (streamer_tree_cache_replace_tree): Adjust accordingly.
16887 (streamer_tree_cache_append): Likewise.
16888 (streamer_tree_cache_delete): Likewise.
16889 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
16890 streamer cache map from cache entry to cache index.
16891 * lto-streamer-out.c (create_output_block): Adjust.
16893 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
16895 * config/arm/arm.md (attribute "insn"): Move multiplication and
16896 division attributes to...
16897 (attribute "type"): ... here. Remove mult.
16898 (attribute "mul32"): New attribute.
16899 (attribute "mul64"): Add umaal.
16900 (*arm_mulsi3): Update attributes.
16901 (*arm_mulsi3_v6): Likewise.
16902 (*thumb_mulsi3): Likewise.
16903 (*thumb_mulsi3_v6): Likewise.
16904 (*mulsi3_compare0): Likewise.
16905 (*mulsi3_compare0_v6): Likewise.
16906 (*mulsi_compare0_scratch): Likewise.
16907 (*mulsi_compare0_scratch_v6): Likewise.
16908 (*mulsi3addsi): Likewise.
16909 (*mulsi3addsi_v6): Likewise.
16910 (*mulsi3addsi_compare0): Likewise.
16911 (*mulsi3addsi_compare0_v6): Likewise.
16912 (*mulsi3addsi_compare0_scratch): Likewise.
16913 (*mulsi3addsi_compare0_scratch_v6): Likewise.
16914 (*mulsi3subsi): Likewise.
16915 (*mulsidi3adddi): Likewise.
16916 (*mulsi3addsi_v6): Likewise.
16917 (*mulsidi3adddi_v6): Likewise.
16918 (*mulsidi3_nov6): Likewise.
16919 (*mulsidi3_v6): Likewise.
16920 (*umulsidi3_nov6): Likewise.
16921 (*umulsidi3_v6): Likewise.
16922 (*umulsidi3adddi): Likewise.
16923 (*umulsidi3adddi_v6): Likewise.
16924 (*smulsi3_highpart_nov6): Likewise.
16925 (*smulsi3_highpart_v6): Likewise.
16926 (*umulsi3_highpart_nov6): Likewise.
16927 (*umulsi3_highpart_v6): Likewise.
16928 (mulhisi3): Likewise.
16929 (*mulhisi3tb): Likewise.
16930 (*mulhisi3bt): Likewise.
16931 (*mulhisi3tt): Likewise.
16932 (maddhisi4): Likewise.
16933 (*maddhisi4tb): Likewise.
16934 (*maddhisi4tt): Likewise.
16935 (maddhidi4): Likewise.
16936 (*maddhidi4tb): Likewise.
16937 (*maddhidi4tt): Likewise.
16938 (divsi3): Likewise.
16939 (udivsi3): Likewise.
16940 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
16941 (thumb2_mulsi_short_compare0): Likewise.
16942 (thumb2_mulsi_short_compare0_scratch): Likewise.
16943 * config/arm/arm1020e.md (1020mult1): Update attribute change.
16944 (1020mult2): Likewise.
16945 (1020mult3): Likewise.
16946 (1020mult4): Likewise.
16947 (1020mult5): Likewise.
16948 (1020mult6): Likewise.
16949 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
16951 (cortex_a15_mult64): Likewise.
16952 (cortex_a15_sdiv): Likewise.
16953 (cortex_a15_udiv): Likewise.
16954 * config/arm/arm1026ejs.md (mult1): Update attribute change.
16960 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
16961 (pj4_ir_div): Likewise.
16962 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
16963 (11_mult2): Likewise.
16964 (11_mult3): Likewise.
16965 (11_mult4): Likewise.
16966 (11_mult5): Likewise.
16967 (11_mult6): Likewise.
16968 (11_mult7): Likewise.
16969 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
16970 (cortex_a8_mla): Likewise.
16971 (cortex_a8_mull): Likewise.
16972 (cortex_a8_smulwy): Likewise.
16973 (cortex_a8_smlald): Likewise.
16974 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
16975 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
16976 (cortex_r4_mul_3): Likewise.
16977 (cortex_r4_mla_4): Likewise.
16978 (cortex_r4_mla_3): Likewise.
16979 (cortex_r4_smlald): Likewise.
16980 (cortex_r4_mull): Likewise.
16981 (cortex_r4_sdiv): Likewise.
16982 (cortex_r4_udiv): Likewise.
16983 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
16984 (cortex_a7_idiv): Likewise.
16985 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
16986 (9_mult2): Likewise.
16987 (9_mult3): Likewise.
16988 (9_mult4): Likewise.
16989 (9_mult5): Likewise.
16990 (9_mult6): Likewise.
16991 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
16992 (cortex_a53_sdiv): Likewise.
16993 (cortex_a53_udiv): Likewise.
16994 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
16995 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
16996 (mp626_mult2): Likewise.
16997 (mp626_mult3): Likewise.
16998 (mp626_mult4): Likewise.
16999 * config/arm/fa526.md (526_mult1): Update attribute change.
17000 (526_mult2): Likewise.
17001 * config/arm/arm-generic.md (mult): Update attribute change.
17002 (mult_ldsched_strongarm): Likewise.
17003 (mult_ldsched): Likewise.
17004 (multi_cycle): Likewise.
17005 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
17006 * config/arm/fa606te.md (606te_mult1): Update attribute change.
17007 (606te_mult2): Likewise.
17008 (606te_mult3): Likewise.
17009 (606te_mult4): Likewise.
17010 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
17011 (cortex_a9_mac16): Likewise.
17012 (cortex_a9_multiply): Likewise.
17013 (cortex_a9_mac): Likewise.
17014 (cortex_a9_multiply_long): Likewise.
17015 * config/arm/fa626te.md (626te_mult1): Update attribute change.
17016 (626te_mult2): Likewise.
17017 (626te_mult3): Likewise.
17018 (626te_mult4): Likewise.
17020 2013-06-18 Richard Biener <rguenther@suse.de>
17023 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
17025 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17028 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17029 with next_active_insn.
17031 2013-06-18 Alan Modra <amodra@gmail.com>
17033 * config/rs6000/rs6000.h (enum data_align): New.
17034 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
17035 (DATA_ABI_ALIGNMENT): Define.
17036 (CONSTANT_ALIGNMENT): Correct comment.
17037 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
17038 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
17040 2013-06-17 David Malcolm <dmalcolm@redhat.com>
17042 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
17043 ATTRIBUTE_UNUSED marking.
17045 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
17047 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
17048 alternative and update.
17049 (aarch64_dup_lanedi): Delete.
17050 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
17051 * config/aarch64/aarch64-simd-builtins.def: Update.
17053 2013-06-17 Richard Biener <rguenther@suse.de>
17055 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
17056 (lto_input_scc): Declare.
17057 (lto_input_tree_1): Likewise.
17058 (struct lto_stats_d): Add num_tree_bodies_output and
17059 num_pickle_refs_output.
17060 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
17061 (lto_read_tree_1): Split out from ...
17062 (lto_read_tree): ... this.
17063 (lto_input_scc): New function.
17064 (lto_input_tree_1): Split out from ...
17065 (lto_input_tree): ... this. Handle LTO_tree_scc.
17066 (lto_data_in_create): Create the streamer cache without hashes.
17067 * lto-streamer-out.c (create_output_block): Create the streamer
17068 cache with hashes when not doing WPA.
17069 (lto_write_tree_1): Split out from ...
17070 (lto_write_tree): ... this.
17071 (get_symbol_initial_value): New function.
17072 (lto_output_tree_1): Split out from ...
17073 (lto_output_tree): ... this. Write trees as series of SCCs
17074 using a DFS walk via DFS_write_tree.
17075 (struct sccs, struct scc_entry): New types.
17076 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
17077 (DFS_write_tree_body): New function.
17078 (DFS_write_tree): Likewise.
17079 (hash_tree): Likewise.
17080 (scc_entry_compare): Likewise.
17081 (hash_scc): Likewise.
17082 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
17083 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
17084 TREE_CHAIN as regular reference.
17085 (streamer_read_integer_cst): Remove.
17086 (streamer_get_pickled_tree): Adjust.
17087 * tree-streamer-out.c (streamer_write_chain): Disable streaming
17088 of DECL_EXTERNALs in BLOCK_VARS for now.
17089 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
17091 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
17092 Add hash value argument and record that if hashes are recorded
17094 (streamer_tree_cache_insert_1): Adjust.
17095 (streamer_tree_cache_insert): Likewise.
17096 (streamer_tree_cache_insert_at): Rename to ...
17097 (streamer_tree_cache_replace_tree): ... this and adjust.
17098 (streamer_tree_cache_append): Adjust.
17099 (record_common_node): Likewise.
17100 (streamer_tree_cache_create): Add argument whether to
17101 record hash values together with trees.
17102 (streamer_tree_cache_delete): Adjust.
17103 * tree-streamer.h (struct streamer_tree_cache_d): Add
17105 (streamer_read_integer_cst): Remove.
17106 (streamer_tree_cache_insert): Adjust.
17107 (streamer_tree_cache_append): Likewise.
17108 (streamer_tree_cache_insert_at): Rename to ...
17109 (streamer_tree_cache_replace_tree): ... this and adjust.
17110 (streamer_tree_cache_create): Add argument whether to record hashes.
17111 (streamer_tree_cache_get): Rename to ...
17112 (streamer_tree_cache_get_tree): ... this.
17113 (streamer_tree_cache_get_hash): New function.
17114 * tree.c (cache_integer_cst): New function.
17115 * tree.h (cache_integer_cst): Declare.
17116 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
17117 * lto-symtab.c (lto_varpool_replace_node): Only release
17118 DECL_INITIAL of non-prevailing decls.
17119 * varpool.c (varpool_remove_initializer): Do not release
17120 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
17122 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17124 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
17125 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
17126 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
17127 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
17128 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
17129 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
17130 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
17131 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
17132 instead of TARGET_64BIT.
17133 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
17134 Require ISA_HAS_<D>DIV.
17136 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
17138 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
17139 (mips*-*-linux*): Move default with_llsc setting to where other
17141 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
17143 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
17144 Likewise. Remove default with_tune setting. Move default float
17145 setting to its own block. Handle with_llsc in the same block as above.
17147 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
17149 PR rtl-optimization/57425
17150 PR rtl-optimization/57569
17151 * alias.c (write_dependence_p): Add new parameters mem_mode,
17152 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
17153 Changed all callers.
17154 (canon_anti_dependence): New function.
17155 * cse.c (check_dependence): Use canon_anti_dependence.
17156 * cselib.c (cselib_invalidate_mem): Likewise.
17157 * rtl.h (canon_anti_dependence): Declare.
17159 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17161 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
17162 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
17163 ".set mips3" for 64-bit targets.
17165 2013-06-15 Dehao Chen <dehao@google.com>
17167 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
17168 * gimple-low.c (gimple_check_call_matching_types): Likewise.
17169 (gimple_check_call_args): Likewise.
17170 * value-prof.c (check_ic_target): Likewise.
17171 * ipa-inline.c (early_inliner): Likewise.
17172 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
17173 * cgraph.c (cgraph_create_edge_1): Likewise.
17174 (cgraph_make_edge_direct): Likewise.
17176 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
17179 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
17180 rs6000_output_move_128bit to handle emitting quad memory
17181 operations. Set attribute length to 8 bytes.
17183 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
17185 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
17187 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
17188 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
17189 (aarch64_<su>mlsl<mode>): Likewise.
17191 2013-06-14 Mike Stump <mikestump@comcast.net>
17193 * Makefile.in (TARGET_H): Add insn-codes.h.
17195 2013-06-14 Alan Modra <amodra@gmail.com>
17197 PR middle-end/57134
17198 PR middle-end/57586
17199 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
17200 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
17201 bitfield expansion when EXPAND_MEMORY.
17202 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
17204 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
17206 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
17207 test for clearing quad memory on 32-bit later.
17209 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17211 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
17212 (fold_negate_expr): Likewise.
17213 (fold_real_zero_addition_p): Handle vectors.
17214 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
17216 2013-06-14 Alan Modra <amodra@gmail.com>
17218 * varasm.c (force_const_mem): Revert 2013-06-07 change.
17220 2013-06-13 Jan Hubicka <jh@suse.cz>
17222 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17223 Local comdats are not externally visible.
17224 * symtab.c (dump_symtab_base): Dump externally visible.
17225 (verify_symtab_base): Verify back links in the symtab hash.
17227 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17229 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
17230 CONVERT_EXPR as equal nodes.
17232 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17234 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
17236 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17238 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
17239 Generalize to complex and vector.
17240 * tree.c (build_all_ones_cst): New function.
17241 * tree.h (build_all_ones_cst): Declare it.
17243 2013-06-13 Alan Modra <amodra@gmail.com>
17245 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
17246 * config/rs6000/rs6000.md (signbittf2): New insn.
17247 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
17248 (abstf2_internal, cmptf_internal2): Likewise.
17249 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
17251 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
17252 Pat Haugen <pthaugen@us.ibm.com>
17253 Peter Bergner <bergner@vnet.ibm.com>
17255 * config/rs6000/rs6000.c (emit_load_locked): Add support for
17256 power8 byte, half-word, and quad-word atomic instructions.
17257 (emit_store_conditional): Likewise.
17258 (rs6000_expand_atomic_compare_and_swap): Likewise.
17259 (rs6000_expand_atomic_op): Likewise.
17261 * config/rs6000/sync.md (larx): Add new modes for power8.
17263 (AINT): New mode iterator to include TImode as well as normal
17264 integer modes on power8.
17265 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
17266 that VSX registers are not considered. Use AINT mode iterator
17267 instead of INT1 to allow inclusion of quad word atomic operations
17269 (load_locked<mode>): Likewise.
17270 (store_conditional<mode>): Likewise.
17271 (atomic_compare_and_swap<mode>): Likewise.
17272 (atomic_exchange<mode>): Likewise.
17273 (atomic_nand<mode>): Likewise.
17274 (atomic_fetch_<fetchop_name><mode>): Likewise.
17275 (atomic_nand_fetch<mode>): Likewise.
17276 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
17278 (ATOMIC): On power8, add QImode, HImode modes.
17279 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
17280 modes that promote to SImode.
17281 (load_lockedti): Convert TImode arguments to PTImode, so that we
17282 get a guaranteed even/odd register pair.
17283 (load_lockedpti): Likewise.
17284 (store_conditionalti): Likewise.
17285 (store_conditionalpti): Likewise.
17287 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
17288 atomic load/store instructions.
17291 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
17293 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
17295 (insn_count): New attribute, with most cases extracted from...
17296 (length): ...here. Redefine most cases in terms of insn_count.
17297 (single_insn): Delete.
17298 (can_delay): Use insn_count to check for single instructions.
17299 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
17300 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
17301 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
17302 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
17303 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
17304 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
17305 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
17306 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
17307 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
17308 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
17309 rather than "length".
17310 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
17311 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
17312 Use "insn_count" rather than "length".
17313 * config/mips/mips-dsp.md
17314 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
17315 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
17318 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
17320 PR tree-optimization/57361
17321 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
17323 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
17325 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
17327 (aarch64_simd_combine<mode>): New instruction expansion.
17328 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
17329 function prototype.
17330 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
17331 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
17333 2013-06-12 Jan Hubicka <jh@suse.cz>
17335 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
17336 decl has when in streaming stage.
17337 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
17338 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
17340 2013-06-12 Roland Stigge <stigge@antcom.de>
17343 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
17345 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17347 PR tree-optimization/57537
17348 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
17349 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
17351 2013-06-12 Richard Biener <rguenther@suse.de>
17353 * data-streamer.h (streamer_write_char_stream): CSE
17354 obs->current_pointer.
17355 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
17356 streamer_write_char_stream manually and optimize the resulting loop.
17357 (streamer_write_hwi_stream): Likewise.
17359 2013-06-12 Jan Hubicka <jh@suse.cz>
17361 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
17362 * cgraph.h (varpool_create_empty_node): Declare.
17363 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
17365 * symtab.c (symtab_unregister_node): Be lax about missin entries
17367 (symtab_get_node): Update comment.
17368 * varpool.c (varpool_create_empty_node): Break out from ...
17369 (varpool_node_for_decl): ... here.
17370 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
17372 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
17374 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
17375 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
17376 part. Use straight-line flow at the end.
17377 <COMPONENT_REF>: Remove superfluous else.
17378 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
17380 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17383 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
17384 target hook even for !TREE_PUBLIC decls. If no resolution info
17385 is available, return false for common and external decls.
17387 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17389 * config/rl78/constraints.md (U): New constraint.
17390 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
17393 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17396 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
17397 to allow returning address to AT_PLATFORM name.
17399 2013-06-11 Jan Hubicka <jh@suse.cz>
17401 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
17402 * cgraph.h (symtab_node_base): Add weakref flag.
17403 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
17404 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
17405 (output_weakrefs): Use weakref flag.
17406 * fold-const.c (simple_operand_p): Handle WEAK.
17407 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
17408 * ipa.c (varpool_externally_visible_p): Drop weakref.
17409 (function_and_variable_visibility): Update comment; fix weakref
17410 sanity checks; do not clear DECL_WEAK on them.
17411 * lto-cgraph.c (lto_output_node): update.
17412 (lto_output_varpool_node): Update.
17413 (input_overwrite_node): Update.
17414 (input_node): Update.
17415 (input_varpool_node): Update.
17416 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
17417 (lto_symtab_merge_symbols): Add sanity check.
17418 (lto_symtab_prevailing_decl): Do not special case weakrefs.
17419 * passes.c (rest_of_decl_compilation): Set static flag, too.
17420 * symtab.c (dump_symtab_base): Dump weakref.
17421 (verify_symtab_base): Sanity check weakrefs.
17422 (symtab_make_decl_local): Remove duplicated code.
17423 (symtab_alias_ultimate_target): Simplify.
17424 * varpool.c (varpool_create_variable_alias): Set weakref flag.
17426 2013-06-11 Tom de Vries <tom@codesourcery.com>
17428 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
17429 sequence_vect == NULL.
17431 2013-06-11 DJ Delorie <dj@redhat.com>
17433 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
17434 (rl78_unwind_word_mode): New.
17436 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17438 * final.c (debug_prefix_maps): Make static.
17440 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17442 * function.c (initial_trampoline): Remove stray copy.
17444 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
17446 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
17448 2013-06-11 Martin Jambor <mjambor@suse.cz>
17450 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
17451 within bounds at the beginning of the function.
17453 2013-06-11 Alan Modra <amodra@gmail.com>
17455 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
17457 (get_named_section): Don't NULL !DECL_P decl.
17459 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
17461 * doc/invoke.texi (core-avx2): Document.
17463 (atom): Updated with MOVBE.
17465 2013-06-11 Richard Biener <rguenther@suse.de>
17467 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
17469 2013-06-11 Anton Blanchard <anton@samba.org>
17471 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
17472 correct shift value in little-endian mode.
17474 2013-06-11 Jakub Jelinek <jakub@redhat.com>
17477 * varasm.c (get_variable_align): Move #endif to the right place.
17479 2013-06-10 Cary Coutant <ccoutant@google.com>
17481 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
17482 for hash so that hash table traversal order is deterministic.
17484 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
17485 Pat Haugen <pthaugen@us.ibm.com>
17486 Peter Bergner <bergner@vnet.ibm.com>
17488 * config/rs6000/vector.md (GPR move splitter): Do not split moves
17489 of vectors in GPRS if they are direct moves or quad word load or
17492 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
17494 (direct_move_p): Likewise.
17495 (quad_load_store_p): Likewise.
17497 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
17498 classes into bins based on the physical register type.
17499 (reg_class_to_reg_type): Likewise.
17500 (IS_STD_REG_TYPE): Likewise.
17501 (IS_FP_VECT_REG_TYPE): Likewise.
17502 (reload_fpr_gpr): Arrays to determine what insn to use if we can
17503 use direct move instructions.
17504 (reload_gpr_vsx): Likewise.
17505 (reload_vsx_gpr): Likewise.
17506 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
17507 information that is a simplification of register classes. Also
17508 precalculate direct move reload helpers.
17509 (direct_move_p): New function to return true if the operation can
17510 be done as a direct move instruciton.
17511 (quad_load_store_p): New function to return true if the operation
17512 is a quad memory operation.
17513 (rs6000_legitimize_address): If quad memory, only allow register
17514 indirect for TImode addresses.
17515 (rs6000_legitimate_address_p): Likewise.
17516 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
17517 (rs6000_reload_register_type): Likewise.
17518 (register_to_reg_type): Return register type.
17519 (rs6000_secondary_reload_simple_move): New helper function for
17520 secondary reload and secondary memory needed to identify anything
17521 that is a simple move, and does not need reloading.
17522 (rs6000_secondary_reload_direct_move): New helper function for
17523 secondary reload to identify cases that can be done with several
17524 instructions via the direct move instructions.
17525 (rs6000_secondary_reload_move): New helper function for secondary
17526 reload to identify moves between register types that can be done.
17527 (rs6000_secondary_reload): Add support for quad memory operations
17528 and for direct move.
17529 (rs6000_secondary_memory_needed): Likewise.
17530 (rs6000_debug_secondary_memory_needed): Change argument names.
17531 (rs6000_output_move_128bit): New function to return the move to
17532 use for 128-bit moves, including knowing about the various
17533 limitations of quad memory operations.
17535 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
17536 memory operations. call rs6000_output_move_128bit for the actual
17537 instruciton(s) to generate.
17538 (vsx_movti_64bit): Likewise.
17540 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
17541 (UNSPEC_P8V_MTVSRWZ): Likewise.
17542 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
17543 (UNSPEC_P8V_MTVSRD): Likewise.
17544 (UNSPEC_P8V_XXPERMDI): Likewise.
17545 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
17546 (UNSPEC_FUSION_GPR): Likewise.
17547 (FMOVE128_GPR): New iterator for direct move.
17548 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
17549 (f32_sv): Likewise.
17550 (f32_dm): Likewise.
17551 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
17552 loads and direct move instructions.
17553 (zero_extendsidi2_lfiwzx): Likewise.
17554 (extendsidi2_lfiwax): Likewise.
17555 (extendsidi2_nocell): Likewise.
17556 (floatsi<mode>2_lfiwax): Likewise.
17557 (lfiwax): Likewise.
17558 (floatunssi<mode>2_lfiwzx): Likewise.
17559 (lfiwzx): Likewise.
17560 (fix_trunc<mode>_stfiwx): Likewise.
17561 (fixuns_trunc<mode>_stfiwx): Likewise.
17562 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
17563 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
17564 (parity<mode>2_cmpb): Set length/type attr.
17565 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
17566 for 'mr.' to fast_compare.
17567 (bpermd_<mode>): Change type attr to popcnt.
17568 (p8_fmrgow_<mode>): New insns for power8 direct move support.
17569 (p8_mtvsrwz_1): Likewise.
17570 (p8_mtvsrwz_2): Likewise.
17571 (reload_fpr_from_gpr<mode>): Likewise.
17572 (p8_mtvsrd_1): Likewise.
17573 (p8_mtvsrd_2): Likewise.
17574 (p8_xxpermdi_<mode>): Likewise.
17575 (reload_vsx_from_gpr<mode>): Likewise.
17576 (reload_vsx_from_gprsf): Likewise.
17577 (p8_mfvsrd_3_<mode>): LIkewise.
17578 (reload_gpr_from_vsx<mode>): Likewise.
17579 (reload_gpr_from_vsxsf): Likewise.
17580 (p8_mfvsrd_4_disf): Likewise.
17581 (multi-word GPR splits): Do not split direct moves or quad memory
17584 2013-06-10 David Malcolm <dmalcolm@redhat.com>
17586 * tree-into-ssa.c (interesting_blocks): Make static.
17588 2013-06-10 Jakub Jelinek <jakub@redhat.com>
17591 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
17592 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
17593 Use DATA_ABI_ALIGNMENT for that case instead if defined.
17594 (get_variable_align): New function.
17595 (get_variable_section, emit_bss, emit_common,
17596 assemble_variable_contents, place_block_symbol): Use
17597 get_variable_align instead of DECL_ALIGN.
17598 (assemble_noswitch_variable): Add align argument, use it
17599 instead of DECL_ALIGN.
17600 (assemble_variable): Adjust caller. Use get_variable_align
17601 instead of DECL_ALIGN.
17602 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
17604 (DATA_ABI_ALIGNMENT): Define.
17605 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
17606 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
17607 opt is false, only return the psABI mandated alignment increase.
17608 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
17609 (DATA_ABI_ALIGNMENT): ... this.
17610 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
17611 (DATA_ABI_ALIGNMENT): ... this.
17612 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
17613 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
17614 (DATA_ABI_ALIGNMENT): ... this.
17615 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
17616 * doc/tm.texi: Regenerated.
17618 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
17620 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
17621 cmp_code to construct REG_EQUAL note.
17623 2013-06-09 Jakub Jelinek <jakub@redhat.com>
17626 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
17627 that operands[2] doesn't overlap with operands[0].
17629 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
17630 Jan Hubicka <jh@suse.cz>
17632 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
17633 hack to mark symbols as used.
17635 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
17637 PR rtl-optimization/57559
17638 * lra-constraints.c (process_alt_operands): Don't discourage
17639 memory with known offset for offsetable memory constraint.
17640 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
17642 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
17644 * varasm.c (struct oc_local_state): Reorder fields.
17645 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
17646 and adjust accordingly.
17647 (output_constructor): Reorder initialization code and adjust call to
17648 output_constructor_bitfield.
17650 2013-06-07 Jan Hubicka <jh@suse.cz>
17652 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
17654 2013-06-07 David Malcolm <dmalcolm@redhat.com>
17656 * tree-object-size.c (unknown): Make const.
17658 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17660 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
17661 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
17662 for last alternative in the cpu_facility attribute.
17664 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17667 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
17668 (xordi3): Change operand 2 constraint to arm_xordi_operand.
17669 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
17670 * config/arm/constraints.md (Dg): New constraint.
17671 * config/arm/neon.md (xordi3_neon): Remove.
17672 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
17673 * config/arm/predicates.md (arm_xordi_operand): New predicate.
17675 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17677 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
17678 Clean up alternatives.
17680 2013-06-07 Alan Modra <amodra@gmail.com>
17682 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
17685 2013-06-07 Alan Modra <amodra@gmail.com>
17687 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
17689 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17691 * config/arm/constraints.md (Df): New constraint.
17692 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
17693 Correct length attribute for last two alternatives.
17695 2013-06-07 Alan Modra <amodra@gmail.com>
17697 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
17698 override user -mfp-in-toc.
17699 (offsettable_ok_by_alignment): Consider just the current access
17700 rather than the whole object, unless BLKmode. Handle
17701 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
17702 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
17703 for -mcmodel=medium.
17704 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
17705 override user -mfp-in-toc or -msum-in-toc. Default to
17706 -mno-fp-in-toc for -mcmodel=medium.
17708 2013-06-06 DJ Delorie <dj@redhat.com>
17710 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
17711 TARGET_VALID_POINTER_MODE.
17713 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
17714 Pat Haugen <pthaugen@us.ibm.com>
17715 Peter Bergner <bergner@vnet.ibm.com>
17717 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
17718 Document new power8 builtins.
17720 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
17721 condition code register, to allow 128-bit logical operations to be
17722 done in the VSX or GPR registers.
17723 (nor<mode>3): Use the canonical form for nor.
17724 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
17725 vclz*, and vpopcnt* vector instructions.
17726 (nand<mode>3): Likewise.
17727 (orc<mode>3): Likewise.
17728 (clz<mode>2): LIkewise.
17729 (popcount<mode>2): Likewise.
17731 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
17732 that only the GPRs are recognized.
17734 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
17735 support for new power8 builtins.
17737 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
17739 (xscvdpspn): Likewise.
17745 (vpopcnt): Likewise.
17746 (vpopcntb): Likewise.
17747 (vpopcnth): Likewise.
17748 (vpopcntw): Likewise.
17749 (vpopcntd): Likewise.
17751 (vmrgew): Likewise.
17752 (vmrgow): Likewise.
17754 (eqv_v16qi3): Likewise.
17755 (eqv_v8hi3): Likewise.
17756 (eqv_v4si3): Likewise.
17757 (eqv_v2di3): Likewise.
17758 (eqv_v4sf3): Likewise.
17759 (eqv_v2df3): Likewise.
17761 (nand_v16qi3): Likewise.
17762 (nand_v8hi3): Likewise.
17763 (nand_v4si3): Likewise.
17764 (nand_v2di3): Likewise.
17765 (nand_v4sf3): Likewise.
17766 (nand_v2df3): Likewise.
17768 (orc_v16qi3): Likewise.
17769 (orc_v8hi3): Likewise.
17770 (orc_v4si3): Likewise.
17771 (orc_v2di3): Likewise.
17772 (orc_v4sf3): Likewise.
17773 (orc_v2df3): Likewise.
17775 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
17776 allow power8 quad mode in 64-bit.
17777 (rs6000_builtin_vectorized_function): Add support to vectorize
17778 ISA 2.07 count leading zeros, population count builtins.
17779 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
17780 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
17781 (builtin_function_type): Add vgbbd builtin function which takes an
17783 (altivec_expand_vec_perm_const): Add support for new power8 merge
17786 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
17787 that does not include TImdoe for use with 32-bit.
17788 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
17790 (UNSPEC_VSX_CVDPSPN): Likewise.
17791 (vsx_xscvdpspn): Likewise.
17792 (vsx_xscvspdpn): Likewise.
17793 (vsx_xscvdpspn_scalar): Likewise.
17794 (vsx_xscvspdpn_directmove): Likewise.
17795 (vsx_and<mode>3): Split logical operations into 32-bit and
17796 64-bit. Add support to do logical operations on TImode as well as
17797 VSX vector types. Allow logical operations to be done in either
17798 VSX registers or in general purpose registers in 64-bit mode. Add
17799 splitters if GPRs were used. For AND, add clobber of CCmode to
17800 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
17802 (vsx_and<mode>3_32bit): Likewise.
17803 (vsx_and<mode>3_64bit): Likewise.
17804 (vsx_ior<mode>3): Likewise.
17805 (vsx_ior<mode>3_32bit): Likewise.
17806 (vsx_ior<mode>3_64bit): Likewise.
17807 (vsx_xor<mode>3): Likewise.
17808 (vsx_xor<mode>3_32bit): Likewise.
17809 (vsx_xor<mode>3_64bit): Likewise.
17810 (vsx_one_cmpl<mode>2): Likewise.
17811 (vsx_one_cmpl<mode>2_32bit): Likewise.
17812 (vsx_one_cmpl<mode>2_64bit): Likewise.
17813 (vsx_nor<mode>3): Likewise.
17814 (vsx_nor<mode>3_32bit): Likewise.
17815 (vsx_nor<mode>3_64bit): Likewise.
17816 (vsx_andc<mode>3): Likewise.
17817 (vsx_andc<mode>3_32bit): Likewise.
17818 (vsx_andc<mode>3_64bit): Likewise.
17819 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
17820 and xxlorc instructions.
17821 (vsx_eqv<mode>3_64bit): Likewise.
17822 (vsx_nand<mode>3_32bit): Likewise.
17823 (vsx_nand<mode>3_64bit): Likewise.
17824 (vsx_orc<mode>3_32bit): Likewise.
17825 (vsx_orc<mode>3_64bit): Likewise.
17827 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
17829 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
17831 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
17832 (p8_vmrgow): Likewise.
17833 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
17834 GPRs to be split under VSX.
17835 (p8v_clz<mode>2): Add power8 count leading zero support.
17836 (p8v_popcount<mode>2): Add power8 population count support.
17837 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
17840 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
17843 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
17845 (vec_nand): Likewise.
17846 (vec_vclz): Likewise.
17847 (vec_vclzb): Likewise.
17848 (vec_vclzd): Likewise.
17849 (vec_vclzh): Likewise.
17850 (vec_vclzw): Likewise.
17851 (vec_vgbbd): Likewise.
17852 (vec_vmrgew): Likewise.
17853 (vec_vmrgow): Likewise.
17854 (vec_vpopcnt): Likewise.
17855 (vec_vpopcntb): Likewise.
17856 (vec_vpopcntd): Likewise.
17857 (vec_vpopcnth): Likewise.
17858 (vec_vpopcntw): Likewise.
17860 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17862 PR rtl-optimization/57468
17863 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
17866 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17868 PR rtl-optimization/57459
17869 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
17870 type when setting live regs.
17872 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17874 * config/s390/s390.opt (mlra): New option.
17875 * config/s390/s390.c (s390_decompose_address): Check displacement
17876 for all registers for LRA.
17877 (s390_secondary_reload): Don't used secondary reloads for LRA.
17878 (s390_lra_p): New function.
17879 (TARGET_LRA_P): Define.
17880 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
17881 of attribute cpu_facility to zarch for the last alternative.
17882 (*cmpmem_short): Ditto.
17884 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
17886 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
17887 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
17888 (arm_expand_prologue): Likewise.
17890 2013-06-06 Teresa Johnson <tejohnson@google.com>
17893 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
17894 as this is now done by redirect_edge_and_branch_force.
17895 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
17896 barriers, and fix interaction with splitting.
17897 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
17898 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
17899 reflect changes made in the routine.
17900 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
17901 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
17902 since this is called in cfglayout mode, and replace partition fixup
17903 with assert as that is now done by force_nonfallthru_and_redirect.
17904 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
17905 already be marked with region crossing note.
17906 (insert_section_boundary_note): Make non-static, gate on flag
17907 has_bb_partition, rewrite to also check for multiple partitions.
17908 (rest_of_handle_reorder_blocks): Remove call to
17909 insert_section_boundary_note, now done later during free_cfg.
17910 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
17911 * bb-reorder.h (insert_section_boundary_note): Declare.
17912 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
17913 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
17914 invoke insert_section_boundary_note.
17915 (try_redirect_by_replacing_jump): Remove unnecessary
17916 check for region crossing note.
17917 (fixup_partition_crossing): New function.
17918 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
17919 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
17920 in non-cfglayout mode.
17921 (force_nonfallthru_and_redirect): Fixup partition boundaries,
17922 remove old code that tried to do this. Emit barrier correctly
17923 when we are in cfglayout mode.
17924 (last_bb_in_partition): New function.
17925 (rtl_split_edge): Correctly fixup partition boundaries.
17926 (commit_one_edge_insertion): Remove old code that tried to
17927 fixup region crossing edge since this is now handled in
17928 split_block, and set up insertion point correctly since
17929 block may now end in a jump.
17930 (verify_hot_cold_block_grouping): Guard against checking when not in
17931 linearized RTL mode.
17932 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
17934 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
17935 rtl_verify_flow_info, so not called in cfglayout mode.
17936 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
17937 (fixup_reorder_chain): Remove old code that attempted to fixup region
17938 crossing note as this is now handled in force_nonfallthru_and_redirect.
17939 (duplicate_insn_chain): Don't duplicate switch section notes.
17940 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
17942 * basic-block.h (emit_barrier_after_bb): Declare.
17944 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17946 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
17947 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
17948 arm_usatsihi): Adjust alternatives for arm_restrict_it.
17950 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17952 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
17954 * config/arm/ldmstm.md: Regenerate.
17956 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17958 * config/arm/sync.md (atomic_loaddi_1):
17959 Disable predication for arm_restrict_it.
17960 (arm_load_exclusive<mode>): Likewise.
17961 (arm_load_exclusivesi): Likewise.
17962 (arm_load_exclusivedi): Likewise.
17963 (arm_load_acquire_exclusive<mode>): Likewise.
17964 (arm_load_acquire_exclusivesi): Likewise.
17965 (arm_load_acquire_exclusivedi): Likewise.
17966 (arm_store_exclusive<mode>): Likewise.
17967 (arm_store_exclusive<mode>): Likewise.
17968 (arm_store_release_exclusivedi): Likewise.
17969 (arm_store_release_exclusive<mode>): Likewise.
17971 2013-06-06 Richard Biener <rguenther@suse.de>
17973 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
17975 (lto_tag_is_tree_code_p): Adjust.
17976 (lto_tag_is_gimple_code_p): Likewise.
17977 (lto_gimple_code_to_tag): Likewise.
17978 (lto_tag_to_gimple_code): Likewise.
17979 (lto_tree_code_to_tag): Likewise.
17980 (lto_tag_to_tree_code): Likewise.
17981 * data-streamer.h (streamer_write_hwi_in_range): Use
17982 uhwi streaming to stream the normalized range.
17983 (streamer_read_hwi_in_range): Likewise.
17985 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17987 * config/arm/arm.md (enabled_for_depr_it): New attribute.
17988 (predicable_short_it): Likewise.
17989 (predicated): Likewise.
17990 (enabled): Handle above.
17991 (define_cond_exec): Set predicated attribute to yes.
17993 2013-06-05 Mike Stump <mikestump@comcast.net>
17995 * gdbinit.in (__FUNCTION__): Add.
17997 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
17999 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
18000 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
18002 2013-06-05 Jan Hubicka <jh@suse.cz>
18004 * varasm.c (mark_decl_referenced): Revert the removal until targets
18007 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
18009 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
18010 instead of mark_decl_referenced.
18012 2013-06-05 Jan Hubicka <jh@suse.cz>
18014 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
18015 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
18016 and symtab_used_from_object_file_p.
18017 (cgraph_make_node_local_1): Clear forced_by_abi.
18018 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
18019 * cgraph.h (symtab_node_base): Add forced_by_abi.
18020 (decide_is_variable_needed): Remove.
18021 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
18022 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
18023 (decide_is_symbol_needed): ... this one; handle symbols in general;
18024 always analyze virtuals; honnor forced_by_abi.
18025 (cgraph_finalize_function): Update.
18026 (varpool_finalize_decl): Update.
18027 (symbol_defined_and_needed): Remove.
18028 (analyze_functions): Update.
18029 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18030 output_refs, input_overwrite_node): Handle forced_by_abi.
18031 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
18032 (address_taken_from_non_vtable_p): ... this one.
18033 (comdat_can_be_unshared_p_1): New function.
18034 (cgraph_comdat_can_be_unshared_p): Rename to ...
18035 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
18036 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
18037 (function_and_variable_visibility): Clear forced_by_abi as needed.
18038 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
18039 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
18040 * symtab.c (dump_symtab_base): Dump forced_by_abi.
18041 * varpool.c (decide_is_variable_needed): Remove.
18043 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18045 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
18046 (arm_option_override): Override arm_restrict_it where appropriate.
18047 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
18048 * config/arm/arm.opt (mrestrict-it): New command-line option.
18049 * doc/invoke.texi: Document -mrestrict-it.
18051 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18053 * tsan.c (tsan_atomic_table): Make const.
18055 2013-06-05 Richard Biener <rguenther@suse.de>
18057 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
18058 index associated with the tree we are supposed to replace.
18059 * tree-streamer-out.c (pack_ts_base_value_fields): Output
18060 TREE_ASM_WRITTEN as zero for everything but SSA names.
18062 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18064 * tree-ssa-structalias.c (call_stmt_vars): Make static.
18066 2013-06-04 Jan Hubicka <jh@suse.cz>
18068 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
18069 (input_node, input_varpool_node): Handle correctly external same
18071 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
18072 nodes at ltrans stage.
18074 2013-06-04 Jan Hubicka <jh@suse.cz>
18076 * ipa-inline.c (update_caller_keys): Fix availability test.
18077 (update_callee_keys): Likewise.
18078 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
18079 to follow ELF standard.
18081 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
18083 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
18084 (mips64r5900el-*-elf*): New configurations.
18085 * config/mips/mips-cpus.def (r5900): New processor.
18086 * config/mips/mips-tables.opt: Regenerate.
18087 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
18088 (mips_issue_rate): Handle PROCESSOR_R5900.
18089 (mips_reorg_process_insns): Force reorder mode for the R5900.
18090 * config/mips/mips.h (TARGET_MIPS5900): Define.
18091 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
18093 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
18095 * config/mips/mips.md (processor): Add r5900.
18096 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
18098 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18100 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
18101 into function to generate MOVI instruction.
18102 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
18103 (aarch64_preferred_simd_mode): Turn into wrapper.
18104 (aarch64_output_scalar_simd_mov_immediate): New function.
18105 * config/aarch64/aarch64-protos.h: Add prototype for above.
18107 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18109 * config/aarch64/aarch64.c (simd_immediate_info): Remove
18110 element_char member.
18111 (sizetochar): Return signed char.
18112 (aarch64_simd_valid_immediate): Remove elchar and other
18113 unnecessary variables.
18114 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
18115 Calculate element_char as required.
18116 * config/aarch64/aarch64-protos.h: Update and move prototype
18117 for aarch64_output_simd_mov_immediate.
18118 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
18121 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18123 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
18124 information completed by aarch64_simd_valid_immediate.
18125 (aarch64_legitimate_constant_p): Update arguments.
18126 (aarch64_simd_valid_immediate): Work with struct rather than many
18128 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
18129 (aarch64_simd_make_constant): Update arguments.
18130 (aarch64_output_simd_mov_immediate): Work with struct rather than
18131 many pointers. Output immediate directly rather than as operand.
18132 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18134 * config/aarch64/constraints.md (Dn): Update arguments.
18136 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18138 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
18140 (aarch64_simd_immediate_valid_for_move): Remove.
18141 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
18142 (aarch64_simd_make_constant): Update call.
18143 (aarch64_output_simd_mov_immediate): Update call.
18144 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18146 * config/aarch64/constraints.md (Dn): Update call.
18148 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18150 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
18151 return type to bool for prototype.
18152 (aarch64_legitimate_constant_p): Check for true instead of not -1.
18153 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
18154 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
18156 2013-06-04 Catherine Moore <clm@codesourcery.com>
18158 * config/mips/mips.opt (meva): New.
18159 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
18160 (ASM_SPEC): Handle -meva.
18161 * doc/invoke.texi (meva): Document.
18163 2013-06-04 Alan Modra <amodra@gmail.com>
18165 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
18168 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18170 * rtl.def: Add extra fourth optional field to define_cond_exec.
18171 * gensupport.c (process_one_cond_exec): Process attributes from
18173 * doc/md.texi: Document fourth field in define_cond_exec.
18175 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
18177 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
18178 out the processing order as in store_bit_field_1.
18180 2013-06-04 Jan Hubicka <jh@suse.cz>
18182 PR middle-end/57500
18183 * cgraphunit.c (cgraph_process_same_body_aliases): Create
18184 non-VAR_DECL node if it does not exist yet.
18186 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
18188 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
18189 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
18190 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
18191 target_cpu_default setting.
18193 2013-06-03 Teresa Johnson <tejohnson@google.com>
18195 * dumpfile.c (opt_info_switch_p): Change -fopt-info
18196 default to -fopt-info=optimized instead of all.
18197 * doc/invoke.texi: Ditto.
18198 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
18199 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
18200 (execute_vect_slp): Emit BB vectorization success under
18201 MSG_OPTIMIZED_LOCATIONS.
18202 * tree-vect-slp.c (vect_slp_transform_bb): Change
18203 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
18204 * tree-vect-loop.c (vect_transform_loop): Ditto.
18206 2013-06-03 Jason Merrill <jason@redhat.com>
18209 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
18210 Use TARGET_EXPR for C++.
18212 2013-06-03 Jakub Jelinek <jakub@redhat.com>
18214 PR rtl-optimization/57268
18215 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
18216 if DEBUG_INSN_P (insn).
18219 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18221 PR rtl-optimization/57268
18222 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18223 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18225 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18227 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
18228 (ix86_avoid_lea_for_addr): Likewise.
18229 (exact_dependency_1): Likewise.
18230 (ix86_adjust_cost): Likewise.
18231 (swap_top_of_ready_list): Fix formatting and !reload_completed check
18233 (do_reorder_for_imul): Fix typo, formatting and
18234 !reload_completed check removed.
18235 (ix86_sched_reorder): Fix typo and formatting.
18236 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
18239 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
18241 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
18243 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
18245 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
18246 <STRING_CST>: Likewise.
18247 <VECTOR_CST>: Likewise.
18249 2013-06-01 Janus Weil <janus@gcc.gnu.org>
18250 Mikael Morin <mikael@gcc.gnu.org>
18252 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
18253 * config.in: Regenerated.
18254 * configure: Regenerated.
18256 2013-06-01 Jan Hubicka <jh@suse.cz>
18258 PR middle-end/57366
18259 * cgraphunit.c (compile): When weakref is not supported,
18260 set up transparent aliases before final output pass.
18261 * varasm.c (assemble_alias): Do not try to do it here.
18263 2013-06-01 Jan Hubicka <jh@suse.cz>
18265 PR middle-end/57467
18266 * passes.c (for_per_function): Skip unanalyzed functions.
18268 2013-06-01 Jan Hubicka <jh@suse.cz>
18270 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
18271 (lto_symtab_merge_symbols_1): ... this one.
18272 (lto_symtab_merge_cgraph_nodes): Rename to ...
18273 (lto_symtab_merge_symbols): ... this one; simplify.
18274 * cgraph.c (same_body_aliases_done): Rename to ...
18275 (cpp_implicit_aliases_done): ... this one.
18276 (cgraph_create_function_alias): Update.
18277 (cgraph_same_body_alias): Update.
18278 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
18279 (verify_edge_corresponds_to_fndecl): Simplify.
18280 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
18281 (cgraph_node): Remove same_body_alias.
18282 (varpool_node): Remove alias_of and extra_name_alias.
18283 (same_body_aliases_done): Rename to ..
18284 (cpp_implicit_aliases_done): ... this one.
18285 (symtab_alias_ultimate_target): Add default parameter.
18286 (symtab_resolve_alias): New function.
18287 (fixup_same_cpp_alias_visibility): Declare.
18288 (cgraph_function_node): Add default parameter.
18289 (cgraph_node_asm_name): Likewise.
18290 (cgraph_function_or_thunk_node): Add default parameter; do
18291 not ICE when it is NULL.
18292 (varpool_variable_node): Likewise.
18293 * tree-emutls.c (create_emultls_var): Update.
18294 (ipa_lower_emutls): Update.
18295 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
18296 (cgraph_reset_node): Reset alias info.
18297 (cgraph_finalize_function): Update.
18298 (fixup_same_cpp_alias_visibility): Move to symtab.c.
18299 (analyze_function): Simplify.
18300 (cgraph_process_same_body_aliases): Simplify.
18301 (analyze_functions): Fixup same body aliases.
18302 (handle_alias_pairs): Simplify.
18303 (assemble_thunk): Update.
18304 (assemble_thunks_and_aliases): Update.
18305 (output_weakrefs): Rewrite.
18306 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
18307 (lto_output_varpool_node): Likewise.
18308 (compute_ltrans_boundary): Remve assert.
18309 (get_alias_symbol): New functoin.
18310 (input_node): Rewrite alias handling.
18311 (input_varpool_node): Likewise.
18312 * ipa-pure-const.c (propagate_pure_const): Fix formating.
18313 * ipa.c (process_references): Handle weakrefs correctly.
18314 (symtab_remove_unreachable_nodes): Likewise.
18315 * trans-mem.c (get_cg_data): Update.
18316 (ipa_tm_create_version_alias): Update.
18317 (ipa_tm_execute): Update.
18318 * symtab.c (dump_symtab_base): Dump aliases.
18319 (verify_symtab_base): Verify aliases.
18320 (symtab_node_availability): New function.
18321 (symtab_alias_ultimate_target): Simplify.
18322 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
18323 handle all the fixup cases.
18324 (symtab_resolve_alias): New function.
18325 * passes.c (ipa_write_summaries): Handle weakrefs.
18326 * varpool.c (varpool_analyze_node): Simplify.
18327 (assemble_aliases): Update.
18328 (varpool_create_variable_alias): Simplify.
18329 (varpool_extra_name_alias): Simplify.
18330 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
18331 (lto_symtab_merge_symbols): ... this one.
18333 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
18336 PR rtl-optimization/57268
18337 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18338 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18340 2013-06-01 Tobias Burnus <burnus@net-b.de>
18342 Partially reverted:
18343 2013-05-31 Tobias Burnus <burnus@net-b.de>
18345 PR middle-end/57073
18346 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
18349 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18351 PR rtl-optimization/57268
18352 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18353 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18355 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18357 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
18358 unordered comparison operators when -fno-trapping-math is in effect
18360 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
18361 and implement unordered comparison operators properly on the e500.
18363 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18365 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
18366 for constant scalar integers.
18367 (simplify_relational_operation_1): Likewise.
18369 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
18371 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
18372 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
18375 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18376 Igor Zamyatin <igor.zamyatin@intel.com>
18378 Silvermont (SLM) architecture performance tuning.
18379 * config/i386/i386.h (enum ix86_tune_indices): Add
18380 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
18381 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
18383 * config/i386/i386.c (initial_ix86_tune_features)
18384 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
18385 (ix86_lea_outperforms): Handle Silvermont tuning.
18386 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
18388 (ix86_use_lea_for_mov): Likewise.
18389 (ix86_avoid_lea_for_addr): Likewise.
18390 (ix86_lea_for_add_ok): Likewise.
18391 (exact_dependency_1): New function.
18392 (exact_store_load_dependency): Likewise.
18393 (ix86_adjust_cost): Handle Silvermont tuning.
18394 (do_reoder_for_imul): Likewise.
18395 (swap_top_of_ready_list): New function.
18396 (ix86_sched_reorder): Changed to handle Silvermont tuning.
18398 * config/i386/i386.md (peepholes that split memory operand in fp
18401 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18403 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
18404 Remove un-necessary braces.
18406 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18408 * config/aarch64/aarch64.c (aarch64_classify_symbol):
18409 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
18411 2013-05-31 Tobias Burnus <burnus@net-b.de>
18413 PR middle-end/57073
18414 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
18416 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18419 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
18420 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
18421 * config/arm/neon.md (iordi3_neon): Remove.
18422 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
18423 * config/arm/predicates.md (imm_for_neon_logic_operand):
18424 Move to earlier in the file.
18425 (neon_logic_op2): Likewise.
18426 (arm_iordi_operand_neon): New predicate.
18428 2013-05-31 Richard Biener <rguenther@suse.de>
18430 PR tree-optimization/57478
18431 PR tree-optimization/57453
18432 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
18435 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18437 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
18438 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
18440 2013-05-30 Tobias Burnus <burnus@net-b.de>
18441 Thomas Koenig <tkoenig@gcc.gnu.org>
18443 PR middle-end/57073
18444 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
18445 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
18447 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
18449 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
18451 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
18453 * target.def (register_usage_leveling_p): New hook.
18454 * targhooks.c (default_register_usage_leveling_p): New.
18455 * targhooks.h (default_register_usage_leveling_p): New prototype.
18456 * lra-assigns.c (register_usage_leveling_p): Use the hook.
18457 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
18458 * doc/tm.texi: Update.
18459 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
18461 2013-05-30 Ian Bolton <ian.bolton@arm.com>
18463 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
18464 (*insv_reg<mode>): New define_insn.
18466 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
18468 PR rtl-optimization/57439
18469 * postreload.c (move2add_valid_value_p): Check that we have
18470 a zero subreg_regno_offset when accessing the register in
18471 the requested mode.
18473 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18474 Igor Zamyatin <igor.zamyatin@intel.com>
18476 Silvermont (SLM) architecture pipeline model, tuning and
18478 * config.gcc: Add slm config options and target.
18480 * config/i386/slm.md: New.
18482 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
18484 * config/i386/i386-c.c (ix86_target_macros_internal): New case
18486 (ix86_target_macros_internal): Likewise.
18488 * config/i386/i386.c (slm_cost): New cost.
18489 (m_SLM): New macro flag.
18490 (initial_ix86_tune_features): Set m_SLM.
18491 (x86_accumulate_outgoing_args): Likewise.
18492 (x86_arch_always_fancy_math_387): Likewise.
18493 (processor_target_table): Add slm cost.
18494 (cpu_names): Add slm cpu name.
18495 (x86_option_override_internal): Set SLM ISA.
18496 (ix86_issue_rate): New case PROCESSOR_SLM.
18497 (ia32_multipass_dfa_lookahead): Likewise.
18498 (fold_builtin_cpu): Add slm.
18500 * config/i386/i386.h (TARGET_SLM): New target macro.
18501 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
18502 (processor_type): Add PROCESSOR_SLM.
18504 * config/i386/i386.md (cpu): Add new value "slm".
18505 (slm.md): Include slm.md.
18507 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
18508 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18510 * config/arm/arm-protos.h: Add and update function protos.
18511 * config/arm/arm.c (use_simple_return_p): New added.
18512 (thumb2_expand_return): Check simple_return flag.
18513 * config/arm/arm.md: Add simple_return and conditional simple_return.
18514 * config/arm/iterators.md: Add iterator for return and simple_return.
18516 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18518 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
18519 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
18520 (arm_emit_vfp_multi_reg_pop): Likewise.
18521 (thumb2_emit_ldrd_pop): Likewise.
18522 (arm_expand_epilogue): Add misc REG_CFA notes.
18523 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
18525 2013-05-29 Lawrence Crowl <crowl@google.com>
18527 * config/arm/t-arm: Update for below.
18529 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
18530 Change type to hash_table. Update dependent calls and types.
18532 * config/i386/t-cygming: Update for below.
18534 * config/i386/t-interix: Update for below.
18536 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
18537 Change type to hash_table. Update dependent calls and types.
18538 (i386_find_on_wrapper_list::wrappers): Likewise.
18540 * config/ia64/t-ia64: Update for below.
18542 * config/ia64/ia64.c (bundle_state_table):
18543 Change type to hash_table. Update dependent calls and types.
18545 * config/mips/mips.c (mips_reorg_process_insns::htab):
18546 Change type to hash_table. Update dependent calls and types.
18548 * config/sol2.c (solaris_comdat_htab):
18549 Change type to hash_table. Update dependent calls and types.
18551 * config/t-sol2: Update for above.
18553 2013-05-29 Teresa Johnson <tejohnson@google.com>
18555 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
18556 functions are not yet marked as defined.
18558 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
18559 Pat Haugen <pthaugen@us.ibm.com>
18560 Peter Bergner <bergner@vnet.ibm.com>
18562 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
18566 (vrotl<mode>3): Likewise.
18567 (vashl<mode>3): Likewise.
18568 (vlshr<mode>3): Likewise.
18569 (vashr<mode>3): Likewise.
18571 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18572 support for power8 V2DI builtins.
18574 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
18575 power8 V2DI builtins.
18576 (vupkhsw): Likewise.
18577 (vupklsw): Likewise.
18578 (vaddudm): Likewise.
18579 (vminsd): Likewise.
18580 (vmaxsd): Likewise.
18581 (vminud): Likewise.
18582 (vmaxud): Likewise.
18583 (vpkudum): Likewise.
18584 (vpksdss): Likewise.
18585 (vpkudus): Likewise.
18586 (vpksdus): Likewise.
18591 (vsubudm): Likewise.
18592 (vcmpequd): Likewise.
18593 (vcmpgtsd): Likewise.
18594 (vcmpgtud): Likewise.
18595 (vcmpequd_p): Likewise.
18596 (vcmpgtsd_p): Likewise.
18597 (vcmpgtud_p): Likewise.
18598 (vupkhsw): Likewise.
18599 (vupklsw): Likewise.
18600 (vaddudm): Likewise.
18601 (vmaxsd): Likewise.
18602 (vmaxud): Likewise.
18603 (vminsd): Likewise.
18604 (vminud): Likewise.
18605 (vpksdss): Likewise.
18606 (vpksdus): Likewise.
18607 (vpkudum): Likewise.
18608 (vpkudus): Likewise.
18613 (vsubudm): Likewise.
18615 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
18616 support for power8 V2DI instructions.
18618 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
18619 power8 V2DI instructions. Combine pack and unpack insns to use an
18620 iterator for each mode. Check whether a particular mode supports
18621 Altivec instructions instead of just checking TARGET_ALTIVEC.
18622 (UNSPEC_VPKUWUM): Likewise.
18623 (UNSPEC_VPKSHSS): Likewise.
18624 (UNSPEC_VPKSWSS): Likewise.
18625 (UNSPEC_VPKUHUS): Likewise.
18626 (UNSPEC_VPKSHUS): Likewise.
18627 (UNSPEC_VPKUWUS): Likewise.
18628 (UNSPEC_VPKSWUS): Likewise.
18629 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
18630 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
18631 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
18632 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
18633 (UNSPEC_VUPKHSB): Likewise.
18634 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
18635 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
18636 (UNSPEC_VUPKHSH): Likewise.
18637 (UNSPEC_VUPKLSB): Likewise.
18638 (UNSPEC_VUPKLSH): Likewise.
18640 (VI_char): Likewise.
18641 (VI_scalar): Likewise.
18642 (VI_unit): Likewise.
18644 (VP_small): Likewise.
18645 (VP_small_lc): Likewise.
18646 (VU_char): Likewise.
18647 (add<mode>3): Likewise.
18648 (altivec_vaddcuw): Likewise.
18649 (altivec_vaddu<VI_char>s): Likewise.
18650 (altivec_vadds<VI_char>s): Likewise.
18651 (sub<mode>3): Likewise.
18652 (altivec_vsubcuw): Likewise.
18653 (altivec_vsubu<VI_char>s): Likewise.
18654 (altivec_vsubs<VI_char>s): Likewise.
18655 (altivec_vavgs<VI_char>): Likewise.
18656 (altivec_vcmpbfp): Likewise.
18657 (altivec_eq<mode>): Likewise.
18658 (altivec_gt<mode>): Likewise.
18659 (altivec_gtu<mode>): Likewise.
18660 (umax<mode>3): Likewise.
18661 (smax<mode>3): Likewise.
18662 (umin<mode>3): Likewise.
18663 (smin<mode>3): Likewise.
18664 (altivec_vpkuhum): Likewise.
18665 (altivec_vpkuwum): Likewise.
18666 (altivec_vpkshss): Likewise.
18667 (altivec_vpkswss): Likewise.
18668 (altivec_vpkuhus): Likewise.
18669 (altivec_vpkshus): Likewise.
18670 (altivec_vpkuwus): Likewise.
18671 (altivec_vpkswus): Likewise.
18672 (altivec_vpks<VI_char>ss): Likewise.
18673 (altivec_vpks<VI_char>us): Likewise.
18674 (altivec_vpku<VI_char>us): Likewise.
18675 (altivec_vpku<VI_char>um): Likewise.
18676 (altivec_vrl<VI_char>): Likewise.
18677 (altivec_vsl<VI_char>): Likewise.
18678 (altivec_vsr<VI_char>): Likewise.
18679 (altivec_vsra<VI_char>): Likewise.
18680 (altivec_vsldoi_<mode>): Likewise.
18681 (altivec_vupkhsb): Likewise.
18682 (altivec_vupkhs<VU_char>): Likewise.
18683 (altivec_vupkls<VU_char>): Likewise.
18684 (altivec_vupkhsh): Likewise.
18685 (altivec_vupklsb): Likewise.
18686 (altivec_vupklsh): Likewise.
18687 (altivec_vcmpequ<VI_char>_p): Likewise.
18688 (altivec_vcmpgts<VI_char>_p): Likewise.
18689 (altivec_vcmpgtu<VI_char>_p): Likewise.
18690 (abs<mode>2): Likewise.
18691 (vec_unpacks_hi_v16qi): Likewise.
18692 (vec_unpacks_hi_v8hi): Likewise.
18693 (vec_unpacks_lo_v16qi): Likewise.
18694 (vec_unpacks_hi_<VP_small_lc>): Likewise.
18695 (vec_unpacks_lo_v8hi): Likewise.
18696 (vec_unpacks_lo_<VP_small_lc>): Likewise.
18697 (vec_pack_trunc_v8h): Likewise.
18698 (vec_pack_trunc_v4si): Likewise.
18699 (vec_pack_trunc_<mode>): Likewise.
18701 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
18703 (vec_vmaxsd): Likewise.
18704 (vec_vmaxud): Likewise.
18705 (vec_vminsd): Likewise.
18706 (vec_vminud): Likewise.
18707 (vec_vpksdss): Likewise.
18708 (vec_vpksdus): Likewise.
18709 (vec_vpkudum): Likewise.
18710 (vec_vpkudus): Likewise.
18711 (vec_vrld): Likewise.
18712 (vec_vsld): Likewise.
18713 (vec_vsrad): Likewise.
18714 (vec_vsrd): Likewise.
18715 (vec_vsubudm): Likewise.
18716 (vec_vupkhsw): Likewise.
18717 (vec_vupklsw): Likewise.
18719 2013-05-29 Jan Hubicka <jh@suse.cz>
18721 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
18722 flags; reorder rest of fields in more consistent way.
18723 (varpool_node): Remove analyzed, finalized and alias.
18724 (cgraph_ndoe): Likewise.
18725 (symtab_alias_ultimate_target): New function.
18726 (cgraph_function_node): Move offline.
18727 (cgraph_reset_node): Declare.
18728 (cgraph_comdat_can_be_unshared_p): Remove.
18729 (varpool_remove_initializer): Declare.
18730 (varpool_first_defined_variable, varpool_next_defined_variable
18731 cgraph_first_defined_function, cgraph_next_defined_function): Update.
18732 (cgraph_function_with_gimple_body_p): Update.
18733 (varpool_all_refs_explicit_p): Update.
18734 (symtab_alias_target): New function.
18735 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
18736 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
18737 (cgraph_function_or_thunk_node): Simplify using
18738 symtab_alias_ultimate_target.
18739 (varpool_variable_node): Likewise.
18740 * cgraph.c (cgraph_create_function_alias): Update.
18741 (cgraph_add_thunk): Update.
18742 (cgraph_remove_node): Update.
18743 (dump_cgraph_node): Do not dump removed flags.
18744 (cgraph_function_body_availability): Update.
18745 (cgraph_propagate_frequency): Update.
18746 (verify_cgraph_node): Check sanity of local flag.
18747 (cgraph_function_node): Move here from cgraph.h; revamp for
18748 cgraph_function_or_thunk_node.
18749 * lto-symtab.c (lto_varpool_replace_node): Update.
18750 (lto_symtab_resolve_can_prevail_p): Update.
18751 (lto_symtab_merge_cgraph_nodes): Update.
18752 * ipa-cp.c (determine_versionability, initialize_node_lattices,
18753 propagate_constants_accross_call, devirtualization_time_bonus,
18754 ipcp_propagate_stage): Update.
18755 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
18756 * ipa-inline-transform.c (clone_inlined_nodes,
18757 preserve_function_body_p): Update.
18758 * ipa-reference.c (propagate): Update.
18759 (write_node_summary_p): Update.
18760 * toplev.c (wrapup_global_declaration_2): Update.
18761 * cgraphunit.c (cgraph_analyze_function): Rename to ...
18762 (analyze_function) ... this one.
18763 (cgraph_process_new_functions): Update.
18764 (cgraph_reset_node): Export.
18765 (cgraph_finalize_function): Update.
18766 (cgraph_add_new_function): Update.
18767 (process_function_and_variable_attributes): Update.
18768 (varpool_finalize_decl): Update.
18769 (symbol_finalized): Remove.
18770 (symbol_finalized_and_needed): Rename to ...
18771 (symbol_defined_and_needed): ... update.
18772 (cgraph_analyze_functions): Update.
18773 (handle_alias_pairs): Update.
18774 (mark_functions_to_output): Update.
18775 (assemble_thunk): Update.
18776 (output_in_order): Update.
18777 (output_weakrefs): Update.
18778 (finalize_compilation_unit): Update.
18779 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
18780 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
18781 input_node, input_varpool_node): Update.
18782 * dbxout.c (dbxout_expand_expr): Update.
18783 * cgraphclones.c (cgraph_clone_node): Update.
18784 (cgraph_copy_node_for_versioning): Update.
18785 (cgraph_materialize_clone): Update.
18786 (cgraph_materialize_all_clones): Update.
18787 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
18788 propagate_pure_const, propagate_nothrow): Update.
18789 * lto-streamer-out.c (lto_output, write_symbol): Update.
18790 * ipa-utils.c (ipa_reverse_postorder): Update.
18791 * ipa-inline.c (can_inline_edge_p): Update.
18792 (update_caller_keys, ipa_inline): Update.
18793 * dwarf2out.c (reference_to_unused,
18794 premark_types_used_by_global_vars_helper): Update.
18795 * tree-eh.c (tree_could_trap_p): Update.
18796 * ipa-split.c (consider_split, execute_split_functions): Update.
18797 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
18798 has_addr_references_p): Update; move ahead in file for better
18800 (process_references): Simplify.
18801 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
18802 bodies are removed.
18803 (cgraph_comdat_can_be_unshared_p): Make static.
18804 (cgraph_externally_visible_p): Update.
18805 (varpool_externally_visible_p): Update.
18806 (function_and_variable_visibility): Update.
18807 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
18808 ipa_tm_mark_force_output_node): Update.
18809 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
18810 estimate_edge_devirt_benefit, inline_generate_summary,
18811 inline_write_summary): Update.
18812 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
18813 * ipa-prop.c (ipa_compute_jump_functions): Update.
18814 (ipa_print_node_params, ipa_prop_read_section,
18815 ipa_update_after_lto_read, read_replacements_section): Update.
18816 * varasm.c (mark_decl_referenced): Update.
18817 (assemble_alias, dump_tm_clone_pairs): Update.
18818 * tree-inline.c (copy_bb): Update.
18819 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
18821 * symtab.c (dump_symtab_base): Print new flags.
18822 (verify_symtab_base): Verify new flags.
18823 (symtab_alias_ultimate_target): New function.
18824 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
18825 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
18827 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
18829 * i386.c (ix86_get_function_versions_dispatcher,
18830 ix86_generate_version_dispatcher_body): Update.
18831 (fold_builtin_cpu): Use varpool_add_new_variable.
18832 * varpool.c (varpool_remove_initializer): Break out from ...
18833 (varpool_remove_node): ... this one.
18834 (dump_varpool_node, varpool_node_for_asm,
18835 cgraph_variable_initializer_availability, varpool_analyze_node,
18836 varpool_assemble_decl, varpool_remove_unreferenced_decls,
18837 varpool_finalize_named_section_flags, varpool_create_variable_alias):
18840 2013-05-29 Jan Hubicka <jh@suse.cz>
18842 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
18844 2013-05-29 Easwaran Raman <eraman@google.com>
18846 PR tree-optimization/57442
18847 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
18848 when control exits the main loop.
18850 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
18852 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
18854 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
18855 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
18856 * rx/t-rx: Add rx100 under multi library matches option for nofpu
18859 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18861 PR tree-optimization/57441
18862 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
18863 Don't limit size of incr_vec to number of candidates.
18865 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
18867 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
18868 and mips16 directories.
18869 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
18870 (MULTILIB_DIRNAMES): Ditto.
18871 (MULTILIB_EXCEPTIONS): Add new exceptions.
18872 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
18873 (MULTILIB_DIRNAMES): Ditto.
18874 (MULTILIB_EXCEPTIONS): Add new exceptions.
18876 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18877 Marcus Shawcroft <marcus.shawcroft@arm.com>
18879 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
18880 SYMBOL_TINY_ABSOLUTE.
18881 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
18882 SYMBOL_TINY_ABSOLUTE.
18883 (aarch64_expand_mov_immediate): Likewise.
18884 (aarch64_classify_symbol): Likewise.
18885 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
18886 Permit SYMBOL_TINY_ABSOLUTE.
18887 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
18889 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18890 Marcus Shawcroft <marcus.shawcroft@arm.com>
18892 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
18893 Refactor if/switch. Replace gcc_assert with if.
18895 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
18897 * config/i386/i386.c (initial_ix86_tune_features): Enable
18898 FP Reassociation for AMD bdver1 and bdver2.
18900 2013-05-29 Martin Jambor <mjambor@suse.cz>
18902 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
18903 and IMAGPART_EXPR do not occur within other handled_components.
18905 2013-05-29 Richard Biener <rguenther@suse.de>
18907 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
18908 access on whether the use is in the BB we currently try to
18910 (vect_bb_vectorization_profitable_p): Pass the BB we currently
18911 vectorize to vect_bb_slp_scalar_cost.
18913 2013-05-29 Richard Biener <rguenther@suse.de>
18915 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
18916 computing scalar cost offsetted by stmts that are kept live
18918 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
18919 for computation of scalar cost.
18921 2013-05-28 Steve Ellcey <sellcey@mips.com>
18923 * config/mips/mips-cpus.def (mips32r2): Change processor type.
18925 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
18927 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
18928 array notation built-in reduction functions.
18929 * doc/passes.texi (Passes): Added documentation about changes done
18931 * doc/invoke.texi (C Dialect Options): Added documentation about
18932 the -fcilkplus flag.
18933 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
18934 (BUILTINS_DEF): Depend on cilkplus.def.
18935 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
18936 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
18937 * cilkplus.def: New file.
18939 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
18941 PR rtl-optimization/57439
18942 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
18944 2013-05-28 Easwaran Raman <eraman@google.com>
18946 PR tree-optimization/57337
18947 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
18948 (find_insert_point): Correctly identify the insertion point
18949 when two statements with the same UID is compared.
18951 2013-05-28 Richard Biener <rguenther@suse.de>
18953 PR tree-optimization/56787
18954 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
18955 from the list of data references.
18956 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
18958 (vect_analyze_loop_operations): Likewise.
18959 (vect_transform_loop): Remove clobbers.
18961 2013-05-28 Martin Jambor <mjambor@suse.cz>
18963 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
18964 and REALPART_EXPRs have scalar type.
18966 2013-05-28 Richard Biener <rguenther@suse.de>
18968 PR tree-optimization/57411
18969 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
18971 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
18972 virtual operand propagation.
18974 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18976 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
18977 destination register for bmasksi_vis.
18978 (vector_init_bshuffle): Likewise.
18979 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
18981 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18983 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
18984 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
18985 mode if the instruction isn't available in the original mode.
18986 * config/sparc/sparc.opt (mfix-ut699): New option.
18987 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
18988 (divdf3): Turn into expander.
18989 (divdf3_nofix): New insn.
18990 (divdf3_fix): Likewise.
18991 (divsf3): Disable if -mfix-ut699.
18992 (sqrtdf2): Turn into expander.
18993 (sqrtdf2_nofix): New insn.
18994 (sqrtdf2_fix): Likewise.
18995 (sqrtsf2): Disable if -mfix-ut699.
18997 2013-05-27 Richard Biener <rguenther@suse.de>
18999 PR middle-end/57412
19000 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
19001 block for the new loop.
19003 2013-05-27 Richard Biener <rguenther@suse.de>
19005 PR tree-optimization/57343
19006 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
19007 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
19008 (number_of_iterations_cond): Do not build the folded tree.
19010 2013-05-27 Richard Biener <rguenther@suse.de>
19013 PR middle-end/57381
19014 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19015 OEP_CONSTANT_ADDRESS_OF retained.
19017 PR tree-optimization/57417
19018 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
19019 for unchanged base.
19020 (set_ssa_val_to): Compare addresses using
19021 get_addr_base_and_unit_offset.
19023 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
19025 PR rtl-optimization/56833
19026 * postreload.c (move2add_record_mode): New function.
19027 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
19028 (move2add_use_add2_insn): Use move2add_record_sym_value.
19029 (move2add_use_add3_insn): Likewise.
19030 (reload_cse_move2add): Use move2add_valid_value_p and
19031 move2add_record_mode. Invalidate call-clobbered and REG_INC
19032 affected regs by setting reg_mode to VOIDmode.
19033 (move2add_note_store): Don't pretend the inside of a SUBREG is
19034 the actual destination. Invalidate single/leading registers by
19035 setting reg_mode to VOIDmode.
19036 Use move2add_record_sym_value, move2add_valid_value_p and
19037 move2add_record_mode.
19039 2013-05-27 Richard Biener <rguenther@suse.de>
19041 PR tree-optimization/57396
19042 * tree-affine.c (double_int_constant_multiple_p): Properly
19043 return false for val == 0 and div != 0.
19045 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19047 * config/mips/mips.h: Use #elif in preprocessor conditions.
19049 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19052 * config/mips/constraints.md (kl): New constraint.
19053 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
19054 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
19055 constraint for operand 0. Split after CSE for MIPS16. Emit a move
19056 from LO for MIPS16.
19057 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
19059 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19062 * config/mips/mips.c (mips_can_inline_p): New function.
19063 (TARGET_CAN_INLINE_P): Define.
19065 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
19067 * sched-int.h (ds_t, dw_t): Make unsigned int.
19068 Fix documentation that describes how all the ds_t bits are used.
19069 Reserve the last bit for delayed-branch scheduling.
19070 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
19071 (BITS_PER_DEP_WEAK): Fix definition and documentation.
19072 (gen_dep_weak_1): Remove prototype.
19073 * sched-deps.c (get_dep_weak_1): Make static.
19074 * target.def (speculate_insn, needs_block_p, gen_spec_check,
19075 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
19076 * doc/tm.texi: Regenerate.
19077 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
19079 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
19082 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
19084 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
19085 Sandra Loosemore <sandra@codesourcery.com>
19087 * config.gcc (powerpc-*): Allow native for with-cpu.
19089 2013-05-24 Jeff Law <law@redhat.com>
19091 PR tree-optimization/57124
19092 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
19093 conversion feeding a condition if the range has an overflow
19094 if -fstrict-overflow. Add warnings for when we do make the
19097 2013-05-24 Dehao Chen <dehao@google.com>
19099 * tree-cfg.c (locus_discrim_map): Fix the typo.
19100 (locus_discrim_hasher): Likewise.
19101 (locus_discrim_hasher::hash): Likewise.
19102 (locus_discrim_hasher::equal): Likewise.
19104 2013-05-24 Martin Jambor <mjambor@suse.cz>
19106 PR tree-optimization/57294
19107 * cgraph.h (ipa_record_stmt_references): Declare.
19108 * cgraphbuild.c (ipa_record_stmt_references): New function.
19109 (build_cgraph_edges): Use ipa_record_stmt_references.
19110 (rebuild_cgraph_edges): Likewise.
19111 (cgraph_rebuild_references): Likewise.
19112 * ipa-prop.c (ipa_modify_call_arguments): Discard references
19113 associated with the old statement and build references from the
19114 newly built statements.
19115 * ipa-ref.c (ipa_remove_stmt_references): New function.
19116 * ipa-ref.h (ipa_remove_stmt_references): Declare.
19118 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
19120 * lra-constraints.c (emit_spill_move): Use smaller mode for
19122 (check_and_process_move): Consider mem-reg moves for secondary
19124 (curr_insn_transform): Don't lose insns emitted before for
19125 secondary memory moves.
19126 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
19127 reg set up in the current insn.
19129 2013-05-24 Dehao Chen <dehao@google.com>
19131 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
19133 (locus_descrim_hasher::equal): Likewise.
19134 (build_gimple_cfg): New discriminator assignment algorithm.
19135 (make_edges): Likewise.
19136 (next_discriminator_for_locus): Likewise.
19137 (same_line_p): Likewise.
19138 (assign_discriminators): Likewise.
19139 (make_cond_expr_edges): Likewise.
19140 (make_gimple_switch_edges): Likewise.
19141 (make_goto_expr_edges): Likewise.
19142 (make_gimple_asm_edges): Likewise.
19144 2013-05-24 Ian Bolton <ian.bolton@arm.com>
19146 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
19147 X format specifier to only display bottom 16 bits.
19148 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
19149 immediate to match for operand 2, since it will be masked.
19151 2013-05-24 Richard Biener <rguenther@suse.de>
19153 PR tree-optimization/57287
19154 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
19155 all SSA names that occur in abnormal PHIs.
19157 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
19160 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
19161 that index is not negative.
19163 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
19165 PR rtl-optimization/55177
19166 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
19167 (simplify_byte_swapping_operation): New.
19168 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
19169 (simplify_relational_operation_1): Deal with BSWAP.
19171 2013-05-23 Richard Henderson <rth@redhat.com>
19174 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
19175 (ix86_reorg): Call it.
19177 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
19180 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
19181 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
19182 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
19184 2013-05-23 Christian Bruel <christian.bruel@st.com>
19187 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
19189 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19190 Marcus Shawcroft <marcus.shawcroft@arm.com>
19192 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
19193 * config/aarch64/constraints.md (Usa): Remove.
19194 * doc/md.texi (AArch64 Usa): Remove.
19196 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19197 Marcus Shawcroft <marcus.shawcroft@arm.com>
19199 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
19200 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
19201 * config/aarch64/predicates.md (aarch64_const_address): Remove.
19202 (aarch64_mov_operand): Use aarch64_mov_operand_p.
19204 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
19206 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
19207 instruction (AdvSIMD).
19208 * config/aarch64/aarch64-builtins.c
19209 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
19210 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
19212 2013-05-23 Martin Jambor <mjambor@suse.cz>
19214 PR middle-end/57347
19215 * tree.h (contains_bitfld_component_ref_p): Declare.
19216 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
19217 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
19219 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
19220 not access a bit-field. Assert all final offsets are byte-aligned.
19222 2013-05-23 Richard Biener <rguenther@suse.de>
19224 PR tree-optimization/57380
19225 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
19226 least one invariant or re-used load.
19227 * passes.c (init_optimization_passes): Move pass_phiprop before
19230 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
19232 * config/aarch64/aarch64-simd.md
19233 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
19235 2013-05-23 Richard Biener <rguenther@suse.de>
19237 PR middle-end/57381
19238 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19239 OEP_CONSTANT_ADDRESS_OF retained.
19241 2013-05-23 Jakub Jelinek <jakub@redhat.com>
19243 PR middle-end/57344
19244 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
19245 don't lower unit. Handle unit not being always BITS_PER_WORD.
19247 2013-05-23 Richard Biener <rguenther@suse.de>
19249 PR rtl-optimization/57341
19250 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
19251 instead of true_dependence.
19253 2013-05-22 David Malcolm <dmalcolm@redhat.com>
19255 * bb-reorder.c (branch_threshold): Make const.
19256 (exec_threshold): Ditto.
19258 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19259 Pat Haugen <pthaugen@us.ibm.com>
19260 Peter Bergner <bergner@vnet.ibm.com>
19262 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
19263 documentation for the power8 crypto builtins.
19265 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
19267 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
19268 macros for defining power8 builtin functions.
19269 (BU_P8V_AV_2): Likewise.
19270 (BU_P8V_AV_P): Likewise.
19271 (BU_P8V_VSX_1): Likewise.
19272 (BU_P8V_OVERLOAD_1): Likewise.
19273 (BU_P8V_OVERLOAD_2): Likewise.
19274 (BU_CRYPTO_1): Likewise.
19275 (BU_CRYPTO_2): Likewise.
19276 (BU_CRYPTO_3): Likewise.
19277 (BU_CRYPTO_OVERLOAD_1): Likewise.
19278 (BU_CRYPTO_OVERLOAD_2): Likewise.
19279 (XSCVSPDP): Fix typo, point to the correct instruction.
19280 (VCIPHER): Add power8 crypto builtins.
19281 (VCIPHERLAST): Likewise.
19282 (VNCIPHER): Likewise.
19283 (VNCIPHERLAST): Likewise.
19284 (VPMSUMB): Likewise.
19285 (VPMSUMH): Likewise.
19286 (VPMSUMW): Likewise.
19287 (VPERMXOR_V2DI): Likewise.
19288 (VPERMXOR_V4SI: Likewise.
19289 (VPERMXOR_V8HI: Likewise.
19290 (VPERMXOR_V16QI: Likewise.
19291 (VSHASIGMAW): Likewise.
19292 (VSHASIGMAD): Likewise.
19293 (VPMSUM): Likewise.
19294 (VPERMXOR): Likewise.
19295 (VSHASIGMA): Likewise.
19297 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
19298 __CRYPTO__ if the crypto instructions are available.
19299 (altivec_overloaded_builtins): Add support for overloaded power8
19302 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
19303 support for power8 crypto builtins.
19304 (builtin_function_type): Likewise.
19305 (altivec_init_builtins): Add support for builtins that take vector
19306 long long (V2DI) arguments.
19308 * config/rs6000/crypto.md: New file, define power8 crypto
19311 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19312 Pat Haugen <pthaugen@us.ibm.com>
19313 Peter Bergner <bergner@vnet.ibm.com>
19315 * doc/invoke.texi (Option Summary): Add power8 options.
19316 (RS/6000 and PowerPC Options): Likewise.
19318 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
19319 constraints.md instead of rs6000.h. Reorder w* constraints. Add
19320 wm, wn, wr documentation.
19322 * config/rs6000/constraints.md (wm): New constraint for VSX
19323 registers if direct move instructions are enabled.
19324 (wn): New constraint for no registers.
19325 (wq): New constraint for quad word even GPR registers.
19326 (wr): New constraint if 64-bit instructions are enabled.
19327 (wv): New constraint if power8 vector instructions are enabled.
19328 (wQ): New constraint for quad word memory locations.
19330 * config/rs6000/predicates.md (const_0_to_15_operand): New
19331 constraint for 0..15 for crypto instructions.
19332 (gpc_reg_operand): If VSX allow registers in VSX registers as well
19333 as GPR and floating point registers.
19334 (int_reg_operand): New predicate to match only GPR registers.
19335 (base_reg_operand): New predicate to match base registers.
19336 (quad_int_reg_operand): New predicate to match even GPR registers
19337 for quad memory operations.
19338 (vsx_reg_or_cint_operand): New predicate to allow vector logical
19339 operations in both GPR and VSX registers.
19340 (quad_memory_operand): New predicate for quad memory operations.
19341 (reg_or_indexed_operand): New predicate for direct move support.
19343 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
19344 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
19345 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
19346 (POWERPC_MASKS): Add power8 options.
19347 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
19350 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
19351 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
19353 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
19354 (-mpower8-fusion): New power8 options.
19355 (-mpower8-fusion-sign): Likewise.
19356 (-mpower8-vector): Likewise.
19357 (-mcrypto): Likewise.
19358 (-mdirect-move): Likewise.
19359 (-mquad-memory): Likewise.
19361 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
19362 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
19364 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
19365 (rs6000_debug_vector_unit): Add p8_vector.
19366 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
19367 definitions. Also print fusion state.
19368 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
19369 (rs6000_builtin_mask_calculate): Add power8 builtin support.
19370 (rs6000_option_override_internal): Add support for power8.
19371 (rs6000_common_init_builtins): Add debugging for skipped builtins
19372 if -mdebug=builtin.
19373 (rs6000_adjust_cost): Add power8 support.
19374 (rs6000_issue_rate): Likewise.
19375 (insn_must_be_first_in_group): Likewise.
19376 (insn_must_be_last_in_group): Likewise.
19377 (force_new_group): Likewise.
19378 (rs6000_register_move_cost): Likewise.
19379 (rs6000_opt_masks): Likewise.
19381 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
19382 power8 capable assembler, default to power7 options.
19383 (TARGET_DIRECT_MOVE): Likewise.
19384 (TARGET_CRYPTO): Likewise.
19385 (TARGET_P8_VECTOR): Likewise.
19386 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
19387 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
19388 (VECTOR_MEM_P8_VECTOR_P): Likewise.
19389 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
19390 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
19391 (TARGET_XSCVDPSPN): Likewise.
19392 (TARGET_XSCVSPDPN): Likewsie.
19393 (TARGET_SYNC_HI_QI): Likewise.
19394 (TARGET_SYNC_TI): Likewise.
19395 (MASK_CRYPTO): Likewise.
19396 (MASK_DIRECT_MOVE): Likewise.
19397 (MASK_P8_FUSION): Likewise.
19398 (MASK_P8_VECTOR): Likewise.
19399 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
19400 temporary used by some of the direct move instructions to get two FP
19401 temporary registers does not force creation of a stack frame.
19402 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
19403 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
19404 that any VSX registers are tieable, even if they are also an
19405 Altivec vector mode.
19406 (r6000_reg_class_enum): Add wm, wr, wv constraints.
19407 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
19408 (RS6000_BTM_CRYPTO): Likewise.
19409 (RS6000_BTM_COMMON): Likewise.
19411 * config/rs6000/rs6000.md (cpu attribute): Add power8.
19412 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
19413 (enum rs6000_vector): Add power8 vector support.
19415 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19419 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
19420 (any_sibcall_could_use_r3): this and handle indirect calls.
19421 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
19423 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19425 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
19427 2013-05-22 Richard Biener <rguenther@suse.de>
19429 PR middle-end/57349
19430 * profile.c (branch_prob): Do not split blocks that are
19431 abnormally receiving from ECF_RETURNS_TWICE functions.
19433 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19435 * recog.c (offsettable_address_addr_space_p): Fix calculation of
19436 address mode. Move pointer mode initialization to the same place.
19438 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
19440 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
19441 while it has any effect.
19443 2013-05-21 Easwaran Raman <eraman@google.com>
19445 PR tree-optimization/57322
19446 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
19447 UID of the statement added to the BB to be 1.
19449 2013-05-21 Jakub Jelinek <jakub@redhat.com>
19451 PR tree-optimization/57331
19452 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
19453 of conversion from pointer type to integral type with integer.
19455 2013-05-21 Martin Jambor <mjambor@suse.cz>
19458 * ipa-prop.c (ipa_read_node_info): Process param_used and
19459 controlled_uses in the same order as when writing.
19461 2013-05-21 Magnus Granberg <baldrick@free.fr>
19464 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
19466 2013-05-21 Richard Biener <rguenther@suse.de>
19468 PR tree-optimization/57318
19469 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
19470 estimate stmts with side-effects as likely eliminated.
19472 2013-05-21 Richard Biener <rguenther@suse.de>
19474 PR tree-optimization/57330
19475 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
19476 preserve the call stmts fntype.
19478 2013-05-21 Richard Biener <rguenther@suse.de>
19480 PR tree-optimization/57303
19481 * tree-ssa-sink.c (statement_sink_location): Improve killing
19482 stmt detection and properly handle self-assignments.
19484 2013-05-21 Christian Bruel <christian.bruel@st.com>
19486 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
19487 spanning registers. LEAF_REG_REMAP is supported only for contiguous
19488 registers. Set register size out of the PARALLEL loop.
19490 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
19493 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
19494 (*fmasf4, *fmasf4_media): New insns.
19496 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19498 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
19499 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
19500 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
19501 (mips_idiv_insns): Update the comments to say that the returned
19502 instruction counts are in units of BASE_INSN_LENGTH.
19503 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
19504 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
19505 using 2 rather than 4 as the length of indirect MIPS16 and
19506 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
19507 length of a NOP. Don't divide MIPS16 lengths by 2.
19508 (mips16_split_long_branches): Assume a branch is long if the
19509 length is greater than 4 rather than 8.
19510 * config/mips/mips.md (length): Give MIPS16 lengths directly,
19511 rather than multiplying them by 2. Multiply instruction counts
19512 by BASE_INSN_LENGTH rather than 4.
19513 (*jump_mips16, tls_get_tp_mips16_<mode>)
19514 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
19516 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19518 * config/mips/mips.md (extended_mips16): Remove branch case.
19519 (length): Remove duplicated extended_mips16 test.
19521 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19523 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
19525 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
19527 * recog.h (Recog_data): Rename to...
19528 (recog_data_d): ...this.
19529 (recog_data): Update accordingly.
19530 * recog.c (recog_data): Likewise.
19531 * reload.c (save_recog_data): Likewise.
19532 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
19533 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
19535 2013-05-17 Julian Brown <julian@codesourcery.com>
19537 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
19538 found in a REG_EQUAL note, invalidate it.
19540 2013-05-17 Easwaran Raman <eraman@google.com>
19542 * tree-ssa-reassoc.c (find_insert_point): New function.
19543 (insert_stmt_after): Likewise.
19544 (get_def_stmt): Likewise.
19545 (ensure_ops_are_available): Likewise.
19546 (not_dominated_by): Likewise.
19547 (rewrite_expr_tree): Do not move statements beyond what is
19548 necessary. Remove call to swap_ops_for_binary_stmt...
19549 (reassociate_bb): ... and move it here.
19550 (build_and_add_sum): Assign UIDs for new statements.
19551 (linearize_expr): Likewise.
19552 (do_reassoc): Renumber gimple statement UIDs.
19554 2013-05-17 Jan Hubicka <jh@suse.cz>
19556 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
19558 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
19559 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
19561 (output_weakrefs): Update.
19563 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
19564 Martin Jambor <mjambor@suse.cz>
19566 PR middle-end/57276
19567 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
19568 value that corresponds to the given aggval is found in values vector.
19570 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
19572 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
19573 sse, sse2, sse3, ssse3 and sse4a flags to options.
19575 2013-05-17 David Malcolm <dmalcolm@redhat.com>
19577 * gengtype-state.c: (s_expr_writer): New class, to handle
19578 prettifying of output layout of s-expressions.
19579 (state_writer): New class, to write out gtype.state.
19580 (state_written_type_count): Move this variable into member data of
19582 (s_expr_writer::s_expr_writer): New code: constructor for new class
19583 (state_writer::state_writer(): ditto
19584 (s_expr_writer::write_new_line): New function
19585 (s_expr_writer::write_any_indent): ditto
19586 (s_expr_writer::begin_s_expr): ditto
19587 (s_expr_writer::end_s_expr): ditto
19588 (write_state_fileloc): convert to method of state_writer...
19589 (state_writer:: write_state_fileloc): ...and use methods of
19590 s_expr_writer to write indentation into the gtype.state output file
19591 to visually represent the hierarchical structure of the list
19593 (write_state_fields): ditto, renaming to...
19594 (state_writer::write_state_fields)
19595 (write_state_a_string): ditto, renaming to...
19596 (state_writer::write_state_a_string)
19597 (write_state_string_option): ditto, renaming to...
19598 (state_writer::write_state_string_option)
19599 (write_state_type_option): ditto, renaming to...
19600 (state_writer::write_state_type_option)
19601 (write_state_nested_option): ditto, renaming to...
19602 (state_writer::write_state_nested_option)
19603 (write_state_option): ditto, renaming to...
19604 (state_writer::write_state_option)
19605 (write_state_options): ditto, renaming to...
19606 (state_writer::write_state_options)
19607 (write_state_lang_bitmap): ditto, renaming to...
19608 (state_writer::write_state_lang_bitmap)
19609 (write_state_version): ditto, renaming to...
19610 (state_writer::write_state_version)
19611 (write_state_scalar_type): ditto, renaming to...
19612 (state_writer::write_state_scalar_type)
19613 (write_state_string_type): ditto, renaming to...
19614 (state_writer::write_state_string_type)
19615 (write_state_undefined_type): ditto, renaming to...
19616 (state_writer::write_state_undefined_type)
19617 (write_state_struct_union_type): ditto, renaming to...
19618 (state_writer::write_state_struct_union_type)
19619 (write_state_struct_type): ditto, renaming to...
19620 (state_writer::write_state_struct_type)
19621 (write_state_user_struct_type): ditto, renaming to...
19622 (state_writer::write_state_user_struct_type)
19623 (write_state_lang_struct_type): ditto, renaming to...
19624 (state_writer::write_state_lang_struct_type)
19625 (write_state_param_struct_type): ditto, renaming to...
19626 (state_writer::write_state_param_struct_type)
19627 (write_state_pointer_type): ditto, renaming to...
19628 (state_writer::write_state_pointer_type)
19629 (write_state_array_type): ditto, renaming to...
19630 (state_writer::write_state_array_type)
19631 (write_state_gc_used): ditto, renaming to...
19632 (state_writer::write_state_gc_used)
19633 (write_state_common_type_content): ditto, renaming to...
19634 (state_writer::write_state_common_type_content)
19635 (write_state_type): ditto, renaming to...
19636 (state_writer::write_state_type)
19637 (write_state_pair_list): ditto, renaming to...
19638 (state_writer::write_state_pair_list)
19639 (write_state_pair): ditto, renaming to...
19640 (state_writer::write_state_pair)
19641 (write_state_typedefs): ditto, renaming to...
19642 (state_writer::write_state_typedefs)
19643 (write_state_structures): ditto, renaming to...
19644 (state_writer::write_state_structures)
19645 (write_state_param_structs): ditto, renaming to...
19646 (state_writer::write_state_param_structs)
19647 (write_state_variables): ditto, renaming to...
19648 (state_writer::write_state_variables)
19649 (write_state_srcdir): ditto, renaming to...
19650 (state_writer::write_state_srcdir)
19651 (write_state_files_list): ditto, renaming to...
19652 (state_writer::write_state_files_list)
19653 (write_state_languages): ditto, renaming to...
19654 (state_writer::write_state_languages)
19655 (write_state): create a state_writer instance and use it when
19656 writing out the state file
19658 2013-05-17 Mike Stump <mikestump@comcast.net>
19660 PR rtl-optimization/57304
19661 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
19662 accessing DF_REF_REAL_LOC.
19664 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19666 PR rtl-optimization/57281
19667 PR rtl-optimization/57300
19668 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
19669 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
19670 what the other splitter did if the registers are dead.
19672 2013-05-17 Richard Biener <rguenther@suse.de>
19674 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
19677 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19679 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
19682 2013-05-17 Marek Polacek <polacek@redhat.com>
19684 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
19685 length when doing non-zero store of storing '\0' to '\0'.
19687 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19689 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
19690 vect_external_def oprnd1 with loop_vinfo, try to emit
19691 optional cast, negation and and stmts on the loop preheader
19692 edge instead of into the pattern def seq.
19694 PR tree-optimization/57051
19695 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
19696 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
19698 2013-05-16 Nick Clifton <nickc@redhat.com>
19700 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
19701 (rl78_is_naked_func): New function.
19702 (rl78_expand_prologue): Skip prologue generation for naked functions.
19703 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
19704 * doc/extend.texi (naked): Add RL78 to the list of processors
19705 that supports this attribute.
19707 2013-05-16 Jeff Law <law@redhat.com>
19709 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
19711 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19713 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
19714 cache parameters using detect_caches_amd also for CYRIX,
19715 NSC and TM2 signatures.
19717 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19718 Dzianis Kahanovich <mahatma@eu.by>
19722 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
19723 VIA/Centaur processors and determine their cache parameters
19724 using detect_caches_amd.
19726 2013-05-16 Teresa Johnson <tejohnson@google.com>
19728 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
19729 (rtl_verify_edges): New function.
19730 (rtl_verify_bb_insns): Ditto.
19731 (rtl_verify_bb_pointers): Ditto.
19732 (rtl_verify_bb_insn_chain): Ditto.
19733 (rtl_verify_fallthru): Ditto.
19734 (rtl_verify_bb_layout): Ditto.
19735 (rtl_verify_flow_info_1): Outline checks into new functions.
19736 (rtl_verify_flow_info): Ditto.
19738 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
19740 * cfghooks.c (copy_bbs): Add update_dominance argument.
19741 * cfghooks.h (copy_bbs): Update prototype.
19742 * tree-cfg.c (gimple_duplicate_sese_region):
19743 Add update_dominance argument.
19744 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
19745 * tree-ssa-loop-ch.c (copy_loop_headers): Update
19746 gimple_duplicate_sese_region call.
19747 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
19748 Update copy_bbs call.
19749 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
19750 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
19752 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19754 * tree-vectorizer.h (NUM_PATTERNS): Increment.
19755 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
19756 vect_recog_rotate_pattern.
19757 (vect_recog_rotate_pattern): New function.
19759 2013-05-16 Jason Merrill <jason@redhat.com>
19761 * Makefile.in (LLINKER): New variable.
19762 (mostlyclean): Remove link mutex.
19763 * configure.ac: Handle --enable-link-mutex.
19764 * lock-and-run.sh: New script.
19766 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19769 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
19772 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19774 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
19776 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
19778 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
19779 * config/arm/arm.c (next_consecutive_mem): New function.
19780 (gen_movmem_ldrd_strd): Likewise.
19781 * config/arm/arm.md (movmemqi): Update condition and code.
19782 (unaligned_loaddi, unaligned_storedi): New patterns.
19784 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19786 * config.gcc: Obsolete *-*-solaris2.9*.
19787 * doc/install.texi (Specific, *-*-solaris2*): Document it.
19789 2013-05-16 Richard Biener <rguenther@suse.de>
19791 * passes.c (init_optimization_passes): Move pass_parallelize_loops
19792 earlier, after GRAPHITE transforms and IV canonicalization.
19794 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19796 * omp-low.c (extract_omp_for_data): For collapsed loops,
19797 if at least one of the loops is known at compile time to
19798 iterate zero times, set count to 0.
19799 (expand_omp_regimplify_p): New function.
19800 (expand_omp_for_generic): For collapsed loops, if at least
19801 one of the loops isn't known to iterate at least once,
19802 add runtime check with setting count to 0.
19803 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
19804 For unsigned types if it isn't known at compile time that
19805 the loop will iterate at least once, add runtime check to bypass
19806 the whole loop if initial condition isn't true.
19808 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
19810 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
19812 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
19814 PR middle-end/57286
19815 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
19816 transformations to avoid an infinite loop.
19818 2013-05-16 Marek Polacek <polacek@redhat.com>
19820 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
19822 2013-05-15 Leif Ekblad <leif@rdos.net>
19824 * config/i386/i386.c (ix86_decompose_address): Use
19825 DEFAULT_TLS_SEG_REG to access TLS segment register.
19826 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
19827 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
19828 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
19830 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
19833 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
19834 sibling calls to functions that would normally be lazily bound,
19835 unless $gp is call-clobbered.
19837 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
19839 * config/i386/i386.c (ix86_option_override_internal): Update
19840 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
19841 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
19842 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
19843 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
19845 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
19847 2013-05-15 Andreas Schwab <schwab@suse.de>
19849 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
19850 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
19853 2013-05-15 Teresa Johnson <tejohnson@google.com>
19855 * loop-unroll.c (report_unroll_peel): Check decision before
19856 emitting unroll/peel message.
19858 2013-05-15 Teresa Johnson <tejohnson@google.com>
19860 * function.h (has_bb_partition): New rtl_data flag.
19861 (bb_reorder_complete): Ditto.
19862 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
19863 instead of flag_reorder_blocks_and_partition.
19864 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
19865 with some enhancements.
19866 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
19867 * bb-reorder.c (connect_traces): Check for has_bb_partition
19868 instead of flag_reorder_blocks_and_partition.
19869 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
19870 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
19871 verify_hot_cold_block_grouping.
19872 (partition_hot_cold_basic_blocks): Set has_bb_partition.
19874 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19877 * config/arm/predicates.md (call_insn_operand): New predicate.
19878 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
19879 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
19880 if insn is not a tail call.
19881 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
19883 * config/arm/arm.h (enum reg_class): New caller save register class.
19884 (REG_CLASS_NAMES): Likewise.
19885 (REG_CLASS_CONTENTS): Likewise.
19886 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
19889 2013-05-15 Richard Biener <rguenther@suse.de>
19891 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
19892 of MSG_OPTIMIZED_LOCATIONS.
19893 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
19894 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
19896 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
19897 of MSG_OPTIMIZED_LOCATIONS.
19898 (execute_vect_slp): Likewise.
19899 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
19900 (vect_create_cond_for_alias_checks): Likewise.
19901 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
19902 (vect_recog_widen_mult_pattern): Likewise.
19903 (vect_recog_widen_sum_pattern): Likewise.
19904 (vect_recog_over_widening_pattern): Likewise.
19905 (vect_recog_widen_shift_pattern): Likewise.
19906 (vect_recog_vector_vector_shift_pattern): Likewise.
19907 (vect_recog_divmod_pattern): Likewise.
19908 (vect_recog_mixed_size_cond_pattern): Likewise.
19909 (vect_recog_bool_pattern): Likewise.
19910 (vect_pattern_recog_1): Likewise.
19912 2013-05-15 Martin Jambor <mjambor@suse.cz>
19914 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
19915 non-functions to builtin_unreachable.
19916 * ipa-inline-transform.c (inline_call): Do not assert estimates were
19917 correct when new direct edges were discovered.
19919 2013-05-15 Martin Jambor <mjambor@suse.cz>
19921 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
19922 header, print symbol order instead of node uid, print more information
19923 about indirect edge targets.
19924 (ipa_make_edge_direct_to_target): Print symbol order instead of node
19926 (ipa_make_edge_direct_to_target): Likewise.
19927 (remove_described_reference): Likewise.
19928 (propagate_controlled_uses): Likewise.
19929 (ipa_print_node_params): Also print symbol order.
19930 (ipcp_transform_function): Print symbol order instead of node uids.
19931 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
19932 (cgraph_get_create_real_symbol_node): Likewise.
19933 * ipa-cp.c (print_lattice): Likewise.
19934 (print_all_lattices): Likewise.
19935 (determine_versionability): Likewise.
19936 (initialize_node_lattices): Likewise.
19937 (estimate_local_effects): Likewise.
19938 (update_profiling_info): Likewise.
19939 (create_specialized_node): Likewise.
19940 (perhaps_add_new_callers): Likewise.
19941 (decide_about_value): Likewise.
19942 (decide_whether_version_node): Likewise.
19943 (identify_dead_nodes): Likewise.
19944 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
19945 (dump_inline_summary): Likewise.
19946 (estimate_node_size_and_time): Likewise.
19947 (inline_analyze_function): Likewise.
19948 * ipa-inline.c (report_inline_failed_reason): Likewise.
19949 (want_early_inline_function_p): Likewise.
19950 (edge_badness): Likewise.
19951 (update_edge_key): Likewise.
19952 (inline_small_functions): Likewise. Add dumping of order to two other
19954 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
19955 instead of node uids.
19956 (propagate_pure_const): Likewise.
19957 (propagate_pure_const): Likewise.
19958 * ipa-utils.c (dump_cgraph_node_set): Likewise.
19959 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
19960 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
19962 * tree-pretty-print.c (dump_function_header): Likewise.
19963 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
19964 Print symbol order instead of node uids.
19966 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19968 * config/s390/s390.c (s390_register_move_cost): Don't impose the
19969 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
19971 2013-05-15 Richard Biener <rguenther@suse.de>
19973 PR tree-optimization/57275
19974 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
19975 return value for fail to do runtime alias checks for gather loads.
19977 2013-05-15 Jan Hubicka <jh@suse.cz>
19981 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
19982 weakrefs are not external.
19983 (lto_symtab_merge_decls): Fix thinko when dealing with
19984 non-lto_symtab decls.
19985 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
19986 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
19987 * varpool.c (dump_varpool_node): Dump more flags.
19989 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19991 * config/i386/i386.c (processor_alias_table): Add instruction
19992 FSGSBASE for AMD bdver3 architecture.
19994 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19996 * tree.c (warn_deprecated_use): Print file:line using locus color.
19997 * diagnostic.c (diagnostic_report_current_module): Print file:line
19998 and file:line:column using locus color.
20000 2013-05-14 Mike Stump <mikestump@comcast.net>
20002 * gdbinit.in: Add __null.
20004 2013-05-14 Mike Stump <mikestump@comcast.net>
20006 * recog.h: Rename struct recog_data to Recog_data.
20007 * recog.c: Likewise.
20008 * reload.c (can_reload_into): Likewise.
20009 * config/picochip/picochip.c: Likewise.
20011 2013-05-14 Mike Stump <mikestump@comcast.net>
20013 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
20015 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
20017 * resource.h (struct resources): Remove unch_memory member.
20018 (CLEAR_RESOURCE): Don't clear unch_memory.
20019 * resource.c (mark_referenced_resources): Don't set it.
20020 (mark_set_resources): Likewise.
20021 (mark_target_live_regs): Don't clear it.
20022 (init_resource_info): Likewise.
20023 * reorg.c (resource_conflicts_p): Don't compare it.
20024 (redundant_insn): Don't set it.
20026 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
20028 * emit-rtl.c (next_label): Remove unused function.
20029 (skip_consecutive_labels, link_cc0_insns): Move to ...
20030 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
20031 only place where these functions are used, and make them static.
20033 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20035 * fold-const.c (fold_negate_expr): Handle vectors.
20036 (fold_truth_not_expr): Make it static.
20037 (fold_invert_truthvalue): New static function.
20038 (invert_truthvalue_loc): Handle vectors. Do not call
20039 fold_truth_not_expr directly.
20040 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
20041 <TRUTH_NOT_EXPR>: Do not cast to boolean.
20042 (fold_comparison): Handle vector constants.
20043 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
20044 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
20045 * tree.h (fold_truth_not_expr): Remove declaration.
20047 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
20049 * config/aarch64/aarch64-simd.md
20050 (aarch64_vcond_internal<mode>): Rename to...
20051 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
20052 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
20053 float modes. Clarify all iterator modes.
20054 (vcond<mode><mode>): Use new name for vcond expanders.
20055 (vcond<v_cmp_result><mode>): Likewise.
20056 (vcondu<mode><mode>: Likewise.
20057 * config/aarch64/iterators.md (VDQF_COND): New.
20059 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20062 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
20063 variable for the shift amount. Check that we shift by non-negative
20066 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
20069 * config/arm/arm.h (EPILOGUE_USES): Only return true
20070 for LR_REGNUM after epilogue_completed.
20072 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
20074 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
20075 is error_mark_node.
20077 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20080 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
20081 and Solaris 11+/x86 with gld.
20082 * configure: Regenerate.
20084 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20086 * expmed.c (expand_shift_1): Canonicalize rotates by
20087 constant bitsize / 2 to bitsize - 1.
20088 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
20089 case ROTATERT>: Likewise.
20092 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20094 * config/i386/i386.md (rotateinv): New code attr.
20095 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20096 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20097 roll $31, %eax, etc.
20099 2013-05-14 Richard Biener <rguenther@suse.de>
20101 PR middle-end/57235
20102 * tree-eh.c (sink_clobbers): Give up for successors with
20103 multiple predecessors and no virtual uses.
20105 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
20107 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
20108 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
20110 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20112 PR middle-end/57251
20113 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
20114 the case when both op0 and op1 have VOIDmode.
20116 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20118 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
20119 in multiply-accumulate mode.
20121 2013-05-13 Guozhi Wei <carrot@google.com>
20123 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
20125 2013-05-13 Kai Tietz <ktietz@redhat.com>
20128 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
20129 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
20130 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
20131 * config/i386/i386.c (ix86_option_override_internal): Likewise.
20132 (ix86_expand_prologue): Likewise.
20133 (ix86_expand_split_stack_prologue): Likewise.
20134 (legitimate_pic_address_disp_p): Likewise.
20135 (legitimize_pic_address): Likewise.
20136 (legitimize_tls_address): Likewise.
20137 (legitimize_pe_coff_symbol): Likewise.
20138 (output_pic_addr_const): Likewise.
20139 (construct_plt_address): Likewise.
20140 (ix86_expand_call): Likewise.
20141 (x86_output_mi_thunk): Likewise.
20142 (x86_function_profiler): Likewise.
20144 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
20146 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
20147 similar switch cases.
20148 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
20149 (aarch64_simd_mov_to_<mode>low): Delete.
20150 (aarch64_simd_mov_to_<mode>high): Delete.
20151 (move_lo_quad_<mode>): Add w<-r alternative.
20152 (aarch64_simd_move_hi_quad_<mode>): Likewise.
20153 (aarch64_simd_mov_from_*): Update type attribute.
20154 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
20157 2013-05-13 Jan Hubicka <jh@suse.cz>
20159 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
20160 * config/i386/i386.c (ix86_compute_frame_layout,
20161 ix86_expand_epilogue, emit_i387_cw_initialization,
20162 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
20163 ix86_local_alignment): Fix use of size/speed predicates.
20165 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20167 PR tree-optimization/45216
20168 PR tree-optimization/57157
20169 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
20170 the (-Y) & (B - 1) variant if OP is |.
20171 * expmed.c (expand_shift_1): For rotations by const0_rtx just
20172 return shifted. Use (-op1) & (prec - 1) as other_amount
20173 instead of prec - op1.
20175 2013-05-13 Martin Jambor <mjambor@suse.cz>
20177 PR middle-end/42371
20178 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
20179 (ipa_constant_data): New type.
20180 (ipa_jump_func): Use ipa_constant_data to hold information about
20181 constant jump functions.
20182 (ipa_get_jf_constant): Adjust to jump function type changes.
20183 (ipa_get_jf_constant_rdesc): New function.
20184 (ipa_param_descriptor): New field controlled_uses.
20185 (ipa_get_controlled_uses): New function.
20186 (ipa_set_controlled_uses): Likewise.
20187 * ipa-ref.h (ipa_find_reference): Declare.
20188 * ipa-prop.c (ipa_cst_ref_desc): New type.
20189 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
20191 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
20192 New parameter cs. Adjust all callers.
20193 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
20194 (remove_described_reference): New function.
20195 (jfunc_rdesc_usable): Likewise.
20196 (try_make_edge_direct_simple_call): Decrement controlled use count,
20197 attempt to remove reference if it hits zero.
20198 (combine_controlled_uses_counters): New function.
20199 (propagate_controlled_uses): Likewise.
20200 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
20201 (ipa_edge_duplication_hook): Duplicate reference descriptions.
20202 (ipa_print_node_params): Print described use counter.
20203 (ipa_write_jump_function): Adjust to jump function type changes.
20204 (ipa_read_jump_function): New parameter CS, pass it to
20205 ipa_set_jf_constant. Adjust caller.
20206 (ipa_write_node_info): Stream controlled use count
20207 (ipa_read_node_info): Likewise.
20208 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
20210 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
20211 count. Remove cloning-added reference if it reaches zero.
20212 * ipa-ref.c (ipa_find_reference): New function.
20214 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20216 * config/i386/i386.c (processor_target_table): Modified default
20217 alignment values for AMD BD and BT architectures.
20219 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
20221 * tree-vect-generic.c (uniform_vector_p): Move ...
20222 * tree.c (uniform_vector_p): ... here.
20223 * tree.h (uniform_vector_p): Declare it.
20224 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
20227 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20229 PR tree-optimization/57230
20230 * tree-ssa-strlen.c (handle_char_store): Record length for
20231 array store from STRING_CST.
20233 PR tree-optimization/57230
20234 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
20237 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
20239 * config/epiphany/epiphany.c (epiphany_init): Check size of
20240 NUM_MODES_FOR_MODE_SWITCHING.
20241 (epiphany_expand_prologue):
20242 Remove CONFIG_REGNUM initial value handling code.
20243 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
20244 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
20245 (emit_set_fp_mode, epiphany_mode_after): Likewise.
20246 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
20247 Don't return 1 for FP_MODE_NONE.
20248 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
20249 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
20250 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
20251 * config/epiphany/epiphany.md (save_config): New pattern.
20253 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
20255 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
20257 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
20259 * config/i386/i386.md (memory): Handle sseishft1.
20260 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
20261 (*vec_extractv2di_1): Ditto.
20263 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
20265 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
20268 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
20270 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
20271 Add mthumb/march=armv7-a multilib.
20272 Add mthumb/march=armv7-r multilib.
20273 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
20275 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
20277 * config/v850/t-rtems: Add more multilibs.
20279 2013-05-10 Richard Biener <rguenther@suse.de>
20281 PR tree-optimization/57214
20282 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
20283 not propagate from SSA names that occur in abnormal PHI nodes.
20285 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
20287 * stor-layout.c (element_precision): New function.
20288 * machmode.h (element_precision): Declare it.
20289 * tree.c (build_minus_one_cst): New function.
20290 (element_precision): Likewise.
20291 * tree.h (build_minus_one_cst): Declare new function.
20292 (element_precision): Likewise.
20293 * fold-const.c (operand_equal_p): Use element_precision.
20294 (fold_binary_loc): Handle vector types.
20295 * convert.c (convert_to_integer): Use element_precision.
20296 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
20299 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
20301 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
20302 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
20303 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
20304 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20305 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
20306 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
20307 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
20308 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20309 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
20310 (Uuw8): New constraints.
20311 (Usb4): Move into alphabetical order.
20312 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
20313 (sd8_operand, ub8_operand, uw8_operand): New predicates.
20314 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
20315 previously unnamed patterns.
20316 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
20317 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
20318 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
20319 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
20320 of set_attr_alternative/if_then_else. Use extended_mips16 instead
20321 of specific lengths.
20323 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20325 * config/i386/i386.md (rotateinv): New code attr.
20326 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20327 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20328 roll $31, %eax, etc.
20330 PR tree-optimization/45216
20331 PR tree-optimization/57157
20332 * tree-ssa-forwprop.c (simplify_rotate): New function.
20333 (ssa_forward_propagate_and_combine): Call it.
20335 2013-05-10 Richard Biener <rguenther@suse.de>
20337 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
20338 disable peeling when we version for aliasing.
20339 (vector_alignment_reachable_p): Honor explicit user alignment.
20340 (vect_supportable_dr_alignment): Likewise.
20341 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
20342 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
20343 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
20344 then peeling to arrange for the cost-model check to come first.
20346 2013-05-10 Alan Modra <amodra@gmail.com>
20348 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
20349 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
20350 * configure: Regenerate.
20352 2013-05-10 Alan Modra <amodra@gmail.com>
20355 * varasm.c (default_elf_select_section): Move !DECL_P check..
20356 (get_named_section): ..to here before calling get_section_name.
20358 (default_section_type_flags): Add DECL_P check.
20359 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
20360 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
20362 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
20364 * config/epiphany/epiphany.c (epiphany_expand_prologue):
20365 When using gen_stack_adjust_str with a register offset, add a
20366 REG_FRAME_RELATED_EXPR note.
20368 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
20370 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
20371 (*vec_extractv4si_zext_mem): Ditto.
20372 (*vec_extractv2di): Add 0->x and x->x alternatives.
20373 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
20374 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
20376 2013-05-09 Jason Merrill <jason@redhat.com>
20378 N3639 C++1y VLA support
20379 * gimplify.c (gimplify_vla_decl): Don't touch an existing
20382 * tree.c (build_constructor_va): New.
20383 * tree.h: Declare it.
20385 2013-05-09 Martin Jambor <mjambor@suse.cz>
20388 * gimple-fold.c (canonicalize_constructor_val): Call
20389 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
20391 2013-05-09 Jan Hubicka <jh@suse.cz>
20392 Richard Biener <rguenther@suse.de>
20395 * symtab.c (symtab_make_decl_local): Do not add private names.
20397 2013-05-09 Jan Hubicka <jh@suse.cz>
20400 * symtab.c (insert_to_assembler_name_hash): Handle clones.
20401 (unlink_from_assembler_name_hash): Likewise.
20402 (symtab_prevail_in_asm_name_hash, symtab_register_node,
20403 symtab_unregister_node, symtab_initialize_asm_name_hash,
20404 change_decl_assembler_name): Update.
20406 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
20408 * config/aarch64/aarch64.md: New movtf split.
20409 (*movtf_aarch64): Update.
20410 (aarch64_movdi_tilow): Handle TF modes and rename to
20411 aarch64_movdi_<mode>low.
20412 (aarch64_movdi_tihigh): Handle TF modes and rename to
20413 aarch64_movdi_<mode>high
20414 (aarch64_movtihigh_di): Handle TF modes and rename to
20415 aarch64_mov<mode>high_di
20416 (aarch64_movtilow_di): Handle TF modes and rename to
20417 aarch64_mov<mode>low_di
20418 (aarch64_movtilow_tilow): Remove spurious whitespace.
20419 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
20421 (aarch64_print_operand): Update.
20423 2013-05-09 Alan Modra <amodra@gmail.com>
20425 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
20427 * configure: Regenerate.
20429 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20431 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
20432 splitter preparation statements.
20433 * config/i386/sse.md (*vec_extract* splitters): Ditto.
20434 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
20437 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20439 * gimple-ssa-strength-reduction.c (count_candidates): Change
20440 return value to int.
20441 (analyze_candidates_and_replace): Change type of length to int.
20443 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20445 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
20446 (*vec_extract<mode>): Use VI12_128 mode iterator.
20447 (*vec_extract<mode>_mem): Ditto.
20448 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
20451 2013-05-08 Diego Novillo <dnovillo@google.com>
20456 2012-08-17 Diego Novillo <dnovillo@google.com>
20459 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
20460 * config.in: Regenerate.
20461 * configure: Regenerate.
20462 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
20464 2013-05-08 Jan Hubicka <jh@suse.cz>
20467 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
20468 * cgraph.h (symtab_node_base): Add unique_name.
20469 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20470 input_overwrite_node, input_varpool_node): Stream unique_name.
20471 * cgraphclones.c (cgraph_create_virtual_clone,
20472 cgraph_function_versioning): Set unique_name.
20473 * ipa.c (function_and_variable_visibility): Set unique_name.
20475 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20477 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
20478 (alloc_cand_and_find_basis): Restrict conditional candidate
20479 processing to CAND_MULTs.
20481 2013-05-08 Jan Hubicka <jh@suse.cz>
20484 lto-symtab.c (lto_symtab_symbol_p): New function.
20485 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
20486 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
20487 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
20488 Skip static symbols.
20490 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
20492 PR tree-optimization/57200
20493 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
20494 Only call inform if the preceding warning_at returns true.
20496 2013-05-07 Han Shen <shenhan@google.com>
20498 * cfgexpand.c (record_or_union_type_has_array_p): New function.
20499 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
20500 * common.opt (fstack-protector-strong): New option.
20501 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
20502 * doc/invoke.texi (Optimization Options): Document
20503 "-fstack-protector-strong".
20504 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
20506 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20508 * config/mips/mips.c (mips_machine_reorg2): Return 0.
20510 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
20512 * ira.c (update_equiv_regs): Add insn having equiv memory even if
20513 it is not lhs of the insn.
20514 (setup_reg_equiv): Remove insn having equiv memory which it is not
20516 * lra-constraints.c (process_address): Try to improve generation
20517 code for address base + disp.
20518 (lra_constraints): Make correct the code for checking insn setting
20519 up backward equivalence. Remove insn only if it is in the init
20521 * lra-eliminations.c (update_reg_eliminate): Change return value.
20522 (lra_eliminate): Use the result.
20524 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
20526 * config/i386/sse.md (ssescalarnummask): New mode attribute.
20527 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
20528 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
20529 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
20530 register target operands.
20531 (*vec_extractv8hi_sse2): New pattern.
20532 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
20533 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
20534 (*vec_extract<mode>_mem): New insn and split pattern.
20536 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
20538 * config/arm/arm.c (arm_asan_shadow_offset): New function.
20539 (TARGET_ASAN_SHADOW_OFFSET): Define.
20540 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
20541 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
20543 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20545 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
20546 (incr_vec_index): Return -1 if increment not found.
20547 (create_add_on_incoming_edge): Assert if increment not found.
20548 (record_increment): Limit number of increments recorded.
20549 (all_phi_incrs_profitable): Return false if an increment not found.
20550 (replace_profitable_candidates): Don't process increments that were
20552 (analyze_candidates_and_replace): Limit size of incr_vec.
20554 2013-05-07 Richard Biener <rguenther@suse.de>
20556 * calls.c (special_function_p): setjmp-like functions are leaf.
20557 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
20558 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
20560 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
20562 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
20563 (aarch64_simd_mov<mode>): New expander.
20564 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
20565 (aarch64_simd_mov_to_<mode>high): Likewise.
20566 (aarch64_simd_mov_from_<mode>low): Likewise.
20567 (aarch64_simd_mov_from_<mode>high): Likewise.
20568 (aarch64_dup_lane<mode>): Update.
20569 (aarch64_dup_lanedi): New instruction pattern.
20570 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
20571 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
20573 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20575 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
20576 (replace_mult_candidate): Remove unnecessary argument; remove
20577 unnecessary parameter from call to introduce_cast_before_cand.
20578 (replace_unconditional_candidate): Remove unnecessary parameter
20579 from call to replace_mult_candidate.
20580 (replace_conditional_candidate): Likewise.
20581 (insert_initializers): Use make_temp_ssa_name.
20582 (introduce_cast_before_cand): Remove unnecessary argument; use
20583 make_temp_ssa_name.
20584 (replace_one_candidate): Remove unnecessary argument; remove
20585 unnecessary parameter from calls to introduce_cast_before_cand.
20586 (replace_profitable_candidates): Remove unnecessary parameters
20587 from calls to replace_one_candidate.
20589 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20591 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
20592 phi def as possibly hiding a basis for a CAND_ADD whose operands
20593 have been commuted in the analysis.
20594 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
20596 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20598 * config/aarch64/aarch64.md
20599 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
20600 shift value between 0-4.
20602 2013-05-07 Richard Biener <rguenther@suse.de>
20604 * double-int.h (rshift): New overload.
20605 * double-int.c (rshift): New function.
20606 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
20607 (create_reference_ops_from_ref): Remove.
20608 (vn_reference_insert): Use shared ops for constructing the
20609 reference and copy it.
20611 2013-05-07 Richard Biener <rguenther@suse.de>
20613 PR middle-end/57190
20614 * tree-eh.c (sink_clobbers): Properly propagate
20615 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
20617 2013-05-07 Jakub Jelinek <jakub@redhat.com>
20619 PR tree-optimization/57149
20620 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
20621 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
20622 collect_phi_def_edges, execute_late_warn_uninitialized): Use
20623 uninit_undefined_value_p instead of ssa_undefined_value_p.
20626 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
20627 for modifier == EXPAND_INITIALIZER.
20629 2013-05-07 Anton Blanchard <anton@samba.org>
20631 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
20632 for powerpc64 little endian.
20633 * configure: Regenerate.
20635 2013-05-06 Graham Stott <grahams@btinternet.com>
20637 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
20638 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
20639 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
20640 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
20642 2013-05-06 Graham Stott <grahams@btinternet.com>
20644 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
20645 codes which allow non-lvalues.
20647 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
20649 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
20650 components are all 1s.
20651 (integer_minus_onep): New function.
20652 * tree.h (integer_minus_onep): Declare it.
20653 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
20654 integer_minus_onep instead of integer_all_onesp.
20656 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20659 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
20660 variations of these patterns.
20662 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20664 * config/i386/i386.md (isa): Add x64_sse4 member.
20665 (enabled): Handle x64_sse4.
20666 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
20667 instruction for 64bit SSE4_1 targets. Update insn attributes.
20668 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
20669 instruction for SSE4_1 targets. Update insn attributes.
20670 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20671 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
20672 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
20674 (*vec_extractv4si): Rename from *sse4_1_pextrd.
20675 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
20676 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
20678 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20681 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
20683 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
20685 * final.c (do_assembler_dialects): Don't handle curly braces and
20686 vertical bar escaped by % as dialect delimiters.
20687 (output_asm_insn): Print curly braces and vertical bar if escaped
20688 by % and ASSEMBLER_DIALECT defined.
20689 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
20690 * doc/tm.texi: Regenerated.
20692 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20694 * config/mips/mips.c: Include tree-pass.h.
20695 (mips_reorg): Split in pre- and post-dbr_schedule parts.
20696 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
20697 (pass_mips_machine_reorg2): New machine specific pass.
20698 (insert_pass_mips_machine_reorg2): New pass plugin definition.
20699 (mips_option_override): Register the new pass.
20700 * rtl.h (cleanup_barriers): Remove prototype.
20701 (dbr_schedule): Likewise.
20702 * jump.c (cleanup_barriers): Make static.
20703 * reorg.c (dbr_schedule): Likewise.
20705 2013-05-06 Richard Biener <rguenther@suse.de>
20707 PR tree-optimization/57185
20708 * tree-parloops.c (add_field_for_reduction): Handle anonymous
20709 SSA names properly.
20711 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20714 * config/i386/i386.c (add_parameter_dependencies): Add dependence
20715 between "first_arg" and "insn", not "last" and "insn".
20717 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20719 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
20720 (find_candidates_in_block): Re-enable slsr_process_phi.
20721 (create_phi_basis): Fix double counting of candidate adjustment.
20723 2013-05-06 Richard Biener <rguenther@suse.de>
20725 PR middle-end/57147
20726 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
20727 the edge is also fallthru, preserve it and just clear the
20729 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
20730 also complex, preserve that and just clear the fallthru flag.
20731 * tree-inline.c (update_ssa_across_abnormal_edges): Also
20732 update virtual operands.
20734 2013-05-06 Alan Modra <amodra@gmail.com>
20736 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
20737 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
20738 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
20739 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
20740 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20741 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
20743 2013-05-06 Alan Modra <amodra@gmail.com>
20745 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
20746 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
20747 (DEFAULT_ASM_ENDIAN): Define.
20748 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20749 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
20750 Update -K PIC clause from sysv4.h.
20751 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
20752 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
20754 2013-05-06 Alan Modra <amodra@gmail.com>
20756 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
20757 twice for little-endian.
20758 (ashrdi3_no_power, ashrdi3): Support little-endian.
20760 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20763 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
20764 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
20766 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
20767 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
20768 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
20771 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
20772 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20774 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
20775 * config/pa/pa.opt: Make mbig-switch a no-op.
20776 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
20777 (CASE_VECTOR_MODE): Always return SImode.
20778 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
20779 for the !TARGET_BIG_SWITCH case.
20780 * config/pa/pa-linux.h: Likewise.
20781 * config/pa/pa-openbsd.h: Likewise.
20782 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
20783 * config/pa/pa.md (short_jump): Remove define_insn.
20784 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
20785 (casesi0): Remove define_insn.
20786 (type): Remove btable_branch.
20787 (pa_combine_type): Likewise.
20788 (in_nullified_branch_delay): Likewise.
20789 (in_call_delay): Likewise.
20790 (define_delay): Likewise.
20791 (define_insn_reservation "Z3"): Likewise.
20792 (define_insn_reservation "Z4"): Likewise.
20793 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
20794 (pa_adjust_insn_length): Remove adjustment for btable branches.
20795 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
20798 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
20800 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20801 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
20802 Add m->r,x alternatives.
20803 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
20804 splitters using SWI48x mode iterator.
20805 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
20806 TARGET_64BIT. Add m->x alternative.
20807 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
20808 Add o->x alternative. Enable for TARGET_SSE.
20809 (sse_storeq): Remove expander.
20810 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
20811 with memory input operand.
20812 (*vec_extractv2di_1 splitter): New.
20813 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
20814 * config/i386/i386.md (ssevecmodelower): New mode attribute.
20816 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20818 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
20819 (INT_LOWPART): Delete.
20820 (extract_MB): Adjust.
20821 (extract_ME): Adjust.
20822 (print_operand): Adjust.
20824 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20826 * config/rs6000/predicates.md (reg_or_add_cint_operand,
20827 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20828 (reg_or_logical_cint_operand, easy_fp_constant,
20829 logical_const_operand): Delete "CONST_DOUBLE" case.
20830 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
20831 "HOST_BITS_PER_WIDE_INT == 64" test.
20832 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
20833 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
20834 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
20835 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
20837 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
20838 CONST_DOUBLE DImode/VOIDmode case.
20839 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
20840 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20841 CONST_DOUBLE VOIDmode case.
20842 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20843 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
20844 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
20845 Delete CONST_DOUBLE case.
20846 (splitters for mov FMOVE64 const_double): Delete
20847 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20848 "HOST_BITS_PER_WIDE_INT >= 64" test.
20849 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
20851 (mov DI const_double): Delete.
20853 2013-05-04 Jakub Jelinek <jakub@redhat.com>
20855 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
20856 on op shows all bits zero in mode of a lowpart subreg, return zero.
20858 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
20861 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
20862 to save TFmode registers and DImode to save TImode registers for
20863 caller save operations.
20864 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
20865 mark being partially clobbered since they only use the first
20868 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
20869 and TDmode only use the upper 64-bits of each VSX register.
20871 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20873 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
20874 (find_candidates_in_block): Disable slsr_process_phi.
20876 2013-05-03 Guozhi Wei <carrot@google.com>
20878 * coverage.c (coverage_obj_init): Move the construction of gcov
20880 (build_init_ctor): ... here.
20882 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20884 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
20885 (slsr_cand_d): Redefine def_phi.
20886 (stride_status, phi_adjust_status, count_phis_status): New enums.
20887 (find_phi_def): New.
20888 (find_basis_for_base_expr): New.
20889 (find_basis_for_candidate): Handle hidden bases.
20890 (alloc_cand_and_find_basis): Handle phi candidates.
20891 (slsr_process_phi): New.
20892 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
20893 (create_mul_imm_cand): Likewise.
20894 (create_add_ssa_cand): Exclude phi base candidates.
20895 (create_add_imm_cand): Likewise.
20896 (slsr_process_cast): Likewise.
20897 (slsr_process_copy): Likewise.
20898 (find_candidates_in_block): Handle phi candidates.
20899 (dump_candidate): Likewise.
20900 (unconditional_cands): Delete.
20901 (unconditional_cands_with_known_stride_p): Delete.
20902 (phi_dependent_cand_p): New.
20903 (cand_increment): Handle phi-dependent candidates.
20904 (replace_dependent): Delete.
20905 (replace_mult_candidate): New.
20906 (replace_unconditional_candidate): New.
20907 (incr_vec_index): Move to avoid forward reference.
20908 (create_add_on_incoming_edge): New.
20909 (create_phi_basis): New.
20910 (replace_dependents): Delete.
20911 (replace_conditional_candidate): New.
20912 (phi_add_costs): New.
20913 (replace_uncond_cands_and_profitable_phis): New.
20914 (record_increment): Handle phi adjustments.
20915 (record_phi_increments): New.
20916 (record_increments): Handle phi adjustments.
20917 (phi_incr_cost): New.
20918 (lowest_cost_path): Handle phis.
20919 (total_savings): Likewise.
20920 (analyze_increments): Likewise.
20921 (ncd_with_phi): New.
20922 (ncd_of_cand_and_phis): New.
20923 (nearest_common_dominator_for_cands): Handle phi increments.
20924 (all_phi_incrs_profitable): New.
20925 (replace_profitable_candidates): Handle phi-dependent candidates.
20926 (analyze_candidates_and_replace): Likewise.
20928 2013-05-03 Teresa Johnson <tejohnson@google.com>
20931 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
20932 do not exceed REG_BR_PROB_BASE.
20934 2013-05-03 Jeff Law <law@redhat.com>
20936 PR tree-optimization/57144
20937 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
20938 operand of the condition will bit into the new type when eliminating
20939 a cast feeding a condition.
20941 2013-05-03 Jakub Jelinek <jakub@redhat.com>
20943 PR rtl-optimization/57130
20944 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
20945 of COMPARE as in_code to the recursive call if needed.
20947 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
20949 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
20950 (enabled): Handle new members.
20951 * config/i386/sse.md (*vec_concatv2si): Merge from
20952 *vec_concatv2si_sse2 and vec_concatv2si_sse.
20953 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
20955 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
20957 PR tree-optimization/57027
20958 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
20959 for fnms opportunity, check we got the prerequisite kind
20960 of tree / gimple before using accessor functions.
20962 2013-05-03 Richard Biener <rguenther@suse.de>
20964 * double-int.h (lshift): New overload without precision
20965 and arith argument.
20966 (operator *=, operator +=, operator -=): Move ...
20967 * double-int.c (operator *=, operator +=, operator -=): ... here
20968 and implement more efficiently.
20969 (mul_double_with_sign): Remove.
20970 (lshift_double): Adjust to take unsinged shift argument, push
20971 dispatching code to callers.
20972 (mul_double_wide_with_sign): Add early out for callers that
20973 are not interested in high parts or overflow.
20974 (lshift): New function.
20975 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
20976 dispatch code here.
20977 (lrotate, rrotate): Use logical shifts.
20978 * expr.c (get_inner_reference): Use lshift.
20979 * fixed-value.c (do_fixed_divide): Likewise.
20980 * tree-dfa.c (get_ref_base_and_extent): Likewise.
20981 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
20982 (indirect_refs_may_alias_p): Likewise.
20983 (stmt_kills_ref_p_1): Likewise.
20985 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20987 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
20989 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20991 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
20992 scalar form of FABD instruction.
20994 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
20996 * lra-constraints.c (process_alt_operands): Add checking alt
20997 number to choose the best alternative.
20999 2013-05-02 Richard Biener <rguenther@suse.de>
21001 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
21002 bitmap and its handling.
21003 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
21005 2013-05-02 Richard Biener <rguenther@suse.de>
21007 PR middle-end/57140
21008 * tree-inline.c (copy_loops): Properly handle removed loops.
21009 (copy_cfg_body): Mark destination loops for fixup if source
21010 loops needed fixup.
21012 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
21015 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
21016 generating simple_return for naked functions.
21018 2013-05-02 Martin Jambor <mjambor@suse.cz>
21020 PR middle-end/56988
21021 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
21022 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
21024 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
21025 ipa_agg_replacement_value structures.
21026 (known_aggs_to_agg_replacement_list): Likewise.
21027 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
21028 (read_agg_replacement_chain): Likewise.
21029 (ipcp_transform_function): Also check that by_ref flags match.
21031 2013-05-02 Richard Biener <rguenther@suse.de>
21033 * graphds.h (struct graph): Add obstack member.
21034 * graphds.c (new_graph): Initialize obstack and allocate
21036 (add_edge): Allocate edge from the obstack.
21037 (free_graph): Free the obstack instead of all edges and vertices.
21039 2013-05-02 Teresa Johnson <tejohnson@google.com>
21041 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
21043 * cfg.c (update_bb_profile_for_threading): Ditto.
21044 * tree-inline.c (copy_bb): Ditto.
21045 (copy_edges_for_bb): Ditto.
21046 (initialize_cfun): Ditto.
21047 (copy_cfg_body): Ditto.
21048 (expand_call_inline): Ditto.
21049 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
21050 (estimate_node_size_and_time): Ditto.
21051 (inline_merge_summary): Ditto.
21052 * cgraphclones.c (cgraph_clone_edge): Ditto.
21053 (cgraph_clone_node): Ditto.
21054 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21055 (compute_trg_info): Ditto.
21057 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21059 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
21060 S reg when fp attribute set.
21061 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
21063 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21065 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
21067 (*and_one_cmplsi3_compare0_uxtw): Likewise.
21068 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
21069 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
21071 2013-05-02 Richard Biener <rguenther@suse.de>
21073 * tree-scalar-evolution.c (scev_info_hasher): Remove.
21074 (struct instantiate_cache_entry): New type.
21075 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
21076 (struct instantiate_cache_type): New type.
21077 (set_instantiated_value, get_instantiated_value): Remove.
21078 (get_instantiated_value_entry): New function.
21079 (instantiate_scev_name): Use the new cache and adjust.
21080 (instantiate_scev_poly): Adjust.
21081 (instantiate_scev_binary): Likewise.
21082 (instantiate_array_ref): Likewise.
21083 (instantiate_scev_convert): Likewise.
21084 (instantiate_scev_not): Likewise.
21085 (instantiate_scev_3): Likewise.
21086 (instantiate_scev_2): Likewise.
21087 (instantiate_scev_r): Likewise.
21088 (instantiate_scev): Likewise.
21089 (resolve_mixers): Likewise.
21091 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
21094 * lra-constraints.c (best_small_class_operands_num): Remove.
21095 (process_alt_operands): Remove small_class_operands_num. Take
21096 small classes operands into losers and only if the operand is not
21097 matched. Modify debugging output.
21098 (curr_insn_transform): Remove best_small_class_operands_num.
21101 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21103 * config/aarch64/aarch64-builtins.c
21104 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
21105 * config/aarch64/aarch64-simd-builtins.def
21106 (reduc_splus_): Add new modes.
21107 (reduc_uplus_): New.
21108 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
21109 (reduc_uplus_v4sf): Likewise.
21110 (reduc_splus_v4sf): Likewise.
21111 (aarch64_addv<mode>): Likewise.
21112 (reduc_uplus_<mode>): Likewise.
21113 (reduc_splus_<mode>): Likewise.
21114 (aarch64_addvv2di): Likewise.
21115 (reduc_uplus_v2di): Likewise.
21116 (reduc_splus_v2di): Likewise.
21117 (aarch64_addvv2si): Likewise.
21118 (reduc_uplus_v2si): Likewise.
21119 (reduc_splus_v2si): Likewise.
21120 (reduc_<sur>plus_<mode>): New.
21121 (reduc_<sur>plus_v2di): Likewise.
21122 (reduc_<sur>plus_v2si): Likewise.
21123 (reduc_<sur>plus_v4sf): Likewise.
21124 (aarch64_addpv4sf): Likewise.
21125 * config/aarch64/arm_neon.h
21126 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
21127 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
21128 add UNSPEC_SADDV, UNSPEC_UADDV.
21130 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
21132 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21134 * config/aarch64/arm_neon.h
21135 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
21137 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21139 * config/aarch64/aarch64-builtins
21140 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
21142 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21144 * config/aarch64/aarch64-simd-builtins.def
21145 (reduc_smax_): New.
21146 (reduc_smin_): Likewise.
21147 (reduc_umax_): Likewise.
21148 (reduc_umin_): Likewise.
21149 (reduc_smax_nan_): Likewise.
21150 (reduc_smin_nan_): Likewise.
21153 (smax): Update for V2SF, V4SF and V2DF modes.
21156 (smin_nan): Likewise.
21157 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
21158 (<su><maxmin><mode>3): ...This, refactor.
21159 (s<maxmin><mode>3): New.
21160 (<maxmin_uns><mode>3): Likewise.
21161 (reduc_<maxmin_uns>_<mode>): Refactor.
21162 (reduc_<maxmin_uns>_v4sf): Likewise.
21163 (reduc_<maxmin_uns>_v2si): Likewise.
21164 (aarch64_<fmaxmin><mode>: Remove.
21165 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
21167 (vmin<q>_f<32,64>): Likewise.
21168 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21170 (su): Add mappings for smax, smin, umax, umin.
21172 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21173 (FMAXMIN): Rename as...
21174 (FMAXMIN_UNS): ...This.
21176 (fmaxminv): Likewise.
21177 (fmaxmin): Likewise.
21179 (maxmin_uns_op): Likewise.
21181 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21183 * config/aarch64/arm_neon.h
21184 (vac<ge, gt><sd>_f<32, 64>): Rename to...
21185 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
21186 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
21188 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21190 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
21191 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
21193 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21195 * config/aarch64/aarch64-simd.md
21196 (vcond<mode>_internal): Handle special cases for constant masks.
21197 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
21198 (vcondu<mode><mode>): Likewise.
21199 (vcond<v_cmp_result><mode>): New.
21201 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21203 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
21204 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
21205 * config/aarch64/aarch64-simd-builtins.def
21206 (cmeq): Update to BUILTIN_VALLDI.
21211 * config/aarch64/arm_neon.h
21212 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
21213 to builtins or C as appropriate.
21215 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21217 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
21219 (cmhi): Rename to...
21221 * config/aarch64/aarch64-simd.md
21222 (simd_mode): Add SF.
21223 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
21224 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
21225 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
21226 (cstore<mode>_neg): ...This.
21227 * config/aarch64/iterators.md
21229 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
21230 (COMPARISONS): New.
21231 (UCOMPARISONS): Likewise.
21232 (optab): Add missing comparisons.
21238 (VCMP_S): Likewise.
21239 (VCMP_U): Likewise.
21240 (V_cmp_result): Add DF, SF modes.
21241 (v_cmp_result): Likewise.
21243 (vmtype): Likewise.
21244 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
21246 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
21248 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
21249 define_insn to define_insn_and_split.
21250 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
21251 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
21252 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
21253 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
21254 (thumb2_negscc): Likewise.
21256 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21258 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
21260 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21262 * config/arm/thumb2.md: Remove trailing whitespaces.
21264 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21266 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
21267 Use gen_int_mode rather than GEN_INT.
21269 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
21271 * value-prof.c (stream_in_histogram_value): Remove the strayed
21274 2013-04-30 Richard Biener <rguenther@suse.de>
21276 PR middle-end/57122
21277 * cfghooks.c (split_edge): Properly check for the loop latch edge.
21279 2013-04-30 Richard Biener <rguenther@suse.de>
21281 PR middle-end/57107
21282 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
21284 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
21286 PR rtl-optimization/56957
21287 PR rtl-optimization/57105
21288 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
21289 variable. Use just INSN_UID for determining whether an insn
21290 should be only disconnected from the insn stream.
21291 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
21293 2013-04-30 Jakub Jelinek <jakub@redhat.com>
21295 PR tree-optimization/57104
21296 * tsan.c (instrument_expr): Don't instrument accesses to
21297 DECL_HARD_REGISTER VAR_DECLs.
21299 2013-04-30 Richard Biener <rguenther@suse.de>
21301 * function.h (loops_for_fn): New inline function.
21302 (set_loops_for_fn): Likewise.
21303 * cfgloop.h (place_new_loop): Add struct function parameter.
21304 (get_loop): Likewise.
21305 (get_loops): Likewise.
21306 (number_of_loops): Likewise.
21307 (fel_next): Adjust.
21308 (fel_init): Likewise.
21309 * cfg.c (get_loop_copy): Adjust.
21310 * cfgloop.c (flow_loops_dump): Likewise.
21311 (record_loop_exits): Likewise.
21312 (verify_loop_structure): Likewise.
21313 * cfgloopanal.c (mark_irreducible_loops): Likewise.
21314 (estimate_reg_pressure_cost): Likewise.
21315 (mark_loop_exit_edges): Likewise.
21316 * cfgloopmanip.c (place_new_loop): Likewise.
21317 (add_loop): Likewise.
21318 (duplicate_loop): Likewise.
21319 * graph.c (draw_cfg_nodes): Likewise.
21320 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
21321 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
21322 (extract_affine_chrec): Likewise.
21323 (build_scop_iteration_domain): Likewise.
21324 * graphite.c (graphite_initialize): Likewise.
21325 * ira-build.c (create_loop_tree_nodes): Likewise.
21326 (more_one_region_p): Likewise.
21327 (rebuild_regno_allocno_maps): Likewise.
21328 (mark_loops_for_removal): Likewise.
21329 (mark_all_loops_for_removal): Likewise.
21330 (remove_unnecessary_regions): Likewise.
21331 (ira_build): Likewise.
21332 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21333 * loop-init.c (fix_loop_structure): Likewise.
21334 (gate_rtl_move_loop_invariants): Likewise.
21335 (gate_rtl_unswitch): Likewise.
21336 (gate_rtl_unroll_and_peel_loops): Likewise.
21337 (rtl_doloop): Likewise.
21338 * lto-streamer-in.c (input_cfg): Likewise.
21339 * lto-streamer-out.c (output_cfg): Likewise.
21340 * modulo-sched.c (sms_schedule): Likewise.
21341 * predict.c (tree_estimate_probability): Likewise.
21342 (tree_estimate_probability_driver): Likewise.
21343 (estimate_loops): Likewise.
21344 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
21345 (move_sese_region_to_fn): Likewise.
21346 (debug_loop_num): Likewise.
21347 * tree-chrec.c (chrec_evaluate): Likewise.
21348 (hide_evolution_in_other_loops_than_loop): Likewise.
21349 (chrec_component_in_loop_num): Likewise.
21350 (reset_evolution_in_loop): Likewise.
21351 (evolution_function_is_invariant_rec_p): Likewise.
21352 * tree-if-conv.c (main_tree_if_conversion): Likewise.
21353 * tree-inline.c (copy_loops): Likewise.
21354 (copy_cfg_body): Likewise.
21355 (tree_function_versioning): Likewise.
21356 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
21357 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
21359 (add_to_evolution_1): Likewise.
21360 (scev_const_prop): Likewise.
21361 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
21362 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
21363 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
21364 (tree_ssa_lim_initialize): Likewise.
21365 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
21366 (verify_loop_closed_ssa): Likewise.
21367 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
21368 (tree_ssa_loop_im): Likewise.
21369 (tree_ssa_loop_unswitch): Likewise.
21370 (tree_vectorize): Likewise.
21371 (check_data_deps): Likewise.
21372 (tree_ssa_loop_ivcanon): Likewise.
21373 (tree_ssa_loop_bounds): Likewise.
21374 (tree_complete_unroll): Likewise.
21375 (tree_complete_unroll_inner): Likewise.
21376 (tree_parallelize_loops): Likewise.
21377 (tree_ssa_loop_prefetch): Likewise.
21378 (tree_ssa_loop_ivopts): Likewise.
21379 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
21380 * tree-vectorizer.c (vectorize_loops): Likewise.
21382 2013-04-29 Mike Frysinger <vapier@gentoo.org>
21384 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
21385 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
21386 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
21387 with EABI_LINK_SPEC.
21389 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21392 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
21395 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
21398 * lra-constraints.c (process_alt_operands): Discourage a bit more
21399 using memory for pseudos. Print cost dump for alternatives.
21400 Modify cost values for conflicts with early clobbers.
21401 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
21403 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21406 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
21408 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21410 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
21411 from/to S register.
21412 (movdi_aarch64): Support LDR/STR from/to D register.
21414 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21416 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
21417 or higher by default.
21419 2013-04-29 Richard Biener <rguenther@suse.de>
21421 PR middle-end/57075
21422 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
21423 even if not adding abnormal edges for calls that can make
21426 2013-04-29 Richard Biener <rguenther@suse.de>
21428 PR middle-end/57103
21429 * tree-cfg.c (move_stmt_op): Fix condition under which to update
21431 (move_stmt_r): Remove redundant checking.
21433 2013-04-29 Teresa Johnson <tejohnson@google.com>
21436 * basic-block.h (apply_scale): New function.
21437 (apply_probability): Use apply_scale.
21438 * gimple-streamer-in.c (input_bb): Ditto.
21439 * lto-streamer-in.c (input_cfg): Ditto.
21440 * lto-cgraph.c (merge_profile_summaries): Ditto.
21441 * tree-optimize.c (execute_fixup_cfg): Ditto.
21442 * tree-inline.c (copy_bb): Update comment to use apply_scale.
21443 (copy_edges_for_bb): Ditto.
21444 (copy_cfg_body): Ditto.
21446 2013-04-29 Tom de Vries <tom@codesourcery.com>
21448 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
21449 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
21450 (tail_merge_optimize): Handle current_loops == NULL.
21452 2013-04-26 Jeff Law <law@redhat.com>
21454 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
21455 (simplify_cond_using_ranges): Generalize code to simplify
21456 COND_EXPRs where one argument is a constant and the other
21457 is an SSA_NAME created by an integral type conversion.
21459 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21461 * config/arm/arm.md (store_minmaxsi): Use only when
21462 optimize_insn_for_size_p.
21464 2013-04-29 Christian Bruel <christian.bruel@st.com>
21467 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
21469 2013-04-29 Richard Biener <rguenther@suse.de>
21471 PR middle-end/57089
21472 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
21473 loop tree make sure to schedule a fixup for the child as well.
21474 (expand_omp_for_generic): Properly add loops.
21475 (expand_omp_for_static_nochunk): Likewise.
21476 (expand_omp_for_static_chunk): Likewise.
21477 (expand_omp_for): For the degenerate case fixup loops.
21478 (expand_omp_sections): Fix default bb placement in loops.
21479 (expand_omp_atomic_pipeline): Properly add loops.
21481 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21483 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
21485 2013-04-29 Tom de Vries <tom@codesourcery.com>
21487 * tree-ssa-tail-merge.c: Update header comment.
21489 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21491 * config/aarch64/arm_neon.h
21492 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
21493 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
21494 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
21495 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
21496 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
21497 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
21498 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
21499 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
21501 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21503 * config/aarch64/aarch64-simd.md
21504 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
21505 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
21506 fix_trunc, fixuns_trunc.
21507 (ftrunc<VDQF:mode>2): New.
21508 * config/aarch64/iterators.md (optab): Add fix, fixuns.
21509 (fix_trunc_optab): New.
21511 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21513 * config/aarch64/aarch64-builtins.c
21514 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
21515 iceilf, lround, iroundf.
21517 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21520 * config/i386/i386.h (enum ix86_tune_indices)
21521 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
21522 New, split from X86_TUNE_INTER_UNIT_MOVES.
21523 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
21524 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
21525 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
21526 (TARGET_INTER_UNIT_MOVES): Remove.
21527 * config/i386/i386.c (initial_ix86_tune_features): Update.
21528 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
21529 (ix86_expand_convert_uns_didf_sse): Use
21530 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21531 (ix86_expand_vector_init_one_nonzero): Ditto.
21532 (ix86_expand_vector_init_interleave): Ditto.
21533 (inline_secondary_memory_needed): Return true for moves from SSE class
21534 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
21535 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
21536 * config/i386/constraints.md (Yi, Ym): Depend on
21537 TARGET_INTER_UNIT_MOVES_TO_VEC.
21538 (Yj, Yn): New constraints.
21539 * config/i386/i386.md (*movdi_internal): Change constraints of
21540 operand 1 from Yi to Yj and from Ym to Yn.
21541 (*movsi_internal): Ditto.
21542 (*movdf_internal): Ditto.
21543 (*movsf_internal): Ditto.
21544 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
21545 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21546 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
21547 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
21548 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
21549 * config/i386/sse.md (movdi_to_sse): Ditto.
21550 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
21551 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
21552 TARGET_INTER_UNIT_MOVES.
21553 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
21554 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
21555 instead of TARGET_INTER_UNIT_MOVES.
21556 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
21557 operand 1 from Yi to Yj and from Ym to Yn.
21559 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21561 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
21562 (float_truncate_hi_): Likewise.
21563 (float_extend_lo_): Likewise.
21564 (float_truncate_lo_): Likewise.
21565 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
21566 (aarch64_float_extend_lo_v2df): Likewise.
21567 (vec_unpacks_hi_v4sf): Likewise.
21568 (aarch64_float_truncate_lo_v2sf): Likewise.
21569 (aarch64_float_truncate_hi_v4sf): Likewise.
21570 (vec_pack_trunc_v2df): Likewise.
21571 (vec_pack_trunc_df): Likewise.
21573 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21575 * config/aarch64/aarch64-builtins.c
21576 (aarch64_fold_builtin): Fold float conversions.
21577 * config/aarch64/aarch64-simd-builtins.def
21578 (floatv2si, floatv4si, floatv2di): New.
21579 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
21580 * config/aarch64/aarch64-simd.md
21581 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
21582 * config/aarch64/iterators.md (FLOATUORS): New.
21583 (optab): Add float, floatuns.
21584 (su_optab): Likewise.
21586 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21588 * config/aarch64/aarch64-builtins.c
21589 (aarch64_builtin_vectorized_function): Use new names for
21591 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
21592 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
21593 (fcvtzu): Split as...
21594 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
21595 (fcvtas): Split as...
21596 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
21597 (fcvtau): Split as...
21598 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
21599 (fcvtps): Split as...
21600 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
21601 (fcvtpu): Split as...
21602 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
21603 (fcvtms): Split as...
21604 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
21605 (fcvtmu): Split as...
21606 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
21607 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
21608 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
21609 (lfrintnusf, lfrintnudf): Likewise.
21610 * config/aarch64/aarch64-simd.md
21611 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
21613 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
21614 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
21615 (fcvt_pattern): Likewise.
21617 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21619 * config/aarch64/aarch64-simd.md
21620 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
21621 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
21623 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21625 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
21626 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
21627 (vrnd<a,m,n,p>_f32): Implement using builtins.
21628 (vrnd<i,x><q>_f<32, 64>): New.
21630 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21632 * config/aarch64/aarch64-builtins.c
21633 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
21634 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
21635 (frintz): Rename to...
21637 (frintp): Rename to...
21639 (frintm): Rename to...
21641 (frinti): Rename to...
21642 (nearbyint): ...this.
21643 (frintx): Rename to...
21645 (frinta): Rename to...
21647 * config/aarch64/aarch64-simd.md
21648 (aarch64_frint<frint_suffix><mode>): Delete.
21649 (<frint_pattern><mode>2): Convert to insn.
21650 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
21651 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
21652 (frint_pattern): Likewise.
21653 (frint_suffix): Likewise.
21655 2013-04-29 Richard Biener <rguenther@suse.de>
21657 PR tree-optimization/57081
21658 * loop-init.c: Include tree-flow.h.
21659 (loop_optimizer_finalize): Free number of iteration estimates.
21660 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
21662 2013-04-29 Jakub Jelinek <jakub@redhat.com>
21664 PR tree-optimization/57083
21665 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
21666 non-singleton shift count range, zero extend low_bound for uns case.
21668 * config/i386/predicates.md (general_vector_operand): New predicate.
21669 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
21670 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
21671 if they aren't nonimmediate operands. If their original values
21672 satisfy const_vector_equal_evenodd_p, don't shift them.
21673 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
21674 predicates. For the SSE4.1 case force operands[{1,2}] into registers
21675 if not nonimmediate_operand.
21676 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
21677 instead of register_operand.
21678 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
21680 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
21682 * stor-layout.c (finalize_size_functions): Allocate a structure and
21683 reset cfun before dumping the functions.
21685 2013-04-27 Jakub Jelinek <jakub@redhat.com>
21687 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
21690 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
21691 use xop_pmacsdqh if uns_p.
21692 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
21693 the immediate rotate count.
21695 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
21697 * rtl.h (struct rtx_def): Add comment for field jump.
21698 (LRA_SUBREG_P): New macro.
21699 * recog.c (register_operand): Check LRA_SUBREG_P.
21700 * lra.c (lra): Add note at the end of RTL code. Align non-empty
21702 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
21703 (lra_final_code_change): Skip subreg change for operators.
21704 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
21705 if there are no operand changes.
21706 * lra-constraints.c (curr_insn_set): New.
21707 (match_reload): Set LRA_SUBREG_P.
21708 (emit_spill_move): Ditto.
21709 (check_and_process_move): Use curr_insn_set. Process only single
21710 set insns. Don't initialize sec_mem_p and change_p.
21711 (simplify_operand_subreg): Use LRA_SUBREG_P.
21712 (reg_in_class_p): New function.
21713 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
21714 of #ifdef. Add code to remove cycling.
21715 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
21716 non-null disp. Reload inner instead of disp when base and index
21717 are null. Try to put lo_sum into register.
21718 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
21719 (check_and_process_move): Move code for move cost check to
21720 simple_move_p. Remove equiv_substitution.
21721 (simple_move_p): New function.
21722 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
21723 curr_insn_set. Call check_and_process_move only for single set
21724 insns. Use the new function. Move call of check_and_process_move
21725 after operand equiv substitution and address process.
21727 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21730 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
21731 with nonlocal goto receivers or returns twice calls, ignore
21732 unininitialized values from abnormal edges to nl goto receiver
21733 or returns twice call.
21735 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21737 PR tree-optimization/57051
21738 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
21739 and VEC_RSHIFT_EXPR if shift count is a multiple of element
21742 2013-04-26 Richard Biener <rguenther@suse.de>
21744 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
21745 (expand_omp_taskreg): Likewise. Mark loops for fixup.
21746 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
21747 (fixup_loop_arrays_after_move): New function.
21748 (move_sese_region_to_fn): Properly outline the loop tree parts
21749 of the SESE region.
21751 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
21753 * config/i386/i386.md (type, unit): Fix long lines.
21755 2013-04-26 Richard Biener <rguenther@suse.de>
21757 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
21758 (lto-streamer-out.o): Likewise.
21759 * cfgloop.c (init_loops_structure): Export, add struct function
21760 argument and adjust.
21761 (flow_loops_find): Adjust.
21762 * cfgloop.h (enum loop_estimation): Add EST_LAST.
21763 (init_loops_structure): Declare.
21764 * lto-streamer-in.c: Include cfgloop.h.
21765 (input_cfg): Input the loop tree.
21766 * lto-streamer-out.c: Include cfgloop.h.
21767 (output_cfg): Output the loop tree.
21768 (output_struct_function_base): Do not drop PROP_loops.
21770 2013-03-26 Richard Biener <rguenther@suse.de>
21772 * tree-cfg.c (execute_build_cfg): Build the loop tree.
21773 (pass_build_cfg): Provide PROP_loops.
21774 (move_sese_region_to_fn): Remove loops that are outlined into fn
21776 * tree-inline.c: Include cfgloop.h.
21777 (initialize_cfun): Do not drop PROP_loops.
21778 (copy_loops): New function.
21779 (copy_cfg_body): Copy loop structure.
21780 (tree_function_versioning): Initialize destination loop tree.
21781 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
21782 (pass_parallelize_loops): Do IL verification.
21783 * loop-init.c (loop_optimizer_init): Fixup loops if required.
21784 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
21785 the CFG make sure we fixup loops as well.
21786 * tree-ssa-tail-merge.c: Include cfgloop.h.
21787 (replace_block_by): When merging loop latches mark loops for fixup.
21788 * lto-streamer-out.c (output_struct_function_base): Drop
21789 PROP_loops for now.
21790 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
21791 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
21792 * ipa-split.c: Include cfgloop.h.
21793 (split_function): Add the new return block to the loop tree root.
21794 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
21795 whether we have removed the forwarder block.
21796 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
21797 * cfgloop.h (place_new_loop): Declare.
21798 * cfgloopmanip.c (place_new_loop): Export.
21799 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
21800 (tree-switch-conversion.o): Likewise.
21801 (tree-complex.o): Likewise.
21802 (tree-inline.o): Likewise.
21803 (tree-ssa-tailmerge.o): Likewise.
21804 (ipa-split.o): Likewise.
21805 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
21806 (tree-ssa-copy.o): Likewise.
21807 * tree-switch-conversion.c: Include cfgloop.h
21808 (process_switch): If we emit a bit-test cascade, schedule loops
21810 * tree-complex.c: Include cfgloop.h.
21811 (expand_complex_div_wide): Properly add new basic-blocks to loops.
21812 * asan.c: Include cfgloop.h.
21813 (create_cond_insert_point): Properly add new basic-blocks to
21814 loops, schedule loop fixup.
21815 * cfgloop.c (verify_loop_structure): Check that looks are not
21817 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
21819 (expand_omp_for_generic): Likewise.
21820 (expand_omp_sections): Likewise.
21821 (expand_omp_atomic_pipeline): Schedule loops for fixup.
21822 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
21823 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
21824 is initialized, not when loops are present.
21825 * tree-parloops.c (parallelize_loops): Remove checking here.
21826 * passes.c (init_optimization_passes): Schedule a copy-propagation
21827 pass before complete unrolling of inner loops.
21829 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21831 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
21832 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
21833 (colorize_init): Add argument to _WIN32 version.
21834 * toplev.c: Include diagnostic-color.h.
21835 (process_options): Default to -fdiagnostics-color=auto if
21836 GCC_COLORS env var is in the environment.
21837 * common.opt (fdiagnostics-color=): Add Var and Init.
21838 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
21839 env var is in the environment, the default is auto rather than never.
21841 * diagnostic.h (file_name_as_prefix): Add context argument.
21842 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
21843 the string as locus.
21844 * langhooks.c (lhd_print_error_function): Adjust caller.
21846 2013-04-25 Lawrence Crowl <crowl@google.com>
21848 * var-tracking.c (shared_hash_def::htab):
21849 Change type to hash_table. Update dependent calls and types.
21851 2013-04-25 Lawrence Crowl <crowl@google.com>
21853 * Makefile.in: Update as needed below.
21855 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
21856 Move declaration to after the type's method definitons.
21858 * attribs.c (htab_t scoped_attributes::attribute_hash):
21859 Change type to hash_table. Update dependent calls and types.
21861 * bitmap.c (htab_t bitmap_desc_hash):
21862 Change type to hash_table. Update dependent calls and types.
21864 * cselib.c (htab_t cselib_hash_table):
21865 Change type to hash_table. Update dependent calls and types.
21867 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
21868 (hash_string_slot_node): Move implementation into lto-streamer.h
21869 struct string_slot_hasher.
21870 (eq_string_slot_node): Likewise.
21872 * data-streamer-out.c: Update output_block::string_hash_table
21873 dependent calls and types.
21875 * dwarf2cfi.c (htab_t trace_index):
21876 Change type to hash_table. Update dependent calls and types.
21878 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
21879 Change type to hash_table. Update dependent calls and types.
21880 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
21881 (htab_t optimize_external_refs::map): Likewise.
21882 (htab_t output_comp_unit::extern_map): Likewise.
21883 (htab_t output_comdat_type_unit::extern_map): Likewise.
21884 (htab_t output_macinfo::macinfo_htab): Likewise.
21885 (htab_t optimize_location_lists::htab): Likewise.
21886 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
21888 * except.c (htab_t ehspec_hash_type):
21889 Change type to hash_table. Update dependent calls and types.
21890 (assign_filter_values::ttypes): Likewise.
21891 (assign_filter_values::ehspec): Likewise.
21892 (sjlj_assign_call_site_values::ar_hash): Likewise.
21893 (convert_to_eh_region_ranges::ar_hash): Likewise.
21895 * gcse.c (htab_t pre_ldst_table):
21896 Change type to hash_table. Update dependent calls and types.
21898 * ggc-common.c (htab_t saving_htab):
21899 Change type to hash_table. Update dependent calls and types.
21900 (htab_t loc_hash): Likewise.
21901 (htab_t ptr_hash): Likewise.
21902 (call_count): Rename ggc_call_count.
21903 (call_alloc): Rename ggc_call_alloc.
21904 (loc_descriptor): Rename make_loc_descriptor.
21905 (add_statistics): Rename ggc_add_statistics.
21907 * ggc-common.c (saving_htab):
21908 Change type to hash_table. Update dependent calls and types.
21910 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
21911 (push_gimplify_context): Likewise.
21912 (pop_gimplify_context): Likewise.
21913 (struct gimple_temp_hash_elt): Added.
21914 (struct gimplify_hasher): Likewise.
21915 (struct gimplify_ctx.temp_htab):
21916 Change type to hash_table. Update dependent calls and types.
21918 * gimple-fold.c: Include gimplify-ctx.h.
21920 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
21921 Change type to hash_table. Update dependent calls and types.
21922 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
21923 avoid potential global name collision.
21925 * gimplify.c: Include gimplify-ctx.h.
21926 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
21927 (htab_t gimplify_ctx::temp_htab):
21928 Update dependent calls and types for new type hash_table.
21929 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
21930 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
21932 * gimplify-ctx.h: New.
21933 (struct gimple_temp_hash_elt): Move from gimplify.c.
21934 (class gimplify_hasher): New.
21935 (struct gimplify_ctx): Move from gimple.h.
21936 (htab_t gimplify_ctx::temp_htab):
21937 Change type to hash_table. Update dependent calls and types.
21939 * graphite-clast-to-gimple.c: Include graphite-htab.h.
21940 (htab_t ivs_params::newivs_index):
21941 Change type to hash_table. Update dependent calls and types.
21942 (htab_t ivs_params::params_index): Likewise.
21943 (htab_t print_generated_program::params_index): Likewise.
21944 (htab_t gloog::newivs_index): Likewise.
21945 (htab_t gloog::params_index): Likewise.
21947 * graphite.c: Include graphite-htab.h.
21948 4htab_t graphite_transform_loops::bb_pbb_mapping):
21949 Change type to hash_table. Update dependent calls and types.
21951 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
21952 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
21953 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
21955 * graphite-dependences.c: Include graphite-htab.h.
21956 (loop_is_parallel_p): Change hash table type of parameter.
21958 * graphite-htab.h: New.
21959 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
21960 (extern find_pbb_via_hash): Move from graphite-poly.h.
21961 (extern loop_is_parallel_p): Move from graphite-poly.h.
21962 (extern get_loop_body_pbbs): Move from graphite-poly.h.
21964 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
21965 (extern loop_is_parallel_p): Move to graphite-htab.h.
21966 (extern get_loop_body_pbbs): Move to graphite-htab.h.
21968 * haifa-sched.c (htab_t delay_htab):
21969 Change type to hash_table. Update dependent calls and types.
21970 (htab_t delay_htab_i2): Likewise.
21972 * ira-color.c (htab_t allocno_hard_regs_htab):
21973 Change type to hash_table. Update dependent calls and types.
21975 * ira-costs.c (htab_t cost_classes_htab):
21976 Change type to hash_table. Update dependent calls and types.
21978 * loop-invariant.c (htab_t merge_identical_invariants::eq):
21979 Change type to hash_table. Update dependent calls and types.
21981 * loop-iv.c (htab_t bivs):
21982 Change type to hash_table. Update dependent calls and types.
21984 * loop-unroll.c (htab_t opt_info::insns_to_split):
21985 Change type to hash_table. Update dependent calls and types.
21986 (htab_t opt_info::insns_with_var_to_expand): Likewise.
21988 * lto-streamer.h (struct string_slot): Move from data-streamer.h
21989 (struct string_slot_hasher): New.
21990 (htab_t output_block::string_hash_table):
21991 Change type to hash_table. Update dependent calls and types.
21993 * lto-streamer-in.c (freeing_string_slot_hasher): New.
21994 (htab_t file_name_hash_table):
21995 Change type to hash_table. Update dependent calls and types.
21997 * lto-streamer-out.c: Update output_block::string_hash_table dependent
22000 * lto-streamer.c (htab_t tree_htab):
22001 Change type to hash_table. Update dependent calls and types.
22003 * omp-low.c: Include gimplify-ctx.h.
22005 * passes.c (htab_t name_to_pass_map):
22006 Change type to hash_table. Update dependent calls and types.
22007 (pass_traverse): Rename to passes_pass_traverse.
22009 * plugin.c (htab_t event_tab):
22010 Change type to hash_table. Update dependent calls and types.
22012 * postreload-gcse.c (htab_t expr_table):
22013 Change type to hash_table. Update dependent calls and types.
22014 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
22016 * sese.c (debug_rename_map_1): Make extern.
22017 (htab_t copy_bb_and_scalar_dependences::rename_map):
22018 Change type to hash_table. Update dependent calls and types.
22020 * sese.h (extern debug_rename_map): Move to .c file.
22022 * store-motion.c (htab_t store_motion_mems_table):
22023 Change type to hash_table. Update dependent calls and types.
22025 * trans-mem.c (htab_t tm_new_mem_hash):
22026 Change type to hash_table. Update dependent calls and types.
22028 * tree-browser.c (htab_t TB_up_ht):
22029 Change type to hash_table. Update dependent calls and types.
22031 * tree-cfg.c (htab_t discriminator_per_locus):
22032 Change type to hash_table. Update dependent calls and types.
22034 * tree-complex.c: Include tree-hasher.h
22035 (htab_t complex_variable_components):
22036 Change type to hash_table. Update dependent calls and types.
22038 * tree-eh.c (htab_t finally_tree):
22039 Change type to hash_table. Update dependent calls and types.
22041 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
22042 struct int_tree_hasher.
22043 (extern int_tree_map_eq): Likewise.
22044 (uid_decl_map_hash): Removed.
22045 (extern decl_tree_map_eq): Likewise.
22047 * tree-hasher.h: New.
22048 (struct int_tree_hasher): New.
22049 (typedef int_tree_htab_type): New.
22051 * tree-inline.c: Include gimplify-ctx.h.
22053 * tree-mudflap.c: Include gimplify-ctx.h.
22055 * tree-parloops.c: Include tree-hasher.h.
22056 (htab_t eliminate_local_variables_stmt::decl_address):
22057 Change type to hash_table. Update dependent calls and types.
22058 (htab_t separate_decls_in_region::decl_copies): Likewise.
22060 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
22061 Change type to hash_table. Update dependent calls and types.
22063 * tree-sra.c (candidates):
22064 Change type to hash_table. Update dependent calls and types.
22066 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
22068 (int_tree_map_hash): Likewise.
22070 * tree-ssa-dom.c (htab_t avail_exprs):
22071 Change type to hash_table. Update dependent calls and types.
22073 * tree-ssa-live.c (var_map_base_init::tree_to_index):
22074 Change type to hash_table. Update dependent calls and types.
22076 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
22077 Change type to hash_table. Update dependent calls and types.
22079 * tree-ssa-phiopt.c (seen_ssa_names):
22080 Change type to hash_table. Update dependent calls and types.
22082 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
22083 Change type to hash_table. Update dependent calls and types.
22085 * tree-ssa-uncprop.c (equiv):
22086 Change type to hash_table. Update dependent calls and types.
22088 2013-04-25 Jakub Jelinek <jakub@redhat.com>
22090 PR rtl-optimization/57003
22091 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
22092 call note_stores with kill_clobbered_value callback again after
22093 killing regs_invalidated_by_call.
22095 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22097 * config/aarch64/aarch64-simd.md
22098 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
22099 (aarch64_simd_bsl<mode>): Likewise.
22100 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
22102 2013-04-25 Marek Polacek <polacek@redhat.com>
22104 PR tree-optimization/57066
22105 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
22107 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22109 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
22111 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22113 * config/aarch64/aarch64-builtins.c
22114 (aarch64_fold_builtin): New.
22115 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
22116 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
22117 * config/aarch64/aarch64-simd-builtins.def (abs): New.
22118 * config/aarch64/arm_neon.h
22119 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
22121 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22122 Tejas Belagod <tejas.belagod@arm.com>
22124 * config/aarch64/aarch64-builtins.c
22125 (aarch64_gimple_fold_builtin): New.
22126 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
22127 * config/aarch64/aarch64-simd-builtins.def (addv): New.
22128 * config/aarch64/aarch64-simd.md (addpv4sf): New.
22129 (addvv4sf): Update.
22130 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
22132 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22134 * config/aarch64/aarch64.md
22135 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
22137 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22139 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
22140 (*ngcsi_uxtw): New pattern.
22142 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22143 Julian Brown <julian@codesourcery.com>
22145 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
22146 (TB_DREG): Add T_V4HF.
22147 (v4hf_UP): New macro.
22148 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22149 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22150 Handle initialisation of V4HF. Adjust initialisation of reinterpret
22152 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22153 (arm_vector_mode_supported_p): Handle V4HF.
22154 (arm_mangle_map): Handle V4HFmode.
22155 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
22156 * config/arm/arm_neon_builtins.def: Add entries for
22157 vcvtv4hfv4sf, vcvtv4sfv4hf.
22158 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
22159 (neon_vcvtv4hfv4sf): Likewise.
22160 * config/arm/neon-gen.ml: Handle half-precision floating point
22162 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
22163 * config/arm/arm_neon.h: Regenerate.
22164 * config/arm/neon.ml (type elts): Add F16.
22165 (type vectype): Add T_float16x4, T_floatHF.
22166 (type vecmode): Add V4HF.
22167 (type features): Add Requires_FP_bit feature.
22168 (elt_width): Handle F16.
22169 (elt_class): Likewise.
22170 (elt_of_class_width): Likewise.
22171 (mode_of_elt): Refactor.
22172 (type_for_elt): Handle F16, fix error messages.
22173 (vectype_size): Handle T_float16x4.
22174 (vcvt_sh): New function.
22175 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
22176 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
22177 (string_of_mode): Handle V4HF.
22178 * doc/arm-neon-intrinsics.texi: Regenerate.
22180 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22182 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
22183 format specifier in 'X' case.
22185 2013-04-25 Alan Modra <amodra@gmail.com>
22188 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
22189 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
22190 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
22191 Repeat for many other rotate/shift and mask patterns using subregs.
22192 Name lshiftrt insns.
22193 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
22194 on WORDS_BIG_ENDIAN.
22196 2013-04-25 Alan Modra <amodra@gmail.com>
22198 * config.gcc: Support little-endian powerpc-linux targets.
22199 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
22200 (LINK_OS_LINUX_SPEC): Define.
22201 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
22202 Preserve MASK_LITTLE_ENDIAN.
22203 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
22204 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
22205 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
22206 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
22207 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
22208 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
22209 Correct fp word order for little-endian. Don't shift toc entries
22210 smaller than a word for little-endian.
22211 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
22212 (bswapdi2 splits): Correct low-part subreg for little-endian.
22213 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
22214 low/high where such is correct only for be.
22215 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
22216 little-endian for -mcall-aixdesc.
22218 2013-04-25 Alan Modra <amodra@gmail.com>
22220 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
22221 replace_equiv_address_nv.
22223 2013-04-25 Alan Modra <amodra@gmail.com>
22225 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
22227 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22230 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22231 * rtl.h (struct rtx_def): ...
22233 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22235 PR rtl-optimizations/57046
22236 * lra-constraints (split_reg): Set up lra_risky_transformations_p
22237 for multi-reg splits.
22239 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
22241 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
22243 2013-04-24 Sterling Augustine <saugustine@google.com>
22245 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
22246 (comp_dir_string, debug_str_dwo_section): New.
22247 (DEBUG_STR_DWO_SECTION): Rename to ...
22248 (DEBUG_DWO_STR_SECTION): ... this.
22249 (DEBUG_NORM_STR_SECTION): Delete.
22250 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
22251 (DEBUG_STR_DWO_SECTION_FLAGS): New.
22252 (find_AT_string): Move most logic to ...
22253 (find_AT_string_in_table): ... here. New.
22254 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
22255 add_skeleton_AT_string. Delete logic.
22256 (output_skeleton_debug_sections): Remove call to
22257 add_top_level_skeleton_die_attrs.
22258 (add_comp_dir_attribute): Move logic to comp_dir_string.
22259 (dwarf2out_init): Initialize debug_str_dwo_section.
22260 (output_indirect_string): Call find_string_form.
22261 (output_indirect_strings): Rewrite.
22262 (prune_unused_types): Empty skeleton_debug_str_hash.
22263 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
22264 (dwarf2out_finish): Call output_indirect_strings.
22266 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22268 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
22270 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22272 * rtl.h (struct rtx_def): Add comment for field jump.
22273 (LRA_SUBREG_P): New macro.
22274 * recog.c (register_operand): Check LRA_SUBREG_P.
22275 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22277 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22278 (lra_final_code_change): Skip subreg change for operators.
22279 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22280 if there are no operand changes.
22281 * lra-constraints.c (curr_insn_set): New.
22282 (match_reload): Set LRA_SUBREG_P.
22283 (emit_spill_move): Ditto.
22284 (check_and_process_move): Use curr_insn_set. Process only single
22285 set insns. Don't initialize sec_mem_p and change_p.
22286 (simplify_operand_subreg): Use LRA_SUBREG_P.
22287 (reg_in_class_p): New function.
22288 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22289 of #ifdef. Add code to remove cycling.
22290 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22291 non-null disp. Reload inner instead of disp when base and index
22292 are null. Try to put lo_sum into register.
22293 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22294 (check_and_process_move): Move code for move cost check to
22295 simple_move_p. Remove equiv_substitution.
22296 (simple_move_p): New function.
22297 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22298 curr_insn_set. Call check_and_process_move only for single set
22299 insns. Use the new function. Move call of check_and_process_move
22300 after operand equiv substitution and address process.
22302 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
22304 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
22305 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
22306 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
22308 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22310 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
22312 2013-04-24 Marek Polacek <polacek@redhat.com>
22314 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
22315 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
22316 (select_loops_exit_conditions): Likewise.
22317 (number_of_iterations_for_all_loops): Likewise.
22318 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
22319 (scev_analysis): Likewise.
22321 2013-04-02 Catherine Moore <clm@codesourcery.com>
22322 Chao-ying Fu <fu@mips.com>
22324 * config/mips/micromips.md (jraddiusp): New pattern.
22325 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
22326 instruction if possible.
22328 2013-04-24 Alan Modra <amodra@gmail.com>
22330 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
22332 2013-04-24 Julian Brown <julian@codesourcery.com>
22333 Chung-Lin Tang <cltang@codesourcery.com>
22335 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
22336 dependency behavior in enumeration type DIE generation. Add TODO note
22337 to comments about future DW_FORM_sdata/udata re-work of related code.
22339 2013-04-23 Lawrence Crowl <crowl@google.com>
22341 * Makefile.in: Update as needed below.
22343 * hash-table.h (class hash_table):
22344 Correct many methods with parameter types compare_type to the correct
22345 value_type. (Correct code was unlikely to notice the change.)
22346 (hash_table::elements_with_deleted) New.
22347 (class hashtable::iterator): New.
22348 (hashtable::begin()): New.
22349 (hashtable::end()): New.
22350 (FOR_EACH_HASH_TABLE_ELEMENT): New.
22352 * statistics.c (statistics_hashes):
22353 Change type to hash_table. Update dependent calls and types.
22355 * tree-into-ssa.c (var_infos):
22356 Change type to hash_table. Update dependent calls and types.
22358 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
22359 Change type to hash_table. Update dependent calls and types.
22361 * tree-ssa-loop-im.c (struct mem_ref.refs):
22362 Change type to hash_table. Update dependent calls and types.
22364 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
22365 Change type to hash_table. Update dependent calls and types.
22367 * tree-ssa-sccvn.c (vn_tables_s::nary):
22368 Change type to hash_table. Update dependent calls and types.
22369 (vn_tables_s::phis): Likewise.
22370 (vn_tables_s::references): Likewise.
22372 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
22373 (vn_reference_eq): Update parameter and return types.
22375 * tree-ssa-structalias.c (pointer_equiv_class_table):
22376 Change type to hash_table. Update dependent calls and types.
22377 (location_equiv_class_table): Likewise.
22379 * tree-vect-data-refs.c: Consequential changes for making
22380 peeling a hash_table.
22382 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
22383 (destroy_loop_vec_info): Dependent hash_table update.
22385 * tree-vectorizer.h (peeling_htab):
22386 Change type to hash_table. Update dependent calls and types.
22388 2013-04-23 Shiva Chen <shiva0217@gmail.com>
22390 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
22391 to check the register content is equal or not.
22392 * lra-constraints.c (match_reload): Use lra_assign_reg_val
22393 to assign register content record.
22394 * lra-eliminations.c (update_reg_eliminate): Use
22395 lra_update_reg_val_offset to update register content offset.
22396 * lra-int.h (struct lra_reg): Add offset member.
22397 (lra_reg_val_equal_p): New static inline function.
22398 (lra_update_reg_val_offset): New static inline function.
22399 (lra_assign_reg_val): New static inline function.
22400 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
22401 to assign register content record.
22402 (initialize_lra_reg_info_element): Initial offset to zero.
22404 2013-04-23 Catherine Moore <clm@codesourcery.com>
22406 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
22407 operands. Record compression.
22409 2013-04-23 Xinliang David Li <davidxl@google.com>
22411 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
22413 2013-04-23 Richard Biener <rguenther@suse.de>
22415 PR middle-end/57036
22416 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
22417 parameter, only add abnormal goto edges from the copied body
22418 if the call could perform abnormal gotos.
22419 (copy_cfg_body): Adjust.
22421 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
22423 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
22425 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
22427 * coretypes.h (gimple_stmt_iterator): Add struct to make
22430 2013-04-23 Richard Biener <rguenther@suse.de>
22432 PR tree-optimization/57026
22433 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
22434 from SSA names occuring in abnormal PHI nodes.
22436 2013-04-22 Andi Kleen <ak@linux.intel.com>
22438 * lto/lto.c (print_lto_report_1): Fix LTO report names.
22440 2013-04-22 Andi Kleen <ak@linux.intel.com>
22442 * lto/lto.c (print_lto_report_1): Declare early.
22443 (read_cgraph_and_symbols): Call print_lto_report_1 early.
22445 2013-04-22 Andi Kleen <ak@linux.intel.com>
22447 * common.opt (-flto-report-wpa): Add.
22448 * doc/invoke.texi (-flto-report-wpa): Add.
22449 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
22452 2013-04-22 Xinliang David Li <davidxl@google.com>
22454 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
22455 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
22456 * Makefile.in: New dependency
22458 David Daney <ddaney.cavm@gmail.com>
22460 * configure.ac (gcc_cv_as_micromips_support): Use the
22461 --fatal-warnings option.
22462 * configure: Regenerate.
22464 2013-04-22 Marek Polacek <polacek@redhat.com>
22467 * tsan.c (instrument_expr): Don't instrument expression
22468 in case its size is zero.
22470 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
22474 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22476 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22478 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22480 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
22481 (gimple_stmt_iterator): New typedef.
22482 * gimple.h (gimple_stmt_iterator): Rename to...
22483 (gimple_stmt_iterator_d): ... This.
22484 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
22485 trees be valid for GIMPLE and GENERIC.
22486 (TARGET_GIMPLE_FOLD_BUILTIN): New.
22487 * gimple-fold.c (gimple_fold_call): Call target hook
22488 gimple_fold_builtin.
22489 * hooks.c (hook_bool_gsiptr_false): New.
22490 * hooks.h (hook_bool_gsiptr_false): New.
22491 * target.def (fold_stmt): New.
22492 * doc/tm.texi: Regenerate.
22494 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
22497 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
22498 a set sp if no stack realignment.
22500 2013-04-22 Nick Clifton <nickc@redhat.com>
22502 * config.gcc (tilegx-linux): Extend extra_objs rather than
22504 (tilepro-linux): Likewise.
22506 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22508 * config/aarch64/aarch64-builtins.c
22510 (CF0, CF1, CF2, CF3, CF4, CF10): New.
22511 (VAR<1-12>): Add MAP parameter.
22512 (BUILTIN_*): Likewise.
22513 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
22514 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
22515 (aarch64_ushl_n<mode>): Likewise.
22516 (aarch64_sshr_n<mode>): Likewise.
22517 (aarch64_ushr_n<mode>): Likewise.
22518 (aarch64_<maxmin><mode>): Likewise.
22519 (aarch64_sqrt<mode>): Likewise.
22520 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
22521 (vshr<q>_n_*): Likewise.
22523 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22525 * config/aarch64/aarch64-builtins.c
22526 (aarch64_simd_builtin_type_mode): Handle SF types.
22528 (BUILTIN_GPF): Define.
22529 (aarch64_init_simd_builtins): Handle SF types.
22530 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
22531 (frecps): Likewise.
22532 (frecpx): Likewise.
22533 * config/aarch64/aarch64-simd.md
22534 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
22535 (aarch64_frecpe<mode>): New.
22536 (aarch64_frecps<mode>): Likewise.
22537 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
22538 (v8type): Add frecp<esx>.
22539 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
22540 (aarch64_frecps<mode>): Likewise.
22541 * config/aarch64/iterators.md (FRECP): New.
22542 (frecp_suffix): Likewise.
22543 * config/aarch64/arm_neon.h
22544 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
22546 2013-04-22 Christian Bruel <christian.bruel@st.com>
22549 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
22550 (REG_CLASS_NAMES): Idem.
22551 (REG_CLASS_CONTENTS): Idem.
22552 (REGCLASS_HAS_FP_REG): Idem.
22553 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
22554 (sh_conditional_register_usage): Idem.
22556 2013-04-21 Jeff Law <law@redhat.com>
22558 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
22559 (ssa_forward_propagate_and_combine): Use it.
22561 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22563 * lra.c: Update the flow chart diagram.
22565 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22567 PR rtl-optimization/56847
22568 * lra-constraints.c (process_alt_operands): Discourage alternative
22569 with non-matche doffsettable memory constraint fro memory with
22572 2013-04-19 Richard Biener <rguenther@suse.de>
22574 PR tree-optimization/56982
22575 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
22577 * gimplify.c (gimplify_call_expr): Notice special calls.
22578 (gimplify_modify_expr): Likewise.
22579 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
22580 abnormal control flow receivers.
22581 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
22582 in the same way as cfun->has_nonlocal_labels.
22583 (gimple_purge_dead_abnormal_call_edges): Likewise.
22584 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
22585 receivers start a basic-block.
22587 2013-04-19 Richard Biener <rguenther@suse.de>
22589 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
22591 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
22592 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
22593 (SLP_TREE_LOAD_PERMUTATION): Add.
22594 (vect_transform_slp_perm_load): Adjust prototype.
22595 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
22596 (vect_free_slp_instance): Likewise.
22597 (vect_create_new_slp_node): Likewise.
22598 (vect_supported_slp_permutation_p): Remove.
22599 (vect_slp_rearrange_stmts): Adjust.
22600 (vect_supported_load_permutation_p): Likewise. Inline
22601 vect_supported_slp_permutation_p here.
22602 (vect_analyze_slp_instance): Compute load permutations per
22603 slp node instead of per instance.
22604 (vect_get_slp_defs): Adjust.
22605 (vect_transform_slp_perm_load): Likewise.
22606 (vect_schedule_slp_instance): Remove redundant code.
22607 (vect_schedule_slp): Remove hack for PR56270, add it ...
22608 * tree-vect-stmts.c (vectorizable_load): ... here, do not
22609 CSE loads for SLP. Adjust.
22611 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22613 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
22614 spelling in two comments.
22616 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22619 * config/arm/arm.c (load_multiple_sequence): Require SP
22620 as base register for loads if SP is in the register list.
22622 2013-04-19 Martin Jambor <mjambor@suse.cz>
22624 PR tree-optimization/56718
22625 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
22626 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
22627 and made public. Adjusted all callers.
22628 (ipa_intraprocedural_devirtualization): New function.
22629 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
22630 (ipa_intraprocedural_devirtualization): Likewise.
22631 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
22633 2013-04-19 Richard Biener <rguenther@suse.de>
22635 PR tree-optimization/57000
22636 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
22638 2013-04-19 Terry Guo <terry.guo@arm.com>
22640 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
22642 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
22643 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
22644 (cortex_m4_fmacs): Use new reservations.
22645 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
22647 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
22649 PR rtl-optimization/56999
22650 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
22652 (lra_coalesce): Remove split_origin_bitmap and related code.
22653 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
22654 ranges if necessary.
22656 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
22658 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
22660 (ix86_expand_call): Remove clobbered_registers array and use
22661 x86_64_ms_sysv_extra_clobbered_registers instead.
22662 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
22664 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
22666 * config/i386/i386.md (*call_rex64_ms_sysv): Use
22667 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
22668 (*call_value_rex64_ms_sysv): Ditto.
22670 2013-04-18 Cary Coutant <ccoutant@google.com>
22672 * dwarf2out.c (output_pubnames): Check die_perennial_p of
22673 parent instead of die_mark.
22675 2013-04-18 Diego Novillo <dnovillo@google.com>
22677 * gimple.c (create_gimple_tmp): New.
22678 (get_expr_type): New.
22679 (build_assign): New.
22680 (build_type_cast): New.
22681 * gimple.h (enum ssa_mode): Define.
22682 (gimple_seq_set_location): New.
22683 * asan.c (build_check_stmt): Change some gimple_build_* calls
22684 to use build_assign and build_type_cast.
22686 2013-04-18 Richard Biener <rguenther@suse.de>
22688 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
22689 handle negative step. Remove redundant checks.
22690 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
22691 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
22692 for negative step and grouped loads fail to vectorize.
22694 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22696 * emit-rtl.c (reset_insn_used_flags): New function.
22697 (reset_all_used_flags): Use it.
22698 (verify_insn_sharing): New function.
22699 (verify_rtl_sharing): Fix verification for SEQUENCEs.
22701 2013-04-18 Jakub Jelinek <jakub@redhat.com>
22703 PR tree-optimization/56984
22704 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
22705 and (x >> M) >= N don't register any assertion if N << M is the
22708 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22710 * lower-subreg.c (resolve_simple_move): If called self-recursive,
22711 do not delete_insn insns that have not yet been emitted, only
22712 unlink them with remove_insn.
22713 * df-scan.c (df_insn_delete): Revert r197492.
22715 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
22717 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
22718 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
22720 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22722 * config/arm/arm.md (movsicc_insn): Convert define_insn into
22723 define_insn_and_split.
22724 (and_scc,ior_scc,negscc): Likewise.
22725 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
22727 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22729 * config/arm/arm.c (use_return_insn): Return 0 for targets that
22730 can benefit from using a sequence of LDRD instructions in epilogue
22731 instead of a single LDM instruction.
22733 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
22736 * doc/extend.texi: Fix typo.
22738 2013-04-17 Richard Biener <rguenther@suse.de>
22740 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
22741 (vect_build_slp_tree): ... here.
22742 (vect_build_slp_tree_1): Compute which stmts of the SLP group
22743 match. Remove special-casing of mismatched complex loads.
22744 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
22745 re-try the match with swapped commutative operands.
22746 (vect_supported_load_permutation_p): Remove special-casing of
22747 mismatched complex loads.
22748 (vect_analyze_slp_instance): Adjust.
22750 2013-04-17 Richard Biener <rguenther@suse.de>
22752 PR rtl-optimization/56921
22753 * cfgloop.h (struct loop): Add simple_loop_desc member.
22754 (struct niter_desc): Mark with GTY(()).
22755 (simple_loop_desc): Do not use aux field but simple_loop_desc.
22756 * loop-iv.c (get_simple_loop_desc): Likewise.
22757 (free_simple_loop_desc): Likewise.
22760 2013-04-16 Richard Biener <rguenther@suse.de>
22762 PR rtl-optimization/56921
22763 * loop-init.c (pass_rtl_move_loop_invariants): Add
22764 TODO_do_not_ggc_collect to todo_flags_finish.
22765 (pass_rtl_unswitch): Same.
22766 (pass_rtl_unroll_and_peel_loops): Same.
22767 (pass_rtl_doloop): Same.
22769 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
22771 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
22772 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
22773 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
22775 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
22776 * tree-streamer.c (record_common_node): Adjust reference in comment.
22778 2013-04-17 Terry Guo <terry.guo@arm.com>
22780 * config/arm/cortex-m4.md: Add a new bypass.
22782 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22784 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
22786 (*subs_<optab><mode>_multp2): New pattern.
22787 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22788 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22790 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22792 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
22793 (*subs_mul_imm_<mode>): New pattern.
22795 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
22798 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
22799 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
22800 (vsx_movti_32bit): Change j->wa to O->wa.
22802 2013-04-16 Richard Biener <rguenther@suse.de>
22804 PR rtl-optimization/56921
22805 * loop-init.c (pass_rtl_move_loop_invariants): Add
22806 TODO_do_not_ggc_collect to todo_flags_finish.
22807 (pass_rtl_unswitch): Same.
22808 (pass_rtl_unroll_and_peel_loops): Same.
22809 (pass_rtl_doloop): Same.
22811 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
22813 * config/arm/arm.c (emit_multi_reg_push): New declaration
22814 for an existing function.
22815 (arm_emit_strd_push): New function.
22816 (arm_expand_prologue): Used here.
22817 (arm_emit_ldrd_pop): New function.
22818 (arm_expand_epilogue): Used here.
22819 (arm_get_frame_offsets): Update condition.
22820 (arm_emit_multi_reg_pop): Add a special case for load of a single
22821 register with writeback.
22823 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
22825 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
22828 2013-04-16 Richard Biener <rguenther@suse.de>
22830 PR tree-optimization/56756
22831 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
22832 (first_mem_ref_loc): New.
22833 (execute_sm): Place the load temporarily before a previous
22834 access instead of in the latch edge to ensure its SSA dependencies
22835 are defined at points dominating the load.
22837 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
22839 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
22840 correct fix by moving header and footer insn to the footer of
22841 the merged basic block. Clear BB_END of the merged-away block.
22843 PR middle-end/43631
22844 * emit-rtl.c (make_note_raw): New function.
22845 (link_insn_into_chain): New static inline function.
22846 (add_insn): Use it.
22847 (add_insn_before, add_insn_after): Factor insn chain linking code...
22848 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
22849 using link_insn_into_chain.
22850 (note_outside_basic_block_p): New helper function for emit_note_after
22851 and emit_note_before.
22852 (emit_note_after): Use nobb variant of add_insn_after if the note
22853 should not be contained in a basic block.
22854 (emit_note_before): Use nobb variant of add_insn_before if the note
22855 should not be contained in a basic block.
22856 (emit_note_copy): Use make_note_raw.
22857 (emit_note): Likewise.
22858 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
22859 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
22860 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
22861 the moved barrier the tail of the basic block it follows.
22862 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
22864 2013-04-15 Jakub Jelinek <jakub@redhat.com>
22866 PR tree-optimization/56962
22867 * gimple-ssa-strength-reduction.c (record_increment): Only set
22868 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
22869 either rhs1 or rhs2 is equal to c->base_expr.
22871 2013-04-15 Richard Biener <rguenther@suse.de>
22873 PR tree-optimization/56933
22874 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
22876 (GROUP_READ_WRITE_DEPENDENCE): Remove.
22877 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
22878 * tree-vect-data-refs.c (vect_analyze_group_access): Move
22879 dependence check ...
22880 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
22882 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
22883 GROUP_READ_WRITE_DEPENDENCE.
22885 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22887 * emit-rtl.c (reset_all_used_flags): New function.
22888 (verify_rtl_sharing): Call reset_all_used_flags before and after
22889 performing the checks.
22891 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22893 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
22894 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
22895 * config/arm/constraints.md (De): New constraint.
22896 * config/arm/neon.md (anddi3_neon): Delete.
22897 (neon_vand<mode>): Expand to standard anddi3 pattern.
22898 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
22899 Move earlier in the file.
22900 (neon_inv_logic_op2): Likewise.
22901 (arm_anddi_operand_neon): New predicate.
22903 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22905 * configure.ac (gcc_cv_ld_as_needed): Set
22906 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
22907 Use -z ignore, -z record on *-*-solaris2*.
22908 (HAVE_LD_AS_NEEDED): Update comment.
22909 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
22910 * configure: Regenerate.
22911 * config.in: Regenerate.
22912 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
22913 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
22914 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
22915 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
22916 equivalents. Fix markup.
22917 * doc/tm.texi: Regenerate.
22919 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
22921 * config/i386/i386.opt: New option mstack-protector-guard=.
22922 * config/i386/i386-opts.h: Add enum stack_protector_guard.
22923 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
22924 TARGET_SSP_TLS_GUARD.
22925 * config/i386/i386.c (ix86_option_override_internal): Set
22926 ix86_stack_protector_guard.
22927 * config/i386/i386.md (stack_protect_set): Enable for
22928 TARGET_SSP_TLS_GUARD only.
22929 (stack_protect_set_<mode>): Ditto.
22930 (stack_protect_test): Ditto.
22931 (stack_protect_test_<mode>): Ditto.
22932 * doc/invoke.texi (i386 Option): Document.
22934 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
22937 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
22938 (S_MODES): Set H_MODE bit.
22939 (SF_MODES): Set only S_MODE and SF_MODE bits.
22940 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
22941 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
22942 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
22943 <MODE_FLOAT>: Likewise.
22945 2013-04-15 Joey Ye <joey.ye@arm.com>
22947 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
22949 2013-04-15 Joey Ye <joey.ye@arm.com>
22951 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
22953 (thumb_far_jump_used_p): Count instruction size and set
22956 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
22958 * reorg.c (fill_simple_delay_slots): Reindent block of code.
22959 * resource.c (mark_target_live_regs): Reformat conditional block.
22961 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22963 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
22964 notes, they are emitted only just before final.
22965 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
22967 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22969 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
22970 * cfgrtl.c (delete_insn): Call it here instead.
22971 * lra-spills.c (lra_final_code_change): Use delete_insn.
22972 * haifa-sched.c (sched_remove_insn): Likewise.
22973 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
22974 returning to the nop pool.
22975 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
22976 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
22978 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22980 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
22981 * doc/tm.texi: Regenerated.
22983 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
22985 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
22988 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22990 * df-core.c (df_find_def): Compare register numbers.
22991 (df_find_use): Likewise.
22993 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
22996 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
22997 lra_in_progress for return.
22999 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23001 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
23002 define_insn into define_insn_and_split and emit movsicc patterns.
23004 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23006 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
23008 2013-04-12 Richard Biener <rguenther@suse.de>
23010 * tree-pass.h (TODO_do_not_ggc_collect): New.
23011 * passes.c (execute_one_ipa_transform_pass): Honor
23012 TODO_do_not_ggc_collect.
23013 (execute_one_pass): Likewise.
23016 2013-04-10 Richard Biener <rguenther@suse.de>
23018 * passes.c (init_optimization_passes): Remove reload pass.
23019 * ira.c (do_reload): Merge into ...
23021 (rest_of_handle_reload): Remove.
23022 (pass_reload): Likewise.
23023 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23024 of reload for vzeroupper pass placement.
23026 2013-04-12 Jakub Jelinek <jakub@redhat.com>
23028 PR tree-optimization/56918
23029 PR tree-optimization/56920
23030 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
23031 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
23032 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
23033 use wide_mul_with_sign method.
23035 2013-04-12 Richard Biener <rguenther@suse.de>
23037 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
23038 not be considered a gimple constant.
23040 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
23042 * fold-const.c (const_binop): Handle vector shifts by a scalar.
23043 (fold_binary_loc): Call const_binop also for mixed vector-scalar
23046 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
23047 Jakub Jelinek <jakub@redhat.com>
23049 * opts.c: Include diagnostic-color.h.
23050 (common_handle_option): Handle OPT_fdiagnostics_color_.
23051 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
23052 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
23053 (diagnostic-color.o): New.
23054 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
23055 (diagnostic_color_rule): New enum.
23056 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
23057 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
23058 the location string.
23059 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
23060 either NULL, or color kind.
23061 * diagnostic-color.c: New file.
23062 * diagnostic-color.h: New file.
23063 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
23065 * doc/invoke.texi (-fdiagnostics-color): Document.
23066 * pretty-print.h (pp_show_color): Define.
23067 (struct pretty_print_info): Add show_color field.
23068 * diagnostic.c: Include diagnostic-color.h.
23069 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
23070 macros. Colorize error:, warning: etc. strings and also the location
23072 (diagnostic_show_locus): Colorize the caret line.
23073 * pretty-print.c: Include diagnostic-color.h.
23074 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
23075 inside of %< %> quotes or quoted through q format modifier.
23077 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23079 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
23081 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23083 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
23084 code in CC_NZ mode.
23085 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
23088 2013-04-11 Marek Polacek <polacek@redhat.com>
23090 PR tree-optimization/48184
23091 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
23093 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
23095 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
23096 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
23097 (skip_simple_arithmetic): Tidy up.
23098 * tree.h (skip_simple_constant_arithmetic): Declare.
23100 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23102 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
23104 2013-04-11 Richard Biener <rguenther@suse.de>
23106 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23107 generate vector constants.
23109 2013-04-11 Richard Biener <rguenther@suse.de>
23111 PR tree-optimization/56878
23112 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
23113 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
23115 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
23116 Prefer to align the DR with the most invariant base address.
23118 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23120 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
23123 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
23125 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
23126 floating-point vector comparisons against 0.
23128 2013-04-11 Jakub Jelinek <jakub@redhat.com>
23130 PR tree-optimization/56899
23131 * fold-const.c (extract_muldiv_1): Apply distributive law
23132 only if TYPE_OVERFLOW_WRAPS (ctype).
23134 2013-04-11 Bin Cheng <bin.cheng@arm.com>
23137 * ira-costs.c (scan_one_insn): Check whether the source rtx of
23138 loading has side effect.
23140 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
23142 * config/sparc/sparc.c: Include tree-pass.h.
23143 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
23144 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
23145 head of file. Change return type. Split off gate function.
23146 (sparc_gate_work_around_errata): New function.
23147 (pass_work_around_errata): New pass definition.
23148 (insert_pass_work_around_errata) New pass insert definition to
23149 insert pass_work_around_errata just after delayed-branch scheduling.
23150 (sparc_option_override): Insert the pass.
23151 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
23153 2013-04-10 David S. Miller <davem@davemloft.net>
23155 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
23156 or -mcpu=hypersparc.
23158 * target.def (cstore_mode): New hook.
23159 * target.h: Include insn-codes.h
23160 * targhooks.c: Likewise.
23161 (default_cstore_mode): New function.
23162 * targhooks.h: Declare it.
23163 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
23164 * doc/tm.texi: Rebuild.
23165 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
23166 target hook, rather than inspecting the insn_data.
23167 * config/sparc/sparc.c (sparc_cstore_mode): New function.
23168 (TARGET_CSTORE_MODE): Redefine.
23169 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
23171 * config/sparc/predicates.md (cstore_result_operand): New special
23173 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
23174 Use it for operand 0.
23175 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
23176 (*snesi_special): Likewise.
23177 (*snesi_zero): Likewise.
23178 (*seqsi_zero): Likewise.
23179 (*sltu_insn): Likewise.
23180 (*sgeu_insn): Likewise.
23181 (*seqdi_special): Make operand 0 and comparison operation be of
23183 (*snedi_special): Likewise.
23184 (*snedi_special_vis3): Likewise.
23185 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
23186 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
23187 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
23188 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
23189 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
23190 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
23191 (*sltu_extend_sp64): Likewise.
23192 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
23193 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
23194 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
23195 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
23196 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
23198 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
23200 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
23201 (aarch64_start_file): Use the new function.
23203 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23204 Jason Merrill <jason@redhat.com>
23206 * common.opt: Add -gdwarf.
23207 * opts.c (common_handle_option): Handle it.
23208 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
23210 2013-04-10 Richard Biener <rguenther@suse.de>
23212 * passes.c (execute_todo): Do not call ggc_collect conditional here.
23213 (execute_one_ipa_transform_pass): But unconditionally here.
23214 (execute_one_pass): And here.
23215 (init_optimization_passes): Remove reload pass.
23216 * tree-pass.h (TODO_ggc_collect): Remove.
23217 (pass_reload): Likewise.
23218 * ira.c (do_reload): Merge into ...
23220 (rest_of_handle_reload): Remove.
23221 (pass_reload): Likewise.
23222 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23223 of reload for vzeroupper pass placement.
23224 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
23225 and todo_flags_finish of all passes.
23227 2013-04-10 Richard Biener <rguenther@suse.de>
23229 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
23230 first_const_oprnd field, rename first_def_type to first_op_type.
23231 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
23232 (vect_get_and_check_slp_defs): Always use the type of the
23233 operand. Allow mixed vect_external_def, vect_constant_def types.
23234 (vect_get_constant_vectors): Handle mixed vect_external_def,
23235 vect_constant_def types.
23237 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
23239 PR tree-optimization/55524
23240 * tree-ssa-math-opts.c
23241 (convert_mult_to_fma): Don't use an fms construct
23242 when we don't have an fms operation, but fnma, and it looks
23243 likely that we'll be able to use the latter.
23245 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
23247 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
23249 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
23250 inline fail caused by overwritable functions.
23252 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
23254 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
23255 unnecessary bits in the constant power of two case.
23257 2013-04-10 Richard Biener <rguenther@suse.de>
23259 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
23260 broken code swapping operands.
23261 (vect_build_slp_tree): Do not compute load permutations here.
23262 (vect_analyze_slp_instance): Compute load permutations here,
23263 after building the SLP tree.
23265 2013-04-09 Christian Bruel <christian.bruel@st.com>
23267 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
23268 of next/prev_real_insn.
23270 2013-04-09 Jan Hubicka <jh@suse.cz>
23272 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
23273 Drop aliased parameter.
23274 (function_and_variable_visibility): Do not handle alias pairs.
23275 * cgraph.c (varpool_externally_visible_p): Update prototype.
23276 * varpool.c (varpool_add_new_variable): Update.
23278 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23280 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
23282 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
23284 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
23286 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
23288 2013-04-09 Marek Polacek <polacek@redhat.com>
23290 PR tree-optimization/48762
23291 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
23293 2013-04-09 Richard Biener <rguenther@suse.de>
23295 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
23297 (vect_build_slp_tree): Likewise.
23298 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
23299 calculating the cost of a SLP instance.
23300 (vect_analyze_slp_instance): Use it from here, after building
23303 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23305 PR middle-end/56883
23306 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
23307 expand_omp_for_static_chunk): Use simple_p = true in
23308 force_gimple_operand_gsi calls when assigning to addressable decls.
23310 2013-04-09 Jeff Law <law@redhat.com>
23312 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
23313 when the boolean was created by converting a wider object which
23314 had a boolean range.
23316 2013-04-09 Richard Biener <rguenther@suse.de>
23318 * tree-vectorizer.h (slp_void_p): Remove.
23319 (slp_tree): Typedef before _slp_tree declaration.
23320 (struct _slp_tree): Use a vector of slp_tree as children.
23321 (vect_get_place_in_interleaving_chain): Remove.
23322 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
23324 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
23326 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
23327 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
23328 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
23329 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
23330 Use slp_node instead of slp_void_p and adjust.
23332 2013-04-09 Richard Biener <rguenther@suse.de>
23334 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
23335 work that is not necessary.
23337 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23339 PR tree-optimization/56854
23340 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
23341 forward into clobber stmts if it would change MEM_REF lhs into
23344 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
23346 * tree.c (type_hash_lookup, type_hash_add): Make static.
23347 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
23349 2013-04-09 Richard Biener <rguenther@suse.de>
23351 * tree.h (unsave_expr_now): Remove.
23352 * tree-inline.c (mark_local_for_remap_r): Remove.
23353 (unsave_expr_1): Likewise.
23354 (unsave_r): Likewise.
23355 (unsave_expr_now): Likewise.
23356 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
23357 (propagate_tree_value): Likewise.
23359 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
23361 * doc/rtl.texi (sequence): Rewrite documentation to match the
23362 current use of SEQUENCE rtl objects.
23363 * rtl.def (SEQUENCE): Likewise.
23365 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
23366 Update documentation.
23367 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
23368 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
23370 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
23372 2013-04-08 Teresa Johnson <tejohnson@google.com>
23374 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
23375 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
23377 (estimate_edge_size_and_time): Add comment to suggest using rounding
23379 (estimate_node_size_and_time): Ditto.
23380 (remap_edge_change_prob): Use helper rounding divide methods.
23381 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
23382 (gimple_mod_pow2_value_transform): Ditto.
23383 (gimple_mod_subtract_transform): Ditto.
23384 (gimple_ic_transform): Ditto.
23385 (gimple_stringops_transform): Ditto.
23386 * stmt.c (conditional_probability): Ditto.
23387 (emit_case_dispatch_table): Ditto.
23388 * lto-cgraph.c (merge_profile_summaries): Ditto.
23389 * tree-optimize.c (execute_fixup_cfg): Ditto.
23390 * cfgcleanup.c (try_forward_edges): Ditto.
23391 * cfgloopmanip.c (scale_loop_profile): Ditto.
23393 (duplicate_loop_to_header_edge): Ditto.
23394 (lv_adjust_loop_entry_edge): Ditto.
23395 * tree-vect-loop.c (vect_transform_loop): Ditto.
23396 * profile.c (compute_branch_probabilities): Ditto.
23397 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
23398 * lto-streamer-in.c (input_cfg): Ditto.
23399 * gimple-streamer-in.c (input_bb): Ditto.
23400 * ipa-cp.c (update_profiling_info): Ditto.
23401 (update_specialized_profile): Ditto.
23402 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
23403 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
23405 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23406 (compute_trg_info): Ditto.
23407 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
23408 (purge_dead_edges): Ditto.
23409 * loop-unswitch.c (unswitch_loop): Ditto.
23410 * cgraphclones.c (cgraph_clone_edge): Ditto.
23411 (cgraph_clone_node): Ditto.
23412 * tree-inline.c (copy_bb): Ditto.
23413 (copy_edges_for_bb): Ditto.
23414 (initialize_cfun): Ditto.
23415 (copy_cfg_body): Ditto.
23416 (expand_call_inline): Ditto.
23418 2013-04-08 Kai Tietz <ktietz@redhat.com>
23420 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
23421 TARGET_CYGWIN64 by TARGET_64BIT.
23423 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23425 * config/epiphany/epiphany.md (GPR_1): New constant.
23426 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
23427 * config/epiphany/epiphany.c (gen_compare_reg):
23428 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
23429 is already in place.
23430 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
23431 Don't require being called during rtl expansion; If y operlaps r0,
23433 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
23434 (epiphany_expand_epilogue): Likewise.
23436 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
23437 Don't use CC_FPmode for ORDERED / UNORDERED.
23438 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
23440 * config/epiphany/constraints.md (CnL): New constraint.
23441 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
23442 * config/epiphany/predicates.md (add_operand): Allow 1024.
23444 * config/epiphany/epiphany.md (logical_op): New code iterator.
23445 (op_mnc): New code attribute.
23446 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
23447 (mov_f+1, mov_f+2): New peephole2 patterns.
23449 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
23450 (cstoresi4): Also allow re-use of zero result when doing a NE
23451 comparison to a non-zero operand.
23452 Use (clobber (scratch)) for first insn if the gpr output is not needed.
23454 * config/epiphany/epiphany.md (<insn_opname>v2si3):
23455 Use gen_addsi3_i / gen_subsi3_i.
23457 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23461 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
23462 contain anything but clobbers, at most one __builtin_stack_restore,
23463 optionally debug stmts and final resx, and if it has at least one
23464 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
23465 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
23466 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
23467 which isn't defaut definition, remove them.
23468 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
23469 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
23470 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
23471 with MEM_REF LHS with SSA_NAME address.
23473 2013-04-08 Jeff Law <law@redhat.com>
23475 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
23477 2013-04-08 Richard Biener <rguenther@suse.de>
23479 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
23481 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
23482 determined vector type.
23483 (vect_analyze_data_refs): Likewise.
23484 (vect_get_new_vect_var): Adjust.
23485 (vect_create_destination_var): Preserve SSA name versions.
23486 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
23487 not dump anything here.
23489 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23491 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
23492 Add member lr_slot_known.
23493 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
23495 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
23496 Remove code that sets lr_slot_offset according to what a previous
23497 version of epiphany_emit_save_restore used to do.
23498 (epiphany_emit_save_restore): When doing an lr save or restore,
23499 set/verify lr_slot_known and lr_slot_offset.
23501 2013-04-08 Xinyu Qi <xyqi@marvell.com>
23504 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
23507 2013-04-08 Richard Biener <rguenther@suse.de>
23509 * alias.c (find_base_term): Fix thinko in previous change.
23511 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23513 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
23514 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
23515 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
23516 if possible to compute val.
23517 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
23518 For QImode integers don't require anything about precision. Use
23519 const_with_all_bytes_same to find out if the constant doesn't have
23520 repeated bytes in it.
23522 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23524 * config/s390/s390.c (s390_expand_insv): Only accept insertions
23527 2013-04-08 Marek Polacek <polacek@redhat.com>
23529 PR rtl-optimization/48182
23530 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
23533 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23536 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
23537 nuses, make sure we have a label.
23539 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23542 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
23543 (rs6000_emit_swdiv_low_precision): Remove.
23544 (rs6000_emit_swdiv): Rewrite to handle between one and four
23545 iterations of Newton-Raphson generally; modify required number of
23546 iterations for some cases.
23547 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
23549 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
23551 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
23552 set-but-unused variable.
23554 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
23555 basic blocks of released function bodies garbage-collectable.
23557 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
23558 (struct rtl_opt_pass): Add TODO_df_finish.
23560 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
23562 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23564 * config/arm/constraints.md (q): New constraint.
23565 * config/arm/ldrdstrd.md: New file.
23566 * config/arm/arm.md (ldrdstrd.md) New include.
23567 (arm_movdi): Use "q" instead of "r" constraint
23568 for double-word memory access.
23569 (movdf_soft_insn): Likewise.
23570 * config/arm/vfp.md (movdi_vfp): Likewise.
23571 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
23572 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
23573 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
23574 (mem_ok_for_ldrd_strd): Likewise.
23575 (output_move_double): Update assertion.
23577 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23579 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
23581 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23583 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
23584 define_insn_and_split.
23585 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
23587 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23589 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
23590 define_insn_and_split.
23591 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
23592 (shiftsi3_compare): New pattern.
23593 (rrx): New pattern.
23594 * config/arm/unspecs.md (UNSPEC_RRX): New.
23596 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23598 * config/arm/arm.md (negdi_extendsidi): New pattern.
23599 (negdi_zero_extendsidi): Likewise.
23601 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23603 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
23604 define_insn_and_split.
23605 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
23606 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
23608 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23610 * config/arm/arm.md (arm_subdi3): Convert define_insn into
23611 define_insn_and_split.
23612 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
23613 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
23615 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23617 * config/arm/arm.md (subsi3_carryin): New pattern.
23618 (subsi3_carryin_const): Likewise.
23619 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
23620 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
23622 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23624 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
23626 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23628 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
23629 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
23631 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23633 * config/arm/arm.c (arm_expand_builtin): Change fcode
23634 type to unsigned int.
23636 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23638 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
23640 2013-04-04 Ian Lance Taylor <iant@google.com>
23642 * doc/standards.texi (Standards): The Go frontend supports the Go 1
23645 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
23647 PR middle-end/56729
23648 * df-scan.c (df_insn_delete): Disable failing assert.
23650 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23652 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
23653 New function prototype.
23654 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
23655 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
23656 (arm_builtin_vectorized_function): New function.
23658 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23660 * config/arm/arm_neon_builtins.def: New file.
23661 * config/arm/arm.c (neon_builtin_data): Move contents to
23662 arm_neon_builtins.def.
23663 (enum arm_builtins): Include neon builtin definitions.
23664 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
23665 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
23667 2013-04-04 Marek Polacek <polacek@redhat.com>
23669 PR tree-optimization/48186
23670 * predict.c (maybe_hot_frequency_p): Return false if
23671 HOT_BB_FREQUENCY_FRACTION is 0.
23672 (cgraph_maybe_hot_edge_p): Likewise.
23674 2013-04-04 Richard Biener <rguenther@suse.de>
23676 PR tree-optimization/56826
23677 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
23680 2013-04-04 Richard Biener <rguenther@suse.de>
23682 PR tree-optimization/56213
23683 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
23684 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
23686 2013-04-04 Richard Biener <rguenther@suse.de>
23688 PR tree-optimization/56837
23689 * tree-loop-distribution.c (classify_partition): For non-zero
23690 values require that the value has the same precision as its
23691 mode to be useful as memset value.
23693 2013-04-03 Nick Clifton <nickc@redhat.com>
23695 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
23696 (fmssf4): Use fmsf.s on E3V5 architectures.
23697 (fnmasf4): Use fnmaf.s on E3V5 architectures.
23698 (fnmssf4): Use fnmsf.s on E3V5 architectures.
23700 2013-04-03 Jeff Law <law@redhat.com>
23702 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
23703 (lra-eliminations.o): Likewise.
23705 2013-04-03 Teresa Johnson <tejohnson@google.com>
23707 * gcov-io.c (compute_working_sets): Moved most of body of old
23708 compute_working_sets here from profile.c.
23709 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
23710 (gcov_working_set_t): Moved typedef here from basic-block.h
23711 (compute_working_set): Declare.
23712 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
23713 (get_working_sets): Renamed from compute_working_set,
23714 replace most of body with call to new compute_working_sets.
23715 (get_exec_counts): Replace call to compute_working_sets
23716 to get_working_sets.
23717 * profile.h (get_working_sets): Renamed from compute_working_set.
23718 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
23719 to get_working_sets.
23720 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
23721 * gcov-dump.c (dump_working_sets): New function.
23723 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
23725 * hwint.c (sext_hwi, zext_hwi): New functions.
23726 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
23727 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
23728 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
23729 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
23730 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
23731 (sext_hwi, zext_hwi): New functions.
23733 2013-04-03 Jeff Law <law@redhat.com>
23735 PR tree-optimization/56799
23736 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
23737 back test for widening conversion erroneously dropped in prior change.
23739 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23742 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
23743 instead of next_real_insn.
23745 2013-04-03 Marek Polacek <polacek@redhat.com>
23748 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
23750 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23753 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
23755 (thumb1_output_casesi): Likewise.
23756 (thumb2_output_casesi): Likewise.
23758 2013-04-03 Richard Biener <rguenther@suse.de>
23760 PR tree-optimization/56817
23761 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
23763 (tree_unroll_loops_completely_1): ... new function to manually
23764 walk the loop tree, properly defering outer loops of unrolled
23765 loops to later iterations.
23767 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23769 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
23770 (vectorizable_load): Likewise.
23771 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
23772 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
23774 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23776 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
23779 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
23781 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
23783 2013-04-03 Bin Cheng <bin.cheng@arm.com>
23785 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
23787 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23789 PR tree-optimization/56790
23790 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
23793 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23795 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
23797 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
23798 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
23801 2013-04-03 Jakub Jelinek <jakub@redhat.com>
23804 * tree.h (force_folding_builtin_constant_p): New decl.
23805 * builtins.c (force_folding_builtin_constant_p): New variable.
23806 (fold_builtin_constant_p): Fold immediately also if
23807 force_folding_builtin_constant_p.
23809 2013-04-03 Richard Biener <rguenther@suse.de>
23811 PR tree-optimization/56812
23812 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
23813 DRs of the same interleaving chain are independent.
23815 2013-04-02 Jason Merrill <jason@redhat.com>
23817 * gdbinit.in (pbb): Use debug fn.
23819 2013-04-02 Lawrence Crowl <crowl@google.com>
23821 * sese.h (struct ivtype_map_elt_s): Remove unused.
23822 (extern debug_ivtype_map): Remove unused.
23823 (extern eq_ivtype_map_elts): Remove unused.
23824 * sese.c (debug_ivtype_map): Removed unused.
23825 (debug_ivtype_map_1): Removed unused.
23826 (debug_ivtype_elt): Remove unused.
23827 (eq_ivtype_map_elts): Remove unused.
23830 2013-04-02 Kai Tietz <ktietz@redhat.com>
23833 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
23834 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
23835 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
23837 (legitimize_pe_coff_symbol): Likewise.
23838 (is_imported_p): New helper-function.
23839 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
23840 for Windows x64 targets.
23841 (ix86_expand_prologue): Optimize for pe-coff targets.
23842 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
23843 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
23844 medium/large code-model.
23845 (legitimize_pic_address): Likewise.
23846 (legitimize_tls_address): Likewise.
23847 (ix86_expand_call): Likewise.
23848 (x86_output_mi_thunk): Likewise.
23849 (get_dllimport_decl): Add new beimport argument.
23850 (construct_plt_address): Don't assert for x64 pe-coff targets.
23851 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
23853 (SYMBOL_FLAG_STUBVAR): New macro.
23854 (SYMBOL_REF_STUBVAR_P): Likewise.
23855 * config/i386/winnt.c (stub_list): New structure.
23856 (stub_head): New local variable.
23857 (i386_pe_record_stub): New function.
23858 (i386_pe_file_end): Emit refptr-stubs.
23860 2013-04-02 Jakub Jelinek <jakub@redhat.com>
23862 PR rtl-optimization/56745
23863 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
23864 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
23867 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
23868 and both of them are MEM_REFs, just compare first argument for
23869 equality and attempt to deal even with differing offsets.
23872 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
23873 of gimple_clobber_p to be MEM_REF.
23874 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
23875 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
23876 after gimplification.
23877 * asan.c (get_mem_ref_of_assignment): Don't instrument
23878 gimple_clobber_p stmts.
23879 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
23880 gimple_clobber_p stmt if they have MEM_REF lhs and
23881 are dead because of another gimple_clobber_p stmt.
23882 * tree-ssa-live.c (clear_unused_block_pointer): Treat
23883 gimple_clobber_p stmts like debug stmts.
23884 (remove_unused_locals): Remove clobbers with MEM_REF lhs
23885 that refer to unused VAR_DECLs or uninitialized values.
23886 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
23887 gimple_clobber_p stmts if they refer to removed parameters.
23888 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
23891 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
23893 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
23894 using SWI48 mode attribute.
23896 2013-04-02 Wei Mi <wmi@google.com>
23898 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
23899 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
23900 *<rotate_insn><mode>3_mask in i386.md.
23902 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
23904 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
23906 2013-04-02 Richard Biener <rguenther@suse.de>
23908 PR tree-optimization/56778
23909 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
23910 Runtime alias tests are not supported for gather loads.
23911 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
23912 stmts referenced from SSA operands before updating SSA form.
23914 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
23915 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23917 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
23918 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
23919 * config/arm/cortex-a53.md: New file.
23920 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
23921 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
23922 * config/arm/arm.c (arm_issue_rate): Likewise.
23923 * config/arm/arm-tune.md: Regenerate
23924 * config/arm/arm-tables.opt: Regenerate.
23925 * config/arm/arm-cores.def: Add cortex-a53.
23927 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
23929 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
23932 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23934 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
23935 scalar load/store operations using B/H registers.
23936 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
23938 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23940 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
23942 * config/aarch64/aarch64.c
23943 (aarch64_simd_scalar_immediate_valid_for_move): New.
23944 * config/aarch64/aarch64-protos.h
23945 (aarch64_simd_scalar_immediate_valid_for_move): New.
23946 * config/aarch64/constraints.md (Dh, Dq): New.
23947 * config/aarch64/iterators.md (hq): New.
23949 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
23951 * reorg.c (get_branch_condition): Deal with conditional returns.
23952 (fill_simple_delay_slots): Remove dead code dealing with jumps.
23954 2013-04-01 Wei Mi <wmi@google.com>
23956 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
23957 Truncate operand 2 using %b asm operand modifier.
23958 (*<shift_insn><mode>3_mask): Ditto.
23959 (*<rotate_insn><mode>3_mask): Ditto.
23961 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
23963 PR middle-end/56798
23964 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
23966 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
23968 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
23970 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
23972 2013-03-30 Lawrence Crowl <crowl@google.com>
23974 * dse.c (clear_alias_sets): Remove never set.
23975 (disqualified_clear_alias_sets): Remove never set.
23976 (clear_alias_mode_pool): Remove never set.
23977 (dse_step0): Remove condition that is never true.
23978 (canon_address): Remove condition that is never true.
23979 (dse_step7): Remove condition that is never true.
23980 (rest_of_handle_dse): Remove condition that is never true.
23981 (rest_of_handle_dse::did_global): Remove never read from above.
23982 (dse_step2_spill): Remove never called from above.
23983 (dse_step5_spill): Remove never called from above.
23985 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
23987 * doc/md.texi (Standard Names) <casesi>: Update documentation for
23988 JUMP_TABLE_DATA changes.
23989 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
23990 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
23991 (Insns) <jump_table_data>: New entry.
23992 * doc/tm.texi: Regenerate.
23994 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
23996 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
23997 for table jump at the end of a basic block using tablejump_p.
23998 * targhooks.c (default_invalid_within_doloop): Likewise.
23999 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
24000 target hook implementation that is identical to the default hook.
24001 (rs6000_invalid_within_doloop): Remove.
24003 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
24004 unused variable from tablejump_p call.
24006 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
24007 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
24008 (INSN_DELETED_P): Likewise.
24009 (emit_jump_table_data): New prototype.
24010 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
24011 after 4th as unused.
24012 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
24013 * sched-vis.c (print_insn): Likewise.
24014 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
24015 insn for compatibility with back ends that use next_active_insn to
24016 identify jump table data.
24017 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
24018 (remove_insn): Likewise.
24019 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
24021 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
24022 (emit_jump_table_data): New function.
24024 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
24025 basic block, a JUMP_TABLE_DATA never is.
24026 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
24027 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
24028 off from code handling real insns.
24029 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
24030 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
24031 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
24032 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
24033 is not a NONDEBUG_INSN_P.
24034 * ira-costs.c (scan_one_insn): Likewise.
24035 * jump.c (mark_all_labels): Likewise.
24036 (mark_jump_label_1): Likewise.
24037 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24038 * lra.c (get_insn_freq): Expect all insns reaching here to be in
24040 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
24041 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
24042 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
24043 JUMP_TABLE_DATA_P insns.
24044 (calculate_elim_costs_all_insns): Likewise.
24045 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
24046 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
24047 (delete_output_reload): Code style fixups.
24048 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
24049 insn flags on this non-insn.
24050 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
24051 as scheduling barriers, for pre-change compatibility.
24052 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
24053 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
24055 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
24056 redundant JUMP_TABLE_DATA_P test.
24057 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
24058 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24059 (frv_for_each_packet): Likewise.
24060 * config/i386/i386.c (min_insn_size): Likewise.
24061 (ix86_avoid_jump_mispredicts): Likewise.
24062 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24063 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
24064 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24065 (mips16_insn_length): Robustify.
24066 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
24067 (mips16_split_long_branches): Likewise.
24068 * config/pa/pa.c (pa_combine_instructions): Likewise.
24069 * config/rs6000/rs6000.c (get_next_active_insn): Treat
24070 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
24071 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
24072 as contributing to pool range lengths.
24073 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
24074 Remove redundant JUMP_TABLE_DATA_P test.
24075 (sh_loop_align): Likewise.
24076 (split_branches): Likewise.
24077 (sh_insn_length_adjustment): Likewise.
24078 * config/spu/spu.c (get_branch_target): Likewise.
24080 2013-03-29 Jan Hubicka <jh@suse.cz>
24082 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
24083 gcov streaming; stream hot bb threshold to ltrans.
24084 * predict.c (get_hot_bb_threshold): Break out from ....
24085 (maybe_hot_count_p): ... here.
24086 (set_hot_bb_threshold): New function.
24087 * lto-section-in.c (lto_section_name): Add profile.
24088 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
24089 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
24090 and data-streamer.h
24091 (histogram_entry): New structure.
24092 (histogram, histogram_pool): New global vars.
24093 (histogram_hash): New structure.
24094 (histogram_hash::hash): New method.
24095 (histogram_hash::equal): Likewise.
24096 (account_time_size): New function.
24097 (cmp_counts): New function.
24098 (dump_histogram): New function.
24099 (ipa_profile_generate_summary): New function.
24100 (ipa_profile_write_summary): New function.
24101 (ipa_profile_read_summary): New function.
24102 (ipa_profile): Decide on threshold.
24103 (pass_ipa_profile): Add ipa_profile_write_summary and
24104 ipa_profile_read_summary.
24105 * Makefile.in (ipa.o): Update dependencies.
24106 * lto-streamer.h (LTO_section_ipa_profile): New section.
24108 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
24110 * tree.h (VAR_P): New.
24112 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
24115 * doc/invoke.texi ([-fwhole-program]): Fix typo.
24117 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
24119 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
24120 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24121 (control_flow_insn_p): Likewise.
24122 * cfgrtl.c (duplicate_insn_chain): Likewise.
24123 * final.c (get_attr_length_1): Likewise.
24124 (shorten_branches): Likewise.
24125 (final_scan_insn): Likewise.
24126 * function.c (instantiate_virtual_regs): Likewise.
24127 * gcse.c (insert_insn_end_basic_block): Likewise.
24128 * ira-costs.c (scan_one_insn): Likewise.
24129 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24130 * lra.c (check_rtl): Likewise.
24131 * reload1.c (elimination_costs_in_insn): Likewise.
24132 * reorg.c (follow_jumps): Likewise.
24134 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
24135 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24136 (thumb_far_jump_used_p): Likewise.
24137 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
24138 (workaround_speculation): Likewise.
24139 (add_sched_insns_for_speculation): Likewise.
24140 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
24141 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24142 (frv_for_each_packet): Likewise.
24143 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
24144 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
24145 (final_emit_insn_group_barriers): Likewise.
24146 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24147 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24148 (mips16_insn_length): Likewise.
24149 * config/pa/pa.c (pa_reorg): Likewise.
24150 (pa_combine_instructions): Likewise.
24151 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
24152 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
24153 (sh_reorg): Likewise.
24154 (split_branches): Likewise.
24155 * config/spu/spu.c (get_branch_target): Likewise.
24157 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
24160 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
24162 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
24163 Fix declaration name.
24165 2013-03-28 Lawrence Crowl <crowl@google.com>
24167 * graphds.h (struct graph.indicies): Remove unused.
24168 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
24169 (SCOP_ORIGINAL_PDDRS): Remove unused.
24170 * sese.h (extern insert_loop_close_phis): Removed unused.
24171 (extern insert_guard_phis): Removed unused.
24172 (extern ivtype_map_elt_info): Removed unused.
24173 (new_ivtype_map_elt): Removed unused.
24174 * sese.c (ivtype_map_elt_info): Removed unused.
24176 2013-03-28 Lawrence Crowl <crowl@google.com>
24178 * Makefile.in: Add several missing include dependences.
24180 (test-dump.o): New. This object is not added to any executable,
24181 but is present for ad-hoc testing.
24183 (debug (const bitmap_head_def &)): New.
24184 (debug (const bitmap_head_def *)): New.
24186 (extern debug (const bitmap_head_def &)): New.
24187 (extern debug (const bitmap_head_def *)): New.
24189 (debug (edge_def &)): New.
24190 (debug (edge_def *)): New.
24192 (debug (basic_block_def &)): New.
24193 (debug (basic_block_def *)): New.
24195 (dump_node (const_tree, int, FILE *)): Correct source file.
24197 (debug (die_struct &)): New.
24198 (debug (die_struct *)): New.
24200 (extern debug (die_struct &)): New.
24201 (extern debug (die_struct *)): New.
24202 * gimple-pretty-print.c
24203 (debug (gimple_statement_d &)): New.
24204 (debug (gimple_statement_d *)): New.
24205 * gimple-pretty-print.h
24206 (extern debug (gimple_statement_d &)): New.
24207 (extern debug (gimple_statement_d *)): New.
24209 (debug (ira_allocno_copy &)): New.
24210 (debug (ira_allocno_copy *)): New.
24211 (debug (ira_allocno &)): New.
24212 (debug (ira_allocno *)): New.
24214 (extern debug (ira_allocno_copy &)): New.
24215 (extern debug (ira_allocno_copy *)): New.
24216 (extern debug (ira_allocno &)): New.
24217 (extern debug (ira_allocno *)): New.
24219 (debug (live_range &)): New.
24220 (debug (live_range *)): New.
24222 (debug (lra_live_range &)): New.
24223 (debug (lra_live_range *)): New.
24225 (debug (lra_live_range &)): New.
24226 (debug (lra_live_range *)): New.
24228 (debug (omega_pb_d &)): New.
24229 (debug (omega_pb_d *)): New.
24231 (extern debug (omega_pb_d &)): New.
24232 (extern debug (omega_pb_d *)): New.
24234 (debug (const rtx_def &)): New.
24235 (debug (const rtx_def *)): New.
24237 (debug_tree (tree): Move within file.
24238 (debug_raw (const tree_node &)): New.
24239 (debug_raw (const tree_node *)): New.
24240 (dump_tree_via_hooks (const tree_node *, int)): New.
24241 (debug (const tree_node &)): New.
24242 (debug (const tree_node *)): New.
24243 (debug_verbose (const tree_node &)): New.
24244 (debug_verbose (const tree_node *)): New.
24245 (debug_head (const tree_node &)): New.
24246 (debug_head (const tree_node *)): New.
24247 (debug_body (const tree_node &)): New.
24248 (debug_body (const tree_node *)): New.
24249 (debug_vec_tree (tree): Move and reimplement in terms of dump.
24250 (debug (vec<tree, va_gc> &)): New.
24251 (debug (vec<tree, va_gc> *)): New.
24253 (extern debug (const rtx_def &)): New.
24254 (extern debug (const rtx_def *)): New.
24256 (debug_raw (simple_bitmap_def &)): New.
24257 (debug_raw (simple_bitmap_def *)): New.
24258 (debug (simple_bitmap_def &)): New.
24259 (debug (simple_bitmap_def *)): New.
24261 (extern debug (simple_bitmap_def &)): New.
24262 (extern debug (simple_bitmap_def *)): New.
24263 (extern debug_raw (simple_bitmap_def &)): New.
24264 (extern debug_raw (simple_bitmap_def *)): New.
24266 (debug (vinsn_def &)): New.
24267 (debug (vinsn_def *)): New.
24268 (debug_verbose (vinsn_def &)): New.
24269 (debug_verbose (vinsn_def *)): New.
24270 (debug (expr_def &)): New.
24271 (debug (expr_def *)): New.
24272 (debug_verbose (expr_def &)): New.
24273 (debug_verbose (expr_def *)): New.
24274 (debug (vec<rtx> &)): New.
24275 (debug (vec<rtx> *)): New.
24277 (extern debug (vinsn_def &)): New.
24278 (extern debug (vinsn_def *)): New.
24279 (extern debug_verbose (vinsn_def &)): New.
24280 (extern debug_verbose (vinsn_def *)): New.
24281 (extern debug (expr_def &)): New.
24282 (extern debug (expr_def *)): New.
24283 (extern debug_verbose (expr_def &)): New.
24284 (extern debug_verbose (expr_def *)): New.
24285 (extern debug (vec<rtx> &)): New.
24286 (extern debug (vec<rtx> *)): New.
24288 (_list_iter_cond_expr): Make inline instead of static.
24290 (debug (sreal &)): New.
24291 (debug (sreal *)): New.
24293 (extern debug (sreal &)): New.
24294 (extern debug (sreal *)): New.
24296 (extern debug_raw (const tree_node &)): New.
24297 (extern debug_raw (const tree_node *)): New.
24298 (extern debug (const tree_node &)): New.
24299 (extern debug (const tree_node *)): New.
24300 (extern debug_verbose (const tree_node &)): New.
24301 (extern debug_verbose (const tree_node *)): New.
24302 (extern debug_head (const tree_node &)): New.
24303 (extern debug_head (const tree_node *)): New.
24304 (extern debug_body (const tree_node &)): New.
24305 (extern debug_body (const tree_node *)): New.
24306 (extern debug (vec<tree, va_gc> &)): New.
24307 (extern debug (vec<tree, va_gc> *)): New.
24309 (debug (struct loop &)): New.
24310 (debug (struct loop *)): New.
24311 (debug_verbose (struct loop &)): New.
24312 (debug_verbose (struct loop *)): New.
24313 * tree-dump.c: Add header dependence.
24315 (extern debug (struct loop &)): New.
24316 (extern debug (struct loop *)): New.
24317 (extern debug_verbose (struct loop &)): New.
24318 (extern debug_verbose (struct loop *)): New.
24320 (debug (data_reference &)): New.
24321 (debug (data_reference *)): New.
24322 (debug (vec<data_reference_p> &)): New.
24323 (debug (vec<data_reference_p> *)): New.
24324 (debug (vec<ddr_p> &)): New.
24325 (debug (vec<ddr_p> *)): New.
24327 (extern debug (data_reference &)): New.
24328 (extern debug (data_reference *)): New.
24329 (extern debug (vec<data_reference_p> &)): New.
24330 (extern debug (vec<data_reference_p> *)): New.
24331 (extern debug (vec<ddr_p> &)): New.
24332 (extern debug (vec<ddr_p> *)): New.
24334 (debug (pt_solution &)): New.
24335 (debug (pt_solution *)): New.
24337 (extern debug (pt_solution &)): New.
24338 (extern debug (pt_solution *)): New.
24340 (debug (_var_map &)): New.
24341 (debug (_var_map *)): New.
24342 (debug (tree_live_info_d &)): New.
24343 (debug (tree_live_info_d *)): New.
24345 (extern debug (_var_map &)): New.
24346 (extern debug (_var_map *)): New.
24347 (extern debug (tree_live_info_d &)): New.
24348 (extern debug (tree_live_info_d *)): New.
24350 2013-03-28 Jan Hubicka <jh@suse.cz>
24352 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
24354 2013-03-28 Ian Bolton <ian.bolton@arm.com>
24356 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
24357 record only when desired or required.
24359 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
24361 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
24362 *vec_extractv2di_1_rex64. Use x64 isa attribute.
24364 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24366 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
24367 (*andsi3_compare0_uxtw): New pattern.
24368 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
24369 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
24371 2013-03-28 Jan Hubicka <jh@suse.cz>
24373 * data-streamer-in.c (streamer_read_gcov_count): New function.
24374 * gimple-streamer-out.c: Include value-prof.h.
24375 (output_gimple_stmt): Output histogram.
24376 (output_bb): Use streamer_write_gcov_count.
24377 * value-prof.c: Include data-streamer.h
24378 (dump_histogram_value): Add HIST_TYPE_MAX.
24379 (stream_out_histogram_value): New function.
24380 (stream_in_histogram_value): New function.
24381 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
24382 (stream_out_histogram_value, stream_in_histogram_value): Declare.
24383 * data-streamer-out.c (streamer_write_gcov_count): New function.
24384 (streamer_write_gcov_count_stream): New function.
24385 * lto-cgraph.c (lto_output_edge): Update counter streaming.
24386 (lto_output_node): Likewise.
24387 (input_node, input_edge): Likewise.
24388 * lto-streamer-out.c (output_cfg): Update streaming.
24389 * lto-streamer-in.c (input_cfg): Likewise.
24390 * data-streamer.h (streamer_write_gcov_count,
24391 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
24392 * gimple-streamer-in.c: Include value-prof.h
24393 (input_gimple_stmt): Input histograms.
24394 (input_bb): Update profile streaming.
24396 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24398 * genmodes.c (emit_max_int): New function.
24399 (emit_insn_modes_h): Added call to emit_max_function.
24400 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
24402 * machmode.def: Fixed comment.
24404 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24406 * combine.c (try_combine): Removed useless assert.
24407 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
24409 2013-03-28 Marek Polacek <polacek@redhat.com>
24410 Richard Biener <rguenther@suse.de>
24412 PR tree-optimization/56695
24413 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
24414 build signed result of a vector comparison.
24415 * tree-cfg.c (verify_gimple_comparison): Check that a result
24416 of a vector comparison has signed type.
24418 2013-03-28 Richard Biener <rguenther@suse.de>
24420 PR tree-optimization/37021
24421 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
24422 do not restrict gaps between groups.
24423 * tree-vect-stmts.c (vectorizable_load): Properly account for
24424 a gap between groups.
24426 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
24428 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
24429 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
24432 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
24434 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
24435 * doc/extend.texi (Named Address Spaces): Ditto.
24436 (Variable Attributes): Ditto.
24438 2013-03-27 Kai Tietz <ktietz@redhat.com>
24440 * config.build: Add support for cygwin x64 target.
24441 * config.gcc: Likewise.
24442 * config.host: Likewise.
24443 * configure.ac: Likewise
24444 * configure: Regenerated.
24446 2013-03-27 Kai Tietz <ktietz@redhat.com>
24448 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
24449 * config/i386/t-cygwin-w64: New file.
24450 * config/i386/cygwin-w64.h: New file.
24451 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
24452 and add support for x64-cygwin target.
24453 (CPP_SPEC): Likewise.
24454 (CXX_WRAP_SPEC_LIST): Undefine before define.
24455 (LIBGCJ_SONAME): Use 15 as version.
24457 2013-03-27 Richard Biener <rguenther@suse.de>
24459 PR tree-optimization/56716
24460 * tree-ssa-structalias.c (perform_var_substitution): Adjust
24461 dumping for ref nodes.
24463 2013-03-27 Martin Jambor <mjambor@suse.cz>
24465 PR tree-optimization/55334
24466 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
24467 restricted pointers to arrays.
24469 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
24471 * Makefile.in (.SUFFIXES): Add .cc.
24472 (.c.o): Apply same recipe for implicit rule .cc.o.
24474 2013-03-27 Richard Biener <rguenther@suse.de>
24476 PR tree-optimization/37021
24477 * tree-vect-data-refs.c (vect_check_strided_load): Allow
24478 REALPART/IMAGPART_EXPRs around the supported refs.
24479 * tree-ssa-structalias.c (find_func_aliases): Assume that
24480 floating-point values are not used to transfer pointers.
24482 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
24484 * target.def (TARGET_HAS_IFUNC_P): New target hook.
24485 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
24486 * doc/tm.texi: Regenerate.
24487 * targhooks.h (default_has_ifunc_p): New.
24488 * targhooks.c (default_has_ifunc_p): Ditto.
24489 * config/linux-protos.h: New file.
24490 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
24491 hook for linux which disables support of indirect functions in android.
24492 * config/linux-android.c: New file.
24493 * config/t-linux-android.c: Ditto.
24494 * config.gcc: Added new object file linux-android.o.
24495 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
24496 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
24497 * varasm.c (do_assemble_alias): Likewise.
24498 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
24499 doesn't support indirect functions.
24500 * configure: Regenerate.
24502 2013-03-27 Bin Cheng <bin.cheng@arm.com>
24505 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
24506 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
24509 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24511 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
24513 2013-03-27 Terry Guo <terry.guo@arm.com>
24515 * config/arm/arm-cores.def: Added core cortex-r7.
24516 * config/arm/arm-tune.md: Regenerated.
24517 * config/arm/arm-tables.opt: Regenerated.
24518 * doc/invoke.texi: Added entry for core cortex-r7.
24520 2013-03-27 Walter Lee <walt@tilera.com>
24522 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
24523 double-decrement of next_scratch_regno.
24525 2013-03-27 Walter Lee <walt@tilera.com>
24527 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
24529 (insn_v1mulus): Ditto.
24530 (insn_v2muls): Ditto.
24532 2013-03-27 Walter Lee <walt@tilera.com>
24534 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
24535 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
24537 2013-03-27 Walter Lee <walt@tilera.com>
24539 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
24540 (*sibcall_value): Ditto.
24542 2013-03-27 Walter Lee <walt@tilera.com>
24544 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
24545 (insn_mnz_v8qi): ... this ...
24546 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
24548 (insn_v<n>mnz): Replaced by ...
24549 (insn_v1mnz): ... this ...
24550 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
24552 (insn_mz_<mode>): Replaced by ...
24553 (insn_mz_v8qi): ... this ...
24554 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
24556 (insn_v<n>mz): Replaced by ...
24557 (insn_v1mz): ... this ...
24558 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
24561 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
24563 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
24565 2013-03-26 Roland McGrath <mcgrathr@google.com>
24567 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
24568 than fprintf with a non-constant, non-format string.
24570 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24572 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
24573 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
24574 operand 0 predicate.
24575 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
24576 attribute. Use general_x64nomem_operand as operand 1 predicate.
24577 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
24578 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24579 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
24580 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24581 (mov<mode>_insv_1): Remove expander. Merge insn with
24582 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
24583 Use general_x64nomem_operand as operand 1 predicate.
24584 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
24585 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
24586 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24587 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
24588 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
24589 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
24590 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24591 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
24592 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
24593 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
24594 (general_x64nomem_operand): Ditto.
24596 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
24598 * config/rtems.opt: Add -pthread option.
24600 2013-03-26 Richard Biener <rguenther@suse.de>
24602 * alias.c (find_base_term): Avoid redundant and not used recursion.
24603 (base_alias_check): Get the initial base term from the caller.
24604 (true_dependence_1): Compute and pass base terms to base_alias_check.
24605 (write_dependence_p): Likewise.
24606 (may_alias_p): Likewise.
24608 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
24610 * config/aarch64/aarch64.c (aarch64_classify_address): Support
24611 PC-relative load in SI modes and above only.
24613 2013-03-26 Xinyu Qi <xyqi@marvell.com>
24615 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
24616 * config/arm/iwmmxt.md (WCGR0): Update.
24617 (WCGR1, WCGR2, WCGR3): Likewise.
24619 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24621 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
24622 Use x64 and nox64 isa attributes.
24624 2013-03-26 Richard Biener <rguenther@suse.de>
24626 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
24627 alignment computations and rely on get_object_alignment_1
24628 for the !TYPE_P case.
24629 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
24631 2013-03-26 Walter Lee <walt@tilera.com>
24633 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
24634 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
24636 2013-03-25 Jeff Law <law@redhat.com>
24638 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
24639 check for INTEGRAL_TYPE_P that was missing due to checking in
24640 wrong version of prior patch.
24642 2013-03-25 Walter Lee <walt@tilera.com>
24644 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
24645 TILEGX_INSN_SHUFFLEBYTES1.
24646 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
24648 (tilegx_builtins): Ditto.
24649 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
24651 2013-03-25 Walter Lee <walt@tilera.com>
24653 * config/tilegx/tilegx.md (floatsisf2): New pattern.
24654 (floatunssisf2): New pattern.
24655 (floatsidf2): New pattern.
24656 (floatunssidf2): New pattern.
24658 2013-03-25 Walter Lee <walt@tilera.com>
24660 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
24661 tests for constraint J, K, N, P.
24663 2013-03-25 Walter Lee <walt@tilera.com>
24665 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
24666 Use indirect/pcrel encoding.
24667 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
24670 2013-03-25 Steve Ellcey <sellcey@mips.com>
24672 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
24673 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
24674 * config/mips/mips.c (mips_option_override): Set IMADD default.
24675 * config/mips/mips.h (PTF_AVOID_IMADD): New.
24676 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
24677 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
24678 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
24679 * doc/invoke.texi (-mimadd/-mno-imadd): New.
24681 2013-03-25 Jeff Law <law@redhat.com>
24683 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
24684 slightly to avoid creating and folding useless trees. Simplify
24685 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
24687 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
24689 * config/i386/i386.md (*zero_extendsidi2): Merge with
24690 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
24691 * config/i386/predicates.md (x86_64_zext_operand): Rename from
24692 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
24693 targets. Clarify comment.
24695 2013-03-25 Martin Jambor <mjambor@suse.cz>
24697 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
24698 pass-through jump functions differently.
24699 (ipa_read_jump_function): Likewise. Also use setter functions to set
24702 2013-03-25 Martin Jambor <mjambor@suse.cz>
24704 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
24705 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
24707 (ipa_get_indirect_edge_target): New function.
24708 (devirtualization_time_bonus): New parameter known_aggs, pass it to
24709 ipa_get_indirect_edge_target. Update all callers.
24710 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
24711 ipa_get_indirect_edge_target_1 instead of calling
24712 ipa_get_indirect_edge_target.
24713 (create_specialized_node): Pass aggvlas to
24714 ipcp_discover_new_direct_edges.
24716 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24718 * config/arm/arm.md (f_sels, f_seld): New types.
24719 (*cmov<mode>): New pattern.
24720 * config/arm/predicates.md (arm_vsel_comparison_operator): New
24723 2013-03-25 Kai Tietz <ktietz@redhat.com>
24725 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
24726 POSIX-printf for mingw-hosted builds.
24728 2013-03-25 Richard Biener <rguenther@suse.de>
24730 PR middle-end/56694
24731 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
24732 must-not-throw stmt location.
24734 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24736 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
24737 Emit load-acquire versions when acq is true.
24738 (arm_emit_store_exclusive): Add rel parameter.
24739 Emit store-release versions when rel is true.
24740 (arm_split_compare_and_swap): Use acquire-release instructions
24742 of barriers when appropriate.
24743 (arm_split_atomic_op): Likewise.
24744 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
24745 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
24746 (VUNSPEC_SLX): Likewise.
24747 (VUNSPEC_LDA): Likewise.
24748 (VUNSPEC_STL): Likewise.
24749 * config/arm/sync.md (atomic_load<mode>): New pattern.
24750 (atomic_store<mode>): Likewise.
24751 (arm_load_acquire_exclusive<mode>): Likewise.
24752 (arm_load_acquire_exclusivesi): Likewise.
24753 (arm_load_acquire_exclusivedi): Likewise.
24754 (arm_store_release_exclusive<mode>): Likewise.
24756 2013-03-25 Catherine Moore <clm@codesourcery.com>
24758 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
24759 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
24760 * config/mip/predicates.md (lwsp_swsp_operand,
24761 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
24762 sb16_operand, db4_operand, db7_operand, ib3_operand,
24763 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
24764 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
24765 andi16_operand): New predicates.
24766 * config/mips/mips.md (compression): New attribute.
24767 (enabled): New attribute.
24768 (length): Consider compression in computing length.
24769 (shift_compression): New code attribute.
24770 (*add<mode>3): New operands. Record compression.
24771 (sub<mode>3): Likewise.
24772 (one_cmpl<mode>2): Likewise.
24773 (*and<mode>3): Likewise.
24774 (*ior<mode>3): Likewise.
24775 (unnamed pattern for xor): Likewise.
24776 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
24777 (*<optab><mode>3): Likewise.
24778 (*mov<mode>_internal: Likewise.
24779 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
24780 (mips_unsigned_immediate_p): New.
24781 (umips_lwsp_swsp_address_p): New.
24782 (m16_based_address_p): New.
24783 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
24784 (mips_unsigned_immediate_p): New prototype.
24785 (lwsp_swsp_address_p): New prototype.
24786 (m16_based_address_p): New prototype.
24787 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
24788 (mips_signed_immediate_p): New function.
24789 (m16_based_address_p): New function.
24790 (lwsp_swsp_address_p): New function.
24791 (mips_print_operand_punctuation): Recognize short delay slot insns
24792 for microMIPS.add<mode>3"
24794 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24797 * config/arm/iterators.md (v_cmp_result): New mode attribute.
24798 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
24800 2013-03-25 Richard Biener <rguenther@suse.de>
24802 PR tree-optimization/56689
24803 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
24806 2013-03-25 Richard Biener <rguenther@suse.de>
24808 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
24810 (memory_references): Likewise.
24811 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
24812 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
24813 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
24814 (gather_mem_refs_in_loops): Fold into ...
24815 (analyze_memory_references): ... this. Move initialization
24816 to tree_ssa_lim_initialize.
24817 (fill_always_executed_in): Rename to ...
24818 (fill_always_executed_in_1): ... this.
24819 (fill_always_executed_in): Move contains_call computation to
24820 this new function from ...
24821 (tree_ssa_lim_initialize): ... here.
24822 (tree_ssa_lim): Call fill_always_executed_in.
24824 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24826 * postreload.c (reload_combine): Fix code detecting returns.
24828 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24830 * function.c (emit_use_return_register_into_block): On cc0 targets,
24831 do not emit the sequence between cc0 setter and user.
24833 2013-03-25 Kai Tietz <ktietz@redhat.com>
24835 * config/i386/predicates.md (local_symbolic_operand): Interpret
24836 dll-imported symbols as none-local.
24838 2013-03-25 Richard Biener <rguenther@suse.de>
24840 * tree-ssa-loop-im.c (struct depend): Remove.
24841 (struct lim_aux_data): Make depends a vec of gimples.
24842 (free_lim_aux_data): Adjust.
24843 (add_dependency): Likewise.
24844 (set_level): Likewise.
24846 2013-03-25 Richard Biener <rguenther@suse.de>
24848 PR middle-end/56434
24849 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
24850 the pointer returned by calls with ECF_MALLOC set.
24852 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24854 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
24856 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24858 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
24859 using MMXMODE mode iterator.
24860 (*move<mode>_internal): Merge with *movv2sf_internal and
24861 *movv2sf_internal_rex64 using MMXMODE mode iterator.
24863 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24865 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
24866 (record_last_mem_set_info): Likewise.
24868 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
24869 of XNEWVEC followed by memset.
24870 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
24872 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24874 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
24875 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
24876 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
24877 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
24878 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
24879 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
24880 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
24881 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
24882 BARRIER_P instead of GET_CODE.
24884 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24886 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
24887 inaccuracy in the probing code.
24889 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
24890 (ctrapdi4): Likewise.
24892 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24894 * calls.c (expand_call): Add missing guard to code handling return
24895 of non-BLKmode structures in MSB.
24896 * function.c (expand_function_end): Likewise.
24898 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24900 * combine.c (try_combine): Adjust comment. Do not add the set of
24901 insn #0 if the destination indirectly is set or dies in insn #2.
24902 Tidy up code to distribute a new note.
24904 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24906 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
24907 also for alternatives 16 and 17.
24909 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24911 * config/i386/sse.md (*mov<mode>_internal): Merge with
24912 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
24913 Emit insn template depending on type attribute. Use
24914 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24915 movd instead of movq mnemonic for interunit moves. Rewrite mode
24916 attribute calculation. Remove unit attribute calculation.
24917 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
24918 Set prefix_data16 attribute for DImode ssemov types.
24919 Use Ym instead of y for SSE-MMX conversion alternatives.
24920 Reorder operand constraints.
24922 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
24924 * df.h (df_insn_delete): Adjust prototype.
24925 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
24926 and let it decide whether mark the basic block dirty.
24927 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
24928 * df-scan.c (df_insn_info_delete): New helper function, split
24929 off from df_insn_delete.
24930 (df_scan_free_bb_info): Use it.
24931 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
24933 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
24934 that the insn is actually an insn and it has a non-NULL basic block.
24935 Do not mark basic block dirty if only deleting a DEBUG_INSN.
24937 2013-03-22 Richard Biener <rguenther@suse.de>
24939 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
24941 (mem_ref_alloc): Do not allocate them.
24942 (refs_independent_p): Do not query or maintain a cache.
24944 2013-03-22 Richard Biener <rguenther@suse.de>
24946 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
24947 (gather_mem_refs_in_loops): Do not compute it.
24948 (analyze_memory_references): Do not allocate it.
24949 (tree_ssa_lim_finalize): Do not free it.
24950 (for_all_locs_in_loop): Do not query all_refs_in_loop.
24952 2013-03-22 Richard Biener <rguenther@suse.de>
24954 * is-a.h (as_a): Use gcc_checking_assert.
24956 2013-03-22 Ian Bolton <ian.bolton@arm.com>
24958 * config/aarch64/aarch64.c (aarch64_print_operand): New
24959 format specifier for printing a constant in hex.
24960 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
24961 format specifier for printing second operand.
24963 2013-03-22 Richard Biener <rguenther@suse.de>
24965 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
24967 (gather_mem_refs_in_loops): Perform store accumulation here.
24968 (create_vop_ref_mapping_loop): Remove.
24969 (create_vop_ref_mapping): Likewise.
24970 (analyze_memory_references): Initialize refs_stored_in_loop.
24971 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
24972 (record_indep_loop): Remove.
24973 (record_dep_loop): New function.
24974 (ref_indep_loop_p_1): Adjust to only walk over references
24975 in the loop, not its subloops.
24976 (ref_indep_loop_p): Rename to ...
24977 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
24978 maintaining a more fine-grained cache.
24979 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
24980 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
24982 2013-03-22 Richard Biener <rguenther@suse.de>
24984 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
24985 (struct mem_ref): Make accesses_in_loop a vec of a vec of
24986 aggregate mem_ref_loc.
24987 (free_mem_ref_locs): Inline into ...
24988 (memref_free): ... this and adjust.
24989 (mem_ref_alloc): Adjust.
24990 (mem_ref_locs_alloc): Remove.
24991 (record_mem_ref_loc): Adjust.
24992 (get_all_locs_in_loop): Rewrite into ...
24993 (for_all_locs_in_loop): ... this iterator.
24994 (rewrite_mem_ref_loc): New functor.
24995 (rewrite_mem_refs): Use for_all_locs_in_loop.
24996 (sm_set_flag_if_changed): New functor.
24997 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
24998 (ref_always_accessed): New functor.
24999 (ref_always_accessed_p): Use for_all_locs_in_loop.
25001 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25003 * tree-pass.h (PROP_gimple_lvec): New.
25004 * passes.c (dump_properties): Handle PROP_gimple_lvec.
25005 (init_optimization_passes): Move pass_lower_vector.
25006 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
25008 (pass_lower_vector): Provide PROP_gimple_lvec.
25009 (pass_lower_vector_ssa): Likewise.
25010 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
25012 2013-03-21 Mark Wielaard <mjw@redhat.com>
25014 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
25016 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25018 * config/i386/i386.md (*movdi_internal): Disparage slightly
25019 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
25020 conversion alternatives.
25022 2013-03-21 Jakub Jelinek <jakub@redhat.com>
25024 PR middle-end/48087
25025 * diagnostic.def (DK_WERROR): New kind.
25026 * diagnostic.h (werrorcount): Define.
25027 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
25028 promoted to DK_ERROR, increment DK_WERROR counter instead of
25030 * toplev.c (toplev_main): Call print_ignored_options even if
25031 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
25032 even if just werrorcount is non-zero.
25035 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
25037 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
25038 (string_cst_pool_decl): New function.
25039 (optimize_one_addr_into_implicit_ptr): New function.
25040 (resolve_addr_in_expr): Optimize DWARF location expression
25041 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
25042 which doesn't live in memory, but has DW_AT_location or
25043 DW_AT_const_value, or refers to a string literal, into
25044 DW_OP_GNU_implicit_pointer.
25045 (optimize_location_into_implicit_ptr): New function.
25046 (resolve_addr): If removing DW_AT_location of a variable because
25047 it was DW_OP_addr of address of the variable, but the variable doesn't
25048 live in memory, try to emit const value attribute for the initializer.
25050 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25052 * tree.h (VECTOR_TYPE_P): New macro.
25053 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
25054 TYPE_MODE): Use it.
25055 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
25056 VEC_COND_EXPR cannot be lvalues.
25057 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
25059 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25061 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
25062 Restrict the transformation to equal modes.
25064 2013-03-21 Richard Biener <rguenther@suse.de>
25066 PR tree-optimization/39326
25067 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
25068 (MEM_ANALYZABLE): Adjust.
25069 (record_mem_ref_loc): Move bitmap ops ...
25070 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
25071 unanalyzable refs, do not record locations for it.
25072 (analyze_memory_references): Allocate ref zero as shared
25074 (refs_independent_p): Do not test for unanalyzed mems here.
25075 (ref_indep_loop_p_1): Special-case disambiguation against
25076 the unanalyzed ref.
25077 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
25079 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
25081 * config/arm/arm-protos.h (tune_params): Add
25082 prefer_neon_for_64bits field.
25083 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
25084 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
25085 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
25086 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
25087 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
25088 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
25089 (arm_option_override): Handle -mneon-for-64bits new option.
25090 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
25091 (prefer_neon_for_64bits): Declare new variable.
25092 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
25093 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
25094 (arch_enabled): Handle new arch types. Remove support for onlya8
25096 (one_cmpldi2): Use new arch names.
25097 (zero_extend<mode>di2, extend<mode>di2): Ditto.
25098 * config/arm/arm.opt (mneon-for-64bits): Add option.
25099 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
25100 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
25101 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
25103 * doc/invoke.texi (-mneon-for-64bits): Document.
25105 2013-03-21 Richard Biener <rguenther@suse.de>
25107 PR tree-optimization/39326
25108 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
25109 (sort_bbs_in_loop_postorder_cmp): New function.
25110 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
25112 2013-03-21 Richard Biener <rguenther@suse.de>
25114 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
25115 (vect_insert_into_interleaving_chain): Likewise.
25116 (vect_drs_dependent_in_basic_block): Inline ...
25117 (vect_slp_analyze_data_ref_dependence): ... here. New function,
25119 (vect_analyze_data_ref_dependence): ... here. Simplify.
25120 (vect_check_interleaving): Simplify.
25121 (vect_analyze_data_ref_dependences): Likewise. Split out ...
25122 (vect_slp_analyze_data_ref_dependences): ... this new function.
25123 (dr_group_sort_cmp): New function.
25124 (vect_analyze_data_ref_accesses): Compute data-reference groups
25125 here instead of in vect_analyze_data_ref_dependence. Use
25126 a more efficient algorithm.
25127 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
25128 vect_slp_analyze_data_ref_dependences. Call
25129 vect_analyze_data_ref_accesses earlier.
25130 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
25131 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
25132 (vect_slp_analyze_data_ref_dependences): New prototype.
25134 2013-03-21 Richard Biener <rguenther@suse.de>
25136 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
25137 ref is stored in the loop.
25138 (find_refs_for_sm): Walk only over all stores.
25139 (store_motion_loop): Allocate from lim_bitmap_obstack.
25140 (store_motion): Likewise.
25142 2013-03-21 Richard Biener <rguenther@suse.de>
25144 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
25145 Update virtual SSA form.
25147 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25149 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
25150 * configure: Regenerate.
25151 * config.in: Regenerate.
25152 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
25153 if !HAVE_LD_EH_FRAME_CIEV3.
25155 2013-03-21 Richard Biener <rguenther@suse.de>
25157 * tree-cfg.c (verify_expr_no_block): New function.
25158 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
25159 nor DECL_VALUE_EXPR have locations with associated blocks.
25160 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
25161 (clear_unused_block_pointer): Remove code dealing with
25162 blocks in DECL_DEBUG_EXPR locations.
25164 2013-03-21 Richard Biener <rguenther@suse.de>
25166 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
25167 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
25168 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
25169 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
25170 instead of DECL_DEBUG_EXPR_IS_FROM.
25171 * gimplify.c (gimplify_modify_expr): Likewise.
25172 * tree-cfg.c (verify_expr_location_1): Likewise.
25173 * tree-complex.c (create_one_component_var): Likewise.
25174 * tree-sra.c (create_access_replacement): Likewise.
25175 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
25176 (clear_unused_block_pointer): Likewise.
25177 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
25178 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
25179 * var-tracking.c (var_debug_decl): Likewise.
25180 (track_expr_p): Likewise.
25181 * tree-inline.c (add_local_variables): Likewise. Set
25182 DECL_HAS_DEBUG_EXPR_P after copying it.
25183 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
25184 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
25186 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25189 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
25190 * configure: Regenerate.
25191 * config.in: Regenerate.
25192 * config/i386/i386.md (*movdf_internal): Use
25193 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25194 movd instead of movq mnemonic for interunit moves.
25195 (*movdi_internal): Ditto.
25197 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25199 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
25200 (abd<mode>_3): New pattern.
25201 (aba<mode>_3): New pattern.
25202 (fabd<mode>_3): New pattern.
25204 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25206 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
25207 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
25208 occurrence of REGISTER_PREFIX as its empty string.
25210 2013-03-20 Jeff Law <law@redhat.com>
25212 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
25213 addititional equivalences for equality comparisons between an SSA_NAME
25214 and a constant where the SSA_NAME was set from a widening conversion.
25216 2013-03-20 Walter Lee <walt@tilera.com>
25218 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
25220 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25222 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
25223 depending on type attribute.
25224 (*movti_internal): Ditto.
25225 (*movtf_internal): Ditto.
25226 (*movxf_internal): Ditto.
25227 (*movdf_internal): Ditto.
25228 (*movsf_internal): Ditto.
25230 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25232 * config/i386/i386.md (*movti_internal): Set prefix attribute to
25233 maybe_vex for sselog1 and ssemov types.
25234 (*movdi_internal): Reorder operand constraints.
25235 (*movsi_internal): Ditto. Set prefix attribute to
25236 maybe_vex for sselog1 and ssemov types.
25237 (*movtf_internal): Set prefix attribute to maybe_vex
25238 for sselog1 and ssemov types.
25239 (*movdf_internal): Ditto. Set prefix_data16 attribute for
25240 DImode ssemov types. Reorder operand constraints.
25241 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
25242 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
25243 attribute for SImode ssemov types. Reorder operand constraints.
25245 2013-03-20 Martin Jambor <mjambor@suse.cz>
25247 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
25248 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
25250 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
25252 * config/rs6000/predicates.md (indexed_address, update_address_mem
25253 update_indexed_address_mem): New predicates.
25254 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
25255 attribute for load/store instructions.
25256 * config/rs6000/dfp.md (movsd_store): Likewise.
25257 (movsd_load): Likewise.
25258 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
25259 (unnamed HI->DI extend define_insn): Likewise.
25260 (unnamed SI->DI extend define_insn): Likewise.
25261 (unnamed QI->SI extend define_insn): Likewise.
25262 (unnamed QI->HI extend define_insn): Likewise.
25263 (unnamed HI->SI extend define_insn): Likewise.
25264 (unnamed HI->SI extend define_insn): Likewise.
25265 (extendsfdf2_fpr): Likewise.
25266 (movsi_internal1): Likewise.
25267 (movsi_internal1_single): Likewise.
25268 (movhi_internal): Likewise.
25269 (movqi_internal): Likewise.
25270 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
25271 attribute for load/store instructions.
25272 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
25274 (mov<mode>_softfloat): Likewise.
25275 (mov<mode>_hardfloat32): Likewise.
25276 (mov<mode>_hardfloat64): Likewise.
25277 (mov<mode>_softfloat64): Likewise.
25278 (movdi_internal32): Likewise.
25279 (movdi_internal64): Likewise.
25280 (probe_stack_<mode>): Likewise.
25282 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
25284 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
25285 floating point, and decimal floating point to reload iterator.
25287 * config/rs6000/constraints.md (wl constraint): New constraints to
25288 return FLOAT_REGS if certain options are used to reduce the number
25289 of separate patterns that exist in the file.
25290 (wx constraint): Likewise.
25291 (wz constraint): Likewise.
25293 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25294 -mdebug=reg, print wg, wl, wx, and wz constraints.
25295 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25296 Initialize the reload functions for 64-bit binary/decimal floating
25298 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25299 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25300 create the buffer on the stack to overcome not having a 32-bit
25302 (rs6000_emit_move): Likewise.
25303 (rs6000_secondary_memory_needed_rtx): Likewise.
25304 (rs6000_alloc_sdmode_stack_slot): Likewise.
25305 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25306 via xxlxor, just like DFmode 0.0.
25308 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
25309 define as 1 if we are running on a power7 or newer.
25310 (enum r6000_reg_class_enum): Add new constraints.
25312 * config/rs6000/dfp.md (movsd): Delete, combine with binary
25313 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25314 with other moves by using conditional constraits (wg). Use LFIWZX
25315 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
25316 (movsd splitter): Likewise.
25317 (movsd_hardfloat): Likewise.
25318 (movsd_softfloat): Likewise.
25320 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
25321 binary and decimal floating point moves.
25322 (fmove_ok): New attributes to combine binary and decimal floating
25323 point moves, and to combine power6x (mfpgpr) moves along normal
25325 (real_value_to_target): Likewise.
25326 (f32_lr): Likewise.
25327 (f32_lm): Likewise.
25328 (f32_li): Likewise.
25329 (f32_sr): Likewise.
25330 (f32_sm): Likewise.
25331 (f32_si): Likewise.
25332 (movsf): Combine binary and decimal floating point moves. Combine
25333 power6x (mfpgpr) moves with other moves by using conditional
25334 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
25335 (mov<mode> for SFmode/SDmode); Likewise.
25336 (SFmode/SDmode splitters): Likewise.
25337 (movsf_hardfloat): Likewise.
25338 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
25339 (movsf_softfloat): Likewise.
25340 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
25342 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
25343 wx and wz constraints.
25345 * config/rs6000/constraints.md (wg constraint): New constraint to
25346 return FLOAT_REGS if -mmfpgpr (power6x) was used.
25348 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
25351 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25352 -mdebug=reg, print wg, wl, wx, and wz constraints.
25353 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25354 Initialize the reload functions for 64-bit binary/decimal floating
25356 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25357 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25358 create the buffer on the stack to overcome not having a 32-bit
25360 (rs6000_emit_move): Likewise.
25361 (rs6000_secondary_memory_needed_rtx): Likewise.
25362 (rs6000_alloc_sdmode_stack_slot): Likewise.
25363 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25364 via xxlxor, just like DFmode 0.0.
25366 * config/rs6000/dfp.md (movdd): Delete, combine with binary
25367 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25368 with other moves by using conditional constraits (wg). Use LFIWZX
25369 and STFIWX for loading SDmode on power7.
25370 (movdd splitters): Likewise.
25371 (movdd_hardfloat32): Likewise.
25372 (movdd_softfloat32): Likewise.
25373 (movdd_hardfloat64_mfpgpr): Likewise.
25374 (movdd_hardfloat64): Likewise.
25375 (movdd_softfloat64): Likewise.
25377 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
25378 64-bit binary and decimal floating point moves.
25379 (FMOVE64X): Likewise.
25380 (movdf): Combine 64-bit binary and decimal floating point moves.
25381 Combine power6x (mfpgpr) moves with other moves by using
25382 conditional constraits (wg).
25383 (mov<mode> for DFmode/DDmode): Likewise.
25384 (DFmode/DDmode splitters): Likewise.
25385 (movdf_hardfloat32): Likewise.
25386 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
25387 (movdf_softfloat32): Likewise.
25388 (movdf_hardfloat64_mfpgpr): Likewise.
25389 (movdf_hardfloat64): Likewise.
25390 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
25391 (movdf_softfloat64): Likewise.
25392 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
25393 (reload_<mode>_load): Move to later in the file so they aren't in
25394 the middle of the floating point move insns.
25395 (reload_<mode>_store): Likewise.
25397 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
25400 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
25401 constraint if -mdebug=reg.
25402 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
25403 Enable using dd reload support if needed.
25405 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
25406 binary and decimal floating point moves in rs6000.md.
25407 (movtd_internal): Likewise.
25409 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
25410 decimal floating point moves.
25412 (movtf_internal): Likewise.
25413 (mov<mode>_internal, TDmode/TFmode): Likewise.
25414 (movtf_softfloat): Likewise.
25415 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
25417 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
25418 movdi_internal64, using wg constraint for move direct operations.
25419 (movdi_internal64): Likewise.
25421 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
25422 MODES_TIEABLE_P for selected modes. Print the numerical value of
25423 the various virtual registers. Use GPR/FPR first/last values,
25424 instead of hard coding the register numbers. Print which modes
25425 have reload functions registered.
25426 (rs6000_option_override_internal): If -mdebug=reg, trace the options
25427 settings before/after setting cpu, target and subtarget settings.
25428 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
25429 and for secondary reload failures in rs6000_secondary_reload_inner.
25430 (rs6000_secondary_reload_fail): Likewise.
25431 (rs6000_secondary_reload_inner): Likewise.
25433 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
25434 macros for first/last GPR and FPR registers.
25435 (LAST_GPR_REGNO): Likewise.
25436 (FIRST_FPR_REGNO): Likewise.
25437 (LAST_FPR_REGNO): Likewise.
25439 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
25440 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
25441 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
25442 (vcond<mode><mode>): Likewise.
25443 (vcondu<mode><mode>): Likewise.
25444 (vector_gtu<mode>): Likewise.
25445 (vector_gte<mode>): Likewise.
25446 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
25447 to prevent the compiler from converting DImode operations to TImode.
25448 (ior<mode>3): Likewise.
25449 (and<mode>3): Likewise.
25450 (one_cmpl<mode>2): Likewise.
25451 (nor<mode>3): Likewise.
25452 (andc<mode>3): Likewise.
25454 * config/rs6000/constraints.md (wt constraint): New constraint
25455 that returns VSX_REGS if TImode is allowed in VSX registers.
25457 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
25458 constant under VSX.
25460 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
25461 similar to TImode, but it is restricted to being in the GPRs.
25463 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
25464 TImode to occupy a single VSX register.
25466 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
25467 -mvsx-timode for power7/power8.
25468 (power7 cpu): Likewise.
25469 (power8 cpu): Likewise.
25471 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
25472 sure that TFmode/TDmode take up two registers if they are ever
25473 allowed in the upper VSX registers.
25474 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
25476 (rs6000_init_hard_regno_mode_ok): Likewise.
25477 (rs6000_debug_reg_global): Add debugging for PTImode and wt
25478 constraint. Print if LRA is turned on.
25479 (rs6000_option_override_internal): Give an error if -mvsx-timode
25480 and VSX is not enabled.
25481 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
25482 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
25483 to reg+offset addressing. Use PTImode when checking offset
25484 addresses for validity.
25485 (reg_offset_addressing_ok_p): Likewise.
25486 (rs6000_legitimate_offset_address_p): Likewise.
25487 (rs6000_legitimize_address): Likewise.
25488 (rs6000_legitimize_reload_address): Likewise.
25489 (rs6000_legitimate_address_p): Likewise.
25490 (rs6000_eliminate_indexed_memrefs): Likewise.
25491 (rs6000_emit_move): Likewise.
25492 (rs6000_secondary_reload): Likewise.
25493 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
25494 reloads to fpr registers to continue to use reg+offset addressing,
25495 but 64-bit reloads to altivec registers need reg+reg addressing.
25496 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
25497 it. Treat LO_SUM like a PLUS operation.
25498 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
25499 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
25500 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
25501 registers to share a register with a smaller sized type, since VSX
25502 puts scalars in the upper 64-bits.
25503 (print_operand): Add support for PTImode.
25504 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
25505 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
25506 registers, but don't have arithmetic support.
25507 (rs6000_memory_move_cost): Add test for VSX.
25508 (rs6000_opt_masks): Add -mvsx-timode.
25510 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
25513 (VSr): Use wt constraint for TImode.
25514 (VSv): Drop TImode support.
25515 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
25516 (vsx_movti_64bit): Likewise.
25517 (vsx_movti_32bit): Likewise.
25518 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
25519 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
25520 one '?' on the appropriate output constraint. Do not allow TImode
25521 logical operations on 32-bit systems.
25522 (vsx_ior<mode>3): Likewise.
25523 (vsx_xor<mode>3): Likewise.
25524 (vsx_one_cmpl<mode>2): Likewise.
25525 (vsx_nor<mode>3): Likewise.
25526 (vsx_andc<mode>3): Likewise.
25527 (vsx_concat_<mode>): Likewise.
25528 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
25530 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
25531 OPTION_MASK_VSX_TIMODE.
25532 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
25533 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
25535 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
25536 (TI2 iterator): New iterator for TImode, PTImode.
25537 (wd mode attribute): Add values for vector types.
25538 (movti_string): Replace TI move operations with operations for TImode
25539 and PTImode. Add support for TImode being allowed in VSX registers.
25540 (mov<mode>_string, TImode/PTImode): Likewise.
25541 (movti_ppc64): Likewise.
25542 (mov<mode>_ppc64, TImode/PTImode): Likewise.
25543 (TI mode splitters): Likewise.
25545 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
25548 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
25550 PR tree-optimization/56355
25551 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
25552 Also handle integers with undefined overflow.
25554 2013-03-20 Catherine Moore <clm@codesourcery.com>
25555 Maciej W. Rozycki <macro@codesourcery.com>
25556 Tom de Vries <tom@codesourcery.com>
25557 Nathan Sidwell <nathan@codesourcery.com>
25558 Iain Sandoe <iain@codesourcery.com>
25559 Nathan Froyd <froydnj@codesourcery.com>
25560 Chao-ying Fu <fu@mips.com>
25562 * doc/extend.texi: (micromips, nomicromips, nocompression):
25563 Document new function attributes.
25564 * doc/invoke.texi (minterlink-compressed, mmicromips,
25565 m14k, m14ke, m14kec): Document new options.
25566 (minterlink-mips16): Update documentation.
25567 * doc/md.texi (ZC, ZD): Document new constraints.
25568 * configure.ac (gcc_cv_as_micromips): Check if linker
25569 supports the .set micromips directive.
25570 * configure: Regenerate.
25571 * config.in: Regenerate.
25572 * config/mips/mips-tables.opt: Regenerate.
25573 * config/mips/micromips.md: New file.
25574 * constraints.md (ZC, ZD): New constraints.
25575 * config/mips/predicates.md (movep_src_register): New predicate.
25576 (movep_src_operand): New predicate.
25577 (non_volatile_mem_operand): New predicate.
25578 * config/mips/mips.md (multimem): New type.
25579 (length): Differentiate between 17-bit and 18-bit branch offsets.
25580 (MOVEP1, MOVEP2): New mode iterator.
25581 (mov_<load>l): Use ZC constraint.
25582 (mov_<load>r): Likewise.
25583 (mov_<store>l): Likewise.
25584 (mov_<store>r): Likewise.
25585 (*branch_equality<mode>_inverted): Add microMIPS support.
25586 (*branch_equality<mode>): Likewise.
25587 (*jump_absolute): Likewise.
25588 (indirect_jump_<mode>): Likewise.
25589 (tablejump_<mode>): Likewise.
25590 (<optab>_internal): Likewise.
25591 (sibcall_internal): Likewise.
25592 (sibcall_value_internal): Likewise.
25593 (prefetch): Use constraint ZD.
25594 * config/mips/mips.opt (minterlink-compressed): New option.
25595 (minterlink-mips16): Now an alias for minterlink-compressed.
25596 (mmicromips): New option.
25597 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
25598 (compare_and_swap_12): Likewise.
25599 (sync_add<mode>): Likewise.
25600 (sync_<optab>_12): Likewise.
25601 (sync_old_<optab>_12): Likewise.
25602 (sync_new_<optab>_12): Likewise.
25603 (sync_nand_12): Likewise.
25604 (sync_old_nand_12): Likewise.
25605 (sync_new_nand_12): Likewise.
25606 (sync_sub<mode>): Likewise.
25607 (sync_old_add<mode>): Likewise.
25608 (sync_old_sub<mode>): Likewise.
25609 (sync_new_add<mode>): Likewise.
25610 (sync_new_sub<mode>): Likewise.
25611 (sync_<optab><mode>): Likewise.
25612 (sync_old_<optab><mode>): Likewise.
25613 (sync_new_<optab><mode>): Likewise.
25614 (sync_nand<mode>): Likewise.
25615 (sync_old_nand<mode>): Likewise.
25616 (sync_new_nand<mode>): Likewise.
25617 (sync_lock_test_and_set<mode>): Likewise.
25618 (test_and_set_12): Likewise.
25619 (atomic_compare_and_swap<mode>): Likewise.
25620 (atomic_exchange<mode>_llsc): Likewise.
25621 (atomic_fetch_add<mode>_llsc): Likewise.
25622 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
25623 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
25624 (umips_save_restore_pattern_p): Likewise.
25625 (umips_load_store_pair_p): Likewise.
25626 (umips_output_load_store_pair): Likewise.
25627 (umips_movep_target_p): Likewise.
25628 (umips_12bit_offset_address_p): Likewise.
25629 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
25630 (mips_base_mips16): Rename this...
25631 (mips_base_compression_flags): ...to this. Update all uses.
25632 (mips_attribute_table): Add micromips, nomicromips and nocompression.
25633 (mips_mips16_decl_p): Delete.
25634 (mips_nomips16_decl_p): Delete.
25635 (mips_get_compress_on_flags): New function.
25636 (mips_get_compress_off_flags): New function.
25637 (mips_get_compress_mode): New function.
25638 (mips_get_compress_on_name): New function.
25639 (mips_get_compress_off_name): New function.
25640 (mips_insert_attributes): Support multiple compression types.
25641 (mips_merge_decl_attributes): Likewise.
25642 (umips_12bit_offset_address_p): New function.
25643 (mips_start_function_definition): Emit .set micromips directive.
25644 (mips_call_may_need_jalx_p): New function.
25645 (mips_function_ok_for_sibcall): Add microMIPS support.
25646 (mips_print_operand_punctuation): Support short delay slots and
25648 (umips_swm_mask, umips_swm_encoding): New.
25649 (umips_build_save_restore): New function.
25650 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
25651 (was_mips16_p): Remove.
25652 (old_compression_mode): New.
25653 (mips_set_compression_mode): New function.
25654 (mips_set_current_function): Add microMIPS support.
25655 (mips_option_override): Likewise.
25656 (umips_save_restore_pattern_p): New function.
25657 (umips_output_save_restore): New function.
25658 (umips_load_store_pair_p_1): New function.
25659 (umips_load_store_pair_p): New function.
25660 (umips_output_load_store_pair_1): New function.
25661 (umips_output_load_store_pair): New function.
25662 (umips_movep_target_p) New function.
25663 (mips_prepare_pch_save): Add microMIPS support.
25664 * config/mips/mips.h (TARGET_COMPRESSION): New.
25665 (TARGET_CPU_CPP_BUILTINS): Update macro
25666 to use new compression flags and to support microMIPS.
25667 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
25668 (MIPS_ARCH_FLOAT_SPEC): Likewise.
25669 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
25670 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
25671 (ASM_SPEC): Support mmicromips and mno-micromips.
25672 (M16STORE_REG_P): New macro.
25673 (MIPS_CALL): Support TARGET_MICROMIPS.
25674 (MICROMIPS_J): New macro.
25675 (mips_base_mips16): Rename this...
25676 (mips_base_compression_flags): ...to this.
25677 (UMIPS_12BIT_OFFSET_P): New macro.
25678 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
25679 (MULTILIB_DIRNAMES): Likewise.
25680 2013-03-20 Richard Biener <rguenther@suse.de>
25682 PR tree-optimization/56661
25683 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
25684 the result does not have to be distinct.
25686 2013-03-20 Richard Biener <rguenther@suse.de>
25688 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
25691 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
25692 Steven Bosscher <steven@gcc.gnu.org>
25694 PR rtl-optimization/56605
25695 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
25697 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25700 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
25701 that require movd instead of movq.
25703 2013-03-20 Richard Biener <rguenther@suse.de>
25705 * tree-ssa-structalias.c (struct variable_info): Add pointer
25706 to the first field of an aggregate with sub-vars. Make
25707 this and the pointer to the next subfield its ID.
25708 (vi_next): New function.
25709 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
25710 storedanything_id, integer_id): Increment by one.
25711 (new_var_info, get_call_vi, lookup_call_clobber_vi,
25712 get_call_clobber_vi): Adjust.
25713 (solution_set_expand): Simplify and speedup.
25714 (solution_set_add): Inline into ...
25715 (set_union_with_increment): ... this. Adjust accordingly.
25716 (do_sd_constraint): Likewise.
25717 (do_ds_constraint): Likewise.
25718 (do_complex_constraint): Simplify.
25719 (build_pred_graph): Adjust.
25720 (solve_graph): Likewise. Simplify and speedup.
25721 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
25722 get_constraint_for_component_ref, get_constraint_for_1,
25723 first_vi_for_offset, first_or_preceding_vi_for_offset,
25724 create_function_info_for, create_variable_info_for_1,
25725 create_variable_info_for, intra_create_variable_infos): Adjust.
25726 (init_base_vars): Push NULL for ID zero.
25727 (compute_points_to_sets): Adjust.
25729 2013-03-20 Richard Biener <rguenther@suse.de>
25731 * cfgloop.c (verify_loop_structure): Streamline and avoid
25732 ICEing on corrupt loop tree.
25733 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
25736 2013-03-20 Richard Biener <rguenther@suse.de>
25738 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
25739 check whether an SSA update is needed.
25741 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
25743 * config/mips/constraints.md (T): Rename to...
25747 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
25748 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
25750 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25752 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
25753 (*subsi3_carryin_uxtw): Likewise.
25755 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25757 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
25758 (*rorsi3_insn_uxtw): Likewise.
25760 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25762 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
25763 (*extrsi5_insn_uxtw): Likewise.
25765 2013-03-19 Richard Biener <rguenther@suse.de>
25767 PR tree-optimization/56273
25768 * passes.c (init_optimization_passes): Move second VRP after DOM.
25770 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
25772 * config/i386/i386.md (*movti_internal): Merge from
25773 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
25774 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
25775 nox64 isa attributes.
25777 2013-03-18 Richard Biener <rguenther@suse.de>
25779 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
25781 (merge_node_constraints): Likewise.
25782 (build_succ_graph): Likewise.
25783 (valid_graph_edge): Inline into single caller.
25784 (unify_nodes): Likewise. Use bitmap_set_bit return value
25786 (scc_visit): Fix formatting and variable use.
25787 (do_sd_constraint): Use gcc_checking_assert.
25788 (do_ds_constraint): Likewise.
25789 (do_complex_constraint): Likewise.
25790 (condense_visit): Likewise. Cleanup.
25791 (dump_pred_graph): New function.
25792 (perform_var_substitution): Dump the pred-graph before
25793 variable substitution.
25794 (find_equivalent_node): Use gcc_checking_assert.
25795 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
25797 2013-03-18 Richard Biener <rguenther@suse.de>
25799 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
25800 Remove cond_expr_stmt_list argument and do not gimplify the
25802 (vect_loop_versioning): Adjust.
25803 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
25804 Cleanup to use less temporaries.
25805 (vect_create_data_ref_ptr): Cleanup.
25807 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25809 PR tree-optimization/56635
25810 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
25811 require types_compatible_p types.
25813 2013-03-18 Nick Clifton <nickc@redhat.com>
25815 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
25816 spurious backslash.
25818 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
25819 Add missing line to comment describing function.
25821 2013-03-18 Richard Biener <rguenther@suse.de>
25823 PR tree-optimization/56210
25824 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
25825 Handle string / character search functions.
25826 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
25828 2013-03-18 Richard Biener <rguenther@suse.de>
25830 PR middle-end/56483
25831 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
25832 and implement properly.
25833 * gimple.h (gimple_cond_single_var_p): Remove.
25835 2013-03-18 Richard Biener <rguenther@suse.de>
25837 * tree-data-ref.h (find_data_references_in_loop): Declare.
25838 * tree-data-ref.c (get_references_in_stmt): Use a stack
25839 vector pre-allocated in the callers.
25840 (find_data_references_in_stmt): Adjust.
25841 (graphite_find_data_references_in_stmt): Likewise.
25842 (create_rdg_vertices): Likewise.
25843 (find_data_references_in_loop): Export.
25844 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
25845 Compute dependences here...
25846 (vect_analyze_data_refs): ...not here. When we encounter
25847 a non-vectorizable data reference in basic-block vectorization
25848 truncate the data reference vector. Do not bother to
25849 fixup data-dependence information for gather loads.
25850 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
25851 of data references, as reported.
25853 2013-03-18 Richard Biener <rguenther@suse.de>
25855 PR tree-optimization/3713
25856 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
25857 has_constants and expr.
25858 (stmt_has_constants): Properly valueize SSA names when deciding
25859 whether the stmt has constants.
25861 2013-03-18 Richard Biener <rguenther@suse.de>
25863 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
25864 whole function when there is nothing to do.
25865 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
25866 * tree-vectorizer.c (vectorize_loops): Update virtual and
25867 loop-closed SSA once.
25868 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
25870 2013-03-18 Richard Biener <rguenther@suse.de>
25872 PR middle-end/56113
25873 * domwalk.c (bb_postorder): New global static.
25874 (cmp_bb_postorder): New function.
25875 (walk_dominator_tree): Replace scheme imposing an order for
25876 visiting dominator sons by one sorting them at the time they
25877 are pushed on the stack.
25879 2013-03-18 Richard Biener <rguenther@suse.de>
25881 PR tree-optimization/39326
25882 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
25883 (struct mem_ref): Replace mem member with ao_ref typed member.
25884 (MEM_ANALYZABLE): Adjust.
25885 (memref_eq): Likewise.
25886 (mem_ref_alloc): Likewise.
25887 (gather_mem_refs_stmt): Likewise.
25888 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
25889 (execute_sm_if_changed_flag_set): Adjust.
25890 (execute_sm): Likewise.
25891 (ref_always_accessed_p): Likewise.
25892 (refs_independent_p): Likewise.
25893 (can_sm_ref_p): Likewise.
25895 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25898 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
25899 return 1 even for !unsignedp.
25901 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25903 * config/i386/i386.md (isa): Add x64 and nox64.
25904 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
25905 (*pushtf): Enable *roF alternative for x64 isa only.
25906 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
25907 mode attribute of integer alternatives to DImode for TARGET_64BIT.
25908 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
25909 (*movtf_internal): Merge from *movtf_internal_rex64 and
25910 *movtf_internal_sse. Use x64 and nox64 isa attributes.
25911 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
25912 nox64 isa attributes.
25913 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
25914 nox64 isa attributes.
25915 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
25917 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25919 * config/alpha/alpha.c (TARGET_LRA_P): New define.
25921 2013-03-17 Jakub Jelinek <jakub@redhat.com>
25924 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
25925 class names. Remove trailing comma after "ALL_REGS".
25927 2013-03-16 Jan Hubicka <jh@suse.cz>
25929 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
25930 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
25931 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
25932 of cgraph_get_create_node.
25933 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
25935 2013-03-16 Jason Merrill <jason@redhat.com>
25938 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
25939 with DW_AT_default_value.
25941 2013-03-16 Jakub Jelinek <jakub@redhat.com>
25943 * BASE-VER: Set to 4.9.0.
25945 2013-03-14 Andi Kleen <ak@linux.intel.com>
25948 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
25949 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
25950 Document _x* TSX intrinsics.
25952 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
25953 David Holsgrove <david.holsgrove@xilinx.com>
25955 * configure.ac: Add MicroBlaze TLS support detection.
25956 * configure: Regenerate.
25957 * config/microblaze/microblaze-protos.h
25958 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
25959 symbol_mentioned_p, label_mentioned_p): Add prototypes.
25960 * config/microblaze/microblaze.c (microblaze_address_type): Add
25961 ADDRESS_TLS and tls_reloc address types.
25962 (microblaze_address_info): Add tls_reloc.
25963 (TARGET_HAVE_TLS): Define.
25964 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
25965 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
25966 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
25967 load_tls_operand, microblaze_call_tls_get_addr,
25968 microblaze_legitimize_tls_address): New functions.
25969 (microblaze_classify_unspec): Handle UNSPEC_TLS.
25970 (get_base_reg): Use microblaze_tls_symbol_p.
25971 (microblaze_classify_address): Handle TLS.
25972 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
25973 label_mentioned_p and microblaze_tls_referenced_p.
25974 (microblaze_legitimize_address): Handle TLS.
25975 (microblaze_address_insns): Handle ADDRESS_TLS.
25976 (pic_address_needs_scratch): Handle TLS.
25977 (print_operand_address): Handle TLS.
25978 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
25979 (microblaze_expand_move): Handle TLS.
25980 (microblaze_legitimate_constant_p): Check
25981 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
25982 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
25983 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
25984 (PIC_OFFSET_TABLE_REGNUM): Set.
25985 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
25986 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
25987 (addsi3, movsi_internal2, movdf_internal): Update constraints
25988 * config/microblaze/predicates.md (arith_plus_operand): Define
25989 (move_operand): Redefine as move_src_operand,
25990 check microblaze_tls_referenced_p.
25992 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25994 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
25995 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
25997 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25999 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
26002 2013-03-14 Jakub Jelinek <jakub@redhat.com>
26004 PR tree-optimization/53265
26005 * common.opt (Waggressive-loop-optimizations): New option.
26006 * tree-ssa-loop-niter.c: Include tree-pass.h.
26007 (do_warn_aggressive_loop_optimizations): New function.
26008 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
26009 if number_of_latch_executions returned constant.
26010 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
26011 early. If number_of_latch_executions returned constant, set
26012 nb_iterations_upper_bound back to it.
26013 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
26015 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
26016 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
26018 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
26019 (MULTILIB_OSDIRNAMES): Set.
26020 * genmultilib: If defaultosdirname doesn't start with :: , set
26021 defaultosdirname2 instead, clear it and emit two . multilib_raw
26022 entries instead of just one.
26024 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
26026 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26027 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
26028 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26029 (SUBTARGET_OVERRIDE_OPTIONS): New.
26031 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
26034 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
26035 (musermode): Convert to Var(TARGET_USERMODE).
26036 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
26037 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
26038 * config/sh/sh.c (sh_option_override): Use
26039 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
26040 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
26042 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
26044 (udivsi3_i4_single, divsi3_i4_single): Use
26045 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
26047 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
26049 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
26052 2013-03-13 Richard Biener <rguenther@suse.de>
26054 PR tree-optimization/56608
26055 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
26056 calls when vectorizing basic-blocks.
26058 2013-03-13 Jakub Jelinek <jakub@redhat.com>
26061 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
26064 2013-03-12 Jakub Jelinek <jakub@redhat.com>
26066 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
26068 2013-03-11 Jan Hubicka <jh@suse.cz>
26071 * lto-streamer-out.c (output_symbol_p): Skip references from
26072 constructors of external variables.
26074 2013-03-11 Jan Hubicka <jh@suse.cz>
26076 PR middle-end/56571
26077 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
26079 * emit-rtl.c (verify_rtx_sharing): Likewise.
26080 (copy_insn_1): Likewise.
26081 * rtl.c (copy_rtx): Likewise.
26083 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
26086 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
26087 output_operand_lossage message.
26089 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
26092 * arm.c (shift_op): Validate RTL pattern on the fly.
26093 (arm_print_operand, case 'S'): Don't use shift_operator to validate
26096 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26099 * config/pa/pa.md (call_value): Check for calls to powf and direct to
26100 new call patterns that clobber %fr12.
26101 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
26102 split and postreload patterns.
26103 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
26104 registers %fr12 and %fr12R as call used.
26106 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
26108 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
26109 (canon_address, record_store, replace_read, check_mem_read_rtx,
26110 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
26111 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
26112 rest_of_handle_dse): Likewise.
26114 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
26116 PR middle-end/56524
26117 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
26119 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
26120 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
26121 (save_optabs_if_changed): Replace with...
26122 (init_tree_optimization_optabs): ...this.
26123 * optabs.c (save_optabs_if_changed): Rename to...
26124 (init_tree_optimization_optabs): ...this. Take the optimization node
26125 as argument. Do nothing if the base optabs are already correct.
26126 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
26127 to recompute optabs.
26128 * function.h (function): Remove optabs field.
26129 * function.c (invoke_set_current_function_hook): Call
26130 init_tree_optimization_optabs. Use the result to initialize
26133 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
26135 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
26136 if GTMA_HAS_NO_INSTRUMENTATION.
26137 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
26138 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
26139 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
26140 * gimple-pretty-print.c (dump_gimple_transaction): Handle
26141 GTMA_HAS_NO_INSTRUMENTATION.
26143 2013-03-08 Jakub Jelinek <jakub@redhat.com>
26145 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
26148 2013-03-08 Marek Polacek <polacek@redhat.com>
26149 Jakub Jelinek <jakub@redhat.com>
26151 PR tree-optimization/56478
26152 * predict.c (is_comparison_with_loop_invariant_p): Change the
26153 type of loop_step to tree.
26154 (predict_loops): Adjust.
26155 (predict_iv_comparison): Perform the computations on double_ints.
26157 2013-03-08 Richard Biener <rguenther@suse.de>
26159 PR tree-optimization/56570
26160 * tree-cfg.c (verify_expr_location_1): Verify locations for
26162 * tree-sra.c (create_access_replacement): Strip locations
26163 from DECL_DEBUG_EXPRs.
26165 2013-03-08 Richard Biener <rguenther@suse.de>
26167 * tree-inline.c (expand_call_inline): Do not associate
26168 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
26169 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
26171 2013-03-08 Richard Biener <rguenther@suse.de>
26173 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
26174 or block changes with -Og. Fix for location / block encoding
26175 changes and PHI arguments with locations.
26177 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
26179 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
26181 (struct output_info): Likewise.
26182 (register_overhead): Remove bad gcc_assert.
26183 (bitmap_find_bit): If there is only a single bitmap element, do not
26184 count a miss as a search.
26185 (print_statistics): Update for counter type changes.
26186 (dump_bitmap_statistics): Likewise. Print headers such that they
26187 are properly lined up with the printed counters.
26189 2013-03-07 Jakub Jelinek <jakub@redhat.com>
26191 PR tree-optimization/56559
26192 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
26193 check that it has only a single use.
26195 2013-03-07 Richard Biener <rguenther@suse.de>
26197 * doc/invoke.texi (fwhole-program): Discourage use in combination
26200 2013-03-06 Jakub Jelinek <jakub@redhat.com>
26202 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
26204 PR tree-optimization/56539
26205 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
26206 instead of GSI_CONTINUE_LINKING as last argument to
26207 force_gimple_operand_gsi. Adjust function comment.
26209 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
26212 PR middle-end/56548
26213 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
26214 promoted mode, convert the result back to the original mode.
26216 2013-03-06 Richard Biener <rguenther@suse.de>
26218 PR middle-end/56294
26219 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
26220 (insert_updated_phi_nodes_compare_uids): New function.
26221 (update_ssa): Sort symbols_to_rename after UID before
26222 traversing it to insert PHI nodes.
26224 2013-03-06 Richard Biener <rguenther@suse.de>
26226 PR middle-end/50494
26227 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
26228 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
26231 2013-02-13 Richard Biener <rguenther@suse.de>
26234 * varasm.c (output_constant_def_1): Get the decl representing
26235 the constant as argument.
26236 (output_constant_def): Wrap output_constant_def_1.
26237 (make_decl_rtl): Use output_constant_def_1 with the decl
26238 representing the constant.
26239 (build_constant_desc): Optionally re-use a decl already
26240 representing the constant.
26241 (tree_output_constant_def): Adjust.
26243 2013-03-06 Joey Ye <joey.ye@arm.com>
26246 * gcc.c (convert_white_space): New function.
26247 (main): Handles white space in function name.
26249 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
26252 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
26253 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
26254 to SH_DIV_CALL_TABLE for TARGET_SH2.
26255 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
26257 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
26258 call-table options.
26260 2013-03-05 Sterling Augustine <saugustine@google.com>
26261 Cary Coutant <ccoutant@google.com>
26264 * dwarf2out.c (resolve_addr): Don't call
26265 remove_loc_list_addr_table_entries a second time for the same
26268 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26271 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
26272 (avoid_complex_debug_insns): New function.
26273 (expand_debug_locations): Call it.
26275 PR rtl-optimization/56484
26276 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
26277 lifetimes of hard registers on small register class machines.
26279 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26281 * config/microblaze/microblaze-protos.h: Rename
26282 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
26283 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
26285 (microblaze_fast_interrupt_function_p): New function.
26286 (microblaze_is_interrupt_handler): Rename to
26287 microblaze_is_interrupt_variant and add fast_interrupt check.
26288 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
26289 (save_restore_insns): Likewise.
26290 (compute_frame_size): Likewise.
26291 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
26292 (microblaze_globalize_label): Likewise.
26293 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
26294 * config/microblaze/microblaze.md: Use wrapper
26295 microblaze_is_interrupt_variant.
26297 2013-03-05 Kai Tietz <ktietz@redhat.com>
26299 * sdbout.c (sdbout_one_type): Switch to current function's section
26300 supporting cold/hot.
26302 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26304 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
26307 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26309 PR middle-end/56461
26310 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
26311 if VALGRIND_GET_VBITS is defined, temporarily make object
26312 memory all defined, and restore previous valgrind addressability
26313 and definability afterwards. Free this_object at the end.
26315 PR middle-end/56461
26316 * lra.c (lra): Call lra_clear_live_ranges if live_p,
26317 right before calling lra_create_live_ranges, also call it
26318 when clearing live_p. Only call lra_clear_live_ranges
26319 at the end if live_p.
26321 PR middle-end/56461
26322 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
26324 2013-03-05 Richard Biener <rguenther@suse.de>
26326 PR tree-optimization/56521
26327 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
26330 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
26333 * except.h (remove_unreachable_eh_regions): New prototype.
26334 * except.c (remove_eh_handler_splicer): New function, split out
26335 of remove_eh_handler.
26336 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
26337 warning about running it on many EH regions one at a time.
26338 (remove_unreachable_eh_regions_worker): New function, walk the
26339 EH tree in depth-first order and remove non-marked regions.
26340 (remove_unreachable_eh_regions): New function.
26341 * tree-eh.c (mark_reachable_handlers): New function, split out
26342 from remove_unreachable_handlers.
26343 (remove_unreachable_handlers): Use mark_reachable_handlers and
26344 remove_unreachable_eh_regions.
26345 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
26346 and remove_unreachable_eh_regions.
26348 2013-03-05 Richard Biener <rguenther@suse.de>
26350 PR middle-end/56525
26351 * loop-init.c (fix_loop_structure): Remove loops in two stages,
26352 not freeing them until the end.
26354 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26356 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
26358 2013-03-05 Richard Biener <rguenther@suse.de>
26360 PR tree-optimization/56270
26361 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
26362 of loads after scheduling an SLP instance.
26364 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26366 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
26368 (check_gcc_parallelize): Run guality.exp as a separate job from
26369 vect.exp with unsorted.exp and $(dg_target_exps) separately from
26370 struct-layout-1.exp with stackalign.exp.
26372 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
26374 PR middle-end/56461
26375 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
26376 load_index sbitmap even if some bit in it isn't set.
26378 PR middle-end/56461
26379 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
26380 (discover_iteration_bound_by_body_walk): Change queues to
26381 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
26382 spelling in comment. Call safe_push on queues[bound_index] directly.
26383 Release queues[queue_index] in every iteration unconditionally.
26384 Release bounds vector.
26386 PR middle-end/56461
26387 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
26388 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
26389 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
26392 2013-03-05 Richard Biener <rguenther@suse.de>
26395 * tree-inline.c (remap_blocks_to_null): New function.
26396 (expand_call_inline): When expanding a call stmt without
26397 an associated block inline remap all callee blocks to NULL.
26399 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26401 PR rtl-optimization/56494
26402 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
26403 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
26404 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
26406 PR middle-end/56461
26407 * sel-sched-ir.c (free_sched_pools): Release
26408 succs_info_pool.stack[succs_info_pool.max_top] vectors too
26409 if succs_info_pool.max_top isn't -1.
26412 * opts.c (opts_obstack, opts_concat): Moved to...
26413 * opts-common.c (opts_obstack, opts_concat): ... here.
26415 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26417 PR middle-end/56461
26418 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
26420 2013-03-04 Martin Jambor <mjambor@suse.cz>
26422 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
26423 all appropriate places.
26425 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
26427 PR tree-optimization/56424
26428 * ipa-split.c (split_function): Do not set the RSO flag if result is
26429 not by reference and its type is a register type.
26431 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
26433 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
26434 (microblaze_legitimate_pic_operand): Likewise
26435 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
26436 new function microblaze_legitimate_pic_operand
26437 * config/microblaze/microblaze-protos.h
26438 (microblaze_legitimate_pic_operand): Declare.
26440 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26442 * config/microblaze/predicates.md (call_insn_simple_operand):
26443 New predicate for supported rtx code types.
26444 * config/microblaze/microblaze.md (call_internal1): Use
26445 call_insn_simple_operand predicate.
26447 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26449 PR middle-end/56461
26450 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
26451 partitions.ordered_remove.
26453 PR middle-end/56461
26454 * tree-vect-stmts.c (vectorizable_conversion): Don't call
26455 vec_oprnds0.create (1) for modifier == NONE.
26457 PR middle-end/56461
26458 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
26459 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
26460 vec_oprnds1 right before pushing anything to it for
26463 PR middle-end/56461
26464 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
26465 set nbbs to 0 instead of having separate code path.
26466 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
26467 instead of false as last argument if returning NULL.
26469 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
26471 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
26472 the attribute is now called "target" instead of "option".
26473 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
26474 * doc/tm.texi.in (Target Attributes): Likewise document the correct
26475 attribute/pragma name for TARGET_OPTION_VALID_P and
26476 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
26477 * doc/tm.texi: Regenerated.
26479 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
26481 * config/microblaze/microblaze.c:
26482 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
26483 * config/microblaze/microblaze.h: Add -mxl-reorder to
26485 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
26486 instructions emitted if TARGET_REORDER.
26487 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
26488 or 0 for -m/-mno case, but initialises as 2 to detect default use case
26491 2013-03-01 Xinliang David Li <davidxl@google.com>
26493 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
26496 2013-03-01 Jakub Jelinek <jakub@redhat.com>
26498 PR middle-end/56461
26499 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
26500 vector even when returning true. Fix up function comment formatting.
26502 PR middle-end/56461
26503 * ira-build.c (ira_loop_nodes_count): New variable.
26504 (create_loop_tree_nodes): Initialize it.
26505 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
26507 PR middle-end/56461
26508 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
26509 method on dr_chain and result_chain.
26510 * tree-vect-stmts.c (vectorizable_store): Only call
26511 result_chain.create if j == 0.
26513 PR middle-end/56461
26514 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
26515 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
26516 before overwriting it.
26518 2013-03-01 Tobias Burnus <burnus@net-b.de>
26520 * doc/extended.texi (C Extensions): Change order in @menu
26522 (Other MIPS Built-in Functions): Move last MIPS entry before
26523 "picoChip Built-in Functions".
26524 (SH Built-in Functions): Move after RX Built-in Functions.
26525 * doc/gcc.texi (Introduction): Change order in @menu
26527 * doc/md.texi (Constraints): Ditto.
26528 * gty.texi (Type Information): Ditto.
26529 (User-provided marking routines for template types): Make
26531 * doc/invoke.texi (AArch64 Options): Move before
26532 "Adapteva Epiphany Options".
26534 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
26535 Jakub Jelinek <jakub@redhat.com>
26538 * asan.c (gate_asan): Lookup no_sanitize_address instead of
26539 no_address_safety_analysis attribute.
26540 * doc/extend.texi (no_address_safety_attribute): Rename to
26541 no_sanitize_address attribute, mention no_address_safety_analysis
26542 attribute as deprecated alias.
26544 2013-02-28 Jakub Jelinek <jakub@redhat.com>
26546 PR middle-end/56461
26547 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
26548 type to vec<vec<tree> > *.
26549 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
26550 to be vec<tree> instead of vec<tree> *, set vec_defs
26551 to vNULL and call vec_defs.create (number_of_vects), adjust other
26553 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
26554 vectorizable_condition): Adjust vect_get_slp_defs callers.
26556 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26558 * config/aarch64/aarch64.c
26559 (aarch64_float_const_representable): Remove unused variable.
26561 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26563 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
26565 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26567 * config/aarch64/aarch64-builtins.c
26568 (aarch64_init_simd_builtins): Make static.
26570 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26572 * config/aarch64/aarch64.c
26573 (aarch64_simd_make_constant): Make static.
26575 2013-02-28 Martin Jambor <mjambor@suse.cz>
26577 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
26578 with no initialization to the RHS of debug statements.
26580 2013-02-28 Martin Jambor <mjambor@suse.cz>
26582 PR tree-optimization/56294
26583 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
26585 (get_access_replacement): Do not call create_access_replacement.
26586 Assert a replacement exists.
26587 (get_repl_default_def_ssa_name): Create the replacement declaration
26590 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26592 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
26593 final_end_function.
26595 2013-02-28 Marek Polacek <polacek@redhat.com>
26597 PR rtl-optimization/56466
26598 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
26599 if we're changing a loop.
26600 (peel_loops_completely): Likewise.
26602 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
26605 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
26607 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26610 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
26611 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
26612 INTX_FTYPE_FX, FX_FTYPE_INTX.
26613 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
26615 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26617 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
26618 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
26619 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
26620 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
26621 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
26622 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
26623 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
26624 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
26625 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
26626 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
26627 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
26628 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
26629 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
26630 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
26631 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
26632 (avrxmega6): Increase max flash segments from 5 to 6.
26633 * config/avr/t-multilib: Regenerate.
26634 * config/avr/avr-tables.opt: Regenerate.
26635 * doc/avr-mmcu.texi: Regenerate.
26637 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26639 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
26640 (avr_device_to_arch): Rename to avr_device_to_ld.
26641 (avr_device_to_as): New prototype.
26642 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
26643 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
26644 * config/avr/driver-avr.c (avr_device_to_as): New.
26645 (avr_device_to_arch): Rename to avr_device_to_ld.
26647 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26649 PR middle-end/56461
26650 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
26651 method on dr_chain and result_chain.
26653 PR middle-end/56461
26654 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
26655 pointer_set_destroy on not_executed_last_iteration.
26657 PR middle-end/56461
26658 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
26660 PR middle-end/56461
26661 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
26662 FOR_EACH_DEFINED_FUNCTION when freeing state.
26664 PR middle-end/56461
26665 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
26667 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
26670 PR middle-end/56461
26671 * ipa-cp.c (decide_whether_version_node): Call vec_free on
26672 known_aggs[i].items and release known_aggs vector.
26674 PR middle-end/56461
26675 * ipa-reference.c (propagate): Free node_info even for alias nodes.
26677 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26679 * config/microblaze/microblaze.c (microblaze_emit_compare):
26680 Use xor for EQ/NE comparisions.
26681 * config/microblaze/microblaze.md (cstoresf4): Add constraints
26682 (cbranchsf4): Adjust operator to comparison_operator.
26684 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26686 PR middle-end/56461
26687 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
26689 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
26690 vec_safe_push, always update *slot.
26691 (redirect_edge_var_map_clear): Use vec_free.
26692 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
26693 (free_var_map_entry): Use vec_free.
26694 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
26695 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
26697 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
26699 PR middle-end/45472
26700 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
26701 when the may_trap_p bit of the exprs being merged differs.
26702 Reorder tests for speculativeness in the logical and operator.
26704 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26706 * incpath.c (add_standard_paths): Use reconcat instead of concat
26707 where appropriate and avoid leaking memory.
26709 * opts.h: Include obstack.h.
26710 (opts_concat): New prototype.
26711 (opts_obstack): New declaration.
26712 * opts.c (opts_concat): New function.
26713 (opts_obstack): New variable.
26714 (init_options_struct): Call gcc_init_obstack on opts_obstack.
26715 (finish_options): Use opts_concat instead of concat
26716 and XOBNEWVEC instead of XNEWVEC.
26717 * opts-common.c (generate_canonical_option, decode_cmdline_option,
26718 generate_option): Likewise.
26719 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
26720 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
26723 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
26724 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
26726 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26728 PR middle-end/56461
26729 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
26731 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26733 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
26734 (arm_block_move_unaligned_straight): Likewise.
26735 (arm_adjust_block_mem): Likewise.
26737 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26740 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
26741 temp, cond and label.
26742 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
26745 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
26746 * config/c6x/c6x.h (dbx_register_map): Update declaration.
26749 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
26750 of prologue/epilogue functions.
26753 * config/tilegx/tilegx.c (tilegx_expand_prologue):
26754 Remove unused variable cfa_offset.
26755 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
26758 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
26759 type promotion to unsigned.
26762 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
26763 for HOST_WIDE_INT of 32 bit / same size as int.
26764 (arm_block_move_unaligned_straight): Likewise.
26765 (arm_adjust_block_mem): Likewise.
26768 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
26771 2013-02-26 Marek Polacek <polacek@redhat.com>
26773 PR tree-optimization/56426
26774 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
26776 2013-02-26 Richard Biener <rguenther@suse.de>
26779 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
26780 unused variable loops.
26782 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26784 PR tree-optimization/56448
26785 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
26786 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
26787 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
26788 later operands of the references, or even first operand for
26789 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
26791 PR tree-optimization/56443
26792 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
26793 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
26794 to type_for_mode langhook.
26796 2013-02-25 Matt Turner <mattst88@gmail.com>
26798 * doc/invoke.texi: Document r4700.
26800 2013-02-25 Richard Biener <rguenther@suse.de>
26802 PR tree-optimization/56175
26803 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
26805 (simplify_bitwise_binary): ... here. Also guard the conversion
26806 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
26808 2013-02-25 Catherine Moore <clm@codesourcery.com>
26811 2013-02-24 Catherine Moore <clm@codesourcery.com>
26812 Maciej W. Rozycki <macro@codesourcery.com>
26813 Tom de Vries <tom@codesourcery.com>
26814 Nathan Sidwell <nathan@codesourcery.com>
26815 Iain Sandoe <iain@codesourcery.com>
26816 Nathan Froyd <froydnj@codesourcery.com>
26817 Chao-ying Fu <fu@mips.com>
26819 * doc/extend.texi: (micromips, nomicromips, nocompression):
26820 Document new function attributes.
26821 * doc/invoke.texi (minterlink-compressed, mmicromips,
26822 m14k, m14ke, m14kec): Document new options.
26823 (minterlink-mips16): Update documentation.
26824 * doc/md.texi (ZC, ZD): Document new constraints.
26825 * configure.ac (gcc_cv_as_micromips): Check if linker
26826 supports the .set micromips directive.
26827 * configure: Regenerate.
26828 * config.in: Regenerate.
26829 * config/mips/mips-tables.opt: Regenerate.
26830 * config/mips/micromips.md: New file.
26831 * constraints.md (ZC, AD): New constraints.
26832 * config/mips/predicates.md (movep_src_register): New predicate.
26833 (movep_src_operand): New predicate.
26834 (non_volatile_mem_operand): New predicate.
26835 * config/mips/mips.md (multimem): New type.
26836 (length): Differentiate between 17-bit and 18-bit branch offsets.
26837 (MOVEP1, MOVEP2): New mode iterator.
26838 (mov_<load>l): Use ZC constraint.
26839 (mov_<load>r): Likewise.
26840 (mov_<store>l): Likewise.
26841 (mov_<store>r): Likewise.
26842 (*branch_equality<mode>_inverted): Add microMIPS support.
26843 (*branch_equality<mode>): Likewise.
26844 (*jump_absolute): Likewise.
26845 (indirect_jump_<mode>): Likewise.
26846 (tablejump_<mode>): Likewise.
26847 (<optab>_internal): Likewise.
26848 (sibcall_internal): Likewise.
26849 (sibcall_value_internal): Likewise.
26850 (prefetch): Use constraint ZD.
26851 * config/mips/mips.opt (minterlink-compressed): New option.
26852 (minterlink-mips16): Now an alias for minterlink-compressed.
26853 (mmicromips): New option.
26854 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26855 (compare_and_swap_12): Likewise.
26856 (sync_add<mode>): Likewise.
26857 (sync_<optab>_12): Likewise.
26858 (sync_old_<optab>_12): Likewise.
26859 (sync_new_<optab>_12): Likewise.
26860 (sync_nand_12): Likewise.
26861 (sync_old_nand_12): Likewise.
26862 (sync_new_nand_12): Likewise.
26863 (sync_sub<mode>): Likewise.
26864 (sync_old_add<mode>): Likewise.
26865 (sync_old_sub<mode>): Likewise.
26866 (sync_new_add<mode>): Likewise.
26867 (sync_new_sub<mode>): Likewise.
26868 (sync_<optab><mode>): Likewise.
26869 (sync_old_<optab><mode>): Likewise.
26870 (sync_new_<optab><mode>): Likewise.
26871 (sync_nand<mode>): Likewise.
26872 (sync_old_nand<mode>): Likewise.
26873 (sync_new_nand<mode>): Likewise.
26874 (sync_lock_test_and_set<mode>): Likewise.
26875 (test_and_set_12): Likewise.
26876 (atomic_compare_and_swap<mode>): Likewise.
26877 (atomic_exchange<mode>_llsc): Likewise.
26878 (atomic_fetch_add<mode>_llsc): Likewise.
26879 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26880 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26881 (umips_save_restore_pattern_p): Likewise.
26882 (umips_load_store_pair_p): Likewise.
26883 (umips_output_load_store_pair): Likewise.
26884 (umips_movep_target_p): Likewise.
26885 (umips_12bit_offset_address_p): Likewise.
26886 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26887 (mips_base_mips16): Rename this...
26888 (mips_base_compression_flags): ...to this. Update all uses.
26889 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26890 (mips_mips16_decl_p): Delete.
26891 (mips_nomips16_decl_p): Delete.
26892 (mips_get_compress_on_flags): New function.
26893 (mips_get_compress_off_flags): New function.
26894 (mips_get_compress_mode): New function.
26895 (mips_get_compress_on_name): New function.
26896 (mips_get_compress_off_name): New function.
26897 (mips_insert_attributes): Support multiple compression types.
26898 (mips_merge_decl_attributes): Likewise.
26899 (umips_12bit_offset_address_p): New function.
26900 (mips_start_function_definition): Emit .set micromips directive.
26901 (mips_call_may_need_jalx_p): New function.
26902 (mips_function_ok_for_sibcall): Add microMIPS support.
26903 (mips_print_operand_punctuation): Support short delay slots and
26905 (umips_swm_mask, umips_swm_encoding): New.
26906 (umips_build_save_restore): New function.
26907 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26908 (was_mips16_p): Remove.
26909 (old_compression_mode): New.
26910 (mips_set_compression_mode): New function.
26911 (mips_set_current_function): Add microMIPS support.
26912 (mips_option_override): Likewise.
26913 (umips_save_restore_pattern_p): New function.
26914 (umips_output_save_restore): New function.
26915 (umips_load_store_pair_p_1): New function.
26916 (umips_load_store_pair_p): New function.
26917 (umips_output_load_store_pair_1): New function.
26918 (umips_output_load_store_pair): New function.
26919 (umips_movep_target_p) New function.
26920 (mips_prepare_pch_save): Add microMIPS support.
26921 * config/mips/mips.h (TARGET_COMPRESSION): New.
26922 (TARGET_CPU_CPP_BUILTINS): Update macro
26923 to use new compression flags and to support microMIPS.
26924 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26925 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26926 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26927 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26928 (ASM_SPEC): Support mmicromips and mno-micromips.
26929 (M16STORE_REG_P): New macro.
26930 (MIPS_CALL): Support TARGET_MICROMIPS.
26931 (MICROMIPS_J): New macro.
26932 (mips_base_mips16): Rename this...
26933 (mips_base_compression_flags): ...to this.
26934 (UMIPS_12BIT_OFFSET_P): New macro.
26935 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26936 (MULTILIB_DIRNAMES): Likewise.
26938 2013-02-25 Tom de Vries <tom@codesourcery.com>
26940 PR rtl-optimization/56131
26941 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
26942 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
26943 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
26945 2013-02-25 Tobias Burnus <burnus@net-b.de>
26947 * doc/invoke.texi (-fsanitize=): Move from optimization
26948 to debugging options.
26950 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26952 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
26954 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26955 Alexander Monakov <amonakov@ispras.ru>
26957 PR middle-end/56077
26958 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
26959 flush pending lists also on non-jumps. Adjust comment.
26961 2013-02-24 Catherine Moore <clm@codesourcery.com>
26962 Maciej W. Rozycki <macro@codesourcery.com>
26963 Tom de Vries <tom@codesourcery.com>
26964 Nathan Sidwell <nathan@codesourcery.com>
26965 Iain Sandoe <iain@codesourcery.com>
26966 Nathan Froyd <froydnj@codesourcery.com>
26967 Chao-ying Fu <fu@mips.com>
26969 * doc/extend.texi: (micromips, nomicromips, nocompression):
26970 Document new function attributes.
26971 * doc/invoke.texi (minterlink-compressed, mmicromips,
26972 m14k, m14ke, m14kec): Document new options.
26973 (minterlink-mips16): Update documentation.
26974 * doc/md.texi (ZC, ZD): Document new constraints.
26975 * configure.ac (gcc_cv_as_micromips): Check if linker
26976 supports the .set micromips directive.
26977 * configure: Regenerate.
26978 * config.in: Regenerate.
26979 * config/mips/mips-tables.opt: Regenerate.
26980 * config/mips/micromips.md: New file.
26981 * constraints.md (ZC, AD): New constraints.
26982 * config/mips/predicates.md (movep_src_register): New predicate.
26983 (movep_src_operand): New predicate.
26984 (non_volatile_mem_operand): New predicate.
26985 * config/mips/mips.md (multimem): New type.
26986 (length): Differentiate between 17-bit and 18-bit branch offsets.
26987 (MOVEP1, MOVEP2): New mode iterator.
26988 (mov_<load>l): Use ZC constraint.
26989 (mov_<load>r): Likewise.
26990 (mov_<store>l): Likewise.
26991 (mov_<store>r): Likewise.
26992 (*branch_equality<mode>_inverted): Add microMIPS support.
26993 (*branch_equality<mode>): Likewise.
26994 (*jump_absolute): Likewise.
26995 (indirect_jump_<mode>): Likewise.
26996 (tablejump_<mode>): Likewise.
26997 (<optab>_internal): Likewise.
26998 (sibcall_internal): Likewise.
26999 (sibcall_value_internal): Likewise.
27000 (prefetch): Use constraint ZD.
27001 * config/mips/mips.opt (minterlink-compressed): New option.
27002 (minterlink-mips16): Now an alias for minterlink-compressed.
27003 (mmicromips): New option.
27004 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27005 (compare_and_swap_12): Likewise.
27006 (sync_add<mode>): Likewise.
27007 (sync_<optab>_12): Likewise.
27008 (sync_old_<optab>_12): Likewise.
27009 (sync_new_<optab>_12): Likewise.
27010 (sync_nand_12): Likewise.
27011 (sync_old_nand_12): Likewise.
27012 (sync_new_nand_12): Likewise.
27013 (sync_sub<mode>): Likewise.
27014 (sync_old_add<mode>): Likewise.
27015 (sync_old_sub<mode>): Likewise.
27016 (sync_new_add<mode>): Likewise.
27017 (sync_new_sub<mode>): Likewise.
27018 (sync_<optab><mode>): Likewise.
27019 (sync_old_<optab><mode>): Likewise.
27020 (sync_new_<optab><mode>): Likewise.
27021 (sync_nand<mode>): Likewise.
27022 (sync_old_nand<mode>): Likewise.
27023 (sync_new_nand<mode>): Likewise.
27024 (sync_lock_test_and_set<mode>): Likewise.
27025 (test_and_set_12): Likewise.
27026 (atomic_compare_and_swap<mode>): Likewise.
27027 (atomic_exchange<mode>_llsc): Likewise.
27028 (atomic_fetch_add<mode>_llsc): Likewise.
27029 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27030 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27031 (umips_save_restore_pattern_p): Likewise.
27032 (umips_load_store_pair_p): Likewise.
27033 (umips_output_load_store_pair): Likewise.
27034 (umips_movep_target_p): Likewise.
27035 (umips_12bit_offset_address_p): Likewise.
27036 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27037 (mips_base_mips16): Rename this...
27038 (mips_base_compression_flags): ...to this. Update all uses.
27039 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27040 (mips_mips16_decl_p): Delete.
27041 (mips_nomips16_decl_p): Delete.
27042 (mips_get_compress_on_flags): New function.
27043 (mips_get_compress_off_flags): New function.
27044 (mips_get_compress_mode): New function.
27045 (mips_get_compress_on_name): New function.
27046 (mips_get_compress_off_name): New function.
27047 (mips_insert_attributes): Support multiple compression types.
27048 (mips_merge_decl_attributes): Likewise.
27049 (umips_12bit_offset_address_p): New function.
27050 (mips_start_function_definition): Emit .set micromips directive.
27051 (mips_call_may_need_jalx_p): New function.
27052 (mips_function_ok_for_sibcall): Add microMIPS support.
27053 (mips_print_operand_punctuation): Support short delay slots and
27055 (umips_swm_mask, umips_swm_encoding): New.
27056 (umips_build_save_restore): New function.
27057 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27058 (was_mips16_p): Remove.
27059 (old_compression_mode): New.
27060 (mips_set_compression_mode): New function.
27061 (mips_set_current_function): Add microMIPS support.
27062 (mips_option_override): Likewise.
27063 (umips_save_restore_pattern_p): New function.
27064 (umips_output_save_restore): New function.
27065 (umips_load_store_pair_p_1): New function.
27066 (umips_load_store_pair_p): New function.
27067 (umips_output_load_store_pair_1): New function.
27068 (umips_output_load_store_pair): New function.
27069 (umips_movep_target_p) New function.
27070 (mips_prepare_pch_save): Add microMIPS support.
27071 * config/mips/mips.h (TARGET_COMPRESSION): New.
27072 (TARGET_CPU_CPP_BUILTINS): Update macro
27073 to use new compression flags and to support microMIPS.
27074 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27075 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27076 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27077 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27078 (ASM_SPEC): Support mmicromips and mno-micromips.
27079 (M16STORE_REG_P): New macro.
27080 (MIPS_CALL): Support TARGET_MICROMIPS.
27081 (MICROMIPS_J): New macro.
27082 (mips_base_mips16): Rename this...
27083 (mips_base_compression_flags): ...to this.
27084 (UMIPS_12BIT_OFFSET_P): New macro.
27085 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27086 (MULTILIB_DIRNAMES): Likewise.
27088 2013-02-24 Jakub Jelinek <jakub@redhat.com>
27091 * target-globals.c (save_target_globals): For init_reg_sets and
27092 target_reinit remporarily set this_fn_optabs to this_target_optabs.
27094 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
27096 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
27097 * config/aarch64/t-aarch64
27098 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
27100 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
27102 PR inline-asm/56148
27103 * lra-constraints.c (process_alt_operands): Reload operand
27104 conflicting with earlier clobber only if no more other conflicting
27107 2013-02-22 Jakub Jelinek <jakub@redhat.com>
27110 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
27111 if not linking a shared library.
27113 2013-02-22 Seth LaForge <sethml@google.com>
27115 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
27117 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
27119 * config/arm/arm.md (split for extendsidi): Update condition.
27120 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
27121 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
27122 (qhs_zextenddi_cstr): Likewise.
27124 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27126 PR middle-end/56420
27127 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
27128 avoid signed wrapping.
27129 (expand_mult): Handle properly multiplication by
27130 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
27131 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
27132 in the compiler if coeff is HOST_WIDE_INT_MIN.
27133 (expand_divmod): Don't make ext_op1 static, change it's type to
27134 uhwi. Avoid undefined behavior in -INTVAL (op1).
27136 PR rtl-optimization/50339
27137 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
27139 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
27140 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
27141 into splitting_ashiftrt field.
27142 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
27144 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
27147 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
27149 PR middle-end/56108
27150 * trans-mem.c (execute_tm_mark): Do not expand transactions that
27151 are sure to go irrevocable.
27153 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
27155 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
27156 scalars are valid operands.
27158 2013-02-21 Martin Jambor <mjambor@suse.cz>
27160 PR tree-optimization/56310
27161 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
27162 only matching indices and non-negative final offsets.
27163 (intersect_aggregates_with_edge): Pass src_idx to
27164 agg_replacements_to_vector. Pass src_idx insstead of index to
27165 intersect_with_agg_replacements.
27167 2013-02-21 Martin Jambor <mjambor@suse.cz>
27169 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
27170 instead of hard-wired defaults.
27172 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
27174 * doc/invoke.texi (MIPS Options): Update documentation of the
27175 floating-point multiply-accumulate instruction restrictions.
27177 2013-02-21 Kostya Serebryany <kcc@google.com>
27179 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
27180 asan_shadow_offset on x86_64 linux.
27182 2013-02-21 Richard Biener <rguenther@suse.de>
27184 PR tree-optimization/56415
27186 2013-02-11 Richard Biener <rguenther@suse.de>
27188 PR tree-optimization/56273
27189 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27192 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27195 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
27198 PR inline-asm/56405
27199 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
27200 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
27202 2013-02-20 Jan Hubicka <jh@suse.cz>
27204 PR tree-optimization/56265
27205 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
27206 when target is referenced for first time.
27208 2013-02-20 Richard Biener <rguenther@suse.de>
27210 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
27211 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
27212 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
27213 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
27214 not return anything.
27215 (rename_ssa_copies): Do not remove unused locals.
27216 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
27217 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
27218 * passes.c (execute_function_todo): Do not schedule unused locals
27219 removal if cleanup_tree_cfg did something.
27220 * tree-ssa-live.c (remove_unused_locals): Dump statistics
27221 about the number of removed locals.
27223 2013-02-20 Richard Biener <rguenther@suse.de>
27225 PR tree-optimization/56398
27226 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
27228 2013-02-20 Martin Jambor <mjambor@suse.cz>
27230 PR tree-optimization/55334
27231 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
27232 restricted pointers to arrays.
27234 2013-02-20 Richard Biener <rguenther@suse.de>
27235 Jakub Jelinek <jakub@redhat.com>
27237 PR tree-optimization/56396
27238 * tree-ssa-ccp.c (n_const_val): New static variable.
27239 (get_value): Return NULL for SSA names we don't have a lattice
27241 (ccp_initialize): Initialize n_const_val.
27242 * tree-ssa-copy.c (n_copy_of): New static variable.
27243 (init_copy_prop): Initialize n_copy_of.
27244 (get_value): Return NULL_TREE for SSA names we don't have a
27247 2013-02-20 Martin Jambor <mjambor@suse.cz>
27249 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
27251 2013-02-20 Richard Biener <rguenther@suse.de>
27253 * genpreds.c (write_lookup_constraint): Do not compare first
27254 letter of the constraint again.
27256 2013-02-20 Richard Biener <rguenther@suse.de>
27258 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
27260 (get_use_iv_cost): Terminate hashtable walk when coming across
27263 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
27265 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
27266 reassociation for avx2 targets.
27268 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27270 * config/microblaze/microblaze.c: microblaze_has_clz = 0
27271 Add version check for v8.10.a to enable microblaze_has_clz
27272 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
27273 version and TARGET_PATTERN_COMPARE check
27274 * config/microblaze/microblaze.md: New clzsi2 instruction
27276 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27278 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
27279 function before branching.
27281 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27283 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
27285 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
27287 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27289 PR middle-end/55889
27290 * sel-sched.c: Include ira.h.
27291 (implicit_clobber_conflict_p): New function.
27292 (moveup_expr): Use it.
27293 * Makefile.in (sel-sched.o): Depend on ira.h.
27295 2013-02-19 Richard Biener <rguenther@suse.de>
27297 PR tree-optimization/56384
27298 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
27299 (vn_hash_type): Split out from ...
27300 (vn_hash_constant_with_type): ... here.
27301 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
27302 (vn_phi_eq): Compare types from vn_phi_s structure.
27303 (vn_phi_lookup): Populate vn_phi_s type.
27304 (vn_phi_insert): Likewise.
27306 2013-02-19 Jakub Jelinek <jakub@redhat.com>
27308 PR tree-optimization/56350
27309 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
27310 if haven't found reduction or nested cycle operand, rather than
27311 asserting we must find it.
27313 PR tree-optimization/56381
27314 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
27317 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
27318 Jakub Jelinek <jakub@redhat.com>
27321 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
27322 (swap_optab_enable): Same.
27323 (init_all_optabs): Use argument instead of global.
27324 * tree.h (struct tree_optimization_option): New field target_optabs.
27325 * expr.h (init_all_optabs): Add argument to prototype.
27326 (TREE_OPTIMIZATION_OPTABS): New.
27327 (save_optabs_if_changed): Protoize.
27328 * optabs.h: Declare this_fn_optabs.
27329 * optabs.c (save_optabs_if_changed): New.
27330 Declare this_fn_optabs.
27331 (init_optabs): Add argument to init_all_optabs() call.
27332 * function.c (invoke_set_current_function_hook): Handle per
27334 * function.h (struct function): New field optabs.
27335 * config/mips/mips.c (mips_set_mips16_mode): Handle when
27336 optimization_current_node has changed.
27337 * target-globals.h (save_target_globals_default_opts): Protoize.
27338 * target-globals.c (save_target_globals_default_opts): New.
27340 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27343 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
27344 registers %fr12 and %fr12R as call used.
27347 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
27348 and HImode, require all displacements to be an integer multiple of
27350 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
27351 only allow QImode and HImode when reload is in progress and strict is
27352 true. Likewise for symbolic addresses. Use base14_operand to check
27353 displacements in REG+BASE addresses.
27355 2013-02-18 Richard Biener <rguenther@suse.de>
27357 PR tree-optimization/56366
27358 * tree-vect-loop.c (get_initial_def_for_induction): Properly
27359 handle sign-conversion of outer-loop initial induction value.
27361 2013-02-18 Richard Biener <rguenther@suse.de>
27363 PR middle-end/56349
27364 * cfghooks.c (merge_blocks): If we merge a latch into another
27365 block adjust references to it.
27366 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
27367 (verify_loop_structure): Verify that a recorded latch is in fact
27370 2013-02-18 Richard Biener <rguenther@suse.de>
27372 PR tree-optimization/56321
27373 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
27374 order SSA name release and virtual operand unlinking.
27376 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27378 * config/microblaze/microblaze.md (save_stack_block): Define.
27379 (restore_stack_block): Likewise.
27381 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27383 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
27384 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
27385 * config/microblaze/microblaze.c (microblaze_option_override):
27386 Bail out early for PIC modes when target does not support PIC.
27388 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27390 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
27391 Replace with a microblaze version.
27392 (microblaze_trampoline_init): Adapt for microblaze.
27393 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
27396 2013-02-16 Jakub Jelinek <jakub@redhat.com>
27397 Dodji Seketeli <dodji@redhat.com>
27400 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
27401 (instrument_mem_region_access): Do not forget to always put
27402 instrumentation of the of 'base' and 'base + len' in a "if (len !=
27403 0) statement, even for cases where either 'base' or 'base + len'
27404 are not instrumented -- because they have been previously
27405 instrumented. Simplify the logic by putting all the statements
27406 instrument 'base + len' inside a sequence, and then insert that
27407 sequence right before the current insertion point. Then, to
27408 instrument 'base + len', just get an iterator on that statement.
27409 And do not forget to update the pointer to iterator the function
27410 received as argument.
27412 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
27414 PR rtl-optimization/56348
27415 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
27417 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
27419 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
27420 (clean_graph_dump_file): Pass base to start_graph_dump.
27422 2013-02-14 Richard Henderson <rth@redhat.com>
27425 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
27427 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
27429 * collect2-aix.h: Define F_LOADONLY.
27431 2013-02-14 Richard Biener <rguenther@suse.de>
27434 * varasm.c (output_constant_def_1): Get the decl representing
27435 the constant as argument.
27436 (output_constant_def): Wrap output_constant_def_1.
27437 (make_decl_rtl): Use output_constant_def_1 with the decl
27438 representing the constant.
27439 (build_constant_desc): Optionally re-use a decl already
27440 representing the constant.
27441 (tree_output_constant_def): Adjust.
27443 2013-02-14 Dodji Seketeli <dodji@redhat.com>
27446 * asan.c (instrument_builtin_call): Really put the length of the
27447 second source argument into src1_len.
27449 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27451 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
27452 argument. If it is false, don't create edge from then_bb to
27454 (insert_if_then_before_iter): Pass true to it.
27455 (build_check_stmt): Pass false to it.
27456 (transform_statements): Flush hash table only on extended basic
27457 block boundaries, rather than at the beginning of every bb.
27458 Don't flush hash table on nonfreeing_call_p calls.
27459 * tree-flow.h (nonfreeing_call_p): New prototype.
27460 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
27462 2013-02-13 David S. Miller <davem@davemloft.net>
27464 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
27466 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
27469 * ira.c (max_regno_before_ira): Move from ...
27471 (fix_reg_equiv_init): Use max_regno_before_ira instead of
27474 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27476 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
27478 2013-02-13 Richard Biener <rguenther@suse.de>
27481 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
27482 globals in MEM_REFs.
27484 2013-02-13 Richard Biener <rguenther@suse.de>
27486 * loop-init.c (loop_optimizer_init): Clear loop state when
27487 re-initializing preserved loops.
27488 * loop-unswitch.c (unswitch_single_loop): Return whether
27489 we unswitched the loop. Do not verify loop state here.
27490 (unswitch_loops): When we unswitched a loop discover new loops.
27492 2013-02-13 Kostya Serebryany <kcc@google.com>
27494 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
27496 * sanitizer.def: Rename __asan_init to __asan_init_v1.
27498 2013-02-12 Dodji Seketeli <dodji@redhat.com>
27500 Avoid instrumenting duplicated memory access in the same basic block
27501 * Makefile.in (asan.o): Add new dependency on hash-table.h
27502 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
27503 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
27504 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
27505 (free_mem_ref_resources, has_mem_ref_been_instrumented)
27506 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
27507 (get_mem_ref_of_assignment): New functions.
27508 (get_mem_refs_of_builtin_call): Extract from
27509 instrument_builtin_call and tweak a little bit to make it fit with
27511 (instrument_builtin_call): Use the new
27512 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
27513 of is_gimple_builtin_call.
27514 (instrument_derefs, instrument_mem_region_access): Insert the
27515 instrumented memory reference into the hash table.
27516 (maybe_instrument_assignment): Renamed instrument_assignment into
27517 this, and change it to advance the iterator when instrumentation
27518 actually happened and return true in that case. This makes it
27519 homogeneous with maybe_instrument_assignment, and thus give a
27520 chance to callers to be more 'regular'.
27521 (transform_statements): Clear the memory reference hash table
27522 whenever we enter a new BB, when we cross a function call, or when
27523 we are done transforming statements. Use
27524 maybe_instrument_assignment instead of instrumentation. No more
27525 need to special case maybe_instrument_assignment and advance the
27526 iterator after calling it; it's now handled just like
27527 maybe_instrument_call. Update comment.
27529 2013-02-13 Richard Biener <rguenther@suse.de>
27531 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
27532 Fix loop discovery code.
27534 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
27536 PR inline-asm/56148
27537 * lra-constraints.c (process_alt_operands): Match early clobber
27538 operand with itself. Check conflicts with earlyclobber only if
27539 the operand is not reloaded. Prefer to reload conflicting operand
27540 if earlyclobber and matching operands are the same.
27542 2013-02-12 Richard Biener <rguenther@suse.de>
27545 * lto-streamer-out.c (write_symbol): Do not output symbols
27546 for hard register variables.
27548 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
27551 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
27552 (umulsidi3_insn, mulsidi3_insn): New insns.
27554 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
27556 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
27557 (struct tune_params): Add vec_costs field.
27558 * config/arm/arm.c (arm_builtin_vectorization_cost)
27559 (arm_add_stmt_cost): New functions.
27560 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
27561 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
27562 (arm_default_vec_cost): New struct of type cpu_vec_costs.
27563 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
27564 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
27565 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
27566 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
27568 2013-02-12 Richard Biener <rguenther@suse.de>
27571 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
27572 decls again if possible.
27574 2013-02-12 Richard Biener <rguenther@suse.de>
27576 PR middle-end/56288
27577 * tree-ssa.c (verify_ssa_name): Fix check, move
27578 SSA_NAME_IN_FREE_LIST check up.
27580 2013-02-12 Jakub Jelinek <jakub@redhat.com>
27581 Steven Bosscher <steven@gcc.gnu.org>
27583 PR rtl-optimization/56151
27584 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
27585 equal to op0 or op1, and last_insn pattern is CODE operation
27586 with MEM dest and one of the operands matches that MEM.
27588 2013-02-11 Sriraman Tallam <tmsriram@google.com>
27590 * doc/extend.texi: Document Function Multiversioning and "default"
27591 parameter string to target attribute.
27592 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
27593 target attribute parameter is "default".
27594 (ix86_compare_version_priority): Remove checks for target attribute.
27595 (ix86_mangle_function_version_assembler_name): Change error to sorry.
27596 Remove check for target attribute equal to NULL. Add assert.
27597 (ix86_generate_version_dispatcher_body): Change error to sorry.
27599 2013-02-11 Iain Sandoe <iain@codesourcery.com>
27600 Jack Howarth <howarth@bromo.med.uc.edu>
27601 Patrick Marlier <patrick.marlier@gmail.com>
27604 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
27605 define ENDFILE_SPEC as TM_DESTRUCTOR.
27606 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
27608 2013-02-11 Alexander Potapenko <glider@google.com>
27609 Jack Howarth <howarth@bromo.med.uc.edu>
27610 Jakub Jelinek <jakub@redhat.com>
27613 * config/darwin.c (cdtor_record): Rename ctor_record.
27614 (sort_cdtor_records): Rename sort_ctor_records.
27615 (finalize_dtors): New routine to sort destructors by
27616 priority before use in assemble_integer.
27617 (machopic_asm_out_destructor): Use finalize_dtors if needed.
27619 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
27621 PR rtl-optimization/56275
27622 * simplify-rtx.c (avoid_constant_pool_reference): Check that
27623 offset is non-negative and less than cmode size before
27624 calling simplify_subreg.
27626 2013-02-11 Richard Biener <rguenther@suse.de>
27628 PR tree-optimization/56264
27629 * cfgloop.h (fix_loop_structure): Adjust prototype.
27630 * loop-init.c (fix_loop_structure): Return the number of
27631 newly discovered loops.
27632 * tree-cfgcleanup.c (repair_loop_structures): When new loops
27633 are discovered, do a full loop-closed SSA rewrite.
27635 2013-02-11 Richard Biener <rguenther@suse.de>
27637 PR tree-optimization/56273
27638 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27640 (check_array_ref): Fix missing newline in dumps.
27641 (search_for_addr_array): Likewise.
27643 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
27645 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
27647 2013-02-09 Jakub Jelinek <jakub@redhat.com>
27650 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
27652 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
27654 PR rtl-optimization/56246
27655 * lra-constraints.c (simplify_operand_subreg): Try to reuse
27657 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
27658 constraints are satisfied.
27660 2013-02-08 Jeff Law <law@redhat.com>
27663 * emit-rtl.c (reg_is_parm_p): New function.
27664 * regs.h (reg_is_parm_p): New prototype.
27665 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
27666 callee-clobbered registers.
27668 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
27671 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
27672 If there is no implicit builtin declaration, just return NULL.
27674 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27676 * config/i386/sse.md (FMAMODEM): New mode iterator.
27677 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
27678 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
27680 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27682 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
27683 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
27684 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
27686 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27688 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
27689 (microblaze*-*-elf): Likewise.
27690 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
27692 * config/microblaze/microblaze-c.c: Add builtin defines for
27693 _LITTLE_ENDIAN and _BIG_ENDIAN.
27694 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
27695 add to TARGET_DEFAULT flags.
27696 Expand ASM_SPEC and LINK_SPEC.
27697 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
27698 * config/microblaze/microblaze.md: Update extendsidi2 and
27699 movdi_internal instructions to use low-order / high-order reg
27701 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
27702 options and inversemask / mask of LITTLE_ENDIAN.
27703 * config/microblaze/t-microblaze: Expand multilib options to
27704 include mlittle-endian (le) and update exceptions patterns.
27706 2013-02-08 Jakub Jelinek <jakub@redhat.com>
27708 PR rtl-optimization/56195
27709 * lra-constraints.c (get_reload_reg): Don't reuse regs
27710 if they have smaller mode than requested, if they have
27711 wider mode than requested, try to return a SUBREG.
27713 PR tree-optimization/56250
27714 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
27715 if type is unsigned and code isn't MULT_EXPR.
27717 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27719 PR tree-optimization/56064
27720 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
27721 bits according to mode.
27722 * fixed-value.h (fixed_from_double_int)
27723 (const_fixed_from_double_int): Adjust comments.
27725 2013-02-08 Richard Biener <rguenther@suse.de>
27728 * lto-streamer.h (struct data_in): Remove current_file, current_line
27729 and current_col members.
27730 * lto-streamer-out.c (lto_output_location): Stream changed bits
27731 en-block for efficiency.
27732 * lto-streamer-in.c (clear_line_info): Remove.
27733 (lto_input_location): Cache current file, line and column
27734 globally via local statics. Read changed bits en-block.
27735 (input_function): Do not call clear_line_info.
27736 (lto_read_body): Likewise.
27737 (lto_input_toplevel_asms): Likewise.
27739 2013-02-08 Michael Matz <matz@suse.de>
27741 PR tree-optimization/52448
27742 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
27743 (nt_call_phase): New static.
27744 (add_or_mark_expr): Only mark accesses with newer phase than any
27746 (nonfreeing_call_p): New.
27747 (nt_init_block): Update nt_call_phase, mark blocks as visited.
27748 (nt_fini_block): Keep blocks marked as visited.
27749 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
27751 2013-02-08 Richard Biener <rguenther@suse.de>
27753 * ira.c (ira): Free broken dominator information.
27755 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27757 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
27759 2013-02-08 Marek Polacek <polacek@redhat.com>
27761 * cfgloop.c (verify_loop_structure): Add more checking of headers.
27763 2013-02-08 Richard Biener <rguenther@suse.de>
27765 PR middle-end/56181
27766 * cfgloop.h (flow_loops_find): Adjust.
27767 (bb_loop_header_p): Declare.
27768 * cfgloop.c (bb_loop_header_p): New function split out from ...
27769 (flow_loops_find): ... here. Adjust function signature,
27770 support incremental loop structure update.
27771 (verify_loop_structure): Cleanup. Verify a loop is a loop.
27772 * cfgloopmanip.c (fix_loop_structure): Move ...
27773 * loop-init.c (fix_loop_structure): ... here.
27774 (apply_loop_flags): Split out from ...
27775 (loop_optimizer_init): ... here.
27776 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
27777 in incremental mode, only remove dead loops here.
27779 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27782 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
27783 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
27784 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
27785 (*round<mode>3.libgcc): New insns for fixed-modes.
27786 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
27787 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
27788 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
27789 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
27790 implementations. Define to __builtin_avr_absFX,
27791 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
27792 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
27793 __builtin_avr_countlsFX, respectively.
27794 * config/avr/avr-c.c (target.h): Include it.
27795 (enum avr_builtin_id): New enum.
27796 (avr_resolve_overloaded_builtin): New static function.
27797 (avr_register_target_pragmas): Use it to set
27798 targetm.resolve_overloaded_builtin.
27799 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
27800 tree nodes used by DEF_BUILTIN.
27801 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
27802 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
27803 <AVR_BUILTIN_xxBITS>: Same.
27805 2013-02-08 Richard Biener <rguenther@suse.de>
27807 * cfgloop.c (verify_loop_structure): Properly handle
27808 a loop exiting to another loop header.
27809 * ira-int.h (ira_loops): Remove.
27810 * ira.c (ira_loops): Remove.
27811 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
27812 (do_reload): Use loop_optimizer_finalize.
27813 * ira-build.c (create_loop_tree_nodes): Use get_loops and
27814 number_of_loops to access the loop tree.
27815 (more_one_region_p): Likewise.
27816 (finish_loop_tree_nodes): Likewise.
27817 (rebuild_regno_allocno_maps): Likewise.
27818 (mark_loops_for_removal): Likewise.
27819 (mark_all_loops_for_removal): Likewise.
27820 (remove_unnecessary_regions): Likewise.
27821 (ira_build): Likewise.
27822 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
27824 2013-02-08 Richard Biener <rguenther@suse.de>
27826 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
27827 * ipa-pure-const.c (analyze_function): Avoid calling
27828 mark_irreducible_loops twice.
27829 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
27831 2013-02-07 David S. Miller <davem@davemloft.net>
27833 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
27835 * var-tracking.c (vt_add_function_parameter): Test the presence of
27836 HAVE_window_save properly and do not remap argument registers when
27837 we have a leaf function.
27839 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
27842 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
27844 * config/i386/i386.c (ix86_print_operand): Ditto.
27846 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27848 * lra-constraints.c (process_alt_operands): Fix recently added comment.
27850 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27852 PR rtl-optimization/56225
27853 * lra-constraints.c (process_alt_operands): Check that reload hard
27854 reg can hold value for strict_low_part.
27856 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27859 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
27860 dwarf2out_end_function.
27861 (in_first_function_p, maybe_at_text_label_p,
27862 first_loclabel_num_not_at_text_label): New variables.
27863 (dwarf2out_var_location): In the first function find out
27864 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
27865 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
27868 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
27870 PR rtl-optimization/56178
27871 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
27872 SUBREG of a register. Tidy up related block of code.
27873 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
27874 note if the source is a register or a SUBREG of a register.
27876 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27879 * config/rs6000/rs6000.md (ptrm): New mode attr.
27880 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
27881 call_value_indirect_aix<pttrsize>,
27882 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
27885 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
27887 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
27888 if -bnortl. Convert to strcmp and strncmp.
27890 2013-02-07 Alan Modra <amodra@gmail.com>
27893 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
27894 addresses won't wrap when offsetting.
27895 (rs6000_secondary_reload): Provide secondary reloads needed for
27896 wrapping LO_SUM addresses.
27898 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
27900 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
27901 MACH, just __MACH__.
27903 2013-02-06 Richard Biener <rguenther@suse.de>
27905 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
27906 instead of calling fix_loop_structure.
27908 2013-02-06 Jakub Jelinek <jakub@redhat.com>
27910 PR middle-end/56217
27911 * omp-low.c (use_pointer_for_field): Return false if
27912 lower_send_shared_vars doesn't generate any copy-out code.
27914 2013-02-06 Tom de Vries <tom@codesourcery.com>
27916 PR rtl-optimization/56131
27917 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
27918 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
27919 of the label is NULL. Add comment.
27921 2013-02-05 Jakub Jelinek <jakub@redhat.com>
27923 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
27926 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
27927 (STATIC_LIBTSAN_LIBS): Likewise.
27928 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
27929 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
27930 is defined, don't add anything else beyond that.
27931 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
27932 (LINK_COMMAND_SPEC): Use them.
27934 PR tree-optimization/56205
27935 * tree-stdarg.c (check_all_va_list_escapes): Return true if
27936 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
27937 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
27939 2013-02-05 Richard Biener <rguenther@suse.de>
27941 PR tree-optimization/53342
27942 PR tree-optimization/53185
27943 * tree-vectorizer.h (vect_check_strided_load): Remove.
27944 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
27945 not disallow peeling for vectorized strided loads.
27946 (vect_check_strided_load): Make static and simplify.
27947 (vect_analyze_data_refs): Adjust.
27948 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
27949 correctly when vectorizing strided loads.
27951 2013-02-05 Richard Biener <rguenther@suse.de>
27953 * doc/install.texi: Refer to ISL, not PPL.
27955 2013-02-05 Jan Hubicka <jh@suse.cz>
27957 PR tree-optimization/55789
27958 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
27960 2013-02-05 Jan Hubicka <jh@suse.cz>
27962 PR tree-optimization/55789
27963 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
27964 the dead call anyway.
27966 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
27969 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
27971 2013-02-04 Alexander Potapenko <glider@google.com>
27972 Jack Howarth <howarth@bromo.med.uc.edu>
27973 Jakub Jelinek <jakub@redhat.com>
27976 * config/darwin.c (sort_ctor_records): Stabilized qsort
27977 on constructor priority by using original position.
27978 (finalize_ctors): New routine to sort constructors by
27979 priority before use in assemble_integer.
27980 (machopic_asm_out_constructor): Use finalize_ctors if needed.
27982 2013-02-04 Jakub Jelinek <jakub@redhat.com>
27985 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
27986 about visibility on artificial decls.
27987 * config/sol2.c (solaris_assemble_visibility): Likewise.
27989 2013-02-04 Kai Tietz <ktietz@redhat.com>
27992 * config/i386/i386.c (function_value_ms_64): Add additional valtype
27993 argument and improve checking of return-argument types for 16-byte
27995 (ix86_function_value_1): Add additional valtype argument on call
27996 of function_value_64.
27997 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
27998 handling infunction_value_64 function.
28000 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
28002 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
28004 2013-02-04 Richard Biener <rguenther@suse.de>
28006 PR tree-optimization/56188
28007 * tree-ssa-structalias.c (label_visit): Consider case with
28008 initially non-empty points-to set.
28009 (perform_var_substitution): Dump node mapping and clean up.
28011 2013-02-04 Richard Guenther <rguenther@suse.de>
28014 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
28015 node prevail as last resort.
28016 (lto_symtab_merge_decls): Remove guard on LTRANS here.
28017 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
28019 2013-02-04 Richard Biener <rguenther@suse.de>
28021 PR tree-optimization/56113
28022 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
28024 (equiv_class_lookup_or_add): ... this.
28025 (label_visit): Adjust and fix error in previous patch.
28026 (perform_var_substitution): Adjust.
28028 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
28030 * config/sh/divtab.c: Fix formatting and comments throughout the file.
28031 * config/sh/sh4-300.md: Likewise.
28032 * config/sh/sh4a.md: Likewise.
28033 * config/sh/constraints.md: Likewise.
28034 * config/sh/sh.md: Likewise.
28035 * config/sh/netbsd-elf.h: Likewise.
28036 * config/sh/predicates.md: Likewise.
28037 * config/sh/sh-protos.h: Likewise.
28038 * config/sh/ushmedia.h: Likewise.
28039 * config/sh/linux.h: Likewise.
28040 * config/sh/sh.c: Likewise.
28041 * config/sh/superh.h: Likewise.
28042 * config/sh/elf.h: Likewise.
28043 * config/sh/sh4.md: Likewise.
28044 * config/sh/sh.h: Likewise.
28046 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28048 * config/pa/constraints.md: Adjust unused letters. Change "T"
28049 constraint to match_test floating_point_store_memory_operand().
28050 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
28051 (base14_operand): New.
28052 (floating_point_store_memory_operand): New.
28053 (integer_store_memory_operand): Revise to use base14_operand and
28054 reg_plus_base_memory_operand.
28055 (move_dest_operand): Allow symbolic_memory_operands.
28056 (symbolic_memory_operand): Check for LO_SOM.
28057 (symbolic_operand): Change default case to break.
28058 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
28059 CONST_DOUBLE values to be reloaded by putting them into memory when
28060 the destination is a floating point register.
28061 (movdf): Remove code to handle CONST_DOUBLE.
28063 (reload_indf_r1): New.
28064 (reload_insf_r1): New.
28065 Consistently use "Q" and "T" constraints with integer and floating
28066 point move instructions, respectively.
28067 (movdi): Remove FAIL.
28068 Change predicate for source operand unamed DImode move from
28069 general_operand to move_src_operand.
28070 (umulsidi3): Change predicate for destination operand to
28072 Likewise for similar unamed patterns.
28073 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
28074 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
28075 (hppa_legitimize_address): Simplify mask calculation.
28076 (pa_emit_move_sequence): Revised handling of secondary reloads from
28077 REG+D addresses for floating point loads and stores. Directly handle
28078 loading CONST0_RTX (mode) to a floating point register.
28079 (pa_secondary_reload): Handle reloading DF and SFmode constant values
28080 to floating point registers. Don't restrict secondary reloads to
28081 floating point registers to integer modes. Revise some comments and
28083 (TARGET_LEGITIMATE_ADDRESS_P): Define.
28084 (pa_legitimate_address_p): New.
28085 (pa_legitimize_reload_address): New.
28086 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
28087 (STRICT_REG_OK_FOR_BASE_P): New.
28088 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
28089 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
28091 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
28092 Andrew Dixie <andrewd@gentrack.com>
28094 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
28097 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
28099 * expmed.c (extract_bit_field_1): Pass the full width of the
28100 structure to get_best_reg_extraction_insn.
28102 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28105 * configure.ac (use_cxa_atexit): Add AIX.
28106 * configure: Regenerate.
28108 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
28110 2013-02-01 Jakub Jelinek <jakub@redhat.com>
28113 * final.c (need_profile_function): New variable.
28114 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
28115 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
28116 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
28117 notes, targetm.asm_out.function_prologue doesn't emit anything,
28118 HAVE_prologue and profiler should be emitted before prologue,
28119 set need_profile_function instead of emitting it.
28120 (final_scan_insn): If need_profile_function, emit
28121 profile_function on the first NOTE_INSN_BASIC_BLOCK or
28122 NOTE_INSN_FUNCTION_BEG note.
28124 2013-02-01 Richard Henderson <rth@redhat.com>
28126 * config/rs6000/rs6000.md (smulditi3): New.
28129 * config/alpha/alpha.md (umulditi3): New.
28131 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28133 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
28134 (ASM_OUTPUT_ALIGNED_LOCAL): New.
28136 2013-02-01 Richard Biener <rguenther@suse.de>
28138 PR tree-optimization/56113
28139 * tree-ssa-structalias.c (label_visit): Reduce work for
28140 single-predecessor nodes.
28142 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
28144 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
28145 range isn't testing for zero.
28147 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
28149 PR middle-end/56113
28150 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
28152 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
28153 Nick Clifton <nickc@redhat.com>
28155 * config/v850/constraints.md (Q): Define as a memory constraint.
28156 * config/v850/predicates.md (label_ref_operand): New predicate.
28157 (e3v5_shift_operand): New predicate.
28158 (ior_operator): New predicate.
28159 * config/v850/t-v850: Add e3v5 multilib.
28160 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
28161 (v850_gen_movdi): Prototype.
28162 * config/v850/v850.c: Add support for e3v5 architecture.
28163 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
28165 (construct_save_jarl): Add e3v5 long JARL support.
28166 (v850_adjust_insn_length): New function. Adjust length of call
28167 insns when using e3v5 instructions.
28168 (v850_gen_movdi): New function: Generate instructions to move a
28170 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
28171 (CPP_SPEC): Define __v850e3v5__ as appropriate.
28172 (TARGET_USE_FPU): Enable for e3v5.
28173 (CONST_OK_FOR_W): New macro.
28174 (ADJUST_INSN_LENGTH): Define.
28175 * config/v850/v850.md (UNSPEC_LOOP): Define.
28176 (attr cpu): Add v850e3v5.
28177 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
28178 (movdi): New pattern.
28179 (movdi_internal): New pattern.
28180 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
28181 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
28182 (cstoresf4): Likewise.
28183 (cstoredf4): Likewise.
28184 (insv): New pattern.
28185 (rotlso3_a): New pattern.
28186 (rotlsi3_b): New pattern
28187 (rotlsi3_v850e3v5): New pattern.
28188 (doloop_begin): New pattern.
28189 (fix_loop_counter): New pattern.
28190 (doloop_end): New pattern.
28191 (branch_normal): Add e3v5 long branch support.
28192 (branch_invert): Likewise.
28193 (branch_z_normal): Likewise.
28194 (branch_z_invert): Likewise.
28195 (branch_nz_normal): Likewise.
28196 (branch_nz_invert): Likewise.
28197 (call_internal_short): Add e3v5 register-indirect JARL support.
28198 (call_internal_long): Likewise.
28199 (call_value_internal_short): Likewise.
28200 (call_value_internal_long): Likewise.
28201 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
28202 (mloop): New option.
28203 * config.gcc: Add support for configuring v840e3v5 target.
28204 * doc/invoke.texi: Document new v850 specific command line options.
28206 2013-01-31 Paul Koning <ni1d@arrl.net>
28210 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
28211 children if parent is a class.
28212 (prune_unused_types_prune): Don't add DW_AT_declaration.
28214 2013-01-31 Richard Biener <rguenther@suse.de>
28216 PR tree-optimization/56157
28217 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
28218 match up operand with SLP child.
28220 2013-01-31 Jason Merrill <jason@redhat.com>
28223 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
28224 parameters the first time.
28225 (gen_scheduled_generic_parms_dies): Check completeness here.
28227 2013-01-31 Richard Biener <rguenther@suse.de>
28229 PR middle-end/53073
28230 * common.opt (faggressive-loop-optimizations): New flag,
28231 enabled by default.
28232 * doc/invoke.texi (faggressive-loop-optimizations): Document.
28233 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
28234 infer_loop_bounds_from_undefined by it.
28236 2013-01-31 Richard Biener <rguenther@suse.de>
28238 PR tree-optimization/56150
28239 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
28240 visit virtual operands.
28241 (find_uses_to_rename_bb): Likewise.
28243 2013-01-31 Richard Biener <rguenther@suse.de>
28245 PR tree-optimization/56150
28246 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
28247 mixed store non-store stmts.
28249 2013-01-30 Jakub Jelinek <jakub@redhat.com>
28252 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
28253 LIBASAN_EARLY_SPEC is defined.
28254 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
28255 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
28257 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
28260 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
28261 invalid args instead of ICEing on it.
28262 (ix86_valid_target_attribute_tree): Return error_mark_node if
28263 ix86_valid_target_attribute_inner_p failed.
28264 (ix86_valid_target_attribute_p): Return false only if
28265 ix86_valid_target_attribute_tree returned error_mark_node. Allow
28266 target("default") attribute.
28267 (sorted_attr_string): Change argument from const char * to tree,
28268 merge in all target attribute arguments rather than just one.
28269 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
28270 instead of free. Avoid using strcat.
28271 (ix86_mangle_function_version_assembler_name): Mangle
28272 target("default") as if no target attribute is present. Adjust
28273 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
28274 instead of xmalloc and XDELETEVEC instead of free.
28275 (ix86_function_versions): Don't return true if one of the decls
28276 doesn't have target attribute. If they don't and one of the decls
28277 is DECL_FUNCTION_VERSIONED, report an error. Adjust
28278 sorted_attr_string caller. Use XDELETEVEC instead of free.
28279 (ix86_supports_function_versions): Remove.
28280 (make_name): Fix up formatting.
28281 (make_dispatcher_decl): Remove resolver_name and its initialization.
28282 Avoid leaking memory.
28283 (is_function_default_version): Return true if there is
28284 target("default") attribute rather than no target attribute at all.
28285 (make_resolver_func): Avoid leaking memory.
28286 (ix86_generate_version_dispatcher_body): Likewise.
28287 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
28288 * target.def (supports_function_versions): Remove.
28289 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
28290 * doc/tm.texi: Regenerated.
28292 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
28294 PR rtl-optimization/56144
28295 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
28296 for values with side effects.
28298 2013-01-30 Richard Biener <rguenther@suse.de>
28300 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
28301 (sparseset_pop): Likewise.
28302 * cfganal.c (compute_idf): Likewise. Increase work-stack size
28303 to be able to use quick_push in the worker loop.
28305 2013-01-30 Marek Polacek <polacek@redhat.com>
28307 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
28309 2013-01-30 Richard Biener <rguenther@suse.de>
28312 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
28314 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
28316 PR tree-optimization/56064
28317 * fixed-value.c (fixed_from_double_int): New function.
28318 * fixed-value.h (fixed_from_double_int): New prototype.
28319 (const_fixed_from_double_int): New static inline function.
28320 * fold-const.c (native_interpret_fixed): New static function.
28321 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
28322 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
28323 (native_encode_fixed): New static function.
28324 (native_encode_expr) <FIXED_CST>: Use it.
28325 (native_interpret_int): Move double_int worker code to...
28326 * double-int.c (double_int::from_buffer): ...this new static method.
28327 * double-int.h (double_int::from_buffer): Prototype it.
28329 2013-01-30 Richard Biener <rguenther@suse.de>
28331 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
28332 New pointer-map and obstack.
28333 (init_alias_vars): Allocate pointer-map and obstack.
28334 (delete_points_to_sets): Free them.
28335 (find_what_var_points_to): Cache result.
28336 (find_what_p_points_to): Adjust for changed interface of
28337 find_what_var_points_to.
28338 (compute_points_to_sets): Likewise.
28339 (ipa_pta_execute): Likewise.
28341 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
28343 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
28344 * configure: Regenerate.
28345 * config.in: Regenerate.
28346 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
28347 #nobits/#progbits if supported.
28349 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
28352 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
28353 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
28354 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
28356 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28358 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
28359 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
28361 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28363 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
28364 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
28366 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28368 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
28370 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
28371 * config/arm/cortex-a7.md: New bypasses using
28372 arm_mac_accumulator_is_result.
28374 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28376 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
28377 (cortex_a7_neon_mla): Likewise.
28378 (cortex_a7_fpfmad): New reservation.
28379 (cortex_a7_fpmacs): Use ffmas and update required units.
28380 (cortex_a7_fpmuld): Update required units and latency.
28381 (cortex_a7_fpmacd): Likewise.
28382 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
28383 (cortex_a7_neon). Likewise.
28384 (bypass) Update participating units.
28386 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28388 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
28389 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
28391 * config/arm/vfp11.md (vfp_farith): Use ffmas.
28392 (vfp_fmul): Use ffmad.
28393 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
28394 (cortex_r4_fmacd): Use ffmad.
28395 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
28396 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
28397 (cortex_a9_fmacd): Use ffmad.
28398 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
28399 (cortex_a8_vfp_macd): Use ffmad.
28400 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
28401 (cortex_a5_fpmacd): Use ffmad.
28402 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
28403 (cortex_a15_vfp_macd): Use ffmad.
28404 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
28406 2013-01-29 Jason Merrill <jason@redhat.com>
28409 * varasm.c (default_assemble_visibility): Don't warn about
28410 visibility on artificial decls.
28412 2013-01-29 Richard Biener <rguenther@suse.de>
28414 PR tree-optimization/56113
28415 * tree-ssa-structalias.c (equiv_class_lookup): Also return
28417 (label_visit): Free duplicate bitmaps and record the leader instead.
28418 (perform_var_substitution): Adjust.
28420 2013-01-29 Richard Biener <rguenther@suse.de>
28422 PR tree-optimization/55270
28423 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
28424 the CFG, schedule loops for fixup.
28426 2013-01-29 Nick Clifton <nickc@redhat.com>
28428 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
28431 2013-01-28 Leif Ekblad <leif@rdos.net>
28433 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
28434 * config/i386/i386.h (TARGET_RDOS): New macro.
28435 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
28436 * config/i386/i386.c (ix86_option_override_internal): For 64bit
28437 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
28438 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
28439 DEFAULT_LARGE_SECTION_THRESHOLD.
28440 * config/i386/i386.md (R14_REG, R15_REG): New constants.
28441 * config/i386/rdos.h: New file.
28442 * config/i386/rdos64.h: New file.
28444 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
28447 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
28450 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28452 PR rtl-optimization/56117
28453 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
28454 call cselib_lookup_from_insn on the MEM before calling
28455 add_insn_mem_dependence.
28457 2013-01-28 Richard Biener <rguenther@suse.de>
28459 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
28460 to a stmt that didn't have one.
28461 (copy_phis_for_bb): Likewise for PHI arguments.
28462 (copy_debug_stmt): Likewise for debug stmts.
28464 2013-01-28 Richard Biener <rguenther@suse.de>
28466 PR tree-optimization/56034
28467 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
28468 (partition_builtin_p): Adjust.
28469 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
28470 it is the last partition.
28471 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
28472 up the vertex for the definition.
28473 (classify_partition): Classify whether a partition is a
28474 PKIND_REDUCTION, thus has uses outside of the loop.
28475 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
28476 Merge all PKIND_REDUCTION partitions into the last partition.
28477 (tree_loop_distribution): Seed partitions from reductions as well.
28479 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28481 PR tree-optimization/56125
28482 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
28483 pow(x,c) into sqrt(x) * powi(x, n/2) or
28484 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
28485 optimizing for size.
28486 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
28487 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
28490 PR tree-optimization/56094
28491 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
28492 to UNKNOWN_LOCATION while gimplifying expr.
28494 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
28497 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
28498 operand 0 in movabs insn template for -masm=intel asm alternative.
28499 (*movabs<mode>_2): Ditto for operand 1.
28501 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
28504 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
28507 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28509 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
28512 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
28514 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
28515 Undef to avoid warning.
28517 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28519 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
28520 * configure: Regenerate.
28522 2013-01-25 Jakub Jelinek <jakub@redhat.com>
28524 PR tree-optimization/56098
28525 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
28526 for stmts with volatile ops.
28527 (cond_store_replacement): Don't optimize if assign has volatile ops.
28528 (cond_if_else_store_replacement_1): Don't optimize if either
28529 then_assign or else_assign have volatile ops.
28530 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
28533 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28535 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
28537 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28539 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
28540 missing ':' in asm example.
28542 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
28544 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
28545 entries into lane and laneq entries.
28546 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
28547 Remove AdvSIMD scalar modes.
28548 (aarch64_sq<r>dmulh_laneq<mode>): New.
28549 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
28551 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
28552 builtin implementations to relfect changes in RTL in aarch64-simd.md.
28553 * config/aarch64/iterators.md (VCOND): New.
28556 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28559 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
28560 Add NULL LIBNAME argument to existing definitions.
28561 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
28562 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
28563 * config/avr/avr.c (DEF_BUILTIN): Same.
28564 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
28565 (avr_expand_builtin): Expand to a vanilla call if a libgcc
28566 implementation is available (DECL_ASSEMBLER_NAME is set).
28567 (avr_fold_absfx): New static function.
28568 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
28569 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
28570 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
28571 AVR_BUILTIN_ABSLLK.
28572 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
28573 (abshk, absk, abslk, absllk): Provide as static inline functions.
28575 2013-01-25 Marek Polacek <polacek@redhat.com>
28577 PR tree-optimization/56035
28578 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
28580 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
28582 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
28583 (*movtf_internal_rex64): Add (!o,C) alternative
28584 (*movxf_internal_rex64): Ditto.
28585 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
28587 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
28589 * doc/invoke.texi: fix typo.
28590 * doc/objc.texi: fix typo.
28592 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
28594 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
28595 for the first two alternatives.
28597 2013-01-24 Diego Novillo <dnovillo@google.com>
28599 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
28600 (ggc-zone.o): Remove.
28601 * configure.ac: Remove option --with-gc.
28602 * configure: Re-generate.
28603 * doc/install.texi: Remove documentation for --with-gc.
28604 * gengtype.c (write_enum_defn): Remove. Update all users.
28605 (write_Types_process_field): Remove generation of gt_e_* argument.
28606 (output_type_enum): Remove. Update all users.
28607 (write_enum_defn): Remove. Update all users.
28608 (enum alloc_zone): Remove. Update all users.
28609 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
28610 * ggc-common.c (ggc_splay_alloc): Remove first argument.
28611 Update all callers.
28612 (struct ptr_data): Remove field TYPE. Update all users.
28613 (gt_pch_note_object): Remove argument TYPE. Update all users.
28614 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
28616 * ggc-none.c (ggc_alloc_typed_stat): Remove.
28617 (struct alloc_zone): Remove.
28618 (ggc_internal_alloc_zone_stat): Remove.
28619 (ggc_internal_cleared_alloc_zone_stat): Remove.
28620 * ggc-page.c (ggc_alloc_typed_stat): Remove.
28621 (ggc_pch_count_object): Remove last argument. Update all users.
28622 (ggc_pch_alloc_object): Remove last argument. Update all users.
28623 (struct alloc_zone): Remove.
28624 * ggc-zone.c: Remove.
28625 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
28626 (struct alloc_zone): Remove.
28627 (ggc_alloc_typed_stat): Remove.
28628 (ggc_alloc_typed): Remove.
28629 (ggc_splay_alloc): Remove first argument.
28630 (rtl_zone): Remove. Update all users.
28631 (tree_zone): Remove. Update all users.
28632 (tree_id_zone): Remove. Update all users.
28633 (ggc_internal_zone_alloc_stat): Remove. Update all users.
28634 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
28635 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
28636 * tree-ssanames.c: Remove references to zone allocator in comments.
28638 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
28640 * config/avr/avr.c (avr_out_fract): Make register numbers that
28641 might be outside of source operand signed.
28643 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
28645 * config/i386/constraints.md (Yf): New constraint.
28646 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
28647 of f constraint to conditionaly disable x87 register preferences.
28648 (*movdf_internal): Ditto.
28649 (*movsf_internal): Ditto.
28651 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
28653 PR inline-asm/55934
28654 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
28655 that have operands with impossible constraints.
28656 Add a FIXME for a speed-up opportunity.
28657 * lra-constraints.c (process_alt_operands): Verify that a class
28658 selected from constraints on asms is valid for the operand mode.
28659 (curr_insn_transform): Remove incorrect comment.
28661 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
28663 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
28664 TOC operand is a valid symbol ref in the constant pool.
28666 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28668 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
28670 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
28673 * config/avr/stdfix.h: New file.
28674 * t-avr (stdfix-gcc.h): New rule to build it.
28675 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
28677 2013-01-23 Kostya Serebryany <kcc@google.com>
28679 * config/darwin.h: remove dependency on
28680 CoreFoundation (asan on Mac OS).
28682 2013-01-23 Jakub Jelinek <jakub@redhat.com>
28685 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
28686 instead of cmpdi_operand for first comparison operand.
28687 Don't assert that comparison operands aren't both constants.
28689 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28691 * doc/install.texi (Downloading the Source): Update references to
28692 downloading separate components.
28694 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28696 * doc/extend.texi (__int128): Improve grammar.
28698 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
28701 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
28702 alternative to (o,r).
28703 (*movdi_internal_rex64): Remove (!o,n) alternative.
28704 (DImode immediate->memory splitter): Remove.
28705 (DImode immediate->memory peephole2): Remove.
28706 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
28707 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
28708 alternative to (!o,*r).
28709 (*movtf_internal_sse): New pattern.
28710 (*movxf_internal_rex64): New pattern.
28711 (*movxf_internal): Disable for TARGET_64BIT.
28712 (*movdf_internal_rex64): Remove (!o,F) alternative.
28714 2013-01-22 Jakub Jelinek <jakub@redhat.com>
28716 PR middle-end/56074
28717 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
28718 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
28719 * tree-vect-loop-manip.c (find_loop_location): Also ignore
28720 stmt locations where LOCATION_LOCUS of the stmt location is
28721 UNKNOWN_LOCATION or BUILTINS_LOCATION.
28724 * config/i386/i386.md (UNSPEC_STOS): New.
28725 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
28726 *strsetqi_1): Add UNSPEC_STOS.
28728 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
28731 * doc/invoke.texi: Remove left over -Wsynth example.
28733 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28735 PR tree-optimization/56051
28736 * fold-const.c (fold_binary_loc): Don't fold
28737 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
28738 a narrowing conversion, or widening conversion from signed
28741 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28743 PR rtl-optimization/56023
28744 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
28745 dependent on debug instruction.
28747 2013-01-21 Martin Jambor <mjambor@suse.cz>
28749 PR middle-end/56022
28750 * function.c (allocate_struct_function): Call
28751 invoke_set_current_function_hook earlier.
28753 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28755 * reload1.c (init_reload): Only initialize reload_obstack
28756 during the first call.
28758 2013-01-21 Marek Polacek <polacek@redhat.com>
28760 * cfgloop.c (verify_loop_structure): Fix up grammar.
28762 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
28764 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
28765 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
28767 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28770 * config/arm/marvell-pj4.md: Update copyright year.
28771 Fix up use of alu to alu_reg and simple_alu_imm.
28773 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28775 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
28777 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
28780 * lra-constraints.c (curr_insn_transform): Don't reuse original
28781 insn for secondary memory move when memory mode should be different.
28783 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28785 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
28786 atomic_storedi_1): New patterns.
28788 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
28790 btver2 pipeline descriptions.
28791 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
28793 * config/i386/i386.md (btver2_decode): New type attributes.
28794 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
28796 * config/i386/btver2.md: New file describing btver2 pipelines.
28798 2013-01-19 Andrew Pinski <apinski@cavium.com>
28800 PR tree-optimization/52631
28801 * tree-ssa-sccvn (visit_use): Before looking up the original
28802 statement, try looking up the simplified expression.
28804 2013-01-19 Anthony Green <green@moxielogic.com>
28806 * config/moxie/moxie.c (moxie_expand_prologue): Set
28807 current_function_static_stack_size.
28809 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28811 PR tree-optimization/56029
28812 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
28813 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
28815 2013-01-18 Sharad Singhai <singhai@google.com>
28817 PR tree-optimization/55995
28818 * dumpfile.c (dump_loc): Print location only if available.
28819 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
28821 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
28824 * lra-constraints.c (curr_insn_transform): Reuse original insn for
28825 secondary memory move.
28826 (inherit_reload_reg): Use rclass instead of cl for
28827 check_secondary_memory_needed_p.
28829 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28831 PR middle-end/56015
28832 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
28833 the case where writing real complex part of target modifies op1.
28835 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
28837 * config/aarch64/aarch64-simd.md
28838 (aarch64_vcond_internal<mode>): Handle unordered cases.
28839 * config/aarch64/iterators.md (v_cmp_result): New.
28841 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
28842 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28844 * config/arm/marvell-pj4.md: New file.
28845 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
28846 * config/arm/arm.md (generic_sched): Add marvell_pj4.
28847 (generic_vfp): Likewise.
28848 * config/arm/arm-cores.def: Add marvell-pj4.
28849 * config/arm/arm-tune.md: Regenerate.
28850 * config/arm/arm-tables.opt: Regenerate.
28851 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
28852 * doc/invoke.texi: Document marvell-pj4.
28854 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
28856 * config/aarch64/arm_neon.h: Map scalar types to standard types.
28858 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28863 * var-tracking.c (negative_power_of_two_p): New.
28864 (global_get_addr_cache, local_get_addr_cache): New.
28865 (get_addr_from_global_cache, get_addr_from_local_cache): New.
28866 (vt_canonicalize_addr): Rewrite using the above. Adjust the
28868 (vt_stack_offset_p): Remove.
28869 (vt_canon_true_dep): Always canonicalize loc's address.
28870 (clobber_overlapping_mems): Make sure we have a MEM.
28871 (local_get_addr_clear_given_value): New.
28872 (val_reset): Clear local cached entries.
28873 (compute_bb_dataflow): Create and release the local cache.
28874 Disable duplicate MEMs clobbering.
28875 (emit_notes_in_bb): Clobber MEMs likewise.
28876 (vt_emit_notes): Create and release the local cache.
28877 (vt_initialize, vt_finalize): Create and release the global
28878 cache, respectively.
28879 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
28881 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28883 PR libmudflap/53359
28884 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
28885 not found in the symtab.
28887 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28890 PR rtl-optimization/55547
28891 PR rtl-optimization/53827
28894 * alias.c (offset_overlap_p): New, factored out of...
28895 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
28896 the conservative special case for symbolic constants. Don't
28897 adjust zero sizes on alignment.
28899 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
28901 PR rtl-optimization/52573
28902 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
28903 REG_UNUSED for the same register.
28905 2013-01-17 Richard Biener <rguenther@suse.de>
28906 Marek Polacek <polacek@redhat.com>
28908 PR rtl-optimization/55833
28909 * loop-unswitch.c (unswitch_loops): Move loop verification...
28910 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
28911 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
28912 Set it to true when we're removing a loop from hierarchy tree in
28913 an irreducible region.
28914 (fix_bb_placements): Adjust caller.
28915 (fix_loop_placements): Likewise.
28917 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
28919 * config/avr/builtins.def (DEF_BUILTIN): Factor out
28920 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
28921 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
28922 Remove ID. Adjust comments.
28923 * config/avr/avr-c.c (avr_builtin_name): Remove.
28924 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
28925 * config/avr/avr.c (avr_tolower): New static function.
28926 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
28927 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
28928 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
28931 2013-01-17 Jan Hubicka <jh@suse.cz>
28933 PR tree-optimization/55273
28934 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
28936 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
28939 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
28940 store through atomic_store<mode>_1.
28941 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
28943 2013-01-17 Martin Jambor <mjambor@suse.cz>
28945 PR tree-optimizations/55264
28946 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
28947 for virtual methods.
28948 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
28949 virtual methods before inlining is over.
28950 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
28952 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
28955 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
28957 PR rtl-optimization/56005
28958 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
28959 pending reads for prefetch.
28961 2013-01-16 Ian Bolton <ian.bolton@arm.com>
28963 * config/aarch64/aarch64.md
28964 (*cstoresi_neg_uxtw): New pattern.
28965 (*cmovsi_insn_uxtw): New pattern.
28966 (*<optab>si3_uxtw): New pattern.
28967 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
28968 (*<optab>si3_insn_uxtw): New pattern.
28969 (*bswapsi2_uxtw): New pattern.
28971 2013-01-16 Richard Biener <rguenther@suse.de>
28973 * tree-inline.c (tree_function_versioning): Remove set but
28974 never used variable.
28976 2013-01-16 Richard Biener <rguenther@suse.de>
28978 PR tree-optimization/55964
28979 * tree-flow.h (rename_variables_in_loop): Remove.
28980 (rename_variables_in_bb): Likewise.
28981 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
28982 (copy_loop_before): Adjust and delete update-ssa status.
28983 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
28984 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
28985 (rename_variables_in_loop): Remove.
28986 (slpeel_update_phis_for_duplicate_loop): Likewise.
28987 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
28988 use available cfg machinery instead of duplicating it.
28989 Update PHI nodes and perform poor-mans SSA update here.
28990 (slpeel_tree_peel_loop_to_edge): Adjust.
28992 2013-01-16 Richard Biener <rguenther@suse.de>
28994 PR tree-optimization/54767
28995 PR tree-optimization/53465
28996 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
28997 (vrp_visit_phi_node): For PHI arguments coming via backedges
28998 drop all symbolical range information.
28999 (execute_vrp): Compute backedges.
29001 2013-01-16 Richard Biener <rguenther@suse.de>
29003 * doc/install.texi: Update CLooG and ISL requirements to
29006 2013-01-16 Christian Bruel <christian.bruel@st.com>
29009 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
29010 (broken_move): Handle UNSPECV_SP_SWITCH_B.
29011 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
29013 2013-01-16 DJ Delorie <dj@redhat.com>
29015 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
29016 (UNSPECV_SP_SWITCH_E): New.
29017 (sp_switch_1): Change to an unspec.
29018 (sp_switch_2): Change to an unspec. Don't use post-inc when we
29021 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
29023 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
29024 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
29025 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
29026 (expand_mem_thread_fence): Ditto.
29027 (expand_mem_signal_fence): Ditto.
29028 (expand_atomic_load): Ditto.
29029 (expand_atomic_store): Ditto.
29031 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
29033 PR rtl-optimization/55547
29034 PR rtl-optimization/53827
29037 * alias.c (memrefs_conflict_p): Set sizes to negative after
29040 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29043 * function.c (thread_prologue_and_epilogue_insns): Always
29044 add crtl->drap_reg to set_up_by_prologue.set, even if
29045 stack_realign_drap is false.
29047 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
29049 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
29050 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
29051 *call): Fix indention.
29053 2013-01-15 Tom de Vries <tom@codesourcery.com>
29056 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
29059 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
29061 PR rtl-optimization/55153
29062 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
29064 2013-01-15 Martin Jambor <mjambor@suse.cz>
29066 PR tree-optimization/55920
29067 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
29068 accesses as grp_to_be_debug_replaced.
29070 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29072 PR tree-optimization/55920
29073 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
29074 there is non-useless type conversion needed from debug rhs to lhs,
29075 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
29077 2013-01-15 Joseph Myers <joseph@codesourcery.com>
29078 Mikael Pettersson <mikpe@it.uu.se>
29081 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
29083 (ASM_OUTPUT_CASE_LABEL): Remove.
29084 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
29085 * final.c (shorten_branches): Update alignment of labels before
29086 jump tables if CASE_VECTOR_SHORTEN_MODE.
29088 2013-01-15 Richard Biener <rguenther@suse.de>
29091 * system.h: Do not include gmp.h for building host tools.
29093 2013-01-15 Richard Biener <rguenther@suse.de>
29095 PR middle-end/55882
29096 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
29097 account for bitpos when computing alignment.
29099 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
29101 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
29102 (ix86_target_macros_internal): Likewise.
29104 * config/i386/i386.c (m_CORE2I7): Removed.
29105 (m_CORE_HASWELL): New macro.
29106 (m_CORE_ALL): Likewise.
29107 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
29108 (initial_ix86_arch_features): Likewise.
29109 (processor_target_table): Initializations for Core avx2.
29110 (cpu_names): New names "core-avx2".
29111 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
29112 PROCESSOR_CORE_HASWELL.
29113 (ix86_issue_rate): New case.
29114 (ia32_multipass_dfa_lookahead): Likewise.
29115 (ix86_sched_init_global): Likewise.
29117 * config/i386/i386.h (TARGET_HASWELL): New macro.
29118 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
29119 (processor_type): New PROCESSOR_HASWELL.
29121 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29123 PR tree-optimization/55955
29124 * tree-vect-loop.c (vectorizable_reduction): Give up early on
29125 *SHIFT_EXPR and *ROTATE_EXPR codes.
29127 PR tree-optimization/48766
29128 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
29129 -ftrapv disable -fwrapv.
29131 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29134 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
29135 etc. to 1 and not to __flash.
29136 Use LL suffix for __INT24_MAX__ with -mint8.
29137 Use ULL suffix for __UINT24_MAX__ with -mint8.
29139 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29141 * config/avr/avr-arch.h
29142 (struct base_arch_s): Use typedef avr_arch_t instead.
29143 (struct arch_info_s): Use typedef avr_arch_info_t instead.
29144 (struct mcu_type_s): Use typedef avr_mcu_t instead.
29145 * config/avr/avr.c: Same.
29146 * config/avr/avr-devices.c: Same.
29147 * config/avr/driver-avr.c: Same.
29148 * config/avr/gen-avr-mmcu-texi.c: Same.
29149 * config/avr/avr-mcus.def: Adjust comment.
29151 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
29153 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
29154 * config/aarch64/iterators.md (VALLDI): New.
29156 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
29157 Andi Kleen <ak@linux.intel.com>
29160 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
29161 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
29164 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29166 * config/avr/avr-stdint.h: Remove trailing blanks.
29167 * config/avr/avr-log.h: Same.
29168 * config/avr/avr-arch.h: Same.
29169 * config/avr/avr-devices.c: Same.
29170 * config/avr/avr-dimode.md: Same.
29171 * config/avr/predicates.md: Same.
29172 * config/avr/avr-c.c: Same. And fix typo.
29174 * config/avr/avr-protos.h: Same. And:
29175 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
29176 (init_cumulative_args): Rename to avr_init_cumulative_args.
29177 (expand_prologue): Rename to avr_expand_prologue.
29178 (expand_epilogue): Rename to avr_expand_epilogue.
29179 (adjust_insn_length): Rename to avr_adjust_insn_length.
29180 (notice_update_cc): Rename to avr_notice_update_cc.
29181 (final_prescan_insn): Rename to avr_final_prescan_insn.
29182 * config/avr/avr.c: Same.
29183 * config/avr/avr.h: Same.
29184 * config/avr/avr.md: Remove trailing blanks.
29185 (prologue): Use avr_expand_prologue.
29186 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
29188 2013-01-14 Richard Biener <rguenther@suse.de>
29190 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
29191 verify_location, collect_subblocks): New functions.
29192 (verify_gimple_in_cfg): Verify that locations only reference
29193 BLOCKs in the functions BLOCK tree.
29195 2013-01-14 Richard Biener <rguenther@suse.de>
29197 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
29199 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
29201 (insert_out_of_ssa_copy_on_edge): Likewise.
29202 (rewrite_close_phi_out_of_ssa): Likewise.
29203 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
29205 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
29206 propagated constants.
29207 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
29210 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29212 * config/avr/avr-modes.def: Add GPL copyright notice.
29214 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
29216 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
29217 MEMMODEL_MASK to determine memory model.
29218 (atomic_store<mode>): Ditto from operands[2].
29219 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
29221 2013-01-13 Jakub Jelinek <jakub@redhat.com>
29224 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
29225 (fold_gimple_assign): Don't call unshare_expr here.
29226 (fold_ctor_reference): Call unshare_expr.
29228 2013-01-13 Terry Guo <terry.guo@arm.com>
29230 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
29231 * doc/fragments.texi: Document MULTILIB_REUSE.
29232 * gcc.c (multilib_reuse): New internal spec.
29233 (set_multilib_dir): Also search multilib from multilib_reuse.
29234 * genmultilib (tmpmultilib3): Refactor code.
29235 (tmpmultilib4): Ditto.
29236 (multilib_reuse): New multilib argument.
29238 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
29240 * Makefile.in: Update copyright.
29242 2013-01-12 Tom de Vries <tom@codesourcery.com>
29244 PR middle-end/55890
29245 * calls.c (expand_call): Check if arg_nr is valid.
29247 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
29249 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
29250 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
29251 documentation. Add missing '__' in front of
29252 __builtin_ia32_packssdw256.
29254 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29257 * config/s390/s390.c (s390_preferred_reload_class): Do not return
29258 NO_REGS for larl operands.
29259 (s390_reload_larl_operand): Use s390_load_address instead of
29262 2013-01-11 Richard Biener <rguenther@suse.de>
29264 * tree-cfg.c (verify_node_sharing_1): Split out from ...
29265 (verify_node_sharing): ... here.
29266 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
29268 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
29270 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
29271 Substitute TREECHECKING.
29272 * configure: Regenerate.
29273 * Makefile.in (TREECHECKING): New.
29275 2013-01-11 Richard Guenther <rguenther@suse.de>
29277 PR tree-optimization/44061
29278 * tree-vrp.c (extract_range_basic): Compute zero as
29279 value-range for __builtin_constant_p of function parameters.
29281 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
29283 Update copyright years.
29285 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
29287 PR rtl-optimization/55672
29288 * lra-eliminations.c (mark_not_eliminable): Permit addition with
29289 const to be eliminable.
29291 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
29293 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
29294 * configure: Regenerate.
29296 2013-01-10 Richard Biener <rguenther@suse.de>
29298 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
29300 2013-01-10 Richard Biener <rguenther@suse.de>
29303 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
29304 locations for virtual PHI arguments.
29305 (rewrite_update_phi_arguments): Likewise.
29307 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
29309 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
29312 2013-01-10 Jakub Jelinek <jakub@redhat.com>
29314 PR tree-optimization/55921
29315 * tree-complex.c (expand_complex_asm): New function.
29316 (expand_complex_operations_1): Call it for GIMPLE_ASM.
29318 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29321 * config/s390/s390.c (s390_symref_operand_p)
29322 (s390_loadrelative_operand_p): Merge the two functions.
29323 (s390_check_qrst_address, print_operand_address): Add parameters
29324 to s390_loadrelative_operand_p invokation.
29325 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
29326 (s390_reload_larl_operand, s390_secondary_reload): Use
29327 s390_loadrelative_operand_p instead of s390_symref_operand_p.
29328 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
29330 2013-01-09 Mike Stump <mikestump@comcast.net>
29332 * dse.c (record_store): Remove unnecessary assert.
29334 2013-01-09 Jan Hubicka <jh@suse.cz>
29336 PR tree-optimization/55569
29337 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
29338 * cfgloop.h (scale_loop_profile): Likewise.
29340 2013-01-09 Jan Hubicka <jh@suse.cz>
29343 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
29345 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
29347 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
29349 PR middle-end/55114
29350 * expr.h (maybe_emit_group_store): Declare.
29351 * expr.c (maybe_emit_group_store): New function.
29352 * builtins.c (expand_builtin_int_roundingfn): Call it.
29353 (expand_builtin_int_roundingfn_2): Likewise.
29355 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
29357 PR rtl-optimization/55829
29358 * lra-constraints.c (match_reload): Add code for absent output.
29359 (curr_insn_transform): Add code for reloads of matched inputs
29362 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
29364 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
29365 attribute of movddup insn to DF.
29366 (*vec_interleave_lowv2df): Ditto.
29367 (vec_dupv2df): Ditto.
29369 2013-01-09 Jan Hubicka <jh@suse.cz>
29371 PR tree-optimiation/55875
29372 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
29373 EVERY_ITERATION parameter.
29374 (number_of_iterations_exit): Check if exit is executed every iteration.
29375 (idx_infer_loop_bounds): Similarly here.
29376 (n_of_executions_at_most): Simplify
29377 to only test for cases where statement is dominated by the
29378 particular bound; handle correctly the "postdominance" test.
29379 (scev_probably_wraps_p): Use max loop iterations info
29380 as a global bound first.
29382 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
29383 Nick Clifton <nickc@redhat.com>
29385 * config/v850/v850.md (cbranchsf4): New pattern.
29386 (cstoresf4): New pattern.
29387 (cbranchdf4): New pattern.
29388 (cstoredf4): New pattern.
29389 (movsicc): Disallow floating point comparisons.
29390 (cmpsf_le_insn): Fix order of operators.
29391 (cmpsf_lt_insn): Likewise.
29392 (cmpsf_eq_insn): Likewise.
29393 (cmpdf_le_insn): Likewise.
29394 (cmpdf_lt_insn): Likewise.
29395 (cmpdf_eq_insn): Likewise.
29396 (cmpsf_ge_insn): Use LE comparison.
29397 (cmpdf_ge_insn): Likewise.
29398 (cmpsf_gt_insn): Use LT comparison.
29399 (cmpdf_gt_insn): Likewise.
29400 (cmpsf_ne_insn): Delete pattern.
29401 (cmpdf_ne_insn): Delete pattern.
29402 * config/v850/v850.c (v850_gen_float_compare): Use
29403 gen_cmpdf_eq_insn for NE comparison.
29404 (v850_float_z_comparison_operator)
29405 (v850_float_nz_comparison_operator): Move from here ...
29406 * config/v850/predicates.md: ... to here. Move GT and GE
29407 comparisons into v850_float_z_comparison_operator.
29408 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
29410 (v850_float_nz_comparison_operator): Likewise.
29412 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29414 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
29415 with calls to gen_insvsi/gen_insvdi.
29417 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29419 * config/i386/i386.c (initial_ix86_tune_features): Set up
29420 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
29422 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
29423 Jakub Jelinek <jakub@redhat.com>
29425 PR tree-optimization/48189
29426 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
29427 If nitercst is 0, don't predict the exit edge.
29429 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29431 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
29432 in asm_fprintf with reg_names.
29433 (aarch64_print_operand_address): Likewise.
29434 (aarch64_return_addr): Likewise.
29435 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
29437 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29439 * config/pa/pa.h (VAL_U6_BITS_P): Define.
29440 (INT_U6_BITS): Likewise.
29441 * config/pa/predicates.md (uint6_operand): New predicate.
29442 (shift5_operand, shift6_operand): Likewise.
29443 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
29445 (lshrdi3): Use shift6_operand.
29446 (shrpsi4, shrpdi4): New insn patterns.
29447 (extzv): Delete expander.
29448 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
29449 predicates in unamed zero extract patterns. Tighten common constraint.
29450 (extv): Delete expander.
29451 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
29452 predicates in unamed sign extract patterns. Tighten common constraint.
29453 (insv): Delete expander.
29454 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
29455 predicates in unamed insert patterns. Tighten common constraint.
29456 Change uint32_operand predicate to uint6_operand predicate in unamed
29457 DImode pattern to insert constant values of type 1...1xxxx.
29459 2013-01-04 Jan Hubicka <jh@suse.cz>
29461 PR tree-optimization/55823
29462 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
29465 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29466 Uros Bizjak <ubizjak@gmail.com>
29468 PR rtl-optimization/55845
29469 * df-problems.c (can_move_insns_across): Stop scanning at
29470 volatile_insn_p source instruction or give up if
29471 across_from .. across_to range contains any volatile_insn_p
29474 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29476 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
29477 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
29479 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
29480 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
29482 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29485 * asan.c (asan_clear_shadow): New function.
29486 (asan_emit_stack_protection): Use it.
29488 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29490 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
29491 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
29492 with tab instead of space.
29494 2013-01-08 Nick Clifton <nickc@redhat.com>
29496 * config/rl78/rl78.c (rl78_expand_prologue): Always select
29497 register bank 0 at the start of an interrupt handler.
29498 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
29501 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29503 * config/aarch64/aarch64-simd.md
29504 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
29505 (aarch64_simd_bsl): Likewise.
29506 (aarch64_vcond_internal<mode>): Likewise.
29507 (vcond<mode><mode>): Likewise.
29508 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
29509 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
29511 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29513 * config/aarch64/aarch64-builtins.c
29514 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
29516 2013-01-08 Martin Jambor <mjambor@suse.cz>
29519 * tree-sra.c (analyze_access_subtree): Return true also after
29520 potentially creating a debug-only replacement.
29522 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29524 PR middle-end/55890
29525 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
29527 PR tree-optimization/54120
29528 * tree-vrp.c (range_fits_type_p): Don't allow
29529 src_precision < precision from signed vr to unsigned_p
29530 if vr->min or vr->max is negative.
29531 (simplify_float_conversion_using_ranges): Test can_float_p
29532 against CODE_FOR_nothing.
29534 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29535 Richard Biener <rguenther@suse.de>
29537 PR middle-end/55851
29538 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
29539 types instead of just INTEGER_TYPE types.
29541 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
29543 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
29546 2013-01-07 Steve Ellcey <sellcey@mips.com>
29549 * config/mips/mips.opt: Change mad to mmad to match documentation.
29551 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29554 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
29555 .progmemx.data now.
29557 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29560 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
29561 (avr_addrspace_t): Add .section_name field.
29562 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
29564 (avr_addrspace): Same. Initialize .section_name. Remove last
29565 NULL entry. Put __memx into .progmemx.data.
29566 (progmem_section_prefix): Remove.
29567 (avr_asm_init_sections): No need to initialize progmem_section.
29568 (avr_asm_named_section): Use avr_addrspace[].section_name to get
29569 section name prefix.
29570 (avr_asm_select_section): Ditto. And use get_unnamed_section to
29571 retrieve the progmem section.
29572 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
29573 boundary to run over avr_addrspace[].
29574 (avr_register_target_pragmas): Ditto.
29576 2013-01-06 Jakub Jelinek <jakub@redhat.com>
29578 * varasm.c (output_constant_def_contents): For asan_protect_global
29579 protected strings, adjust DECL_ALIGN if needed, before testing for
29581 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
29582 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
29584 (output_object_block): For asan protected decls, emit asan padding
29585 after their contents.
29586 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
29587 (asan_finish_file): Test it here instead.
29589 2013-01-07 Nick Clifton <nickc@redhat.com>
29590 Matthias Klose <doko@debian.org>
29591 Doug Kwan <dougkwan@google.com>
29592 H.J. Lu <hongjiu.lu@intel.com>
29595 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
29597 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
29599 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
29601 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
29603 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
29605 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29608 * doc/install.texi (Cross-Compiler-Specific Options): Document
29611 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
29613 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
29614 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
29615 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
29616 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
29617 vqmovun_high_s64): Fix source operand number and update copyright.
29619 2013-01-07 Richard Biener <rguenther@suse.de>
29621 PR middle-end/55890
29622 * gimple.h (gimple_call_builtin_p): New overload.
29623 * gimple.c (validate_call): New function.
29624 (gimple_call_builtin_p): Likewise.
29625 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
29626 Use gimple_call_builtin_p.
29627 (find_func_clobbers): Likewise.
29628 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
29629 (strlen_optimize_stmt): Likewise.
29631 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29633 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
29634 (vld1q_dup_*): Likewise.
29635 (vld1_*): Likewise.
29636 (vld1q_*): Likewise.
29637 (vld1_lane_*): Likewise.
29638 (vld1q_lane_*): Likewise.
29640 2013-01-07 Richard Biener <rguenther@suse.de>
29642 * lto-streamer.h (LTO_minor_version): Bump to 2.
29644 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29646 * config/aarch64/aarch64-protos.h
29647 (aarch64_const_double_zero_rtx_p): Rename to...
29648 (aarch64_float_const_zero_rtx_p): ...this.
29649 (aarch64_float_const_representable_p): New.
29650 (aarch64_output_simd_mov_immediate): Likewise.
29651 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
29652 move immediate case.
29653 * config/aarch64/aarch64.c
29654 (aarch64_const_double_zero_rtx_p): Rename to...
29655 (aarch64_float_const_zero_rtx_p): ...this.
29656 (aarch64_print_operand): Allow printing of new constants.
29657 (aarch64_valid_floating_const): New.
29658 (aarch64_legitimate_constant_p): Check for valid floating-point
29660 (aarch64_simd_valid_immediate): Likewise.
29661 (aarch64_vect_float_const_representable_p): New.
29662 (aarch64_float_const_representable_p): Likewise.
29663 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
29664 (aarch64_output_simd_mov_immediate): New.
29665 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
29666 (*movdf_aarch64): Likewise.
29667 * config/aarch64/constraints.md (Ufc): New.
29668 (Y): call aarch64_float_const_zero_rtx.
29669 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
29671 2013-01-07 Richard Biener <rguenther@suse.de>
29673 PR tree-optimization/55888
29674 PR tree-optimization/55862
29675 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
29676 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
29677 not if it is contained therein.
29679 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29681 * config/avr/t-avr: Typo.
29683 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29686 * config/avr/t-avr: Don't automatically rebuild
29687 $(srcdir)/config/avr/t-multilib
29688 $(srcdir)/config/avr/avr-tables.opt
29689 $(srcdir)/doc/avr-mmcu.texi
29690 (avr-mcus): New phony target to build them on request.
29691 (s-avr-mlib, s-avr-mmcu-texi): Remove.
29692 * avr/avr-mcus.def: Adjust comments.
29694 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
29696 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
29698 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29700 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
29702 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29704 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
29706 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
29708 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
29709 to generate profiling.
29710 * config/rs6000/aix64.h (LIB_SPEC): Same.
29712 2013-01-04 Andrew Pinski <apinski@cavium.com>
29714 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
29716 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
29718 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
29720 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
29722 (ix86_expand_move): Ditto.
29723 (ix86_zero_extend_to_Pmode): Ditto.
29724 (ix86_expand_call): Ditto.
29725 (ix86_expand_special_args_builtin): Ditto.
29726 (ix86_expand_builtin): Ditto.
29728 2013-01-04 Richard Biener <rguenther@suse.de>
29730 PR tree-optimization/55862
29731 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
29732 translating them through PHI nodes.
29734 2013-01-04 Martin Jambor <mjambor@suse.cz>
29736 PR tree-optimization/55755
29737 * tree-sra.c (sra_modify_assign): Do not check that an access has no
29738 children when trying to avoid producing a VIEW_CONVERT_EXPR.
29740 2013-01-04 Marek Polacek <polacek@redhat.com>
29742 PR middle-end/55859
29743 * opts.c (default_options_optimization): Clarify error message.
29745 2013-01-04 Richard Biener <rguenther@suse.de>
29747 PR middle-end/55863
29748 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
29751 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29754 * config/pa/pa.md (movsi): Revert previous change.
29755 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
29758 2013-01-03 Richard Henderson <rth@redhat.com>
29760 * config/i386/i386.c (ix86_expand_move): Always assign to op1
29761 after eliminating TLS symbols.
29763 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
29766 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
29767 * graphite-poly.c (debug_gmp_value): Likewise.
29769 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
29772 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
29773 selected code model, define __code_mode_small__, __code_model_medium__,
29774 __code_model_large__, __code_model_32__ or __code_model_kernel__.
29775 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
29776 xchg temporary register with %k. Declare temporary register as
29778 [__x86_64__]: For medium and large code models, preserve %rbx register.
29780 2013-01-03 Richard Biener <rguenther@suse.de>
29782 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
29783 (dump_subscript): Adjust.
29784 (finalize_ddr_dependent): Do not dump redundant info.
29785 (analyze_siv_subscript): Adjust.
29786 (subscript_dependence_tester): Likewise.
29787 (compute_affine_dependence): Likewise.
29789 2013-01-03 Richard Biener <rguenther@suse.de>
29792 2013-01-03 Richard Biener <rguenther@suse.de>
29794 PR tree-optimization/55857
29795 * tree-vect-stmts.c (vectorizable_load): Do not setup
29796 re-alignment for invariant loads.
29798 2013-01-02 Richard Biener <rguenther@suse.de>
29800 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29801 invariant load do not generate a vector load from the scalar location.
29803 2013-01-03 Richard Biener <rguenther@suse.de>
29805 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
29806 for not vectorizing.
29807 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
29808 not build INDIRECT_REFs, call get_name once only.
29809 (vect_create_data_ref_ptr): Likewise. Dump base object kind
29810 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
29812 2013-01-03 Richard Biener <rguenther@suse.de>
29814 PR tree-optimization/55857
29815 * tree-vect-stmts.c (vectorizable_load): Do not setup
29816 re-alignment for invariant loads.
29818 2013-01-03 Richard Biener <rguenther@suse.de>
29821 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
29822 prefer a built-in decl.
29824 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29826 * gcc.c (process_command): Update copyright notice dates.
29827 * gcov.c (print_version): Likewise.
29828 * gcov-dump.c (print_version): Likewise.
29830 PR rtl-optimization/55838
29831 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
29832 iv0.step, iv1.step and step.
29834 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29835 Marc Glisse <marc.glisse@inria.fr>
29837 PR tree-optimization/55832
29838 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
29839 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
29840 integer_{one,zero}_node.
29842 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29845 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
29846 * var-tracking.c (reverse_op): Don't add reverse ops to
29847 VALUEs that have already
29848 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
29850 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
29852 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
29854 2013-01-02 Teresa Johnson <tejohnson@google.com>
29856 * dumpfile.c (dump_loc): Print filename with location.
29857 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
29858 new location_t parameter to emit complete unroll message with
29859 new dump framework.
29860 (canonicalize_loop_induction_variables): Compute loops location
29861 and pass to try_unroll_loop_completely.
29862 * loop-unroll.c (report_unroll_peel): New function.
29863 (peel_loops_completely): Use new dump format with location
29864 for main dumpfile message, and invoke report_unroll_peel on success.
29865 (decide_unrolling_and_peeling): Ditto.
29866 (decide_peel_once_rolling): Remove old dumpfile message subsumed
29867 by report_unroll_peel.
29868 (decide_peel_completely): Ditto.
29869 (decide_unroll_constant_iterations): Ditto.
29870 (decide_unroll_runtime_iterations): Ditto.
29871 (decide_peel_simple): Ditto.
29872 (decide_unroll_stupid): Ditto.
29873 * cfgloop.c (get_loop_location): New function.
29874 * cfgloop.h (get_loop_location): Declare.
29876 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29878 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
29881 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29883 PR middle-end/55198
29884 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
29885 BLKmode objects when EXPAND_MEMORY is specified.
29887 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29889 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
29891 (fold_builtin_cpu): Do not share cpu model decls across statements.
29893 2013-01-02 Jason Merrill <jason@redhat.com>
29896 * tree.c (build_array_type_1): Revert earlier change.
29898 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
29900 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
29902 * config/aarch64/aarch64-tune.md: Re-generate.
29904 2013-01-02 Richard Biener <rguenther@suse.de>
29906 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29907 invariant load do not generate a vector load from the scalar location.
29909 2013-01-02 Richard Biener <rguenther@suse.de>
29912 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
29913 * configure: Regenerate.
29915 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29917 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
29918 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
29919 (expand_builtin_int_roundingfn_2): Keep the original target around
29920 for the fallback case.
29922 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29924 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
29925 to be clear for sign changes.
29927 2013-01-01 Jan Hubicka <jh@suse.cz>
29929 * ipa-inline-analysis.c: Fix formatting.
29931 2013-01-01 Jakub Jelinek <jakub@redhat.com>
29933 PR tree-optimization/55831
29934 * tree-vect-loop.c (get_initial_def_for_induction): Use
29935 gsi_after_labels instead of gsi_start_bb.
29937 Copyright (C) 2013 Free Software Foundation, Inc.
29939 Copying and distribution of this file, with or without modification,
29940 are permitted in any medium without royalty provided the copyright
29941 notice and this notice are preserved.