1 2013-11-06 Uros Bizjak <ubizjak@gmail.com>
4 * config/i386/i386.c (ix86_avx_u128_mode_needed): Require
5 AVX_U128_DIRTY mode for call_insn RTXes that use AVX256 registers.
6 (ix86_avx_u128_mode_needed): Return AVX_U128_DIRTY mode for call_insn
7 RTXes that return in AVX256 register.
9 2013-11-06 Richard Biener <rguenther@suse.de>
11 PR tree-optimization/58653
12 * tree-predcom.c (ref_at_iteration): Rewrite to generate a MEM_REF.
13 (prepare_initializers_chain): Adjust.
15 2013-11-06 Andrew MacLeod <amacleod@redhat.com>
17 * gimple.h (block_in_transaction): Move to basic-block.h and rename.
18 (gimple_in_transaction): Use bb_in_transaction.
19 * basic-block.h (bb_in_transaction): Relocate here and rename.
20 * tree-ssa-loop-im.c (execute_sm): Use bb_in_transaction.
22 2013-11-06 Richard Biener <rguenther@suse.de>
24 * tree.c (drop_tree_overflow): New function.
25 * tree.h (drop_tree_overflow): Declare.
26 * gimplify.c (gimplify_expr): Drop TREE_OVERFLOW.
27 * tree-vrp.c (range_int_cst_singleton_p): Use
28 is_overflow_infinity instead of testing TREE_OVERFLOW.
29 (extract_range_from_assert): Likewise.
30 (zero_nonzero_bits_from_vr): Likewise.
31 (extract_range_basic): Likewise.
32 (register_new_assert_for): Use drop_tree_overflow.
33 (vrp_visit_phi_node): Likewise.
35 2013-11-06 Dodji Seketeli <dodji@redhat.com>
38 * input.h (location_get_source_line): Take an additional line_size
40 * input.c (get_line): New static function definition.
41 (read_line): Take an additional line_length output parameter to be
42 set to the size of the line. Use the new get_line function do the
44 (location_get_source_line): Take an additional output line_len
45 parameter. Update the use of read_line to pass it the line_len
47 * diagnostic.c (adjust_line): Take an additional input parameter
48 for the length of the line, rather than calculating it with
50 (diagnostic_show_locus): Adjust the use of
51 location_get_source_line and adjust_line with respect to their new
52 signature. While displaying a line now, do not stop at the first
53 null byte. Rather, display the zero byte as a space and keep
54 going until we reach the size of the line.
56 2013-11-06 Eric Botcazou <ebotcazou@adacore.com>
58 * config/i386/i386.c (ix86_expand_prologue): Optimize stack
59 checking for leaf functions without dynamic stack allocation.
60 * config/ia64/ia64.c (ia64_emit_probe_stack_range): Adjust.
61 (ia64_expand_prologue): Likewise.
62 * config/mips/mips.c (mips_expand_prologue): Likewise.
63 * config/rs6000/rs6000.c (rs6000_emit_prologue): Likewise.
64 * config/sparc/sparc.c (sparc_expand_prologue): Likewise.
65 (sparc_flat_expand_prologue): Likewise.
67 2013-11-06 James Greenhalgh <james.greenhalgh@arm.com>
69 * config/aarch64/arm_neon.h (__ST2_LANE_FUNC): Better model data size.
70 (__ST3_LANE_FUNC): Likewise.
71 (__ST4_LANE_FUNC): Likewise.
73 2013-11-06 Nick Clifton <nickc@redhat.com>
75 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define the
76 name returned by msp430_mcu_name.
77 (LIB_SPEC): If a -T option has not been specified then set a
78 default, mcu-specific, linker script.
79 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more mcu names.
80 * config/msp430/msp430.c (msp430x_names): Likewise.
81 Alpha sort the names for ease of comparison.
82 (msp430_mcu_name): New function: Returns a string suitable for
83 use as a C preprocessor symbol based upon the name of the MCU
85 (msp430_option_override): Accept msp430x and msp430xv2 as generic
87 * config/msp430/msp430-protos.h (msp430_mcu_name): Prototype.
89 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
90 unless it is the last part of a spec substring.
91 * doc/invoke.texi (Spec Files): Document space insertion
94 2013-11-06 Christian Bruel <christian.bruel@st.com>
96 * config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
97 Factorize probabilities, Use adjust_address instead of
98 adjust_automodify_address when possible. Enable for optimize.
99 (sh_expand_strlen): New function.
100 * config/sh/sh-protos.h (sh_expand_strlen): Declare.
101 * config/sh/sh.md (strlensi): New pattern.
102 (UNSPEC_BUILTIN_STRLEN): Define.
104 2013-11-06 Jakub Jelinek <jakub@redhat.com>
107 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
108 (expand_assignment): If *bitpos is negative, set *offset
109 and adjust *bitpos, so that it is not negative.
111 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
113 * config/i386/bdver3.md : Added two additional decoder units
114 to support issue rate of 4 and remodeled vector unit.
115 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
116 architectures is set to 4.
117 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
118 lookahead is set to 4 for BD architectures.
120 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
122 * config/rs6000/rs6000.c (rs6000_option_override_internal):
123 Remove restriction against use of VSX instructions when generating
124 code for little endian mode.
126 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
128 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
129 for both big and little endian.
130 (mulv8hi3): Swap input operands for merge high and merge low
131 instructions for little endian.
133 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
135 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
136 define_insn to define_expand that uses even patterns for big
137 endian and odd patterns for little endian.
138 (vec_widen_smult_even_v16qi): Likewise.
139 (vec_widen_umult_even_v8hi): Likewise.
140 (vec_widen_smult_even_v8hi): Likewise.
141 (vec_widen_umult_odd_v16qi): Likewise.
142 (vec_widen_smult_odd_v16qi): Likewise.
143 (vec_widen_umult_odd_v8hi): Likewise.
144 (vec_widen_smult_odd_v8hi): Likewise.
145 (altivec_vmuleub): New define_insn.
146 (altivec_vmuloub): Likewise.
147 (altivec_vmulesb): Likewise.
148 (altivec_vmulosb): Likewise.
149 (altivec_vmuleuh): Likewise.
150 (altivec_vmulouh): Likewise.
151 (altivec_vmulesh): Likewise.
152 (altivec_vmulosh): Likewise.
154 2013-11-05 Mike Stump <mikestump@comcast.net>
156 * Makefile.in (mostlyclean): Remove c-family objects.
158 2013-11-05 Ian Lance Taylor <iant@google.com>
160 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
161 If possible, add .cfi directives to record change to bx.
162 * config/i386/i386.c (ix86_emit_cfi): New function.
163 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
165 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
167 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
168 immediately after a label for a tablejump pattern.
170 * config/arm/arm.c (is_jump_table): Remove.
171 (create_fix_barrier): Use tablejump_p instead.
172 (arm_reorg): Likewise.
173 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
174 (thumb2_output_casesi): Likewise.
175 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
176 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
177 casesi_shift_media, casesi_load_media): Likewise.
178 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
179 * config/microblaze/microblaze.md: Likewise.
181 2013-11-05 Tobias Burnus <burnus@net-b.de>
183 * doc/invoke.texi (-Wdate-time): Document.
185 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
187 * double-int.c (lshift_double, rshift_double): Remove
188 SHIFT_COUNT_TRUNCATED handling.
190 2013-11-05 Jeff Law <law@redhat.com>
192 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
193 * common.opt (-fisolate-erroneous-paths): Add option and documentation.
194 * gimple-ssa-isolate-paths.c: New file.
195 * gimple.c (check_loadstore): New function.
196 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
197 Verify OP is in the argument list and the argument corresponding
198 to OP is a pointer type. Use operand_equal_p rather than
199 pointer equality when testing if OP is on the nonnull list.
200 Use check_loadstore rather than count_ptr_derefs. Handle
201 GIMPLE_RETURN statements.
202 * tree-vrp.c (infer_nonnull_range): Remove.
203 * gimple.h (infer_nonnull_range): Declare.
204 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
205 * passes.def: Add pass_isolate_erroneous_paths.
206 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
207 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
208 * tree-ssa.c (struct count_ptr_d): Remove.
209 (count_ptr_derefs, count_uses_and_derefs): Remove.
210 * tree-ssa.h (count_uses_and_derefs): Remove.
212 2013-11-05 Jakub Jelinek <jakub@redhat.com>
214 PR rtl-optimization/58997
215 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
216 get_iv_value to be in iv->mode rather than iv->extend_mode.
217 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
218 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
219 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
222 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
224 * gimple.h: Move some prototypes to gimple-expr.h and add to include
226 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
227 Move to gimple-expr.h.
228 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
229 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
230 is_gimple_id, virtual_operand_p, is_gimple_addressable,
231 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
233 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
234 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
235 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
236 is_gimple_address, is_gimple_invariant_address,
237 is_gimple_ip_invariant_address, is_gimple_min_invariant,
238 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
239 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
240 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
241 gimple_decl_printable_name, useless_type_conversion_p,
242 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
244 * gimple-expr.c: New File.
245 (useless_type_conversion_p, gimple_set_body, gimple_body,
246 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
247 gimple_can_coalesce_p, extract_ops_from_tree_1,
248 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
249 is_gimple_address, is_gimple_invariant_address,
250 is_gimple_ip_invariant_address, is_gimple_min_invariant,
251 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
252 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
253 is_gimple_mem_ref_addr): Relocate here.
254 * Makefile.in (OBJS): Add gimple-expr.o.
256 2013-11-05 David Malcolm <dmalcolm@redhat.com>
258 * gengtype-parse.c (struct_field_seq): Support empty structs.
260 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
262 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
264 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
266 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
268 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
270 (LONG_DOUBLE_TYPE_SIZE): New define.
271 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
273 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
276 * doc/extend.texi [visibility ("visibility_type")]: Add example
277 about visibility attribute on namespace declaration.
279 2013-11-05 Richard Biener <rguenther@suse.de>
282 * passes.def (all_passes): Start with pass_fixup_cfg again.
284 2013-11-05 Richard Biener <rguenther@suse.de>
286 PR tree-optimization/58955
287 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
290 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
292 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
294 (vec_pack_ufix_trunc_v2df): Likewise.
296 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
299 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
300 pattern, instead of word_mode.
302 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
304 (set_storage_via_setmem): Likewise.
306 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
308 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
309 gimple.h and the rest of the condition in eliminate_build.
310 (eliminate_build): Call new routine.
311 * gimple.h (phi_ssa_name_p): Delete.
313 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
315 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
316 case of no prefix and reserving zero slots, because when that's the
317 case we'll never get here.
318 * vec.h (va_heap::reserve): Don't try and handle
319 vec_prefix::calculate_allocation returning zero because that should
322 2013-11-05 Richard Biener <rguenther@suse.de>
325 * tree-dfa.c (get_ref_base_and_extent): Merge common code
326 in MEM_REF and TARGET_MEM_REF handling. Make sure to
327 process trailing array detection before diving into the
328 view-converted object (and possibly apply some extra offset).
330 2013-11-05 Joseph Myers <joseph@codesourcery.com>
332 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
334 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
336 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
338 PR tree-optimization/58958
339 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
340 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
342 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
344 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
345 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
347 2013-11-05 Jakub Jelinek <jakub@redhat.com>
349 PR tree-optimization/58984
350 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
351 set *SIZE_P if non-NULL on success.
352 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
354 (ipcp_transform_function): Likewise. Punt if size of access
355 is different from TYPE_SIZE on v->value's type.
357 2013-11-05 Tobias Burnus <burnus@net-b.de>
359 * doc/invoke.texi (-fopenmp-simd): Document new option.
360 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
361 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
362 * tree.c (attribute_value_equal): Ditto.
364 2013-11-04 Wei Mi <wmi@google.com>
366 * sched-rgn.c (add_branch_dependences): Keep insns in
367 a SCHED_GROUP at the end of BB to remain their location.
369 2013-11-04 Wei Mi <wmi@google.com>
371 * config/i386/i386.c (memory_address_length): Extract a part
372 of code to rip_relative_addr_p.
373 (rip_relative_addr_p): New Function.
374 (ix86_macro_fusion_p): Ditto.
375 (ix86_macro_fusion_pair_p): Ditto.
376 * config/i386/i386.h: Add new tune features about macro-fusion.
377 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
378 * doc/tm.texi: Generated.
379 * doc/tm.texi.in: Ditto.
380 * haifa-sched.c (try_group_insn): New Function.
381 (group_insns_for_macro_fusion): Ditto.
382 (sched_init): Call group_insns_for_macro_fusion.
383 * target.def: Add two hooks: macro_fusion_p and
386 2013-11-04 Kostya Serebryany <kcc@google.com>
388 Update to match the changed asan API.
389 * asan.c (asan_function_start): New function.
390 (asan_emit_stack_protection): Update the string stored in the
391 stack red zone to match new API. Store the PC of the current
392 function in the red zone.
393 (asan_global_struct): Update the __asan_global definition to match
395 (asan_add_global): Ditto.
396 * asan.h (asan_function_start): New prototype.
397 * final.c (final_start_function): Call asan_function_start.
398 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
401 2013-11-04 Wei Mi <wmi@google.com>
403 * config/i386/i386-c.c (ix86_target_macros_internal): Separate
404 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
405 * config/i386/i386.c (ix86_option_override_internal): Ditto.
406 (ix86_issue_rate): Ditto.
407 (ix86_adjust_cost): Ditto.
408 (ia32_multipass_dfa_lookahead): Ditto.
409 (ix86_sched_init_global): Ditto.
410 (get_builtin_code_for_version): Ditto.
411 * config/i386/i386.h (enum target_cpu_default): Ditto.
412 (enum processor_type): Ditto.
413 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
415 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
417 PR rtl-optimization/58967
418 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
419 !lra_in_progress for mode sizes bigger word.
421 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
423 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
424 arguments to merge instruction for little endian.
425 (vec_widen_umult_lo_v16qi): Likewise.
426 (vec_widen_smult_hi_v16qi): Likewise.
427 (vec_widen_smult_lo_v16qi): Likewise.
428 (vec_widen_umult_hi_v8hi): Likewise.
429 (vec_widen_umult_lo_v8hi): Likewise.
430 (vec_widen_smult_hi_v8hi): Likewise.
431 (vec_widen_smult_lo_v8hi): Likewise.
433 2013-11-04 Ian Lance Taylor <iant@google.com>
435 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
436 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
437 builtins that take pointers.
438 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
439 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
440 (run_gcc): Pass -fnon-call-exceptions.
442 2013-11-04 Jakub Jelinek <jakub@redhat.com>
444 * optabs.c (expand_vec_perm): Revert one incorrect line from
447 PR tree-optimization/58978
448 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
449 use_stmt by single_imm_use directly. Only call single_imm_use
452 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
454 PR rtl-optimization/58968
455 * lra-spills.c (return_regno_p): New function.
456 (lra_final_code_change): Use it.
458 2013-11-04 Joseph Myers <joseph@codesourcery.com>
460 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
461 * target.def (float_exceptions_rounding_supported_p): New hook.
462 * targhooks.c (default_float_exceptions_rounding_supported_p): New
464 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
465 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
467 * doc/tm.texi: Regenerate.
468 * config.gcc (powerpc*-*-linux*): Set extra_objs.
469 * config/rs6000/rs6000-linux.c: New file.
470 * config/rs6000/rs6000-protos.h
471 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
472 * config/rs6000/linux.h
473 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
474 * config/rs6000/linux64.h
475 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
476 * config/rs6000/t-linux (rs6000-linux.o): New rule.
477 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
479 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
481 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
482 Replace the define_insn_and_split with a define_insn and two
483 define_splits, with the split after reload re-permuting the source
484 register to its original value.
485 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
486 (*vsx_le_perm_store_v8hi): Likewise.
487 (*vsx_le_perm_store_v16qi): Likewise.
489 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
491 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
494 2013-11-04 Jakub Jelinek <jakub@redhat.com>
496 PR tree-optimization/58946
497 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
498 bbs with bbinfo[idx].op != NULL before all blocks with
499 bbinfo[idx].op == NULL.
501 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
503 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
504 (avr_dump_double_int_hex): Likewise.
505 (avr_log_vadump): Remove %D and %X handling.
506 * config/avr/avr.c (avr_double_int_push_digit): Delete.
507 (avr_map_op_t): Change map from double_int to unsigned int.
508 (avr_map_op): Update accordingly.
509 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
510 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
511 unsigned ints rather than double_ints.
513 2013-11-03 Marek Polacek <polacek@redhat.com>
515 Implement -fsanitize=vla-bound.
516 * opts.c (common_handle_option): Handle vla-bound.
517 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
518 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
519 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
521 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
523 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
526 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
528 * config/i386/constraints.md (Ts, Tv): New address constrains.
529 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
530 constraint for address_no_seg_operand.
531 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
532 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
533 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
534 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
535 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
536 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
537 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
538 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
539 constraint for vsib_address_operand.
541 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
543 * gcse.c (pre_delete): Remove references to regmove from comments.
544 * recog.c: (validate_replace_rtx_1): Likewise.
545 * config/rl78/rl78.c: Likewise.
546 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
547 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
548 * common/config/mmix/mmix-common.c: Likewise.
550 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
552 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
553 * gimplify.c (gimplify_compound_lval): Likewise.
554 * graphite-clast-to-gimple.c (gloog): Likewise.
555 * graphite-dependences.c (loop_is_parallel_p): Likewise.
556 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
557 (limit_scop); Likewise.
558 (build_scops): Likewise.
559 (dot_scop): Likewise.
560 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
561 (build_scop_drs): Likewise.
562 (insert_stmts): Likewise.
563 (insert_out_of_ssa_copy): Likewise.
564 (remove_phi): Likewise.
565 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
566 * hw-doloop.c (discover_loop): Likewise.
567 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
568 * tree-dfa.c (dump_enumerated_decls): Likewise.
569 * tree-if-conv.c (if_convertable_loop_p): Likewise.
570 * tree-inline.c (tree_function_versioning): Likewise.
571 * tree-loop-distribution.c (build_rdg): Likewise.
572 (rdg_flag_vertex_and_dependent): Likewise.
573 (distribute_loop): Likewise.
574 * tree-parloops.c (loop_parallel_p): Likewise.
575 (eliminate_local_variables): Likewise.
576 (separate_decls_in_region): Likewise.
577 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
578 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
579 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
580 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
581 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
582 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
583 (vectorizable_condition): Likewise.
585 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
587 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
588 * configure: Regenerate.
589 * config/i386/i386.md (*movdi_internal): Change
590 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
591 (*movdf_internal): Ditto.
592 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
593 * config/i386/sse.md (vec_concatv2di): Output interunit movq
594 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
596 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
598 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
601 2013-10-31 David Malcolm <dmalcolm@redhat.com>
603 Automated part of renaming of symtab_node_base to symtab_node.
605 Patch autogenerated by rename_symtab.py from
606 https://github.com/davidmalcolm/gcc-refactoring-scripts
607 revision 58bb219cc090b2f4516a9297d868c245495ee622
608 with ChangeLog entry fixed up by hand.
610 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
612 (cgraph_node_for_asm): Likewise.
613 * cgraph.h (symtab_node_base): Likewise.
614 (cgraph_node): Likewise.
615 (varpool_node): Likewise.
616 (is_a_helper <cgraph_node>::test): Likewise.
617 (is_a_helper <varpool_node>::test): Likewise.
618 (symtab_nodes): Likewise.
619 (symtab_register_node): Likewise.
620 (symtab_unregister_node): Likewise.
621 (symtab_remove_node): Likewise.
622 (symtab_get_node): Likewise.
623 (symtab_node_for_asm): Likewise.
624 (symtab_node_asm_name): Likewise.
625 (symtab_node_name): Likewise.
626 (symtab_insert_node_to_hashtable): Likewise.
627 (symtab_add_to_same_comdat_group): Likewise.
628 (symtab_dissolve_same_comdat_group_list): Likewise.
629 (dump_symtab_node): Likewise.
630 (debug_symtab_node): Likewise.
631 (dump_symtab_base): Likewise.
632 (verify_symtab_node): Likewise.
633 (verify_symtab_base): Likewise.
634 (symtab_used_from_object_file_p): Likewise.
635 (symtab_alias_ultimate_target): Likewise.
636 (symtab_resolve_alias): Likewise.
637 (fixup_same_cpp_alias_visibility): Likewise.
638 (symtab_for_node_and_aliases): Likewise.
639 (symtab_nonoverwritable_alias): Likewise.
640 (availability symtab_node_availability): Likewise.
641 (symtab_semantically_equivalent_p): Likewise.
642 (fixup_same_cpp_alias_visibility): Likewise.
643 (symtab_prevail_in_asm_name_hash): Likewise.
646 (varpool_first_variable): Likewise.
647 (varpool_next_variable): Likewise.
648 (varpool_first_static_initializer): Likewise.
649 (varpool_next_static_initializer): Likewise.
650 (varpool_first_defined_variable): Likewise.
651 (varpool_next_defined_variable): Likewise.
652 (cgraph_first_defined_function): Likewise.
653 (cgraph_next_defined_function): Likewise.
654 (cgraph_first_function): Likewise.
655 (cgraph_next_function): Likewise.
656 (cgraph_first_function_with_gimple_body): Likewise.
657 (cgraph_next_function_with_gimple_body): Likewise.
658 (symtab_alias_target): Likewise.
659 (symtab_real_symbol_p): Likewise.
660 (symtab_can_be_discarded): Likewise.
661 * cgraphbuild.c (mark_address): Likewise.
662 (mark_load): Likewise.
663 (mark_store): Likewise.
664 * cgraphunit.c (decide_is_symbol_needed): Likewise.
666 (enqueue_node): Likewise.
667 (referred_to_p): Likewise.
668 (cgraph_process_same_body_aliases): Likewise.
669 (analyze_functions): Likewise.
670 (handle_alias_pairs): Likewise.
671 (output_weakrefs): Likewise.
673 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
674 * ipa-inline-analysis.c (inline_write_summary): Likewise.
675 * ipa-prop.c (remove_described_reference): Likewise.
676 (try_decrement_rdesc_refcount): Likewise.
677 (ipa_edge_duplication_hook): Likewise.
678 * ipa-ref.c (ipa_record_reference): Likewise.
679 (ipa_maybe_record_reference): Likewise.
680 (ipa_clone_ref): Likewise.
681 (ipa_clone_references): Likewise.
682 (ipa_clone_referring): Likewise.
683 (ipa_find_reference): Likewise.
684 (ipa_remove_stmt_references): Likewise.
685 (ipa_clear_stmts_in_references): Likewise.
686 * ipa-ref.h (symtab_node_base): Likewise.
688 (ipa_record_reference): Likewise.
689 (ipa_maybe_record_reference): Likewise.
690 (ipa_clone_references): Likewise.
691 (ipa_clone_referring): Likewise.
692 (ipa_clone_ref): Likewise.
693 (ipa_find_reference): Likewise.
694 (ipa_remove_stmt_references): Likewise.
695 (ipa_clear_stmts_in_references): Likewise.
696 * ipa-reference.c (ipa_reference_write_optimization_summary):
698 * ipa.c (enqueue_node): Likewise.
699 (process_references): Likewise.
700 (walk_polymorphic_call_targets): Likewise.
701 (symtab_remove_unreachable_nodes): Likewise.
702 (address_taken_from_non_vtable_p): Likewise.
703 (comdat_can_be_unshared_p_1): Likewise.
704 (comdat_can_be_unshared_p): Likewise.
705 (can_replace_by_local_alias): Likewise.
706 (function_and_variable_visibility): Likewise.
707 * is-a.h: Likewise (within example in comment).
708 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
709 (lto_symtab_encoder_encode): Likewise.
710 (lto_symtab_encoder_delete_node): Likewise.
711 (lto_symtab_encoder_in_partition_p): Likewise.
712 (lto_set_symtab_encoder_in_partition): Likewise.
713 (output_refs): Likewise.
714 (compute_ltrans_boundary): Likewise.
715 (output_symtab): Likewise.
716 (input_node): Likewise.
717 (input_ref): Likewise.
718 (input_edge): Likewise.
719 (input_cgraph_1): Likewise.
720 (input_refs): Likewise.
721 (output_cgraph_opt_summary): Likewise.
722 (input_node_opt_summary): Likewise.
723 (input_cgraph_opt_section): Likewise.
724 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
726 * lto-streamer-out.c (lto_output): Likewise.
727 (output_symbol_p): Likewise.
728 (produce_symtab): Likewise.
729 * lto-streamer.h (lto_encoder_entry): Likewise.
730 (lto_free_function_in_decl_state_for_node): Likewise.
731 (lto_symtab_encoder_encode): Likewise.
732 (lto_symtab_encoder_delete_node): Likewise.
733 (lto_symtab_encoder_in_partition_p): Likewise.
734 (lto_set_symtab_encoder_in_partition): Likewise.
735 (lto_symtab_encoder_lookup): Likewise.
736 (lsei_node): Likewise.
737 (lto_symtab_encoder_deref): Likewise.
738 * symtab.c (symtab_hash): Likewise.
739 (assembler_name_hash): Likewise.
740 (symtab_nodes): Likewise.
741 (hash_node): Likewise.
743 (hash_node_by_assembler_name): Likewise.
744 (eq_assembler_name): Likewise.
745 (insert_to_assembler_name_hash): Likewise.
746 (unlink_from_assembler_name_hash): Likewise.
747 (symtab_prevail_in_asm_name_hash): Likewise.
748 (symtab_register_node): Likewise.
749 (symtab_insert_node_to_hashtable): Likewise.
750 (symtab_unregister_node): Likewise.
751 (symtab_get_node): Likewise.
752 (symtab_remove_node): Likewise.
753 (symtab_initialize_asm_name_hash): Likewise.
754 (symtab_node_for_asm): Likewise.
755 (symtab_add_to_same_comdat_group): Likewise.
756 (symtab_dissolve_same_comdat_group_list): Likewise.
757 (symtab_node_asm_name): Likewise.
758 (symtab_node_name): Likewise.
759 (dump_symtab_base): Likewise.
760 (dump_symtab_node): Likewise.
761 (dump_symtab): Likewise.
762 (debug_symtab_node): Likewise.
763 (verify_symtab_base): Likewise.
764 (verify_symtab_node): Likewise.
765 (verify_symtab): Likewise.
766 (symtab_used_from_object_file_p): Likewise.
767 (symtab_node_availability): Likewise.
768 (symtab_alias_ultimate_target): Likewise.
769 (fixup_same_cpp_alias_visibility): Likewise.
770 (symtab_resolve_alias): Likewise.
771 (symtab_for_node_and_aliases): Likewise.
772 (symtab_for_node_and_aliases): Likewise.
773 (symtab_nonoverwritable_alias_1): Likewise.
774 (symtab_nonoverwritable_alias): Likewise.
775 (symtab_semantically_equivalent_p): Likewise.
776 * value-prof.c (init_node_map): Likewise.
777 * varasm.c (find_decl): Likewise.
778 * varpool.c (varpool_node_for_asm): Likewise.
779 (varpool_remove_unreferenced_decls): Likewise.
781 2013-10-31 David Malcolm <dmalcolm@redhat.com>
783 Manual part of renaming of symtab_node_base to symtab_node.
785 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
786 clashes with the preferred name for the base class.
787 (const_symtab_node): Remove redundant typedef.
789 2013-10-31 Jakub Jelinek <jakub@redhat.com>
791 * optabs.c (expand_vec_perm): Avoid vector mode punning
793 * expmed.c (store_bit_field_1): Likewise.
794 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
795 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
796 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
797 * config/i386/i386.c (ix86_expand_vector_move_misalign,
798 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
799 ix86_expand_sse_unpack, ix86_expand_args_builtin,
800 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
801 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
802 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
803 expand_vec_perm_vpshufb2_vpermq,
804 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
805 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
806 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
807 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
810 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
812 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
813 Use gcc_unreachable for unhandled modes. Do not check results of
814 expand_simple_binop. If not expanded to target, move the result.
816 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
817 Shiva Chen <shiva0217@gmail.com>
819 * config.gcc (nds32*-*-*): Add nds32 target.
820 * config/nds32/nds32.c: New file.
821 * config/nds32/nds32.h: New file.
822 * config/nds32/nds32.md: New file.
823 * config/nds32/constants.md: New file.
824 * config/nds32/constraints.md: New file.
825 * config/nds32/iterators.md: New file.
826 * config/nds32/nds32-doubleword.md: New file.
827 * config/nds32/nds32-intrinsic.md: New file.
828 * config/nds32/nds32_intrinsic.h: New file.
829 * config/nds32/nds32-modes.def: New file.
830 * config/nds32/nds32-multiple.md: New file.
831 * config/nds32/nds32.opt: New file.
832 * config/nds32/nds32-opts.h: New file.
833 * config/nds32/nds32-protos.h: New file.
834 * config/nds32/nds32-peephole2.md: New file.
835 * config/nds32/pipelines.md: New file.
836 * config/nds32/predicates.md: New file.
837 * config/nds32/t-mlibs: New file.
838 * common/config/nds32: New directory and files.
840 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
841 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
842 * doc/install.texi (Cross-Compiler-Specific Options): Document
843 --with-nds32-lib for nds32 target.
844 * doc/extend.texi (Function Attributes, Target Builtins): Document
845 nds32 specific attributes.
847 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
849 * lra-constraints (process_alt_operands): Use the result
850 elimination register for operand when matching constraints.
852 2013-10-31 Jakub Jelinek <jakub@redhat.com>
854 * tree-vrp.c (maybe_set_nonzero_bits): New function.
855 (remove_range_assertions): Call it.
857 * tree.c (tree_ctz): New function.
858 * tree.h (tree_ctz): New prototype.
859 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
860 first argument from tree to const_tree.
861 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
862 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
863 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
865 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
866 * tree-vect-loop.c (vect_analyze_loop_operations,
867 vect_transform_loop): Don't force scalar loop for bound just because
868 number of iterations is unknown, only do it if it is not known to be
869 a multiple of vectorization_factor.
870 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
872 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
873 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
874 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
875 (set_nonzero_bits, get_nonzero_bits): New prototypes.
876 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
877 a default def isn't partially constant.
878 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
879 is known to be partially zero.
880 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
881 to see if a default def isn't partially constant.
882 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
883 creation of a range, if VR_RANGE, try to improve nonzero_bits from
885 (set_nonzero_bits, get_nonzero_bits): New functions.
887 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
888 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
889 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
890 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
891 uses but in the condition and ASSERT_EXPR and the other successor of
892 the predecessor bb is __builtin_unreachable (), set_range_info of the
893 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
895 2013-10-31 Martin Jambor <mjambor@suse.cz>
897 PR rtl-optimization/58934
899 2013-10-30 Martin Jambor <mjambor@suse.cz>
900 PR rtl-optimization/10474
901 * ira.c (find_moveable_pseudos): Do not calculate dominance info
903 (interesting_dest_for_shprep): New function.
904 (split_live_ranges_for_shrink_wrap): Likewise.
905 (ira): Calculate dominance info and df analysis. Call
906 split_live_ranges_for_shrink_wrap.
908 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
909 Yury Gribov <y.gribov@samsung.com>
912 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
913 shadow pointer to avoid clobbering the main one.
915 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
917 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
919 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
922 * ira-color.c (update_costs_from_copies): Add new parameter. Use
923 it for calling update_costs_from_allocno.
924 (assign_hard_reg): Call restore_costs_from_copies only for
925 !retry_p. Pass new argument to update_costs_from_copies.
926 (color_pass): Pass new argument to update_costs_from_copies.
927 (ira_mark_allocation_change): Ditto.
929 2013-10-30 Sharad Singhai <singhai@google.com>
932 * opts.c (common_handle_option): Remove deprecated option
933 -ftree-vectorizer-verbose.
934 * doc/invoke.texi (Debugging Options): Ditto.
935 * opts-global.c (handle_common_deferred_options): Ditto.
936 (dump_remap_tree_vectorizer_verbose): Delete.
937 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
939 2013-10-30 DJ Delorie <dj@redhat.com>
941 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
942 that take no arguments.
944 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
947 * reload1.c (update_eliminables_and_spill): New function, broken
949 (reload): Use it. Check for frame size change after frame size
950 alignment, and call update_eliminables_and_spill first if continue-ing.
952 2013-10-30 Cong Hou <congh@google.com>
955 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
956 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
957 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
959 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
961 PR rtl-optimization/58369
962 * reload1.c (compute_reload_subreg_offset): New function.
963 (choose_reload_regs): Use it to pass endian-correct
964 offset to subreg_regno_offset.
966 2013-10-30 Tobias Burnus <burnus@net-b.de>
969 * tree-cfg.c (replace_loop_annotate): Replace warning by
972 2013-10-30 Jason Merrill <jason@redhat.com>
974 * configure.ac (loose_warn): Add -Wno-format if
975 --disable-build-format-warnings.
977 2013-10-30 David Malcolm <dmalcolm@redhat.com>
979 * cgraphunit.c (analyze_functions): Split symtab_node declarations
980 onto multiple lines to make things easier for rename_symtab.py.
982 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
983 (symtab_semantically_equivalent_p): Likewise.
985 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
988 * lra.c (check_rtl): Remove address check before LRA work.
990 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
992 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
993 POINTER_PLUS_EXPR in the defining statement.
995 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
998 * tree-pass.h (make_pass_regmove): Remove.
999 * timevar.def (TV_REGMOVE): Remove.
1000 * passes.def (pass_regmove): Remove.
1001 * opts.c (default_options_table): Remove entry for regmove.
1002 * doc/passes.texi: Remove regmove pass description.
1003 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
1005 (-fdump-rtl-regmove): Ditto.
1006 * common.opt (foptimize-register-move, fregmove): Ignore.
1007 * Makefile.in (OBJS): Remove regmove.o.
1008 * regmove.c: Remove.
1009 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
1011 (struct ira_allocno) New member allocno_prefs.
1012 (ALLOCNO_PREFS): New macro.
1013 (ira_prefs, ira_prefs_num): New external vars.
1014 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
1015 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
1016 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
1018 (ira_add_allocno_copy_to_list): Remove prototype.
1019 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
1020 (ira_pref_iterator): New type.
1021 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
1022 (FOR_EACH_PREF): New macro.
1023 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
1024 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
1026 (ira_setup_alts): New function.
1027 (decrease_live_ranges_number): New function.
1028 (ira): Call the above function.
1029 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
1030 (ira_create_allocno): Initialize allocno prefs.
1031 (pref_pool, pref_vec): New static vars.
1032 (initiate_prefs, find_allocno_pref, ira_create_pref): New
1034 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
1035 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
1036 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
1037 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
1038 (ira_add_allocno_copy_to_list): Make static. Rename to
1039 add_allocno_copy_to_list.
1040 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
1041 swap_allocno_copy_ends_if_necessary.
1042 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
1043 ira_remove_allocno_prefs.
1044 (ira_flattening): Ditto.
1045 (ira_build): Call initiate_prefs, print_prefs.
1046 (ira_destroy): Call finish_prefs.
1047 * ira-color.c (struct update_cost_record): New.
1048 (struct allocno_color_data): Add new member update_cost_records.
1049 (update_cost_record_pool): New static var.
1050 (init_update_cost_records, get_update_cost_record): New functions.
1051 (free_update_cost_record_list, finish_update_cost_records): Ditto.
1052 (struct update_cost_queue_elem): Add member from.
1053 (initiate_cost_update): Call init_update_cost_records.
1054 (finish_cost_update): Call finish_update_cost_records.
1055 (queue_update_cost, get_next_update_cost): Add new param from.
1056 (Update_allocno_cost, update_costs_from_allocno): New functions.
1057 (update_costs_from_prefs): Ditto.
1058 (update_copy_costs): Rename to update_costs_from_copies.
1059 (restore_costs_from_copies): New function.
1060 (update_conflict_hard_regno_costs): Don't go back.
1061 (assign_hard_reg): Call restore_costs_from_copies. Add printing
1063 (pop_allocnos): Add priniting more debug info.
1064 (color_allocnos): Remove prefs for conflicting hard regs.
1065 Call update_costs_from_prefs.
1066 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
1067 (get_dup_num): Rename, modify, and move to ira.c
1068 (process_regs_for_copy): Add prefs.
1069 (add_insn_allocno_copies): Put src as first arg of
1070 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
1071 * ira-costs.c (record_reg_classes): Modify and move code into
1072 record_operands_costs.
1073 (find_costs_and_classes): Create prefs for the hard reg of small
1075 (process_bb_node_for_hard_reg_moves): Add prefs.
1077 2013-10-30 Richard Biener <rguenther@suse.de>
1080 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
1081 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
1083 (pre_and_rev_post_order_compute): ... which now wraps it.
1084 * graph.c (draw_cfg_nodes_no_loops): Use
1085 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
1088 2013-10-30 Christian Bruel <christian.bruel@st.com>
1090 * config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
1091 (sh_expand_cmpstr): Handle known align and schedule improvements.
1092 * config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
1093 * config/sh/sh.md (cmpstrnsi): New pattern.
1095 2013-10-30 Martin Jambor <mjambor@suse.cz>
1097 PR rtl-optimization/10474
1098 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1100 (interesting_dest_for_shprep): New function.
1101 (split_live_ranges_for_shrink_wrap): Likewise.
1102 (ira): Calculate dominance info and df analysis. Call
1103 split_live_ranges_for_shrink_wrap.
1105 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1108 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
1110 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1112 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
1113 * tree.h (POINTER_BOUNDS_P): New.
1114 (BOUNDED_TYPE_P): New.
1116 (pointer_bounds_type_node): New.
1117 * tree.c (build_common_tree_nodes): Initialize
1118 pointer_bounds_type_node.
1119 * gimple.h (gimple_call_get_nobnd_arg_index): New.
1120 (gimple_call_num_nobnd_args): New.
1121 (gimple_call_nobnd_arg): New.
1122 (gimple_return_retbnd): New.
1123 (gimple_return_set_retbnd): New
1124 * gimple.c (gimple_build_return): Increase number of ops
1125 for return statement.
1126 (gimple_call_get_nobnd_arg_index): New.
1127 * gimple-pretty-print.c (dump_gimple_return): Print second op.
1129 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1131 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
1132 with "chkp ctor" and "bnd_legacy" attributes.
1133 * gimplify.c (gimplify_init_constructor): Avoid infinite
1134 loop during gimplification of bounds initializer.
1136 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1138 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
1139 (handle_bnd_legacy): New.
1140 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
1141 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
1144 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
1146 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
1147 (BT_FN_PTR_CONST_PTR): New.
1148 (BT_FN_CONST_PTR_CONST_PTR): New.
1149 (BT_FN_PTR_CONST_PTR_SIZE): New.
1150 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
1151 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
1152 (BT_FN_VOID_CONST_PTR_SIZE): New.
1153 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
1154 * chkp-builtins.def: New.
1155 * builtins.def: include chkp-builtins.def.
1156 (DEF_CHKP_BUILTIN): New.
1157 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
1158 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
1159 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
1160 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
1161 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
1162 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
1163 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
1164 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
1165 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
1166 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
1167 * common.opt (fcheck-pointer-bounds): New.
1168 * toplev.c (process_options): Check Pointer Bounds Checker is
1170 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
1172 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1174 * target.def (builtin_chkp_function): New.
1175 (chkp_bound_type): New.
1176 (chkp_bound_mode): New.
1177 (fn_abi_va_list_bounds_size): New.
1178 (load_bounds_for_arg): New.
1179 (store_bounds_for_arg): New.
1180 * targhooks.h (default_load_bounds_for_arg): New.
1181 (default_store_bounds_for_arg): New.
1182 (default_fn_abi_va_list_bounds_size): New.
1183 (default_chkp_bound_type): New.
1184 (default_chkp_bound_mode): New.
1185 (default_builtin_chkp_function): New.
1186 * targhooks.c (default_load_bounds_for_arg): New.
1187 (default_store_bounds_for_arg): New.
1188 (default_fn_abi_va_list_bounds_size): New.
1189 (default_chkp_bound_type): New.
1190 (default_chkp_bound_mode); New.
1191 (default_builtin_chkp_function): New.
1192 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
1193 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
1194 (TARGET_STORE_BOUNDS_FOR_ARG): New.
1195 (TARGET_BUILTIN_CHKP_FUNCTION): New.
1196 (TARGET_CHKP_BOUND_TYPE): New.
1197 (TARGET_CHKP_BOUND_MODE): New.
1198 * doc/tm.texi: Regenerated.
1199 * langhooks.h (lang_hooks): Add chkp_supported field.
1200 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
1201 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
1203 2013-10-29 Andrew Pinski <apinski@cavium.com>
1205 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
1206 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
1207 fails, combining the branches anyways.
1208 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
1209 increases the number of combinings.
1210 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
1212 2013-10-29 Mike Stump <mikestump@comcast.net>
1214 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
1215 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
1216 (make_vector_mode): Increase namebuf to 16.
1217 (emit_insn_modes_h): When processing BImode, don't
1218 also match partial int modes.
1219 (emit_class_narrowest_mode): Likewise.
1221 * config/bfin/bfin-modes.def: Add precision to PDI.
1222 * config/m32c/m32c-modes.def: Add precision to PSI.
1223 * config/msp430/msp430-modes.def: Add precision to PSI.
1224 * config/rs6000/rs6000-modes.def: Add precision to PTI.
1225 * config/sh/sh-modes.def: Add precision to PSI and PDI.
1227 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
1230 * config/sh/sh.md (*addc): Rename existing variations to ...
1231 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
1232 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
1233 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
1234 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
1236 2013-10-29 Teresa Johnson <tejohnson@google.com>
1239 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
1240 insanities when updating probabilities.
1242 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1244 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
1245 prettyprinter for cgraph_node to reflect the conversion of the
1246 symtable types to a C++ class hierarchy: it now *is* a
1247 symtab_node_base, rather than having one (named "symbol").
1249 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
1251 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
1252 __cilkrts_pop_frame. If matched, then return true for built-in
1254 (expand_builtin): Added BUILT_IN_CILK_DETACH and
1255 BUILT_IN_CILK_POP_FRAME case.
1256 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
1257 (lhs_cilk_detect_spawn): Likewise.
1258 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
1259 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
1260 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
1261 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
1262 (LANG_HOOKS_CILKPLUS): Likewise.
1263 * tree.h (CILK_SPAWN_FN): Likewise.
1264 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
1265 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
1266 (OBJS): Added cilk-common.o.
1267 (BUILTINS_DEF): Added cilk-builtins.def.
1268 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
1269 (lhd_cilk_detect_spawn): Likewise.
1270 * langhooks.h (lang_hooks_for_cilkplus): New struct.
1271 (struct lang_hooks): Added new field called "cilkplus."
1272 * cilk-common.c: New file.
1274 * cilk-builtins.def: Likewise.
1275 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
1276 "__cilk" macro and set it to 200.
1277 * function.h (struct function::cilk_frame_decl): New field.
1278 (struct function::is_cilk_function): Likewise.
1279 (struct function::calls_cilk_spawn): Likewise.
1280 * gimplify.c (gimplify_call_expr): Added a check if the function call
1281 being gimplified is a spawn detach point. If so, then add pop_frame
1282 and detach function calls.
1283 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
1284 for gimplifying _Cilk_spawn and _Cilk_sync statements.
1285 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
1286 function. If so, added a _Cilk_sync and gimplified it.
1287 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
1288 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
1289 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
1291 (can_inline_edge_p): Prevent inling of spawnee function.
1292 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
1293 for functions that use Cilk keywords.
1294 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
1295 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
1296 CILK_SYNC_STMT cases.
1297 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
1299 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
1300 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
1301 * passes.texi (Cilk Keywords): New section that describes the compiler
1302 code changes for handling Cilk Keywords.
1304 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1306 Patch autogenerated by refactor_symtab.py from
1307 https://github.com/davidmalcolm/gcc-refactoring-scripts
1308 revision 58bb219cc090b2f4516a9297d868c245495ee622
1310 * asan.c (asan_finish_file): Update for conversion of symtab types to
1311 a true class hierarchy.
1312 * cfgexpand.c (estimated_stack_frame_size): Likewise.
1313 * cgraph.c (cgraph_get_body): Likewise.
1314 (cgraph_get_create_real_symbol_node): Likewise.
1315 (verify_cgraph_node): Likewise.
1316 (verify_edge_corresponds_to_fndecl): Likewise.
1317 (verify_edge_count_and_frequency): Likewise.
1318 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
1319 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1320 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
1321 (cgraph_node_cannot_return): Likewise.
1322 (cgraph_set_pure_flag_1): Likewise.
1323 (cgraph_set_const_flag_1): Likewise.
1324 (cgraph_set_nothrow_flag_1): Likewise.
1325 (cgraph_make_node_local_1): Likewise.
1326 (cgraph_for_node_and_aliases): Likewise.
1327 (cgraph_for_node_thunks_and_aliases): Likewise.
1328 (cgraph_node_can_be_local_p): Likewise.
1329 (cgraph_node_cannot_be_local_p_1): Likewise.
1330 (cgraph_function_body_availability): Likewise.
1331 (dump_cgraph_node): Likewise.
1332 (cgraph_rtl_info): Likewise.
1333 (cgraph_mark_address_taken_node): Likewise.
1334 (cgraph_remove_node): Likewise.
1335 (cgraph_release_function_body): Likewise.
1336 (cgraph_update_edges_for_call_stmt_node): Likewise.
1337 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
1338 (cgraph_make_edge_direct): Likewise.
1339 (cgraph_resolve_speculation): Likewise.
1340 (cgraph_speculative_call_info): Likewise.
1341 (cgraph_turn_edge_to_speculative): Likewise.
1342 (cgraph_create_edge_1): Likewise.
1343 (cgraph_set_call_stmt): Likewise.
1344 (cgraph_node_for_asm): Likewise.
1345 (cgraph_add_thunk): Likewise.
1346 (cgraph_same_body_alias): Likewise.
1347 (cgraph_create_function_alias): Likewise.
1348 (cgraph_create_node): Likewise.
1349 (cgraph_create_empty_node): Likewise.
1350 (record_function_versions): Likewise.
1351 (used_from_object_file_p): Likewise.
1352 * cgraph.h (symtab_can_be_discarded): Likewise.
1353 (symtab_real_symbol_p): Likewise.
1354 (cgraph_mark_force_output_node): Likewise.
1355 (cgraph_edge_recursive_p): Likewise.
1356 (symtab_alias_target): Likewise.
1357 (varpool_all_refs_explicit_p): Likewise.
1358 (varpool_can_remove_if_no_refs): Likewise.
1359 (cgraph_only_called_directly_or_aliased_p): Likewise.
1360 (cgraph_next_function_with_gimple_body): Likewise.
1361 (cgraph_first_function_with_gimple_body): Likewise.
1362 (cgraph_function_with_gimple_body_p): Likewise.
1363 (cgraph_next_function): Likewise.
1364 (cgraph_first_function): Likewise.
1365 (cgraph_next_defined_function): Likewise.
1366 (cgraph_first_defined_function): Likewise.
1367 (varpool_next_defined_variable): Likewise.
1368 (varpool_first_defined_variable): Likewise.
1369 (varpool_next_static_initializer): Likewise.
1370 (varpool_first_static_initializer): Likewise.
1371 (varpool_next_variable): Likewise.
1372 (varpool_first_variable): Likewise.
1373 (varpool_node_name): Likewise.
1374 (varpool): Likewise.
1376 (is_a_helper <varpool_node>::test): Likewise.
1377 (is_a_helper <cgraph_node>::test): Likewise.
1378 (varpool_variable_node): Likewise.
1379 (cgraph_function_or_thunk_node): Likewise.
1380 (varpool_alias_target): Likewise.
1381 (cgraph_alias_target): Likewise.
1382 (cgraph_node_name): Likewise.
1383 (varpool_node_asm_name): Likewise.
1384 (cgraph_node_asm_name): Likewise.
1385 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
1386 (cgraph_rebuild_references): Likewise.
1387 (rebuild_cgraph_edges): Likewise.
1388 (record_eh_tables): Likewise.
1389 (build_cgraph_edges): Likewise.
1390 (mark_store): Likewise.
1391 (mark_load): Likewise.
1392 (mark_address): Likewise.
1393 (record_type_list): Likewise.
1394 (record_reference): Likewise.
1395 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
1396 (cgraph_materialize_clone): Likewise.
1397 (cgraph_function_versioning): Likewise.
1398 (cgraph_copy_node_for_versioning): Likewise.
1399 (update_call_expr): Likewise.
1400 (cgraph_find_replacement_node): Likewise.
1401 (cgraph_create_virtual_clone): Likewise.
1402 (cgraph_clone_node): Likewise.
1403 * cgraphunit.c (compile): Likewise.
1404 (output_weakrefs): Likewise.
1405 (output_in_order): Likewise.
1406 (expand_function): Likewise.
1407 (assemble_thunks_and_aliases): Likewise.
1408 (expand_thunk): Likewise.
1409 (mark_functions_to_output): Likewise.
1410 (handle_alias_pairs): Likewise.
1411 (analyze_functions): Likewise.
1412 (walk_polymorphic_call_targets): Likewise.
1413 (varpool_finalize_decl): Likewise.
1414 (process_function_and_variable_attributes): Likewise.
1415 (cgraph_process_same_body_aliases): Likewise.
1416 (analyze_function): Likewise.
1417 (cgraph_add_new_function): Likewise.
1418 (cgraph_finalize_function): Likewise.
1419 (referred_to_p): Likewise.
1420 (cgraph_reset_node): Likewise.
1421 (cgraph_process_new_functions): Likewise.
1422 (enqueue_node): Likewise.
1423 (decide_is_symbol_needed): Likewise.
1424 * coverage.c (coverage_compute_profile_id): Likewise.
1425 * dbxout.c (dbxout_expand_expr): Likewise.
1426 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
1427 (reference_to_unused): Likewise.
1428 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1429 * gimplify.c (unvisit_body): Likewise.
1430 (unshare_body): Likewise.
1431 * ipa-cp.c (ipcp_generate_summary): Likewise.
1432 (ipcp_decision_stage): Likewise.
1433 (identify_dead_nodes): Likewise.
1434 (decide_whether_version_node): Likewise.
1435 (decide_about_value): Likewise.
1436 (perhaps_add_new_callers): Likewise.
1437 (create_specialized_node): Likewise.
1438 (update_profiling_info): Likewise.
1439 (ipcp_propagate_stage): Likewise.
1440 (estimate_local_effects): Likewise.
1441 (good_cloning_opportunity_p): Likewise.
1442 (devirtualization_time_bonus): Likewise.
1443 (propagate_constants_accross_call): Likewise.
1444 (initialize_node_lattices): Likewise.
1445 (ipcp_cloning_candidate_p): Likewise.
1446 (determine_versionability): Likewise.
1447 (print_all_lattices): Likewise.
1448 (print_lattice): Likewise.
1449 (ipcp_discover_new_direct_edges): Likewise.
1450 * ipa-devirt.c (ipa_devirt): Likewise.
1451 (likely_target_p): Likewise.
1452 (update_type_inheritance_graph): Likewise.
1453 (possible_polymorphic_call_target_p): Likewise.
1454 (dump_possible_polymorphic_call_targets): Likewise.
1455 (devirt_variable_node_removal_hook): Likewise.
1456 (record_binfo): Likewise.
1457 (maybe_record_node): Likewise.
1458 (build_type_inheritance_graph): Likewise.
1459 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1460 (inline_generate_summary): Likewise.
1461 (inline_analyze_function): Likewise.
1462 (do_estimate_growth): Likewise.
1463 (simple_edge_hints): Likewise.
1464 (estimate_node_size_and_time): Likewise.
1465 (estimate_edge_devirt_benefit): Likewise.
1466 (compute_inline_parameters): Likewise.
1467 (estimate_function_body_sizes): Likewise.
1468 (compute_bb_predicates): Likewise.
1469 (initialize_inline_failed): Likewise.
1470 (dump_inline_summary): Likewise.
1471 (dump_inline_edge_summary): Likewise.
1472 * ipa-inline-transform.c (inline_transform): Likewise.
1473 (preserve_function_body_p): Likewise.
1474 (save_inline_function_body): Likewise.
1475 (inline_call): Likewise.
1476 (clone_inlined_nodes): Likewise.
1477 (can_remove_node_now_p): Likewise.
1478 (can_remove_node_now_p_1): Likewise.
1479 * ipa-inline.c (early_inliner): Likewise.
1480 (early_inline_small_functions): Likewise.
1481 (inline_always_inline_functions): Likewise.
1482 (ipa_inline): Likewise.
1483 (flatten_function): Likewise.
1484 (inline_small_functions): Likewise.
1485 (speculation_useful_p): Likewise.
1486 (recursive_inlining): Likewise.
1487 (update_caller_keys): Likewise.
1488 (reset_edge_caches): Likewise.
1489 (update_edge_key): Likewise.
1490 (edge_badness): Likewise.
1491 (relative_time_benefit): Likewise.
1492 (want_inline_self_recursive_call_p): Likewise.
1493 (want_inline_small_function_p): Likewise.
1494 (want_early_inline_function_p): Likewise.
1495 (num_calls): Likewise.
1496 (can_early_inline_edge_p): Likewise.
1497 (can_inline_edge_p): Likewise.
1498 (report_inline_failed_reason): Likewise.
1499 * ipa-profile.c (ipa_profile): Likewise.
1500 (ipa_propagate_frequency): Likewise.
1501 (ipa_propagate_frequency_1): Likewise.
1502 (ipa_profile_generate_summary): Likewise.
1503 * ipa-prop.c (ipcp_transform_function): Likewise.
1504 (read_replacements_section): Likewise.
1505 (ipa_prop_read_section): Likewise.
1506 (ipa_modify_call_arguments): Likewise.
1507 (ipa_print_node_params): Likewise.
1508 (propagate_controlled_uses): Likewise.
1509 (update_indirect_edges_after_inlining): Likewise.
1510 (remove_described_reference): Likewise.
1511 (ipa_make_edge_direct_to_target): Likewise.
1512 (ipa_analyze_node): Likewise.
1513 (ipa_analyze_params_uses): Likewise.
1514 (ipa_compute_jump_functions): Likewise.
1515 (ipa_get_callee_param_type): Likewise.
1516 (ipa_print_node_jump_functions): Likewise.
1517 (ipa_initialize_node_params): Likewise.
1518 (ipa_populate_param_decls): Likewise.
1519 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
1520 (write_agg_replacement_chain): Likewise.
1521 (ipa_write_node_info): Likewise.
1522 (ipa_edge_duplication_hook): Likewise.
1523 (try_decrement_rdesc_refcount): Likewise.
1524 * ipa-pure-const.c (propagate_nothrow): Likewise.
1525 (propagate_pure_const): Likewise.
1526 (pure_const_read_summary): Likewise.
1527 (pure_const_write_summary): Likewise.
1528 (analyze_function): Likewise.
1529 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
1530 (ipa_ref_referring_ref_list): Likewise.
1531 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
1532 (ipa_remove_stmt_references): Likewise.
1533 (ipa_find_reference): Likewise.
1534 (ipa_dump_referring): Likewise.
1535 (ipa_dump_references): Likewise.
1536 (ipa_record_reference): Likewise.
1537 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
1538 (ipa_reference_write_optimization_summary): Likewise.
1539 (write_node_summary_p): Likewise.
1540 (propagate): Likewise.
1541 (read_write_all_from_decl): Likewise.
1542 (generate_summary): Likewise.
1543 (analyze_function): Likewise.
1544 (propagate_bits): Likewise.
1545 (ipa_reference_get_not_written_global): Likewise.
1546 (ipa_reference_get_not_read_global): Likewise.
1547 * ipa-split.c (execute_split_functions): Likewise.
1548 (split_function): Likewise.
1549 * ipa-utils.c (ipa_merge_profiles): Likewise.
1550 (dump_cgraph_node_set): Likewise.
1551 (ipa_reverse_postorder): Likewise.
1552 (ipa_edge_within_scc): Likewise.
1553 (ipa_get_nodes_in_cycle): Likewise.
1554 (ipa_free_postorder_info): Likewise.
1555 (ipa_reduced_postorder): Likewise.
1556 (searchc): Likewise.
1557 (recursive_call_p): Likewise.
1558 * ipa.c (ipa_cdtor_merge): Likewise.
1559 (record_cdtor_fn): Likewise.
1560 (function_and_variable_visibility): Likewise.
1561 (varpool_externally_visible_p): Likewise.
1562 (cgraph_externally_visible_p): Likewise.
1563 (comdat_can_be_unshared_p): Likewise.
1564 (comdat_can_be_unshared_p_1): Likewise.
1565 (address_taken_from_non_vtable_p): Likewise.
1566 (ipa_discover_readonly_nonaddressable_vars): Likewise.
1567 (symtab_remove_unreachable_nodes): Likewise.
1568 (walk_polymorphic_call_targets): Likewise.
1569 (process_references): Likewise.
1570 (enqueue_node): Likewise.
1571 (has_addr_references_p): Likewise.
1572 (cgraph_non_local_node_p_1): Likewise.
1573 * is-a.h (varpool_analyze_node): Likewise.
1574 * lto-cgraph.c (input_symtab): Likewise.
1575 (merge_profile_summaries): Likewise.
1576 (input_cgraph_1): Likewise.
1577 (input_edge): Likewise.
1578 (input_varpool_node): Likewise.
1579 (input_node): Likewise.
1580 (input_overwrite_node): Likewise.
1581 (compute_ltrans_boundary): Likewise.
1582 (output_refs): Likewise.
1583 (lto_output_varpool_node): Likewise.
1584 (lto_output_node): Likewise.
1585 (reachable_from_other_partition_p): Likewise.
1586 (referenced_from_other_partition_p): Likewise.
1587 (lto_output_edge): Likewise.
1588 (output_node_opt_summary): Likewise.
1589 (add_node_to): Likewise.
1590 (reachable_from_this_partition_p): Likewise.
1591 (lto_set_symtab_encoder_in_partition): Likewise.
1592 (lto_symtab_encoder_in_partition_p): Likewise.
1593 (lto_set_symtab_encoder_encode_initializer): Likewise.
1594 (lto_symtab_encoder_encode_initializer_p): Likewise.
1595 (lto_set_symtab_encoder_encode_body): Likewise.
1596 (lto_symtab_encoder_encode_body_p): Likewise.
1597 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1599 * lto-streamer-in.c (lto_read_body): Likewise.
1600 (fixup_call_stmt_edges): Likewise.
1601 (fixup_call_stmt_edges_1): Likewise.
1602 * lto-streamer-out.c (produce_symtab): Likewise.
1603 (output_symbol_p): Likewise.
1604 (write_symbol): Likewise.
1605 (lto_output): Likewise.
1606 (copy_function): Likewise.
1607 (output_function): Likewise.
1608 * passes.c (function_called_by_processed_nodes_p): Likewise.
1609 (ipa_write_optimization_summaries): Likewise.
1610 (ipa_write_summaries): Likewise.
1611 (do_per_function_toporder): Likewise.
1612 (do_per_function): Likewise.
1613 (dump_passes): Likewise.
1614 * symtab.c (symtab_semantically_equivalent_p): Likewise.
1615 (symtab_nonoverwritable_alias): Likewise.
1616 (symtab_nonoverwritable_alias_1): Likewise.
1617 (symtab_for_node_and_aliases): Likewise.
1618 (symtab_resolve_alias): Likewise.
1619 (fixup_same_cpp_alias_visibility): Likewise.
1620 (symtab_alias_ultimate_target): Likewise.
1621 (symtab_used_from_object_file_p): Likewise.
1622 (verify_symtab_base): Likewise.
1623 (dump_symtab_base): Likewise.
1624 (symtab_node_name): Likewise.
1625 (symtab_node_asm_name): Likewise.
1626 (symtab_dissolve_same_comdat_group_list): Likewise.
1627 (symtab_add_to_same_comdat_group): Likewise.
1628 (symtab_unregister_node): Likewise.
1629 (symtab_insert_node_to_hashtable): Likewise.
1630 (symtab_register_node): Likewise.
1631 (unlink_from_assembler_name_hash): Likewise.
1632 (insert_to_assembler_name_hash): Likewise.
1633 (eq_assembler_name): Likewise.
1634 (hash_node_by_assembler_name): Likewise.
1635 (eq_node): Likewise.
1636 (hash_node): Likewise.
1637 * toplev.c (wrapup_global_declaration_2): Likewise.
1638 * trans-mem.c (ipa_tm_execute): Likewise.
1639 (ipa_tm_transform_clone): Likewise.
1640 (ipa_tm_transform_transaction): Likewise.
1641 (ipa_tm_transform_calls_redirect): Likewise.
1642 (ipa_tm_insert_gettmclone_call): Likewise.
1643 (ipa_tm_insert_irr_call): Likewise.
1644 (ipa_tm_create_version): Likewise.
1645 (ipa_tm_create_version_alias): Likewise.
1646 (ipa_tm_mark_forced_by_abi_node): Likewise.
1647 (ipa_tm_mark_force_output_node): Likewise.
1648 (ipa_tm_diagnose_tm_safe): Likewise.
1649 (ipa_tm_mayenterirr_function): Likewise.
1650 (ipa_tm_scan_irr_function): Likewise.
1651 (ipa_tm_note_irrevocable): Likewise.
1652 (ipa_tm_scan_calls_clone): Likewise.
1653 (get_cg_data): Likewise.
1654 * tree-eh.c (tree_could_trap_p): Likewise.
1655 * tree-emutls.c (ipa_lower_emutls): Likewise.
1656 (create_emultls_var): Likewise.
1657 (lower_emutls_function_body): Likewise.
1658 (gen_emutls_addr): Likewise.
1659 (emutls_decl): Likewise.
1660 (new_emutls_decl): Likewise.
1661 * tree-inline.c (tree_function_versioning): Likewise.
1662 (optimize_inline_calls): Likewise.
1663 (expand_call_inline): Likewise.
1664 (estimate_num_insns): Likewise.
1665 (copy_bb): Likewise.
1666 (delete_unreachable_blocks_update_callgraph): Likewise.
1667 * tree-nested.c (gimplify_all_functions): Likewise.
1668 (create_nesting_tree): Likewise.
1669 (check_for_nested_with_variably_modified): Likewise.
1670 * tree-pretty-print.c (dump_function_header): Likewise.
1671 * tree-profile.c (tree_profiling): Likewise.
1672 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
1673 (modify_function): Likewise.
1674 (convert_callers): Likewise.
1675 (convert_callers_for_node): Likewise.
1676 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
1677 (associate_varinfo_to_alias): Likewise.
1678 (create_variable_info_for): Likewise.
1679 (get_constraint_for_ssa_var): Likewise.
1680 * tree-vectorizer.c (increase_alignment): Likewise.
1681 * tree.c (find_decls_types_in_var): Likewise.
1682 (find_decls_types_in_node): Likewise.
1683 (free_lang_data_in_decl): Likewise.
1684 * value-prof.c (gimple_ic_transform): Likewise.
1685 (gimple_ic): Likewise.
1686 (check_ic_target): Likewise.
1687 (init_node_map): Likewise.
1688 * varasm.c (decl_binds_to_current_def_p): Likewise.
1689 (default_binds_local_p_1): Likewise.
1690 (dump_tm_clone_pairs): Likewise.
1691 (assemble_alias): Likewise.
1692 (find_decl): Likewise.
1693 (mark_decl_referenced): Likewise.
1694 * varpool.c (varpool_for_node_and_aliases): Likewise.
1695 (varpool_extra_name_alias): Likewise.
1696 (varpool_create_variable_alias): Likewise.
1697 (add_new_static_var): Likewise.
1698 (varpool_finalize_named_section_flags): Likewise.
1699 (varpool_remove_unreferenced_decls): Likewise.
1700 (enqueue_node): Likewise.
1701 (varpool_assemble_decl): Likewise.
1702 (assemble_aliases): Likewise.
1703 (varpool_analyze_node): Likewise.
1704 (cgraph_variable_initializer_availability): Likewise.
1705 (varpool_add_new_variable): Likewise.
1706 (ctor_for_folding): Likewise.
1707 (dump_varpool_node): Likewise.
1708 (varpool_remove_initializer): Likewise.
1709 (varpool_remove_node): Likewise.
1710 (varpool_node_for_decl): Likewise.
1711 (varpool_create_empty_node): Likewise.
1712 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
1713 (ix86_get_function_versions_dispatcher): Likewise.
1715 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1717 * cgraph.h (symtab_node_base): Convert to a class;
1718 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
1719 chain_next/prev from symtab_node_def.
1720 (cgraph_node): Inherit from symtab_node; add GTY option
1721 tag ("SYMTAB_FUNCTION").
1722 (varpool_node): Inherit from symtab_node; add GTY option
1723 tag ("SYMTAB_VARIABLE").
1724 (symtab_node_def): Remove.
1725 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
1726 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
1727 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
1728 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
1730 * ipa-ref.h (symtab_node_def): Drop.
1731 (symtab_node): Change underlying type from symtab_node_def to
1733 (const_symtab_node): Likwise.
1735 * is-a.h: Update examples in comment.
1737 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
1738 (assembler_name_hash): Likewise.
1740 2013-10-29 Martin Liska <marxin.liska@gmail.com>
1742 * doc/tree-ssa.texi (gimple_phi_result): Document.
1743 (gimple_phi_num_args, gimple_phi_arg): Likewise.
1744 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
1745 (PHI_RESULT, PHI_NUM_ARGS): Remove.
1746 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
1748 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1750 * expr.h: Revert change and include tree-core.h.
1751 * rtl.h: Revert change and don't include tree-core.h.
1753 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1755 * config/darwin.c: Include gimple.h.
1756 * config/i386/winnt.c: Likewise.
1758 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
1760 PR tree-optimization/19831
1761 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
1763 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1765 * tree-outof-ssa.h: Remove include files.
1766 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
1768 * tree-ssa-coalesce.c: Likewise.
1769 * cfgexpand.c: Likewise.
1770 * tree-ssa-ter.c: Likewise.
1771 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
1772 * lto-streamer.h: Likewise.
1773 * cgraphbuild.c: Add gimple.h to include list.
1774 * data-streamer-in.c: Likewise.
1775 * ipa-cp.c: Likewise.
1776 * tree-streamer.c: Likewise.
1777 * lto-compress.c: Likewise.
1778 * ipa-reference.c: Likewise.
1779 * data-streamer-out.c: Likewise.
1780 * lto-cgraph.c: Likewise.
1781 * cgraphclones.c: Likewise.
1782 * ipa-utils.c: Likewise.
1783 * data-streamer.c: Likewise.
1784 * ipa-split.c: Likewise.
1785 * lto-section-in.c: Likewise.
1786 * tree-streamer-out.c: Likewise.
1787 * ipa-prop.c: Likewise.
1788 * tree-streamer-in.c: Likewise.
1789 * symtab.c: Likewise.
1790 * opts-global.c: Likewise.
1791 * lto-opts.c: Likewise.
1792 * lto-section-out.c: Likewise.
1793 * lto-streamer.c: Likewise.
1794 * rtl.h: Add tree-core.h to include list.
1795 * expr.h: Remove tree-core.h from include list.
1796 * gimple.h: Likewise.
1797 * ipa-utils.h: Likewise.
1798 * streamer-hooks.h: Likewise.
1799 * streamer-hooks.c: Include input.h.
1801 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1803 * config/arm/arm.c (cortexa7_extra_costs): New table.
1804 (arm_cortex_a7_tune): New.
1805 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
1807 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
1809 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
1811 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1813 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
1814 to use autogenerated markers for a class-hierarchy, every class
1815 must have a GTY marker.
1816 * gengtype.h (struct type): Add linked list of subclasses to
1817 the "s" member of the union.
1818 (add_subclass): New decl.
1819 * gengtype-state.c (read_state_struct_type): Set up subclass
1821 * gengtype.c (get_ultimate_base_class): New.
1822 (add_subclass): New.
1823 (new_structure): Set up subclass linked list.
1824 (set_gc_used_type): Propagate usage information to subclasses.
1825 (output_mangled_typename): Use get_ultimate_base_class.
1826 (walk_subclasses): Use the subclass linked list, avoiding an
1827 O(N^2) when writing out all types.
1828 (walk_type): Issue an error if the base class is missing a tag,
1829 rather than generating bogus C code. Add a gcc_unreachable
1830 default case, in case people omit tags from concrete subclasses,
1831 or get the values wrong.
1832 (write_func_for_structure): Issue an error for subclasses for
1833 which the base doesn't have a "desc", since otherwise the
1834 autogenerated routines for the base would silently fail to visit
1835 any subclass fields.
1836 (write_root): Use get_ultimate_base_class, tweaking constness of
1837 tp to match that function's signature.
1839 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1841 * doc/gty.texi (GTY Options): Add note about inheritance to
1842 description of desc and tag.
1843 (Inheritance and GTY): New.
1845 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1847 * gengtype-parse.c (opts_have): Drop "static" so that
1848 we can use this from gengtype.c.
1849 * gengtype.c (set_gc_used_type): Mark any base class as used;
1850 update field traversal to visit inherited fields.
1851 (output_mangled_typename): Convert references to classes within
1852 an inheritance hierarchy to reference the ultimate base class,
1853 since only it will have gt_ functions.
1854 (get_string_option): New.
1855 (walk_subclasses): New.
1856 (walk_type): Treat GTY structs that have a "desc" as being the
1857 root of an inheritance hierarchy. Generate a switch on it
1858 within the marking function which walks all subclasses, adding
1859 cases for them via walk_subclasses. For subclasses, visit all
1860 fields of the type (including inherited ones).
1861 (write_func_for_structure): Don't write fns for subclasses, only
1862 for the ultimate base class within an inheritance hierarchy.
1863 Subclasses-marking will be handled by the base class marking functions.
1864 (write_types): Likewise.
1865 (write_local_func_for_structure): Likewise.
1866 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
1867 a "tag" option (and are thus concrete subclasses).
1868 (write_root): Use the marker function for the ultimate base class.
1869 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
1870 (opts_have): Add declaration.
1872 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
1874 * lra-spills.c (lra_final_code_change): Remove useless move insns
1875 originated from moves of pseudos.
1877 2013-10-28 Jeff Law <law@redhat.com>
1879 * lower-subreg.c (resolve_simple_move): Fix comment typo.
1881 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
1883 * df-scan.c (df_collection_rec): Adjust.
1884 (copy_defs): New constant.
1885 (copy_uses): Likewise.
1886 (copy_eq_uses): Likewise.
1887 (copy_mw): Likewise.
1888 (copy_all): Likewise.
1889 (df_insn_rescan): Adjust.
1890 (df_notes_rescan): Likewise.
1891 (df_swap_refs): Likewise.
1892 (df_sort_and_compress_refs): Likewise.
1893 (df_sort_and_compress_mws): Likewise.
1894 (df_install_refs): Likewise.
1895 (df_install_mws): Likewise.
1896 (df_refs_add_to_chains): Add flags parameter controlling which vectors
1898 (df_bb_refs_record): Adjust.
1899 (df_record_entry_block_defs): Likewise.
1900 (df_record_exit_block_defs): Likewise.
1901 (df_refs_verify): Likewise.
1902 (df_mws_verify): Likewise.
1903 (df_insn_refs_verify): Likewise.
1904 (df_bb_verify): Likewise.
1905 * ipa-pure-const.c (finish_state): Remove.
1906 (propagate): Adjust.
1907 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
1908 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
1909 var-tracking.c: Adjust.
1910 * vec.c (stack_vecs): Remove.
1911 (register_stack_vec): Likewise.
1912 (stack_vec_register_index): Likewise.
1913 (unregister_stack_vec): Likewise.
1914 * vec.h (struct va_stack): Remove.
1915 (struct vec<T, A, vl_ptr>): Specialize as
1916 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
1917 allocation strategy compatable with the vl_ptr layout.
1918 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
1919 specialization anyway.
1920 (class stack_vec): New class.
1921 (vec_stack_alloc): Remove.
1922 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
1924 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1925 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1926 Sergey Lega <sergey.s.lega@intel.com>
1927 Anna Tikhonova <anna.tikhonova@intel.com>
1928 Ilya Tocar <ilya.tocar@intel.com>
1929 Andrey Turetskiy <andrey.turetskiy@intel.com>
1930 Ilya Verbin <ilya.verbin@intel.com>
1931 Kirill Yukhin <kirill.yukhin@intel.com>
1932 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1934 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
1935 (*prefetch_avx512pf_<mode>): New.
1936 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
1937 (avx512f_maskcmp<mode>3): Ditto.
1938 (vashrv16si3): Ditto.
1940 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1941 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1942 Sergey Lega <sergey.s.lega@intel.com>
1943 Anna Tikhonova <anna.tikhonova@intel.com>
1944 Ilya Tocar <ilya.tocar@intel.com>
1945 Andrey Turetskiy <andrey.turetskiy@intel.com>
1946 Ilya Verbin <ilya.verbin@intel.com>
1947 Kirill Yukhin <kirill.yukhin@intel.com>
1948 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1950 * config/i386/i386.md (any_truncate): New.
1951 (trunsuffix): Ditto.
1952 * config/i386/predicates.md (const_8_to_9_operand): New.
1953 (const_10_to_11_operand): Ditto.
1954 (const_12_to_13_operand): Ditto.
1955 (const_14_to_15_operand): Ditto.
1956 (const_16_to_19_operand): Ditto.
1957 (const_20_to_23_operand): Ditto.
1958 (const_24_to_27_operand): Ditto.
1959 (const_28_to_31_operand): Ditto.
1960 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
1961 (cvtusi2<ssescalarmodesuffix>32): New.
1962 (cvtusi2<ssescalarmodesuffix>64): Ditto.
1963 (ufloatv16siv16sf2): Ditto.
1964 (avx512f_fix_notruncv16sfv16si): Ditto.
1965 (avx512f_ufix_notruncv16sfv16si): Ditto.
1966 (avx512f_vcvtss2usi): Ditto.
1967 (avx512f_vcvtss2usiq): Ditto.
1968 (avx512f_vcvttss2usi): Ditto.
1969 (avx512f_vcvttss2usiq): Ditto.
1970 (avx512f_vcvtsd2usi): Ditto.
1971 (avx512f_vcvtsd2usiq): Ditto.
1972 (avx512f_vcvttsd2usi): Ditto.
1973 (avx512f_vcvttsd2usiq): Ditto.
1974 (ufloatv8siv8df): Ditto.
1975 (avx512f_cvtdq2pd512_2): Ditto.
1976 (avx512f_cvtpd2dq512): Ditto.
1977 (avx512f_ufix_notruncv8dfv8si): Ditto.
1978 (avx512f_cvtpd2ps512): Ditto.
1979 (vec_unpacks_lo_v16sf): Ditto.
1980 (vec_unpacks_hi_v16sf): Ditto.
1981 (vec_unpacks_float_hi_v16si): Ditto.
1982 (vec_unpacks_float_lo_v16si): Ditto.
1983 (avx512f_unpckhps512): Ditto.
1984 (avx512f_unpcklps512): Ditto.
1985 (avx512f_movshdup512): Ditto.
1986 (avx512f_movsldup512): Ditto.
1987 (vec_extract_lo_v32hi): Ditto.
1988 (vec_extract_hi_v32hi): Ditto.
1989 (vec_extract_lo_v64qi): Ditto.
1990 (vec_extract_hi_v64qi): Ditto.
1991 (avx512f_unpckhpd512): Ditto.
1992 (avx512f_movddup512): Ditto.
1993 (avx512f_unpcklpd512): Ditto.
1994 (*avx512f_unpcklpd512): Ditto.
1995 (avx512f_shufps512_1): Ditto.
1996 (avx512f_shufpd512_1): Ditto.
1997 (avx512f_interleave_highv8di): Ditto.
1998 (avx512f_interleave_lowv8di): Ditto.
1999 (PMOV_DST_MODE): Ditto.
2000 (pmov_src_mode): Ditto.
2001 (pmov_src_lower): Ditto.
2003 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
2004 (*avx512f_<code>v8div16qi2): Ditto.
2005 (*avx512f_<code>v8div16qi2_store): Ditto.
2006 (vec_widen_umult_even_v16si): Ditto.
2007 (*vec_widen_umult_even_v16si): Ditto.
2008 (vec_widen_smult_even_v16si): Ditto.
2009 (*vec_widen_smult_even_v16si): Ditto.
2010 (avx512f_interleave_highv16si): Ditto.
2011 (avx512f_interleave_lowv16si): Ditto.
2012 (avx512f_<code>v16qiv16si2): Ditto.
2013 (avx512f_<code>v16hiv16si2): Ditto.
2014 (avx512f_<code>v8qiv8di2): Ditto.
2015 (avx512f_<code>v8hiv8di2): Ditto.
2016 (avx512f_<code>v8siv8di2): Ditto.
2017 (avx512cd_maskb_vec_dupv8di): Ditto.
2018 (avx512cd_maskw_vec_dupv16si): Ditto.
2019 (avx512f_vcvtph2ps512): Ditto.
2020 (avx512f_vcvtps2ph512): Ditto.
2021 (VEC_EXTRACT_MODE): Extened with wider modes.
2022 (VEC_PERM_AVX2): Ditto.
2023 (VEC_PERM_CONST): Ditto.
2025 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
2027 * config/arc/arc.c (arc_ccfsm_post_advance):
2028 Add comment about TYPE_RETURN.
2030 2013-10-28 Bin Cheng <bin.cheng@arm.com>
2032 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
2033 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
2034 (strip_offset): Convert offset to unsigned number.
2036 2013-10-27 Tom de Vries <tom@codesourcery.com>
2038 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
2039 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
2041 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
2043 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
2044 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
2045 (TARGET_BIG_ENDIAN): New macro.
2046 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
2047 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
2048 * config/sh/sh.c: Likewise.
2049 * config/sh/sh.h: Likewise.
2051 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
2053 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
2054 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
2056 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2058 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
2060 (cmpstr_t, cmpstrsi): Fix formatting.
2062 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2065 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
2066 addressing, do not use general_operand for memory operands.
2068 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
2071 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2072 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2074 2013-10-26 Jeff Law <law@redhat.com>
2076 * predict.c (PRED_MUDFLAP): Remove.
2077 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
2079 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
2080 (OBJS): Remove tree-nomudflap.o
2081 (GTFILES): Remove tree-mudflap.c
2082 * builtins.c (expand_builtin_alloc): Remove mudflap support.
2083 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2084 (mfwrap_spec, mflib_spec): Likewise.
2085 (cpp_unique_options, cc1_options, static_specs): Likewise.
2086 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
2087 * passes.def: Likewise.
2088 * toplev.c (compile_file, process_options): Likewise.
2089 * tree-inline.c (copy_tree_r): Likewise.
2090 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
2091 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
2092 (build_constant_desc, output_constant_def_contents): Likewise.
2093 (categorize_decl_for_section): Likewise.
2094 * tree-mudflap.c: Removed.
2095 * tree-mudflap.h: Removed.
2096 * tree-nomudflap.c: Removed.
2097 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
2098 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
2099 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2100 * config/sol2.h (MFLIB_SPEC): Likewise.
2101 * doc/install.texi: Remove mudflap references.
2102 * doc/passes.texi: Similarly.
2103 * doc/sourcebuild.texi: Similarly.
2104 * doc/invoke.texi: Remove mudlfap related options.
2106 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2108 PR rtl-optimization/58759
2109 * lra-constraints.c (lra_constraints): Remove wrong condition to
2110 remove insn setting up an equivalent pseudo.
2112 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2114 * config/rs6000/rs6000-protos.h
2115 (rs6000_secondary_memory_needed_mode): New prototype.
2116 * config/rs6000/rs6000.c: Include ira.h.
2117 (TARGET_LRA_P): Redefine.
2118 (rs6000_legitimate_offset_address_p): Call
2119 legitimate_constant_pool_address_p in strict mode for LRA.
2120 (rs6000_legitimate_address_p): Ditto.
2121 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
2122 (rs6000_emit_move): Add LRA version of code to generate load/store
2124 (rs6000_secondary_memory_needed_mode): New.
2125 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
2126 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
2127 constants, memory, and FP registers.
2128 (rs6000_lra_p): New.
2129 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
2130 * config/rs6000/rs6000.opt (mlra): New option.
2131 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2133 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
2135 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
2136 has_single_use () and not do the conversion if has_single_use ()
2137 returns false for the multiplication result.
2139 2013-10-25 David Malcolm <dmalcolm@redhat.com>
2141 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
2142 to "NODE", since this works on a "tree", not an
2144 (CONSTANT_CLASS_P): Likewise.
2147 (INDIRECT_REF_P): Likewise.
2148 (REFERENCE_CLASS_P): Likewise.
2149 (COMPARISON_CLASS_P): Likewise.
2150 (UNARY_CLASS_P): Likewise.
2151 (BINARY_CLASS_P): Likewise.
2152 (STATEMENT_CLASS_P): Likewise.
2153 (VL_EXP_CLASS_P): Likewise.
2154 (EXPRESSION_CLASS_P): Likewise.
2155 (IS_TYPE_OR_DECL_P): Likewise.
2157 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
2159 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
2160 ADDR_EXPR in the defining statement.
2162 2013-10-25 Richard Biener <rguenther@suse.de>
2164 PR tree-optimization/58626
2165 * tree-loop-distribution.c (enum rdg_dep_type): Remove
2166 anti_dd, output_dd and input_dd.
2167 (struct rdg_edge): Remove level and relation members.
2168 (RDGE_LEVEL, RDGE_RELATION): Remove.
2169 (dot_rdg_1): Adjust.
2170 (create_rdg_edge_for_ddr): Remove.
2171 (create_rdg_edges_for_scalar): Adjust.
2172 (create_edge_for_control_dependence): Likewise.
2173 (create_rdg_edges): Split into ...
2174 (create_rdg_flow_edges): ... this
2175 (create_rdg_cd_edges): ... and this.
2177 (build_rdg): Likewise, do not compute data dependences or
2179 (pg_add_dependence_edges): New function.
2181 (distribute_loop): First apply all non-dependence based
2182 partition mergings. Then compute dependences between partitions
2183 and merge and order partitions according to them.
2185 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2187 PR rtl-optimization/58831
2188 * alias.c (init_alias_analysis): At the beginning of each iteration,
2189 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
2191 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2193 * recog.c (search_ofs): New static variable moved from...
2194 (peep2_find_free_register): ...here.
2195 (peephole2_optimize): Initialize it.
2197 2013-10-25 Tobias Burnus <burnus@net-b.de>
2199 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
2201 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
2203 * config/i386/i386.h (TARGET_MPX): New define.
2204 (TARGET_MPX_P): Ditto.
2206 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2208 * config/i386/constraints.md (B): New.
2211 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
2212 * config/i386/i386-modes.def (BND32): New.
2214 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
2215 * config/i386/i386.c (isa_opts): Add mmpx.
2216 (regclass_map): Add bound registers.
2217 (dbx_register_map): Likewise.
2218 (dbx64_register_map): Likewise.
2219 (svr4_dbx_register_map): Likewise.
2221 (ix86_option_override_internal): Support MPX ISA.
2222 (ix86_conditional_register_usage): Support bound registers.
2223 (print_reg): Likewise.
2224 (ix86_code_end): Add MPX bnd prefix.
2225 (output_set_got): Likewise.
2226 (ix86_output_call_insn): Likewise.
2227 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
2228 (ix86_print_operand_punct_valid_p): Likewise.
2229 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
2231 (ix86_class_likely_spilled_p): Add bound regs support.
2232 (ix86_hard_regno_mode_ok): Likewise.
2233 (x86_order_regs_for_local_alloc): Likewise.
2234 (ix86_bnd_prefixed_insn_p): New.
2235 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
2236 (FIXED_REGISTERS): Add bound registers.
2237 (CALL_USED_REGISTERS): Likewise.
2238 (REG_ALLOC_ORDER): Likewise.
2239 (HARD_REGNO_NREGS): Likewise.
2241 (VALID_BND_REG_MODE): New.
2242 (FIRST_BND_REG): New.
2243 (LAST_BND_REG): New.
2244 (reg_class): Add BND_REGS.
2245 (REG_CLASS_NAMES): Likewise.
2246 (REG_CLASS_CONTENTS): Likewise.
2248 (ANY_BND_REG_P): New.
2250 (HI_REGISTER_NAMES): Add bound registers.
2251 * config/i386/i386.md (UNSPEC_BNDMK): New.
2252 (UNSPEC_BNDMK_ADDR): New.
2253 (UNSPEC_BNDSTX): New.
2254 (UNSPEC_BNDLDX): New.
2255 (UNSPEC_BNDLDX_ADDR): New.
2256 (UNSPEC_BNDCL): New.
2257 (UNSPEC_BNDCU): New.
2258 (UNSPEC_BNDCN): New.
2259 (UNSPEC_MPX_FENCE): New.
2262 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
2263 (length_immediate): Likewise.
2264 (prefix_0f): Likewise.
2266 (prefix_rep): Check for bnd prefix.
2267 (length_nobnd): New.
2268 (length): Use length_nobnd if specified.
2273 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
2276 (simple_return_internal): Likewise.
2277 (simple_return_pop_internal): Likewise.
2278 (*indirect_jump): Add MPX bnd prefix.
2279 (*tablejump_1): Likewise.
2280 (simple_return_internal_long): Likewise.
2281 (simple_return_indirect_internal): Likewise.
2285 (*mov<mode>_internal_mpx): New.
2286 (<mode>_<bndcheck>): New.
2287 (*<mode>_<bndcheck>): New.
2292 * config/i386/predicates.md (lea_address_operand): Rename to...
2293 (address_no_seg_operand): ... this.
2294 (address_mpx_no_base_operand): New.
2295 (address_mpx_no_index_operand): New.
2296 (bnd_mem_operator): New.
2297 * config/i386/i386.opt (mmpx): New.
2298 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
2299 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
2301 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2303 * mode-classes.def (MODE_POINTER_BOUNDS): New.
2304 * tree.def (POINTER_BOUNDS_TYPE): New.
2305 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
2306 (POINTER_BOUNDS_MODE): New.
2307 (make_pointer_bounds_mode): New.
2308 * machmode.h (POINTER_BOUNDS_MODE_P): New.
2309 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
2310 (layout_type): Support POINTER_BOUNDS_TYPE.
2311 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
2312 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
2313 (type_contains_placeholder_1): Likewise.
2314 * tree.h (POINTER_BOUNDS_TYPE_P): New.
2315 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
2316 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
2318 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2320 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
2321 mode of address computation when calling memory_address_addr_space.
2323 2013-08-24 Richard Henderson <rth@twiddle.net>
2326 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
2327 instead of create_convert_operand_to.
2328 (maybe_emit_sync_lock_test_and_set): Likewise.
2329 (expand_atomic_compare_and_swap): Likewise.
2330 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
2332 2013-08-24 Sriraman Tallam <tmsriram@google.com>
2334 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
2335 Change param_is to use the struct and not the pointer to the struct.
2337 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
2339 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
2340 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2341 Move to targhooks.c.
2342 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
2343 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2344 Relocate from gimplify.c.
2345 * targhooks.h: Add 2 prototypes.
2346 * tree.h: Delete 2 prototypes.
2348 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2350 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
2351 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
2353 2013-10-24 Cong Hou <congh@google.com>
2355 * convert.c (convert_to_real): Guard those unsafe math function
2356 convertions with flag_unsafe_math_optimizations. Handle sqrt()
2359 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
2362 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
2364 (cgraph_create_edge): Use the new argument.
2365 (cgraph_create_indirect_edge): Likewise.
2367 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
2369 * config/arc/arc.c (arc_ccfsm_post_advance): Also handle
2371 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
2372 changes statep->state.
2374 2013-10-24 Tobias Burnus <burnus@net-b.de>
2377 * tree-cfg.c (replace_loop_annotate): New function.
2378 (execute_build_cfg): Call it.
2379 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
2380 * internal-fn.c (expand_ANNOTATE): New function.
2381 * internal-fn.def (ANNOTATE): Define as new internal function.
2382 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
2383 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
2384 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
2385 * doc/extend.texi (Pragmas): Document #pragma ivdep.
2386 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
2388 2013-10-17 Ian Bolton <ian.bolton@arm.com>
2389 Marcus Shawcroft <marcus.shawcroft@arm.com>
2391 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
2392 Special case reload SP+C into none GENERAL_REGS.
2394 2013-10-24 Michael Matz <matz@suse.de>
2396 * gengtype.c (is_file_equal): Check that files will be same length.
2398 2013-10-25 Christian Bruel <christian.bruel@st.com>
2400 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
2401 * config/sh/t-sh (sh-mem.o): New rule.
2402 * config/sh/sh-mem.cc (expand_block_move): Moved here.
2403 (sh_expand_cmpstr): New function.
2404 * config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
2405 * config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
2406 * config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
2407 (rotlhi3_8): Rename.
2409 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2411 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
2412 * Makefile.in (CXXDEPMODE): Assign and change users.
2413 (CCDEPMODE): Delete.
2414 * configure: Regenerate.
2416 2013-10-23 David Malcolm <dmalcolm@redhat.com>
2418 * gengtype-parse.c (require_without_advance): New.
2419 (type): For GTY-marked types that are not GTY((user)), parse any
2420 base classes, requiring them to be single-inheritance, and not
2421 be templates. For non-GTY-marked types and GTY((user)),
2422 continue to skip over any C++ inheritance specification.
2423 * gengtype-state.c (state_writer::write_state_struct_type):
2424 Write base class of type (if any).
2425 (read_state_struct_type): Read base class of type (if any).
2426 * gengtype.c (new_structure): Add a "base_class" parameter.
2427 (create_optional_field_): Update for new parameter to new_structure.
2428 (adjust_field_rtx_def): Likewise.
2429 (adjust_field_tree_exp): Likewise.
2430 * gengtype.h (struct type): Add "base_class" field to the s
2432 (new_structure): Add "base" parameter.
2434 2013-10-23 Sriraman Tallam <tmsriram@google.com>
2437 * config/i386/i386.c (ix86_option_override_internal):
2438 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
2439 (ix86_valid_target_attribute_tree):
2440 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
2441 Change TARGET_SSE to TARGET_SSE_P (opts->...)
2443 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2445 * tree-ssa-loop.h: Remove include files.
2446 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2447 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
2448 * graphite-clast-to-gimple.c: Likewise.
2449 * graphite-scop-detection.c: Likewise.
2450 * graphite-sese-to-poly.c: Likewise.
2451 * ipa-inline-analysis.c: Likewise.
2452 * ipa-pure-const.c: Likewise.
2453 * loop-init.c: Likewise.
2454 * passes.c: Likewise.
2455 * predict.c: Likewise.
2456 * tree-cfg.c: Likewise.
2457 * tree-cfgcleanup.c: Likewise.
2458 * tree-chrec.c: Likewise.
2459 * tree-data-ref.c: Likewise.
2460 * tree-loop-distribution.c: Likewise.
2461 * tree-parloops.c: Likewise.
2462 * tree-predcom.c: Likewise.
2463 * tree-scalar-evolution.c: Likewise.
2464 * tree-ssa-address.c: Likewise.
2465 * tree-ssa.c: Likewise.
2466 * tree-ssa-dce.c: Likewise.
2467 * tree-ssa-loop.c: Likewise.
2468 * tree-ssa-loop-im.c: Likewise.
2469 * tree-ssa-loop-ivcanon.c: Likewise.
2470 * tree-ssa-loop-ivopts.c: Likewise.
2471 * tree-ssa-loop-manip.c: Likewise.
2472 * tree-ssa-loop-niter.c: Likewise.
2473 * tree-ssa-loop-prefetch.c: Likewise.
2474 * tree-ssa-loop-unswitch.c: Likewise.
2475 * tree-ssa-reassoc.c: Likewise.
2476 * tree-vect-data-refs.c: Likewise.
2477 * tree-vect-loop.c: Likewise.
2478 * tree-vect-loop-manip.c: Likewise.
2479 * tree-vectorizer.c: Likewise.
2480 * tree-vect-stmts.c: Likewise.
2481 * tree-vrp.c: Likewise.
2483 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2485 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
2487 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2489 PR tree-optimization/58775
2490 PR tree-optimization/58791
2491 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
2492 (insert_stmt_after): Rewritten, don't move the stmt, but really
2494 (get_stmt_uid_with_default): Remove.
2495 (build_and_add_sum): Use insert_stmt_after and
2496 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
2497 (update_range_test): Set uid on stmts added by
2498 force_gimple_operand_gsi. Don't immediately modify statements
2499 in inter-bb optimization, just update oe->op values.
2500 (optimize_range_tests): Return bool whether any changed have been made.
2501 (update_ops): New function.
2502 (struct inter_bb_range_test_entry): New type.
2503 (maybe_optimize_range_tests): Perform statement changes here.
2504 (not_dominated_by, appears_later_in_bb, get_def_stmt,
2505 ensure_ops_are_available): Remove.
2506 (find_insert_point): Rewritten.
2507 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
2508 return LHS of the (new resp. old) stmt. Don't call
2509 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
2510 instead of last, move new stmt at the right place.
2511 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
2512 (negate_value): Likewise. Set uids.
2513 (break_up_subtract_bb): Initialize uids.
2514 (reassociate_bb): Adjust rewrite_expr_tree caller.
2515 (do_reassoc): Don't call renumber_gimple_stmt_uids.
2517 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
2520 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
2521 TARGET_32BIT final condition.
2522 (mulsi3_internal2 and splitter): Same.
2524 2013-10-23 Jeff Law <law@redhat.com>
2526 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
2527 through joiner blocks with abnormal outgoing edges.
2529 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
2530 Add parameter JOINERS, to allow/disallow threading through joiner
2532 (thread_block): New. Call thread_block_1.
2533 (mark_threaded_blocks): Remove code to filter out certain cases
2534 of threading through joiner blocks.
2535 (thread_through_all_blocks): Document how we can have a dangling
2536 edge AUX field and clear it.
2538 2013-10-23 Ian Lance Taylor <iant@google.com>
2540 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
2541 (C++ Dialect Options): Likewise.
2542 (Optimize Options): Likewise.
2544 2013-10-23 Tom de Vries <tom@codesourcery.com>
2546 PR tree-optimization/58805
2547 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
2549 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2551 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
2552 sequence based on get_range_info returned range.
2554 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2556 * tree-ssa.h: Remove all #include's
2557 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2558 * alias.c: Move required includes from tree-ssa.h.
2560 * builtins.c: Likewise.
2561 * calls.c: Likewise.
2562 * cfgexpand.c: Likewise.
2563 * cfghooks.c: Likewise.
2564 * cfgloop.c: Likewise.
2565 * cfgloopmanip.c: Likewise.
2566 * cgraph.c: Likewise.
2567 * cgraphbuild.c: Likewise.
2568 * cgraphclones.c: Likewise.
2569 * cgraphunit.c: Likewise.
2571 * except.c: Likewise.
2573 * final.c: Likewise.
2574 * fold-const.c: Likewise.
2575 * ggc-page.c: Likewise.
2576 * gimple-builder.c: Likewise.
2577 * gimple-fold.c: Likewise.
2578 * gimple-iterator.c: Likewise.
2579 * gimple-low.c: Likewise.
2580 * gimple-pretty-print.c: Likewise.
2581 * gimple-ssa-strength-reduction.c: Likewise.
2582 * gimple-streamer-in.c: Likewise.
2583 * gimple-streamer-out.c: Likewise.
2584 * gimplify.c: Likewise.
2585 * graphite-blocking.c: Likewise.
2586 * graphite-clast-to-gimple.c: Likewise.
2587 * graphite-dependences.c: Likewise.
2588 * graphite-interchange.c: Likewise.
2589 * graphite-optimize-isl.c: Likewise.
2590 * graphite-poly.c: Likewise.
2591 * graphite-scop-detection.c: Likewise.
2592 * graphite-sese-to-poly.c: Likewise.
2593 * graphite.c: Likewise.
2594 * ipa-cp.c: Likewise.
2595 * ipa-inline-analysis.c: Likewise.
2596 * ipa-inline-transform.c: Likewise.
2597 * ipa-inline.c: Likewise.
2598 * ipa-prop.c: Likewise.
2599 * ipa-pure-const.c: Likewise.
2600 * ipa-reference.c: Likewise.
2601 * ipa-split.c: Likewise.
2602 * ipa-utils.c: Likewise.
2603 * loop-init.c: Likewise.
2604 * lto-cgraph.c: Likewise.
2605 * lto-section-in.c: Likewise.
2606 * lto-section-out.c: Likewise.
2607 * lto-streamer-in.c: Likewise.
2608 * lto-streamer-out.c: Likewise.
2609 * lto-streamer.c: Likewise.
2610 * omp-low.c: Likewise.
2611 * passes.c: Likewise.
2612 * predict.c: Likewise.
2613 * print-tree.c: Likewise.
2614 * profile.c: Likewise.
2616 * targhooks.c: Likewise.
2617 * tracer.c: Likewise.
2618 * trans-mem.c: Likewise.
2619 * tree-call-cdce.c: Likewise.
2620 * tree-cfg.c: Likewise.
2621 * tree-cfgcleanup.c: Likewise.
2622 * tree-chrec.c: Likewise.
2623 * tree-complex.c: Likewise.
2624 * tree-data-ref.c: Likewise.
2625 * tree-dfa.c: Likewise.
2626 * tree-eh.c: Likewise.
2627 * tree-emutls.c: Likewise.
2628 * tree-if-conv.c: Likewise.
2629 * tree-inline.c: Likewise.
2630 * tree-into-ssa.c: Likewise.
2631 * tree-loop-distribution.c: Likewise.
2632 * tree-mudflap.c: Likewise.
2633 * tree-nested.c: Likewise.
2634 * tree-nrv.c: Likewise.
2635 * tree-object-size.c: Likewise.
2636 * tree-outof-ssa.c: Likewise.
2637 * tree-parloops.c: Likewise.
2638 * tree-phinodes.c: Likewise.
2639 * tree-predcom.c: Likewise.
2640 * tree-pretty-print.c: Likewise.
2641 * tree-profile.c: Likewise.
2642 * tree-scalar-evolution.c: Likewise.
2643 * tree-sra.c: Likewise.
2644 * tree-ssa-address.c: Likewise.
2645 * tree-ssa-alias.c: Likewise.
2646 * tree-ssa-ccp.c: Likewise.
2647 * tree-ssa-coalesce.c: Likewise.
2648 * tree-ssa-copy.c: Likewise.
2649 * tree-ssa-copyrename.c: Likewise.
2650 * tree-ssa-dce.c: Likewise.
2651 * tree-ssa-dom.c: Likewise.
2652 * tree-ssa-dse.c: Likewise.
2653 * tree-ssa-forwprop.c: Likewise.
2654 * tree-ssa-ifcombine.c: Likewise.
2655 * tree-ssa-live.c: Likewise.
2656 * tree-ssa-loop-ch.c: Likewise.
2657 * tree-ssa-loop-im.c: Likewise.
2658 * tree-ssa-loop-ivcanon.c: Likewise.
2659 * tree-ssa-loop-ivopts.c: Likewise.
2660 * tree-ssa-loop-manip.c: Likewise.
2661 * tree-ssa-loop-niter.c: Likewise.
2662 * tree-ssa-loop-prefetch.c: Likewise.
2663 * tree-ssa-loop-unswitch.c: Likewise.
2664 * tree-ssa-loop.c: Likewise.
2665 * tree-ssa-math-opts.c: Likewise.
2666 * tree-ssa-operands.c: Likewise.
2667 * tree-ssa-phiopt.c: Likewise.
2668 * tree-ssa-phiprop.c: Likewise.
2669 * tree-ssa-pre.c: Likewise.
2670 * tree-ssa-propagate.c: Likewise.
2671 * tree-ssa-reassoc.c: Likewise.
2672 * tree-ssa-sccvn.c: Likewise.
2673 * tree-ssa-sink.c: Likewise.
2674 * tree-ssa-strlen.c: Likewise.
2675 * tree-ssa-structalias.c: Likewise.
2676 * tree-ssa-tail-merge.c: Likewise.
2677 * tree-ssa-ter.c: Likewise.
2678 * tree-ssa-threadedge.c: Likewise.
2679 * tree-ssa-threadupdate.c: Likewise.
2680 * tree-ssa-uncprop.c: Likewise.
2681 * tree-ssa-uninit.c: Likewise.
2682 * tree-ssa.c: Likewise.
2683 * tree-ssanames.c: Likewise.
2684 * tree-stdarg.c: Likewise.
2685 * tree-streamer-in.c: Likewise.
2686 * tree-switch-conversion.c: Likewise.
2687 * tree-tailcall.c: Likewise.
2688 * tree-vect-data-refs.c: Likewise.
2689 * tree-vect-generic.c: Likewise.
2690 * tree-vect-loop-manip.c: Likewise.
2691 * tree-vect-loop.c: Likewise.
2692 * tree-vect-patterns.c: Likewise.
2693 * tree-vect-slp.c: Likewise.
2694 * tree-vect-stmts.c: Likewise.
2695 * tree-vectorizer.c: Likewise.
2696 * tree-vrp.c: Likewise.
2699 * value-prof.c: Likewise.
2700 * var-tracking.c: Likewise.
2701 * varpool.c: Likewise.
2702 * vtable-verify.c: Likewise.
2704 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2706 * config/tilegx/tilegx.c: Include "tree.h".
2708 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2710 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
2711 the info, not after it.
2712 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
2713 dump_ssaname_info call.
2714 (pp_gimple_stmt_1): Adjust caller.
2715 (dump_phi_nodes): Likewise. Don't print "# " here.
2717 2013-10-22 Jan Hubicka <jh@suse.cz>
2719 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
2721 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
2722 * i386.c (expand_small_movmem_or_setmem): New function.
2723 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
2725 (alg_usable_p): Add support for value ranges; cleanup.
2726 (ix86_expand_set_or_movmem): Add support for misaligned moves.
2728 2013-10-22 Sterling Augustine <saugustine@google.com>
2730 * doc/invoke.texi: Document -ggnu-pubnames.
2731 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
2733 * dwarf2out.c: Include gdb/gdb-index.h.
2734 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
2735 debug_generate_pub_sections.
2736 (is_java, output_pubtables, output_pubname): New functions.
2737 (include_pubname_in_output): Handle debug_generate_pub_sections at
2739 (size_of_pubnames): Use new local space_for_flags based on
2740 debug_generate_pub_sections.
2741 (output_pubnames): Unify pubnames and pubtypes output logic.
2742 Genericize comments. Call output_pubname.
2743 (dwarf2out_finish): Move logic to output_pubnames and call it.
2745 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
2748 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
2749 Remove CCCmode handling.
2750 <case LTU>: Return 'c' suffix for CCCmode.
2751 <case GEU>: Return 'nc' suffix for CCCmode.
2752 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
2753 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
2754 (*sub<mode>3_cc_overflow): Ditto.
2755 (*subsi3_zext_cc_overflow): Ditto.
2757 2013-10-22 Steve Ellcey <sellcey@mips.com>
2759 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
2760 (AND (NOT OP1) (NOT OP2)).
2762 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
2764 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2765 meaning of merge-high and merge-low masks for little endian; avoid
2766 use of vector-pack masks for little endian for mismatched modes.
2768 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2770 * config/tilepro/tilepro.c: Include "tree.h".
2772 2013-10-22 Andreas Schwab <schwab@suse.de>
2774 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
2777 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
2779 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
2780 [Ed Smith-Rowland]): New entries.
2781 ([Stephen M. Webb]): Update.
2783 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
2785 * tree-ssa-ter.h: Remove duplicate copy of file contents.
2787 2013-10-21 Marek Polacek <polacek@redhat.com>
2790 * fold-const.c (fold_range_test): Return 0 if the type is not
2793 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
2795 * system.h: Move hwint.h include further down.
2796 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
2797 gcc_checking_asserts.
2798 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
2800 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2802 Fix volatile issues in optimize_bit_field_compare.
2803 * fold-const.c (optimize_bit_field_compare): Bail out if
2804 lvolatilep or rvolatilep.
2806 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2808 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
2809 and get_inner_reference returning different pmode for non-volatile
2810 bit-field members dependent on flag_strict_volatile_bitfields.
2811 * stor-layout.c (layout_decl): Remove special handling of
2812 flag_strict_volatile_bitfields.
2813 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
2814 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
2816 2013-10-21 Paulo Matos <pmatos@broadcom.com>
2818 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
2821 2013-10-21 Jeff Law <law@redhat.com>
2823 * tree-ssa-threadedge.c (thread_through_normal_block): New
2824 argument VISITED. Remove VISISTED as a local variable. When we
2825 have a threadable jump, verify the destination of the jump has not
2827 (thread_across_edge): Allocate VISITED bitmap once at function
2828 scope and use it throughout. Make sure to set appropriate bits in
2829 VISITED for E (start of jump thread path).
2830 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
2831 through a joiner if any edge on the path has a recorded jump thread.
2833 2013-10-21 Ian Lance Taylor <iant@google.com>
2835 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
2836 don't imply that attributes can solve all problems.
2837 (Directory Options): Fix typo.
2839 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2841 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
2842 extend and extend_add.
2844 2013-10-21 Richard Biener <rguenther@suse.de>
2846 PR tree-optimization/58794
2847 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
2848 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
2850 2013-10-21 Richard Biener <rguenther@suse.de>
2853 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
2854 to (T) X for sign-changing conversions (or no conversion).
2856 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
2858 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
2860 2013-10-20 Jan Hubicka <jh@suse.cz>
2862 * config/i386/i386-tune.def: Add comment; organize into categories
2864 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2866 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
2867 argument. Update function comment.
2868 (expand_set_or_movmem_via_rep): New function combining
2869 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
2870 (expand_movmem_via_rep_mov): Remove.
2871 expand_setmem_via_rep_stos): Remove.
2872 (expand_movmem_epilogue): Update calls correspondingly.
2873 (expand_setmem_epilogue_via_loop): Likewise.
2875 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
2876 (expand_set_or_movmem_prologue): New function combining
2877 expand_movmem_prologue and expand_setmem_prologue.
2878 (expand_movmem_prologue): Remove.
2879 (expand_setmem_prologue): Remove.
2880 (expand_set_or_movmem_constant_prologue): New function combining
2881 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
2882 (expand_constant_movmem_prologue): Remove.
2883 (expand_constant_setmem_prologue): Remove.
2884 (promote_duplicated_reg): Allow vector-const0 value.
2885 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
2886 and ix86_expand_setmem.
2887 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
2888 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
2890 2013-10-21 Diego Novillo <dnovillo@google.com>
2892 * asan.c: Include tree.h
2893 * bb-reorder.c: Likewise.
2894 * cfgcleanup.c: Likewise.
2895 * cfgloopmanip.c: Likewise.
2896 * data-streamer-in.c: Likewise.
2897 * data-streamer-out.c: Likewise.
2898 * data-streamer.c: Likewise.
2899 * dwarf2cfi.c: Likewise.
2900 * graphite-blocking.c: Likewise.
2901 * graphite-clast-to-gimple.c: Likewise.
2902 * graphite-dependences.c: Likewise.
2903 * graphite-interchange.c: Likewise.
2904 * graphite-optimize-isl.c: Likewise.
2905 * graphite-poly.c: Likewise.
2906 * graphite-scop-detection.c: Likewise.
2907 * graphite-sese-to-poly.c: Likewise.
2908 * graphite.c: Likewise.
2909 * ipa-devirt.c: Likewise.
2910 * ipa-profile.c: Likewise.
2913 * loop-init.c: Likewise.
2914 * loop-unroll.c: Likewise.
2915 * lower-subreg.c: Likewise.
2916 * lto/lto-object.c: Likewise.
2917 * recog.c: Likewise.
2918 * reginfo.c: Likewise.
2919 * tree-loop-distribution.c: Likewise.
2920 * tree-parloops.c: Likewise.
2921 * tree-ssa-strlen.c: Likewise.
2922 * tree-streamer.c: Likewise.
2923 * value-prof.c: Likewise.
2924 * target-globals.c: Likewise.
2925 * expr.h: Include tree-core.h instead of tree.h.
2926 * gimple.h: Likewise.
2927 * ipa-prop.h: Likewise.
2928 * ipa-utils.h: Likewise.
2929 * lto-streamer.h: Likewise.
2930 * streamer-hooks.h: Likewise.
2931 * ipa-reference.h: Include cgraph.h instead of tree.h.
2932 * cgraph.h: Include basic-block.h instead of tree.h.
2933 * tree-streamer.h: Do not include tree.h.
2934 * genattrtab.c (write_header): Generate inclusion of
2936 * genautomata.c (main): Likewise.
2937 * genemit.c: Likewise.
2938 * genopinit.c: Likewise.
2939 * genoutput.c (output_prologue): Likewise.
2940 * genpeep.c: Likewise.
2942 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2944 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
2946 (vec_unpacku_hi_v8hi): Likewise.
2947 (vec_unpacku_lo_v16qi): Likewise.
2948 (vec_unpacku_lo_v8hi): Likewise.
2950 2013-10-20 Jan Hubicka <jh@suse.cz>
2952 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
2953 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
2954 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
2956 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
2958 * config/mips/mips.h (ISA_HAS_WSBH): Define.
2959 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
2961 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
2963 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
2966 * system.h: Undef m_slot.
2968 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2970 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
2971 all elements for both endian flavors.
2973 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2976 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
2977 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
2978 and SI_REG for 64bit SYSV ABI targets.
2980 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2982 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
2983 to multi_reg_return. Clarify that we are skipping USEs of multiple
2984 return registers. Use bool type where appropriate.
2986 2013-10-18 Jan Hubicka <jh@suse.cz>
2988 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
2990 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
2991 (X86_TUNE_PUSH_MEMORY): Likewise.
2992 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
2993 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
2994 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
2996 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
2997 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
2999 (ix86_option_override_internal): Update to use tune features instead
3002 2013-10-18 Cong Hou <congh@google.com>
3004 PR tree-optimization/58508
3005 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
3006 statement that contains data refs with zero-step.
3008 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3010 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
3011 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
3012 * gimple-low.c (gimple_check_call_arg,
3013 gimple_check_call_matching_types): Move to cgraph.c.
3014 * gimple-low.h: Remove prototype.
3015 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
3016 Relocate from gimple-low.c.
3017 * cgraph.h: Add prototype. Don't include basic-block.h.
3018 * gimplify.c: Add gimple-low to include list.
3019 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
3020 * tree-eh.c: Add gimple-low to include list.
3021 * tree-nested.c: Likewise.
3022 * cfgexpand.c: Add tree-ssa-address.h to include list.
3024 * gimple-fold.c: Likewise.
3025 * gimple-ssa-strength-reduction.c: Likewise.
3026 * trans-mem.c: Likewise.
3027 * tree-mudflap.c: Likewise.
3028 * tree-ssa-loop-ivopts.c: Likewise.
3029 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
3030 (degenerate_phi_result): Move to tree-phinodes.c.
3031 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
3032 * tree-ssa-threadedge.c: Likewise.
3033 * tree-vrp.c: Likewise.
3034 * tree-phinodes.c (degenerate_phi_result): Relocate here.
3035 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
3036 * tree-phinodes.h (degenerate_phi_result): Add prototype.
3037 * tree-ssa-copy.c: Include tree-ssa-dom.h.
3038 * tree-ssa-forwprop.c: Likewise.
3039 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
3040 pass_data_cleanup_cfg_post_optimizing,
3041 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
3042 * tree-optimize.c: Delete File.
3043 * graphite.c: Include tree-cfgcleanup.h.
3044 * passes.c: Likewise.
3045 * tree-cfg.c: Likewise.
3046 * tree-profile.c: Likewise.
3047 * tree-ssa-dse.c: Likewise.
3048 * tree-ssa-loop-ivcanon.c: Likewise.
3049 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
3050 * tree-outof-ssa.c: Include sbitmap.h.
3051 * tree-ssa-live.c: Likewise.
3052 * tree-ssa-propagate.c: Likewise.
3053 * tree-ssa-structalias.c: Likewise.
3054 * tree-stdarg.c: Likewise.
3055 * Makefile.in (OBJS): Delete tree-optimize.o.
3056 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
3057 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
3058 * varasm.c: Include basic-block.h.
3059 * cfgloop.h: Include function.h instead of basic-block.h
3060 (bb_loop_depth): Move to cfgloop.c.
3061 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
3063 2013-10-18 Teresa Johnson <tejohnson@google.com>
3065 * predict.c (probably_never_executed): Compare frequency-based
3066 count to number of training runs.
3067 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
3069 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3071 * config/arm/arm.c (cortexa9_extra_costs): New table.
3072 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
3074 2013-10-18 Jeff Law <law@redhat.com>
3076 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
3078 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
3079 (register_jump_thread): Add "registered_jump_thread" debug
3081 * dbgcnt.def (registered_jump_thread): New debug counter.
3083 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3085 * config/rs6000/rs6000.c: Include cgraph.h.
3087 2013-10-18 Teresa Johnson <tejohnson@google.com>
3089 * tree-ssa-tail-merge.c (replace_block_by): Update edge
3090 weights during merging.
3092 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3094 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
3095 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
3096 * tree-ssa-operands.h: Remove prototype.
3097 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
3098 * gimple.c (virtual_operand_p): Relocate from gimple.c.
3099 * gimple.h: Add prototype.
3100 * gimple-ssa.h: Include tree-ssa-operands.h.
3101 * tree-dump.c: Add tree-cfg.h to include list.
3102 * tree-ssa-alias.c: Add ipa-reference.h to include list.
3103 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
3104 * config/i386/i386.c: Don't include tree-flow.h.
3105 * config/rs6000/rs6000.c: Likewise.
3107 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3109 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
3111 2013-10-18 Richard Biener <rguenther@suse.de>
3113 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
3114 or TYPE_UNSIGNED of integral types.
3115 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
3116 NULL_TREE for zero-precision integral types.
3118 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
3120 * config/aarch64/arm_neon.h
3121 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
3123 2013-10-17 Sriraman Tallam <tmsriram@google.com>
3126 * opth-gen.awk: Define target_flags_explicit.
3128 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3130 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
3131 fields to the reg_addr array that describes the valid addressing
3132 mode for any register, general purpose registers, floating point
3133 registers, and Altivec registers.
3134 (FIRST_RELOAD_REG_CLASS): Likewise.
3135 (LAST_RELOAD_REG_CLASS): Likewise.
3136 (struct reload_reg_map_type): Likewise.
3137 (reload_reg_map_type): Likewise.
3138 (RELOAD_REG_VALID): Likewise.
3139 (RELOAD_REG_MULTIPLE): Likewise.
3140 (RELOAD_REG_INDEXED): Likewise.
3141 (RELOAD_REG_OFFSET): Likewise.
3142 (RELOAD_REG_PRE_INCDEC): Likewise.
3143 (RELOAD_REG_PRE_MODIFY): Likewise.
3144 (reg_addr): Likewise.
3145 (mode_supports_pre_incdec_p): New helper functions to say whether
3146 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
3147 (mode_supports_pre_modify_p): Likewise.
3148 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
3149 print the valid address mode bits for each mode.
3150 (rs6000_debug_print_mode): Likewise.
3151 (rs6000_debug_reg_global): Likewise.
3152 (rs6000_setup_reg_addr_masks): New function to set up the address
3153 mask bits for each type.
3154 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
3155 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
3156 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
3157 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
3158 PRE_MODIFY are supported.
3159 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
3160 registers, instead of {src,dest}_av_p.
3161 (rs6000_print_options_internal): Tweak the debug output slightly.
3163 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3165 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
3168 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
3170 * tree-flow.h (struct omp_region): Move to omp-low.c.
3171 Remove omp_ prototypes and variables.
3172 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
3173 (copy_var_decl): Relocate prototype from tree-flow.h.
3174 * gimple.c (copy_var_decl): Relocate from omp-low.c.
3175 * tree.h: Move prototype to omp-low.h.
3176 * omp-low.h: New File. Relocate prototypes here.
3177 * omp-low.c (struct omp_region): Make local here.
3178 (root_omp_region): Make static.
3179 (copy_var_decl) Move to gimple.c.
3180 (new_omp_region): Make static.
3181 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
3182 * tree-cfg.c: Include omp-low.h.
3183 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
3184 * gimplify.c: Include omp-low.h.
3185 * tree-parloops.c: Likewise.
3187 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3189 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
3190 operands are in memory, prefer to force non-matched operand 1 to
3193 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3196 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
3197 restrict TImode addresses to single indirect registers if both
3198 -mquad-memory and -mvsx-timode are used.
3199 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
3200 we should emit load/store quad. Remove using %y for quad memory
3203 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
3204 constraints to allow load/store quad on machines where TImode is
3205 not allowed in VSX registers. Use 'n' instead of 'F' constraint
3206 for TImode to load integer constants.
3208 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3210 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
3212 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
3214 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
3215 handling of STACK_REG.
3217 2013-10-17 Richard Biener <rguenther@suse.de>
3219 PR tree-optimization/58143
3220 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
3222 (rewrite_to_defined_overflow): Likewise.
3223 (move_computations_dom_walker::before_dom): Rewrite stmts
3224 with undefined signed overflow that are not always executed
3225 into unsigned arithmetic.
3227 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3230 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
3231 explicit isa flag to be an options variable, instead of using
3232 global_options_set. Remove define from rs6000.h.
3233 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
3235 * config/rs6000/rs6000.c (rs6000_option_override_internal):
3236 Initialize rs6000_isa_flags_explicit.
3237 (rs6000_function_specific_save): Add gcc_options* parameter, so
3238 that the powerpc builds after the 2013-10-15 changes.
3239 (rs6000_function_specific_restore): Likewise.
3241 2013-10-16 DJ Delorie <dj@redhat.com>
3243 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
3244 op is a REG before checking REGNO.
3245 (rl78_alloc_physical_registers): Verify pattern is a SET before
3248 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3250 * config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
3252 (vec_unpacks_lo_v4sf): Likewise.
3253 (vec_unpacks_float_hi_v4si): Likewise.
3254 (vec_unpacks_float_lo_v4si): Likewise.
3255 (vec_unpacku_float_hi_v4si): Likewise.
3256 (vec_unpacku_float_lo_v4si): Likewise.
3258 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3260 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
3261 (vsx_concat_v2sf): Likewise.
3263 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
3265 * config/aarch64/aarch64.md
3266 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
3268 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
3270 PR tree-optimization/58697
3271 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
3272 estimated_loop_iterations_int.
3273 (max_stmt_executions_int): Call get_max_loop_iterations_int.
3274 (get_max_loop_iterations_int): New. HWINT version of
3275 get_max_loop_iterations.
3276 * cfgloop.h: Add prototypes.
3277 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
3278 * loop-unroll.c (decide_peel_once_rolling): Call
3279 get_estimated_loop_iterations_int.
3280 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
3281 * tree-ssa-loop-niter.h: Tweak prototypes.
3283 2013-10-16 David Malcolm <dmalcolm@redhat.com>
3285 * gengtype-parse.c (struct_field_seq): Ignore access-control
3286 keywords ("public:" etc).
3288 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
3290 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
3291 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
3293 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
3295 * config/arm/arm.opt (mlra): New option.
3296 * config/arm/arm.c (arm_lra_p): New function.
3297 (TARGET_LRA_P): Define.
3299 2013-10-16 Paulo Matos <pmatos@broadcom.com>
3301 * tree-core.h (tree_code_name): Remove.
3302 * tree.h (get_tree_code_name): New prototype.
3303 * tree.c (tree_code_name): Make static.
3304 (get_tree_code_name): New function.
3305 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
3306 tree_class_check_failed, tree_range_check_failed,
3307 tree_not_class_check_failed, omp_clause_check_failed,
3308 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
3309 wrapper get_tree_code_name instead of calling tree_code_name directly.
3310 * tree-vrp.c (dump_asserts_for): Likewise.
3311 * tree-dump.c (dequeue_and_dump): Likewise.
3312 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
3313 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
3314 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
3315 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
3316 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
3317 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
3318 dump_gimple_omp_for): Likewise.
3319 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
3320 * tree-ssa-pre.c (print_pre_expr): Likewise.
3321 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
3322 * print-tree.c (print_node_brief, print_node): Likewise.
3323 * gimple.c (gimple_check_failed): Likewise.
3324 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
3325 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
3326 * config/mep/mep.c (mep_validate_vliw): Likewise.
3327 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
3328 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
3330 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
3332 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
3335 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
3337 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
3338 (MULTILIB_MATCHES): Add multilib for -march=v8.
3340 2013-10-15 Sriraman Tallam <tmsriram@google.com>
3343 * optc-save-gen.awk: Add extra parameter to the save and restore
3345 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
3346 * tree.c (build_optimization_node): New parameter. Add extra parameter
3347 to call to cl_optimization_save.
3348 (build_target_option_node): New parameter. Add extra parameter
3349 to call to cl_target_option_save.
3350 * tree.h (build_optimization_node): New parameter.
3351 (build_target_option_node): New parameter.
3352 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
3353 build_optimization_node and build_target_option_node.
3354 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
3355 (handle_pragma_push_options): Ditto.
3356 * toplev.c (process_options): Ditto.
3357 * opts.c (init_options_struct): Check for opts_set non-null.
3358 * target.def (target_option.save): New parameter.
3359 (target_option.restore): New parameter.
3360 * tm.texi: Generate.
3361 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
3362 (ix86_pragma_target_parse): Ditto.
3363 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
3365 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
3366 to build_optimization_node and build_target_option_node.
3367 (rs6000_valid_attribute_p): Ditto.
3368 (rs6000_pragma_target_parse): Ditto.
3369 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
3371 * config/i386/i386.h:
3372 TARGET_64BIT_P: New Macro
3373 TARGET_MMX_P: New Macro.
3374 TARGET_3DNOW_P: New Macro.
3375 TARGET_3DNOW_A_P: New Macro.
3376 TARGET_SSE_P: New Macro.
3377 TARGET_SSE2_P: New Macro.
3378 TARGET_SSE3_P: New Macro.
3379 TARGET_SSSE3_P: New Macro.
3380 TARGET_SSE4_1_P: New Macro.
3381 TARGET_SSE4_2_P: New Macro.
3382 TARGET_AVX_P: New Macro.
3383 TARGET_AVX2_P: New Macro.
3384 TARGET_AVX512F_P: New Macro.
3385 TARGET_AVX512PF_P: New Macro.
3386 TARGET_AVX512ER_P: New Macro.
3387 TARGET_AVX512CD_P: New Macro.
3388 TARGET_FMA_P: New Macro.
3389 TARGET_SSE4A_P: New Macro.
3390 TARGET_FMA4_P: New Macro.
3391 TARGET_XOP_P: New Macro.
3392 TARGET_LWP_P: New Macro.
3393 TARGET_ABM_P: New Macro.
3394 TARGET_BMI_P: New Macro.
3395 TARGET_BMI2_P: New Macro.
3396 TARGET_LZCNT_P: New Macro.
3397 TARGET_TBM_P: New Macro.
3398 TARGET_POPCNT_P: New Macro.
3399 TARGET_SAHF_P: New Macro.
3400 TARGET_MOVBE_P: New Macro.
3401 TARGET_CRC32_P: New Macro.
3402 TARGET_AES_P: New Macro.
3403 TARGET_PCLMUL_P: New Macro.
3404 TARGET_CMPXCHG16B_P: New Macro.
3405 TARGET_FSGSBASE_P: New Macro.
3406 TARGET_RDRND_P: New Macro.
3407 TARGET_F16C_P: New Macro.
3408 TARGET_RTM_P: New Macro.
3409 TARGET_HLE_P: New Macro.
3410 TARGET_RDSEED_P: New Macro.
3411 TARGET_PRFCHW_P: New Macro.
3412 TARGET_ADX_P: New Macro.
3413 TARGET_FXSR_P: New Macro.
3414 TARGET_XSAVE_P: New Macro.
3415 TARGET_XSAVEOPT_P: New Macro.
3416 TARGET_LP64_P: New Macro.
3417 TARGET_X32_P: New Macro.
3418 TARGET_FPMATH_DEFAULT_P: New Macro.
3419 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
3420 * config/i386/i386.c (ix86_option_override_internal): New parameters.
3422 Change ix86_tune_string to access opts->x_ix86_tune_string.
3423 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
3424 Change ix86_arch_string to access opts->x_ix86_arch_string.
3425 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
3426 Change ix86_pmode to access opts->x_ix86_pmode.
3427 Change ix86_abi to access opts->x_ix86_abi.
3428 Change ix86_cmodel to access opts->x_ix86_cmodel.
3429 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
3430 Change ix86_isa_flags_explicit to access
3431 opts->x_ix86_isa_flags_explicit.
3432 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
3433 Change ix86_regparm to access opts->x_ix86_regparm.
3434 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
3435 Change ix86_preferred_stack_boundary_arg to access
3436 opts->x_ix86_preferred_stack_boundary_arg.
3437 Change ix86_force_align_arg_pointer to access
3438 opts->x_ix86_force_align_arg_pointer.
3439 Change ix86_incoming_stack_boundar_arg to access
3440 opts->x_ix86_incoming_stack_boundar_arg.
3441 Change ix86_fpmath to access opts->x_ix86_fpmath.
3442 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
3443 Change ix86_recip_name to access opts->x_ix86_recip_name.
3444 Change ix86_stack_protector_guard to access
3445 opts->x_ix86_stack_protector_guard.
3446 Change ix86_tune_memcpy_strategy to access
3447 opts->x_ix86_tune_memcpy_strategy.
3448 Change ix86_tune_memset_strategy to access
3449 opts->x_ix86_tune_memset_strategy.
3450 Change global_options to access opts.
3451 Change global_options_set to access opts_set.
3452 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
3453 Change TARGET_MMX to TARGET_MMX_P (opts->...).
3454 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
3455 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
3456 Change TARGET_SSE to TARGET_SSE_P (opts->...).
3457 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
3458 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
3459 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
3460 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
3461 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
3462 Change TARGET_AVX to TARGET_AVX_P (opts->...).
3463 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
3464 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
3465 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
3466 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
3467 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
3468 Change TARGET_FMA to TARGET_FMA_P (opts->...).
3469 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
3470 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
3471 Change TARGET_XOP to TARGET_XOP_P (opts->...).
3472 Change TARGET_LWP to TARGET_LWP_P (opts->...).
3473 Change TARGET_ABM to TARGET_ABM_P (opts->...).
3474 Change TARGET_BMI to TARGET_BMI_P (opts->...).
3475 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
3476 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
3477 Change TARGET_TBM to TARGET_TBM_P (opts->...).
3478 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
3479 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
3480 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
3481 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
3482 Change TARGET_AES to TARGET_AES_P (opts->...).
3483 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
3484 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
3485 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
3486 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
3487 Change TARGET_F16C to TARGET_F16C_P (opts->...).
3488 Change TARGET_RTM to TARGET_RTM_P (opts->...).
3489 Change TARGET_HLE to TARGET_HLE_P (opts->...).
3490 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
3491 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
3492 Change TARGET_ADX to TARGET_ADX_P (opts->...).
3493 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
3494 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
3495 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
3496 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
3497 Change TARGET_X32 to TARGET_X32_P (opts->...).
3498 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
3499 Change TARGET_FLOAT_RETURNS_IN_80387 to
3500 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
3501 (ix86_function_specific_save): New parameter. Use opts-> fields
3502 to replace global fields.
3503 (ix86_function_specific_restore): Ditto.
3504 (ix86_valid_target_attribute_inner_p): New parameters.
3506 Fix call to ix86_handle_option and set_option.
3507 (ix86_valid_target_attribute_tree): New parameters.
3508 Change global_options to access opts.
3509 Change global_options_set to access opts_set.
3510 Fix call to ix86_valid_target_attribute_inner_p.
3511 Change ix86_tune_string to access opts->x_ix86_tune_string.
3512 Change ix86_arch_string to access opts->x_ix86_arch_string.
3513 Change ix86_fpmath to access opts->x_ix86_fpmath
3514 Fix call to ix86_option_override_internal.
3515 Fix call to ix86_add_new_builtins.
3516 Fix calls to build_optimization_node and build_target_option_node.
3517 (ix86_valid_target_attribute_p): Remove access to global_options.
3518 Use new gcc_options structure func_options.
3519 Fix call to ix86_valid_target_attribute_tree.
3520 Fix call to build_optimization_node.
3521 (get_builtin_code_for_version): Fix call to
3522 ix86_valid_target_attribute_tree.
3524 2013-10-15 David Malcolm <dmalcolm@redhat.com>
3526 * Makefile.in (PICFLAG): New.
3527 (enable_host_shared): New.
3528 (INTERNAL_CFLAGS): Use PICFLAG.
3529 (LIBIBERTY): Use pic build of libiberty.a if configured with
3530 --enable-host-shared.
3531 * configure.ac: Add --enable-host-shared, setting up new
3533 * configure: Regenerate.
3534 * doc/install.texi (--enable-shared): Add note contrasting it with ...
3535 (--enable-host-shared): New option.
3537 2013-10-15 Richard Biener <rguenther@suse.de>
3539 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
3540 for built-in functions.
3542 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
3544 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
3545 (optimize_range_tests_1): New function,
3546 extracted from optimize_range_tests.
3547 (optimize_range_tests_xor): Similarly.
3548 (optimize_range_tests_diff): New function.
3549 (optimize_range_tests): Use optimize_range_tests_1.
3551 2013-10-15 Cong Hou <congh@google.com>
3553 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
3554 requirement of the reduction pattern so that one operand of the
3555 reduction operation can come from outside of the loop.
3557 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3559 * config/arm/neon-schedgen.ml: Remove.
3560 * config/arm/cortex-a9-neon.md: Remove comment regarding
3563 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3565 * config/arm/types: Remove old neon types.
3567 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3569 * config/arm/cortex-a7.md
3570 (cortex_a7_neon_type): New.
3571 (cortex_a7_neon_mul): Update for new types.
3572 (cortex_a7_neon_mla): Likewise.
3573 (cortex_a7_neon): Likewise.
3575 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3577 * config/arm/cortex-a15-neon.md
3578 (cortex_a15_neon_type): New,
3580 (cortex_a15_neon_int_1): Remove.
3581 (cortex_a15_neon_int_2): Likewise.
3582 (cortex_a15_neon_int_3): Likewise.
3583 (cortex_a15_neon_int_4): Likewise.
3584 (cortex_a15_neon_int_5): Likewise.
3585 (cortex_a15_neon_vqneg_vqabs): Likewise.
3586 (cortex_a15_neon_vmov): Likewise.
3587 (cortex_a15_neon_vaba): Likewise.
3588 (cortex_a15_neon_vaba_qqq): Likewise.
3589 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3590 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3591 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3593 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3594 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3595 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
3596 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3597 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3598 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3599 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3600 (cortex_a15_neon_shift_1): Likewise.
3601 (cortex_a15_neon_shift_2): Likewise.
3602 (cortex_a15_neon_shift_3): Likewise.
3603 (cortex_a15_neon_vshl_ddd): Likewise.
3604 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3605 (cortex_a15_neon_vsra_vrsra): Likewise.
3606 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3607 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3608 (cortex_a15_neon_bp_3cycle): Likewise.
3609 (cortex_a15_neon_ldm_2): Likewise.
3610 (cortex_a15_neon_stm_2): Likewise.
3611 (cortex_a15_neon_mcr): Likewise.
3612 (cortex_a15_neon_mrc): Likewise.
3613 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3614 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3615 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3616 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3617 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3618 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3619 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3620 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3621 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3622 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3623 (cortex_a15_neon_bp_simple): Likewise.
3624 (cortex_a15_neon_bp_2cycle): Likewise.
3625 (cortex_a15_neon_bp_3cycle): Likewise.
3626 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3627 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3628 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3629 (cortex_a15_neon_vld2_4_regs): Likewise.
3630 (cortex_a15_neon_vld3_vld4): Likewise.
3631 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3632 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3633 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
3634 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
3635 (cortex_a15_neon_vst3_vst4): Rename to...
3636 (cortex_a15_neon_vst4): This, update for new attributes.
3637 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
3638 (cortex_a15_neon_vld3_vld4_lane): Likewise.
3639 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3640 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3641 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3642 (cortex_a15_neon_ldm_2): Likewise.
3643 (cortex_a15_neon_stm_2): Likewise.
3644 (cortex_a15_neon_mcr): Likewise.
3645 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3646 (cortex_a15_neon_mrc): Likewise.
3647 (cortex_a15_neon_mrrc): Likewise.
3649 (cortex_a15_neon_abd): New.
3650 (cortex_a15_neon_abd_q): Likewise.
3651 (cortex_a15_neon_aba): Likewise.
3652 (cortex_a15_neon_aba_q): Likewise.
3653 (cortex_a15_neon_acc): Likewise.
3654 (cortex_a15_neon_acc_q): Likewise.
3655 (cortex_a15_neon_arith_basic): Likewise.
3656 (cortex_a15_neon_arith_complex): Likewise.
3657 (cortex_a15_neon_multiply): Likewise.
3658 (cortex_a15_neon_multiply_q): Likewise.
3659 (cortex_a15_neon_mla): Likewise.
3660 (cortex_a15_neon_mla_q): Likewise.
3661 (cortex_a15_neon_sat_mla_long): Likewise.
3662 (cortex_a15_neon_shift_acc): Likewise.
3663 (cortex_a15_neon_shift_imm_basic): Likewise.
3664 (cortex_a15_neon_shift_imm_complex): Likewise.
3665 (cortex_a15_neon_shift_reg_basic): Likewise.
3666 (cortex_a15_neon_shift_reg_basic_q): Likewise.
3667 (cortex_a15_neon_shift_reg_complex): Likewise.
3668 (cortex_a15_neon_shift_reg_complex_q): Likewise.
3669 (cortex_a15_neon_fp_negabs): Likewise
3670 (cortex_a15_neon_fp_arith): Likewise
3671 (cortex_a15_neon_fp_arith_q): Likewise
3672 (cortex_a15_neon_fp_cvt_int): Likewise
3673 (cortex_a15_neon_fp_cvt_int_q): Likewise
3674 (cortex_a15_neon_fp_cvt_16): Likewise
3675 (cortex_a15_neon_fp_mul): Likewise
3676 (cortex_a15_neon_fp_mul_q): Likewise
3677 (cortex_a15_neon_fp_mla): Likewise
3678 (cortex_a15_neon_fp_mla_q): Likewise
3679 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
3680 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
3681 (cortex_a15_neon_bitops): Likewise.
3682 (cortex_a15_neon_bitops_q): Likewise.
3683 (cortex_a15_neon_from_gp): Likewise.
3684 (cortex_a15_neon_from_gp_q): Likewise.
3685 (cortex_a15_neon_tbl3_tbl4): Likewise.
3686 (cortex_a15_neon_zip_q): Likewise.
3687 (cortex_a15_neon_to_gp): Likewise.
3688 (cortex_a15_neon_load_a): Likewise.
3689 (cortex_a15_neon_load_b): Likewise.
3690 (cortex_a15_neon_load_c): Likewise.
3691 (cortex_a15_neon_load_d): Likewise.
3692 (cortex_a15_neon_load_e): Likewise.
3693 (cortex_a15_neon_load_f): Likewise.
3694 (cortex_a15_neon_store_a): Likewise.
3695 (cortex_a15_neon_store_b): Likewise.
3696 (cortex_a15_neon_store_c): Likewise.
3697 (cortex_a15_neon_store_d): Likewise.
3698 (cortex_a15_neon_store_e): Likewise.
3699 (cortex_a15_neon_store_f): Likewise.
3700 (cortex_a15_neon_store_g): Likewise.
3701 (cortex_a15_neon_store_h): Likewise.
3702 (cortex_a15_vfp_to_from_gp): Likewise.
3704 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3706 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
3708 (cortex_a9_neon_vshl_ddd): Remove.
3709 (cortex_a9_neon_vst3_vst4): Likewise.
3710 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3712 (cortex_a9_neon_bit_ops_q): New.
3714 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
3715 (cortex_a9_neon_int_2): Likewise.
3716 (cortex_a9_neon_int_3): Likewise.
3717 (cortex_a9_neon_int_4): Likewise.
3718 (cortex_a9_neon_int_5): Likewise.
3719 (cortex_a9_neon_vqneg_vqabs): Likewise.
3720 (cortex_a9_neon_vmov): Likewise.
3721 (cortex_a9_neon_vaba): Likewise.
3722 (cortex_a9_neon_vaba_qqq): Likewise.
3723 (cortex_a9_neon_shift_1): Likewise.
3724 (cortex_a9_neon_shift_2): Likewise.
3725 (cortex_a9_neon_shift_3): Likewise.
3726 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3727 (cortex_a9_neon_vsra_vrsra): Likewise.
3728 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3729 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3730 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3732 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3733 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3734 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3736 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3737 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3738 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3739 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3740 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3741 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3742 (cortex_a9_neon_fp_vsum): Likewise.
3743 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3744 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3745 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3746 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3747 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3748 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3749 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3750 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3751 (cortex_a9_neon_bp_simple): Likewise.
3752 (cortex_a9_neon_bp_2cycle): Likewise.
3753 (cortex_a9_neon_bp_3cycle): Likewise.
3754 (cortex_a9_neon_ldr): Likewise.
3755 (cortex_a9_neon_str): Likewise.
3756 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3757 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3758 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3759 (cortex_a9_neon_vld2_4_regs): Likewise.
3760 (cortex_a9_neon_vld3_vld4): Likewise.
3761 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3762 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3763 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3764 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3765 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3766 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3767 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3768 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3769 (cortex_a9_neon_mcr): Likewise.
3770 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3771 (cortex_a9_neon_mrc): Likewise.
3772 (cortex_a9_neon_mrrc): Likewise.
3774 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3776 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
3778 (cortex_a8_neon_vshl_ddd): Remove.
3779 (cortex_a8_neon_vst3_vst4): Likewise.
3780 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3782 (cortex_a8_neon_bit_ops_q): New.
3784 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
3785 (cortex_a8_neon_int_2): Likewise..
3786 (cortex_a8_neon_int_3): Likewise.
3787 (cortex_a8_neon_int_5): Likewise.
3788 (cortex_a8_neon_vqneg_vqabs): Likewise.
3789 (cortex_a8_neon_int_4): Likewise.
3790 (cortex_a8_neon_vaba): Likewise.
3791 (cortex_a8_neon_vaba_qqq): Likewise.
3792 (cortex_a8_neon_shift_1): Likewise.
3793 (cortex_a8_neon_shift_2): Likewise.
3794 (cortex_a8_neon_shift_3): Likewise.
3795 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3796 (cortex_a8_neon_vsra_vrsra): Likewise.
3797 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3798 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3799 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3801 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3802 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3803 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3805 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3806 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3807 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3808 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3809 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3810 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3811 (cortex_a8_neon_fp_vsum): Likewise.
3812 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3813 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3814 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3815 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3816 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3817 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3818 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3819 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3820 (cortex_a8_neon_bp_simple): Likewise.
3821 (cortex_a8_neon_bp_2cycle): Likewise.
3822 (cortex_a8_neon_bp_3cycle): Likewise.
3823 (cortex_a8_neon_ldr): Likewise.
3824 (cortex_a8_neon_str): Likewise.
3825 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3826 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3827 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3828 (cortex_a8_neon_vld2_4_regs): Likewise.
3829 (cortex_a8_neon_vld3_vld4): Likewise.
3830 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3831 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3832 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3833 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3834 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3835 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3836 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3837 (cortex_a8_neon_mcr): Likewise.
3838 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3839 (cortex_a8_neon_mrc): Likewise.
3840 (cortex_a8_neon_mrrc): Likewise.
3842 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3844 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
3846 * config/aarch64/aarch64-simd.md (neon_type): Remove.
3847 (aarch64_simd_dup<mode>): Add "type" attribute.
3848 (aarch64_dup_lane<mode>): Likewise.
3849 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
3850 (*aarch64_simd_mov<mode>): Likewise.
3851 (aarch64_simd_mov_from_<mode>low): Likewise.
3852 (aarch64_simd_mov_from_<mode>high): Likewise.
3853 (orn<mode>3): Likewise.
3854 (bic<mode>3): Likewise.
3855 (add<mode>3): Likewise.
3856 (sub<mode>3): Likewise.
3857 (mul<mode>3): Likewise.
3858 (*aarch64_mul3_elt<mode>): Likewise.
3859 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
3860 (*aarch64_mul3_elt_to_128df): Likewise.
3861 (*aarch64_mul3_elt_to_64v2df): Likewise.
3862 (neg<mode>2): Likewise.
3863 (abs<mode>2): Likewise.
3864 (abd<mode>_3): Likewise.
3865 (aba<mode>_3): Likewise.
3866 (fabd<mode>_3): Likewise.
3867 (*fabd_scalar<mode>3): Likewise.
3868 (and<mode>3): Likewise.
3869 (ior<mode>3): Likewise.
3870 (xor<mode>3): Likewise.
3871 (one_cmpl<mode>2): Likewise.
3872 (aarch64_simd_vec_set<mode>): Likewise.
3873 (aarch64_simd_lshr<mode>): Likewise.
3874 (aarch64_simd_ashr<mode>): Likewise.
3875 (aarch64_simd_imm_shl<mode>): Likewise.
3876 (aarch64_simd_reg_sshl<mode): Likewise.
3877 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
3878 (aarch64_simd_reg_shl<mode>_signed): Likewise.
3879 (aarch64_simd_vec_setv2di): Likewise.
3880 (aarch64_simd_vec_set<mode>): Likewise.
3881 (aarch64_mla<mode>): Likewise.
3882 (*aarch64_mla_elt<mode>): Likewise.
3883 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
3884 (aarch64_mls<mode>): Likewise.
3885 (*aarch64_mls_elt<mode>): Likewise.
3886 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
3887 (<su><maxmin><mode>3): Likewise.
3888 (move_lo_quad_<mode>): Likewise.
3889 (aarch64_simd_move_hi_quad_<mode>): Likewise.
3890 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
3891 (vec_pack_trunc_<mode>): Likewise.
3892 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
3893 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
3894 (*aarch64_<su>mlal_lo<mode>): Likewise.
3895 (*aarch64_<su>mlal_hi<mode>): Likewise.
3896 (*aarch64_<su>mlsl_lo<mode>): Likewise.
3897 (*aarch64_<su>mlsl_hi<mode>): Likewise.
3898 (*aarch64_<su>mlal<mode>): Likewise.
3899 (*aarch64_<su>mlsl<mode>): Likewise.
3900 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
3901 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
3902 (add<mode>3): Likewise.
3903 (sub<mode>3): Likewise.
3904 (mul<mode>3): Likewise.
3905 (div<mode>3): Likewise.
3906 (neg<mode>2): Likewise.
3907 (abs<mode>2): Likewise.
3908 (fma<mode>4): Likewise.
3909 (*aarch64_fma4_elt<mode>): Likewise.
3910 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
3911 (*aarch64_fma4_elt_to_128df): Likewise.
3912 (*aarch64_fma4_elt_to_64v2df): Likewise.
3913 (fnma<mode>4): Likewise.
3914 (*aarch64_fnma4_elt<mode>): Likewise.
3915 (*aarch64_fnma4_elt_<vswap_width_name><mode>
3916 (*aarch64_fnma4_elt_to_128df): Likewise.
3917 (*aarch64_fnma4_elt_to_64v2df): Likewise.
3918 (<frint_pattern><mode>2): Likewise.
3919 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
3920 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
3921 (vec_unpacks_lo_v4sf): Likewise.
3922 (aarch64_float_extend_lo_v2df): Likewise.
3923 (vec_unpacks_hi_v4sf): Likewise.
3924 (aarch64_float_truncate_lo_v2sf): Likewise.
3925 (aarch64_float_truncate_hi_v4sf): Likewise.
3926 (aarch64_vmls<mode>): Likewise.
3927 (<su><maxmin><mode>3): Likewise.
3928 (<maxmin_uns><mode>3): Likewise.
3929 (reduc_<sur>plus_<mode>): Likewise.
3930 (reduc_<sur>plus_v2di): Likewise.
3931 (reduc_<sur>plus_v2si): Likewise.
3932 (reduc_<sur>plus_<mode>): Likewise.
3933 (aarch64_addpv4sf): Likewise.
3934 (clz<mode>2): Likewise.
3935 (reduc_<maxmin_uns>_<mode>): Likewise.
3936 (reduc_<maxmin_uns>_v2di): Likewise.
3937 (reduc_<maxmin_uns>_v2si): Likewise.
3938 (reduc_<maxmin_uns>_<mode>): Likewise.
3939 (reduc_<maxmin_uns>_v4sf): Likewise.
3940 (aarch64_simd_bsl<mode>_internal): Likewise.
3941 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
3942 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
3943 (aarch64_get_lane<mode>): Likewise.
3944 (*aarch64_combinez<mode>): Likewise.
3945 (aarch64_combine<mode>): Likewise.
3946 (aarch64_simd_combine<mode>): Likewise.
3947 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
3948 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
3949 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
3950 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
3951 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
3952 (aarch64_<sur>h<addsub><mode>): Likewise.
3953 (aarch64_<sur><addsub>hn<mode>): Likewise.
3954 (aarch64_<sur><addsub>hn2<mode>): Likewise.
3955 (aarch64_pmul<mode>): Likewise.
3956 (aarch64_<su_optab><optab><mode>): Likewise.
3957 (aarch64_<sur>qadd<mode>): Likewise.
3958 (aarch64_sqmovun<mode>): Likewise.
3959 (aarch64_<sur>qmovn<mode>): Likewise.
3960 (aarch64_s<optab><mode>): Likewise.
3961 (aarch64_sq<r>dmulh<mode>): Likewise.
3962 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3963 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
3964 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3965 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
3966 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3967 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3968 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
3969 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
3970 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
3971 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
3972 (aarch64_sqdmull<mode>): Likewise.
3973 (aarch64_sqdmull_lane<mode>_internal): Likewise.
3974 (aarch64_sqdmull_n<mode>): Likewise.
3975 (aarch64_sqdmull2<mode>_internal): Likewise.
3976 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
3977 (aarch64_sqdmull2_n<mode>_internal): Likewise.
3978 (aarch64_<sur>shl<mode>): Likewise.
3979 (aarch64_<sur>q<r>shl<mode>
3980 (aarch64_<sur>shll_n<mode>): Likewise.
3981 (aarch64_<sur>shll2_n<mode>): Likewise.
3982 (aarch64_<sur>shr_n<mode>): Likewise.
3983 (aarch64_<sur>sra_n<mode>): Likewise.
3984 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
3985 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
3986 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
3987 (aarch64_cm<optab><mode>): Likewise.
3988 (aarch64_cm<optab>di): Likewise.
3989 (aarch64_cm<optab><mode>): Likewise.
3990 (aarch64_cm<optab>di): Likewise.
3991 (aarch64_cmtst<mode>): Likewise.
3992 (aarch64_cmtstdi): Likewise.
3993 (aarch64_cm<optab><mode>): Likewise.
3994 (*aarch64_fac<optab><mode>): Likewise.
3995 (aarch64_addp<mode>): Likewise.
3996 (aarch64_addpdi): Likewise.
3997 (sqrt<mode>2): Likewise.
3998 (vec_load_lanesoi<mode>): Likewise.
3999 (vec_store_lanesoi<mode>): Likewise.
4000 (vec_load_lanesci<mode>): Likewise.
4001 (vec_store_lanesci<mode>): Likewise.
4002 (vec_load_lanesxi<mode>): Likewise.
4003 (vec_store_lanesxi<mode>): Likewise.
4004 (*aarch64_mov<mode>): Likewise.
4005 (aarch64_ld2<mode>_dreg): Likewise.
4006 (aarch64_ld2<mode>_dreg): Likewise.
4007 (aarch64_ld3<mode>_dreg): Likewise.
4008 (aarch64_ld3<mode>_dreg): Likewise.
4009 (aarch64_ld4<mode>_dreg): Likewise.
4010 (aarch64_ld4<mode>_dreg): Likewise.
4011 (aarch64_tbl1<mode>): Likewise.
4012 (aarch64_tbl2v16qi): Likewise.
4013 (aarch64_combinev16qi): Likewise.
4014 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
4015 (aarch64_st2<mode>_dreg): Likewise.
4016 (aarch64_st2<mode>_dreg): Likewise.
4017 (aarch64_st3<mode>_dreg): Likewise.
4018 (aarch64_st3<mode>_dreg): Likewise.
4019 (aarch64_st4<mode>_dreg): Likewise.
4020 (aarch64_st4<mode>_dreg): Likewise.
4021 (*aarch64_simd_ld1r<mode>): Likewise.
4022 (aarch64_frecpe<mode>): Likewise.
4023 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
4024 (aarch64_frecps<mode>): Likewise.
4026 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4028 * config/arm/iterators.md (V_elem_ch): New.
4030 (VQH_type): Likewise.
4031 * config/arm/arm.md (is_neon_type): New.
4032 (conds): Use is_neon_type.
4033 (anddi3_insn): Update type attribute.
4034 (xordi3_insn): Likewise.
4035 (one_cmpldi2): Likewise.
4036 * config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
4037 * config/arm/neon.md (neon_mov): Update type attribute.
4038 (*movmisalign<mode>_neon_store): Likewise.
4039 (*movmisalign<mode>_neon_load): Likewise.
4040 (vec_set<mode>_internal): Likewise.
4041 (vec_set<mode>_internal): Likewise.
4042 (vec_setv2di_internal): Likewise.
4043 (vec_extract<mode>): Likewise.
4044 (vec_extract<mode>): Likewise.
4045 (vec_extractv2di): Likewise.
4046 (*add<mode>3_neon): Likewise.
4047 (adddi3_neon): Likewise.
4048 (*sub<mode>3_neon): Likewise.
4049 (subdi3_neon): Likewise.
4050 (fma<VCVTF:mode>4): Likewise.
4051 (fma<VCVTF:mode>4_intrinsic): Likewise.
4052 (*fmsub<VCVTF:mode>4): Likewise.
4053 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
4054 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
4055 (ior<mode>3): Likewise.
4056 (and<mode>3): Likewise.
4057 (orn<mode>3_neon): Likewise.
4058 (orndi3_neon): Likewise.
4059 (bic<mode>3_neon): Likewise.
4060 (bicdi3_neon): Likewise.
4061 (xor<mode>3): Likewise.
4062 (one_cmpl<mode>2): Likewise.
4063 (abs<mode>2): Likewise.
4064 (neg<mode>2): Likewise.
4065 (negdi2_neon): Likewise.
4066 (*umin<mode>3_neon): Likewise.
4067 (*umax<mode>3_neon): Likewise.
4068 (*smin<mode>3_neon): Likewise.
4069 (*smax<mode>3_neon): Likewise.
4070 (vashl<mode>3): Likewise.
4071 (vashr<mode>3_imm): Likewise.
4072 (vlshr<mode>3_imm): Likewise.
4073 (ashl<mode>3_signed): Likewise.
4074 (ashl<mode>3_unsigned): Likewise.
4075 (neon_load_count): Likewise.
4076 (ashldi3_neon_noclobber): Likewise.
4077 (ashldi3_neon): Likewise.
4078 (signed_shift_di3_neon): Likewise.
4079 (unsigned_shift_di3_neon): Likewise.
4080 (ashrdi3_neon_imm_noclobber): Likewise.
4081 (lshrdi3_neon_imm_noclobber): Likewise.
4082 (<shift>di3_neon): Likewise.
4083 (widen_ssum<mode>3): Likewise.
4084 (widen_usum<mode>3): Likewise.
4085 (quad_halves_<code>v4si): Likewise.
4086 (quad_halves_<code>v4sf): Likewise.
4087 (quad_halves_<code>v8hi): Likewise.
4088 (quad_halves_<code>v16qi): Likewise.
4089 (reduc_splus_v2di): Likewise.
4090 (neon_vpadd_internal<mode>): Likewise.
4091 (neon_vpsmin<mode>): Likewise.
4092 (neon_vpsmax<mode>): Likewise.
4093 (neon_vpumin<mode>): Likewise.
4094 (neon_vpumax<mode>): Likewise.
4095 (*ss_add<mode>_neon): Likewise.
4096 (*us_add<mode>_neon): Likewise.
4097 (*ss_sub<mode>_neon): Likewise.
4098 (*us_sub<mode>_neon): Likewise.
4099 (neon_vadd<mode>_unspec): Likewise.
4100 (neon_vaddl<mode>): Likewise.
4101 (neon_vaddw<mode>): Likewise.
4102 (neon_vhadd<mode>): Likewise.
4103 (neon_vqadd<mode>): Likewise.
4104 (neon_vaddhn<mode>): Likewise.
4105 (neon_vmul<mode>): Likewise.
4106 (neon_vfms<VCVTF:mode>): Likewise.
4107 (neon_vmlal<mode>): Likewise.
4108 (neon_vmls<mode>): Likewise.
4109 (neon_vmlsl<mode>): Likewise.
4110 (neon_vqdmulh<mode>): Likewise.
4111 (neon_vqdmlal<mode>): Likewise.
4112 (neon_vqdmlsl<mode>): Likewise.
4113 (neon_vmull<mode>): Likewise.
4114 (neon_vqdmull<mode>): Likewise.
4115 (neon_vsub<mode>_unspec): Likewise.
4116 (neon_vsubl<mode>): Likewise.
4117 (neon_vsubw<mode>): Likewise.
4118 (neon_vqsub<mode>): Likewise.
4119 (neon_vhsub<mode>): Likewise.
4120 (neon_vsubhn<mode>): Likewise.
4121 (neon_vceq<mode>): Likewise.
4122 (neon_vcge<mode>): Likewise.
4123 (neon_vcgeu<mode>): Likewise.
4124 (neon_vcgt<mode>): Likewise.
4125 (neon_vcgtu<mode>): Likewise.
4126 (neon_vcle<mode>): Likewise.
4127 (neon_vclt<mode>): Likewise.
4128 (neon_vcage<mode>): Likewise.
4129 (neon_vcagt<mode>): Likewise.
4130 (neon_vtst<mode>): Likewise.
4131 (neon_vabd<mode>): Likewise.
4132 (neon_vabdl<mode>): Likewise.
4133 (neon_vaba<mode>): Likewise.
4134 (neon_vabal<mode>): Likewise.
4135 (neon_vmax<mode>): Likewise.
4136 (neon_vmin<mode>): Likewise.
4137 (neon_vpaddl<mode>): Likewise.
4138 (neon_vpadal<mode>): Likewise.
4139 (neon_vpmax<mode>): Likewise.
4140 (neon_vpmin<mode>): Likewise.
4141 (neon_vrecps<mode>): Likewise.
4142 (neon_vrsqrts<mode>): Likewise.
4143 (neon_vqabs<mode>): Likewise.
4144 (neon_vqneg<mode>): Likewise.
4145 (neon_vcls<mode>): Likewise.
4146 (clz<mode>2): Likewise.
4147 (popcount<mode>2): Likewise.
4148 (neon_vrecpe<mode>): Likewise.
4149 (neon_vrsqrte<mode>): Likewise.
4150 (neon_vget_lane<mode>_sext_internal): Likewise.
4151 (neon_vget_lane<mode>_zext_internal): Likewise.
4152 (neon_vdup_n<mode>): Likewise.
4153 (neon_vdup_n<mode>): Likewise.
4154 (neon_vdup_nv2di): Likewise.
4155 (neon_vdup_lane<mode>_interal): Likewise.
4156 (*neon_vswp<mode>): Likewise.
4157 (neon_vcombine<mode>): Likewise.
4158 (float<mode><V_cvtto>2): Likewise.
4159 (floatuns<mode><V_cvtto>2): Likewise.
4160 (fix_trunc<mode><V_cvtto>2): Likewise.
4161 (fixuns_trunc<mode><V_cvtto>2
4162 (neon_vcvt<mode>): Likewise.
4163 (neon_vcvt<mode>): Likewise.
4164 (neon_vcvtv4sfv4hf): Likewise.
4165 (neon_vcvtv4hfv4sf): Likewise.
4166 (neon_vcvt_n<mode>): Likewise.
4167 (neon_vcvt_n<mode>): Likewise.
4168 (neon_vmovn<mode>): Likewise.
4169 (neon_vqmovn<mode>): Likewise.
4170 (neon_vqmovun<mode>): Likewise.
4171 (neon_vmovl<mode>): Likewise.
4172 (neon_vmul_lane<mode>): Likewise.
4173 (neon_vmul_lane<mode>): Likewise.
4174 (neon_vmull_lane<mode>): Likewise.
4175 (neon_vqdmull_lane<mode>): Likewise.
4176 (neon_vqdmulh_lane<mode>): Likewise.
4177 (neon_vqdmulh_lane<mode>): Likewise.
4178 (neon_vmla_lane<mode>): Likewise.
4179 (neon_vmla_lane<mode>): Likewise.
4180 (neon_vmlal_lane<mode>): Likewise.
4181 (neon_vqdmlal_lane<mode>): Likewise.
4182 (neon_vmls_lane<mode>): Likewise.
4183 (neon_vmls_lane<mode>): Likewise.
4184 (neon_vmlsl_lane<mode>): Likewise.
4185 (neon_vqdmlsl_lane<mode>): Likewise.
4186 (neon_vext<mode>): Likewise.
4187 (neon_vrev64<mode>): Likewise.
4188 (neon_vrev32<mode>): Likewise.
4189 (neon_vrev16<mode>): Likewise.
4190 (neon_vbsl<mode>_internal): Likewise.
4191 (neon_vshl<mode>): Likewise.
4192 (neon_vqshl<mode>): Likewise.
4193 (neon_vshr_n<mode>): Likewise.
4194 (neon_vshrn_n<mode>): Likewise.
4195 (neon_vqshrn_n<mode>): Likewise.
4196 (neon_vqshrun_n<mode>): Likewise.
4197 (neon_vshl_n<mode>): Likewise.
4198 (neon_vqshl_n<mode>): Likewise.
4199 (neon_vqshlu_n<mode>): Likewise.
4200 (neon_vshll_n<mode>): Likewise.
4201 (neon_vsra_n<mode>): Likewise.
4202 (neon_vsri_n<mode>): Likewise.
4203 (neon_vsli_n<mode>): Likewise.
4204 (neon_vtbl1v8qi): Likewise.
4205 (neon_vtbl2v8qi): Likewise.
4206 (neon_vtbl3v8qi): Likewise.
4207 (neon_vtbl4v8qi): Likewise.
4208 (neon_vtbl1v16qi): Likewise.
4209 (neon_vtbl2v16qi): Likewise.
4210 (neon_vcombinev16qi): Likewise.
4211 (neon_vtbx1v8qi): Likewise.
4212 (neon_vtbx2v8qi): Likewise.
4213 (neon_vtbx3v8qi): Likewise.
4214 (neon_vtbx4v8qi): Likewise.
4215 (*neon_vtrn<mode>_insn): Likewise.
4216 (*neon_vzip<mode>_insn): Likewise.
4217 (*neon_vuzp<mode>_insn): Likewise.
4218 (neon_vld1<mode>): Likewise.
4219 (neon_vld1_lane<mode>): Likewise.
4220 (neon_vld1_lane<mode>): Likewise.
4221 (neon_vld1_dup<mode>): Likewise.
4222 (neon_vld1_dup<mode>): Likewise.
4223 (neon_vld1_dupv2di): Likewise.
4224 (neon_vst1<mode>): Likewise.
4225 (neon_vst1_lane<mode>): Likewise.
4226 (neon_vst1_lane<mode>): Likewise.
4227 (neon_vld2<mode>): Likewise.
4228 (neon_vld2<mode>): Likewise.
4229 (neon_vld2_lane<mode>): Likewise.
4230 (neon_vld2_lane<mode>): Likewise.
4231 (neon_vld2_dup<mode>): Likewise.
4232 (neon_vst2<mode>): Likewise.
4233 (neon_vst2<mode>): Likewise.
4234 (neon_vst2_lane<mode>): Likewise.
4235 (neon_vst2_lane<mode>): Likewise.
4236 (neon_vld3<mode>): Likewise.
4237 (neon_vld3qa<mode>): Likewise.
4238 (neon_vld3qb<mode>): Likewise.
4239 (neon_vld3_lane<mode>): Likewise.
4240 (neon_vld3_lane<mode>): Likewise.
4241 (neon_vld3_dup<mode>): Likewise.
4242 (neon_vst3<mode>): Likewise.
4243 (neon_vst3qa<mode>): Likewise.
4244 (neon_vst3qb<mode>): Likewise.
4245 (neon_vst3_lane<mode>): Likewise.
4246 (neon_vst3_lane<mode>): Likewise.
4247 (neon_vld4<mode>): Likewise.
4248 (neon_vld4qa<mode>): Likewise.
4249 (neon_vld4qb<mode>): Likewise.
4250 (neon_vld4_lane<mode>): Likewise.
4251 (neon_vld4_lane<mode>): Likewise.
4252 (neon_vld4_dup<mode>): Likewise.
4253 (neon_vst4<mode>): Likewise.
4254 (neon_vst4qa<mode>): Likewise.
4255 (neon_vst4qb<mode>): Likewise.
4256 (neon_vst4_lane<mode>): Likewise.
4257 (neon_vst4_lane<mode>): Likewise.
4258 (neon_vec_unpack<US>_lo_<mode>): Likewise.
4259 (neon_vec_unpack<US>_hi_<mode>): Likewise.
4260 (neon_vec_<US>mult_lo_<mode>): Likewise.
4261 (neon_vec_<US>mult_hi_<mode>): Likewise.
4262 (neon_vec_<US>shiftl_<mode>): Likewise.
4263 (neon_unpack<US>_<mode>): Likewise.
4264 (neon_vec_<US>mult_<mode>): Likewise.
4265 (vec_pack_trunc_<mode>): Likewise.
4266 (neon_vec_pack_trunc_<mode>): Likewise.
4267 (neon_vabd<mode>_2): Likewise.
4268 (neon_vabd<mode>_3): Likewise.
4270 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4272 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
4273 (load_pair): Update type attribute.
4274 (store_pair): Update type attribute.
4275 * config/aarch64/iterators.md (q): New.
4277 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4279 * config/arm/types.md: Add new types for Neon insns.
4281 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4282 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4283 Sergey Lega <sergey.s.lega@intel.com>
4284 Anna Tikhonova <anna.tikhonova@intel.com>
4285 Ilya Tocar <ilya.tocar@intel.com>
4286 Andrey Turetskiy <andrey.turetskiy@intel.com>
4287 Ilya Verbin <ilya.verbin@intel.com>
4288 Kirill Yukhin <kirill.yukhin@intel.com>
4289 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4291 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
4292 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
4293 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
4295 (srcp14<mode>): Ditto.
4296 (rsqrt14<mode>): Ditto.
4297 (rsqrt14<mode>): Ditto.
4298 (avx512f_vmscalef<mode>): Ditto.
4299 (avx512f_scalef<mode>): Ditto.
4300 (avx512f_getexp<mode>): Ditto.
4301 (avx512f_sgetexp<mode>): Ditto.
4302 (avx512f_fixupimm<mode>): Ditto.
4303 (avx512f_sfixupimm<mode>): Ditto.
4304 (avx512f_rndscale<mode>): Ditto.
4305 (*avx512er_exp2<mode>): Ditto.
4306 (*avx512er_rcp28<mode>): Ditto.
4307 (avx512er_rsqrt28<mode>): Ditto.
4308 (avx512f_getmant<mode>): Ditto.
4309 (avx512f_getmant<mode>): Ditto.
4310 (avx512f_rndscale<mode>): Fix formatting.
4312 2013-10-15 Martin Jambor <mjambor@suse.cz>
4314 * ipa-utils.h (ipa_edge_within_scc): Declare.
4315 * ipa-cp.c (edge_within_scc): Moved...
4316 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
4318 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4319 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4320 Sergey Lega <sergey.s.lega@intel.com>
4321 Anna Tikhonova <anna.tikhonova@intel.com>
4322 Ilya Tocar <ilya.tocar@intel.com>
4323 Andrey Turetskiy <andrey.turetskiy@intel.com>
4324 Ilya Verbin <ilya.verbin@intel.com>
4325 Kirill Yukhin <kirill.yukhin@intel.com>
4326 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4328 * config/i386/predicates.md (const_8_to_15_operand): New.
4329 (const_16_to_31_operand): Ditto.
4330 * config/i386/sse.md (V8FI): New.
4332 (reduc_splus_v8df): Ditto.
4333 (reduc_splus_v16sf): Ditto.
4334 (avx512f_vextract<shuffletype>32x4_1): Ditto.
4335 (vec_extract_hi_<mode>): Ditto.
4336 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
4337 (vec_set_lo_<mode>): Ditto.
4338 (vec_set_hi_<mode>): Ditto.
4339 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
4340 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
4341 (avx512f_pshufd_1): Ditto.
4342 (avx512f_broadcast<mode>): Ditto.
4343 (avx512f_broadcast<mode>): Ditto.
4344 (define_split): Split vec_extract_lo into move.
4345 (ssequartermode): Ditto.
4346 (ssedoublemode): Extened with wider modes.
4347 (vec_extract_lo_<mode>): Ditto.
4349 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4350 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4351 Sergey Lega <sergey.s.lega@intel.com>
4352 Anna Tikhonova <anna.tikhonova@intel.com>
4353 Ilya Tocar <ilya.tocar@intel.com>
4354 Andrey Turetskiy <andrey.turetskiy@intel.com>
4355 Ilya Verbin <ilya.verbin@intel.com>
4356 Kirill Yukhin <kirill.yukhin@intel.com>
4357 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4359 * config/i386/predicates.md (register_or_constm1_operand): New.
4360 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
4361 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
4362 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
4363 UNSPEC_SCATTER_PREFETCH
4365 (avx512f_ucmp<mode>3): Ditto.
4366 (avx512f_vternlog<mode>): Ditto.
4367 (avx512f_align<mode>): Ditto.
4368 (<shift_insn><mode>3): Ditto.
4369 (avx512f_<rotate>v<mode>): Ditto.
4370 (avx512f_<rotate><mode>): Ditto.
4371 (avx512f_eq<mode>3): Ditto.
4372 (avx512f_eq<mode>3_1): Ditto.
4373 (avx512f_gt<mode>3): Ditto.
4374 (avx512f_testm<mode>3): Ditto.
4375 (avx512f_testnm<mode>3): Ditto.
4376 (avx512pf_gatherpf<mode>): Ditto.
4377 (*avx512pf_gatherpf<mode>_mask): Ditto.
4378 (*avx512pf_gatherpf<mode>): Ditto.
4379 (avx512pf_scatterpf<mode>): Ditto.
4380 (*avx512pf_scatterpf<mode>_mask): Ditto.
4381 (*avx512pf_scatterpf<mode>): Ditto.
4382 (avx512f_vec_dup_gpr<mode>): Ditto.
4383 (clz<mode>2): Ditto.
4384 (conflict<mode>): Ditto.
4385 (REDUC_SMINMAX_MODE): Extened with wider modes.
4386 (reduc_<code>_<mode>): Ditto.
4387 (vlshr<mode>3): Ditto.
4388 (vashl<mode>3): Ditto.
4390 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4391 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4392 Sergey Lega <sergey.s.lega@intel.com>
4393 Anna Tikhonova <anna.tikhonova@intel.com>
4394 Ilya Tocar <ilya.tocar@intel.com>
4395 Andrey Turetskiy <andrey.turetskiy@intel.com>
4396 Ilya Verbin <ilya.verbin@intel.com>
4397 Kirill Yukhin <kirill.yukhin@intel.com>
4398 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4400 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
4403 (avx512fmaskmode): Ditto.
4404 (bcstscalarsuff): Ditto.
4405 (avx512f_blendm<mode>): Ditto.
4406 (cmp_imm_predicate): Ditto.
4407 (avx512f_cmp<mode>3): Ditto.
4408 (avx512f_vec_dup<mode>): Ditto.
4409 (avx512f_vec_dup_mem<mode>): Ditto.
4410 (avx512f_vpermi2var<mode>3): Ditto.
4411 (avx512f_vpermt2var<mode>3): Ditto.
4412 (vec_init<mode>): Ditto.
4413 (avx512f_gathersi<mode>): Ditto.
4414 (*avx512f_gathersi<mode>): Ditto.
4415 (*avx512f_gathersi<mode>_2): Ditto.
4416 (avx512f_gatherdi<mode>): Ditto.
4417 (*avx512f_gatherdi<mode>): Ditto.
4418 (*avx512f_gatherdi<mode>_2): Ditto.
4419 (avx512f_scattersi<mode>): Ditto.
4420 (*avx512f_scattersi<mode>): Ditto.
4421 (avx512f_scatterdi<mode>): Ditto.
4422 (*avx512f_scatterdi<mode>): Ditto.
4423 (sseintprefix): Extened with wider modes.
4424 (VEC_GATHER_IDXSI): Ditto.
4425 (VEC_GATHER_IDXDI): Ditto.
4426 (VEC_GATHER_SRCDI): Ditto.
4428 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
4429 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4431 * config/arm/t-aprofile: New file.
4432 * config.gcc: Handle --with-multilib-list option.
4434 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
4436 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
4439 2013-10-15 Richard Biener <rguenther@suse.de>
4441 * tree-loop-distribution.c (build_empty_rdg): Inline into
4443 (rdg_flag_vertex): Inline into single user.
4444 (rdg_flag_vertex_and_dependent): Likewise.
4445 (build_rdg_partition_for_vertex): Remove processed bitmap.
4446 (rdg_build_partitions): Simplify.
4448 2013-10-15 Richard Biener <rguenther@suse.de>
4450 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
4451 Restructure forwarding through conversions and copies to
4452 avoid performing copy-propagation the wrong way. Adjust
4453 recursion invocations.
4454 (forward_propagate_addr_expr): Add argument stating if we
4455 are recursing from a single-use.
4456 (ssa_forward_propagate_and_combine): Adjust.
4458 2013-10-14 David Malcolm <dmalcolm@redhat.com>
4460 * dumpfile.h (gcc::dump_manager): New class, to hold state
4461 relating to dumpfile management.
4462 (get_dump_file_name): Remove in favor of method of dump_manager.
4463 (dump_initialized_p): Likewise.
4464 (dump_start): Likewise.
4465 (dump_finish): Likewise.
4466 (dump_switch_p): Likewise.
4467 (dump_register): Likewise.
4468 (get_dump_file_info): Likewise.
4469 * context.c (gcc::context::context): Construct the dump_manager
4471 * context.h (gcc::context::get_dumps): New.
4472 (gcc::context::m_dumps): New.
4473 * coverage.c (coverage_init): Port to dump_manager API.
4474 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
4475 (extra_dump_files_in_use): Likewise.
4476 (extra_dump_files_alloced): Likewise.
4477 (gcc::dump_manager::dump_manager): New.
4478 (dump_register): Convert to...
4479 (gcc::dump_manager::dump_register): ...method, replacing
4480 function-static next_dump with m_next_dump field.
4481 (get_dump_file_info): Convert to...
4482 (gcc::dump_manager::get_dump_file_info): ...method.
4483 (get_dump_file_name): Convert to...
4484 (gcc::dump_manager::get_dump_file_name): ...method.
4485 (dump_start): Convert to...
4486 (gcc::dump_manager::dump_start): ...method.
4487 (dump_finish): Convert to...
4488 (gcc::dump_manager::dump_finish): ...method.
4489 (dump_begin): Replace body with...
4490 (gcc::dump_manager::dump_begin): ...new method.
4491 (dump_phase_enabled_p): Convert to...
4492 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4493 (dump_phase_enabled_p): Convert to...
4494 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4495 (dump_initialized_p): Convert to...
4496 (gcc::dump_manager::dump_initialized_p): ...method.
4497 (dump_flag_name): Replace body with...
4498 (gcc::dump_manager::dump_flag_name): ...new method.
4499 (dump_enable_all): Convert to...
4500 (gcc::dump_manager::dump_enable_all): ...new method.
4501 (opt_info_enable_passes): Convert to...
4502 (gcc::dump_manager::opt_info_enable_passes): ...new method.
4503 (dump_switch_p_1): Convert to...
4504 (gcc::dump_manager::dump_switch_p_1): ...new method.
4505 (dump_switch_p): Convert to...
4506 (gcc::dump_manager::dump_switch_p): ...new method.
4507 (opt_info_switch_p): Port to dump_manager API.
4508 (enable_rtl_dump_file): Likewise.
4509 * opts-global.c (handle_common_deferred_options): Port to new
4511 * passes.c (pass_manager::finish_optimization_passes): Likewise.
4512 (pass_manager::register_one_dump_file): Likewise.
4513 (pass_manager::register_pass): Likewise.
4514 (pass_init_dump_file): Likewise.
4515 (pass_fini_dump_file): Likewise.
4516 * statistics.c (statistics_early_init): Likewise.
4518 2013-10-14 Richard Biener <rguenther@suse.de>
4520 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
4521 iterative_hash_canonical_type, gimple_canonical_type_hash,
4522 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
4523 gimple_register_canonical_type, print_gimple_types_stats,
4524 free_gimple_type_tables): Move to lto/lto.c
4525 (gt-gimple.h): Do not include.
4526 * gimple.h (gimple_register_canonical_type,
4527 print_gimple_types_stats, free_gimple_type_tables): Remove.
4528 * Makefile.in (GTFILES): Remove gimple.c.
4530 2013-10-14 Travis Snoozy <quandary@remstate.com>
4533 * config/msp430/msp430.c (msp430_option_override): Correct thinko
4534 scanning for msp430x targets.
4536 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
4539 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
4540 (registers_ok_for_ldd_peep): Move around.
4541 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
4542 * config/sparc/sparc.md (widening peepholes): Use it.
4544 2013-10-14 Richard Biener <rguenther@suse.de>
4548 * gimple.c (iterative_hash_canonical_type): Make sure to
4549 record the hash into the correct hashtable slot.
4551 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
4553 PR rtl-optimization/58662
4554 * combine.c (try_combine): Take into account death nodes on I2 when
4555 splitting a PARALLEL of two independent SETs. Fix dump message.
4557 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
4560 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
4561 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
4562 * config/sh/t-sh (sh_treg_combine.o): New entry.
4563 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
4564 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
4565 (register_sh_passes): New function. Register sh_treg_combine pass.
4566 (sh_option_override): Invoke it.
4567 (sh_canonicalize_comparison): Handle op0_preserve_value.
4568 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
4569 opportunities. Canonicalize branch condition.
4570 (nott): Allow only if pseudos can be created for non-SH2A.
4572 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
4575 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
4576 (ix86_expand_movmem): Replace copy_addr_to_reg with
4577 ix86_copy_addr_to_reg.
4578 (ix86_expand_setmem): Likewise.
4580 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
4582 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
4583 provided by ix86_fp_compare_mode instead of CCFPUmode.
4585 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
4587 * config/aarch64/arm_neon.h
4588 (vtbx<1,3>_<psu>8): Fix register constriants.
4590 2013-10-11 Jeff Law <law@redhat.com>
4592 PR tree-optimization/58640
4593 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
4594 threading paths that cross over two loop entry points.
4596 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4598 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
4599 handle vector float as well.
4600 (*vsx_le_perm_load_v4si): Likewise.
4601 (*vsx_le_perm_store_v2di): Likewise.
4602 (*vsx_le_perm_store_v4si): Likewise.
4604 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4606 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
4607 directly to circumvent subtract from splat{31} workaround.
4608 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
4610 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
4611 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
4612 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
4613 (altivec_vperm_<mode>): Convert to define_insn_and_split to
4614 separate big and little endian logic.
4615 (*altivec_vperm_<mode>_internal): New define_insn.
4616 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
4617 separate big and little endian logic.
4618 (*altivec_vperm_<mode>_uns_internal): New define_insn.
4619 (vec_permv16qi): Add little endian logic.
4621 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
4623 * doc/extend.texi (returns_nonnull): Remove arguments.
4625 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4626 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4627 Sergey Lega <sergey.s.lega@intel.com>
4628 Anna Tikhonova <anna.tikhonova@intel.com>
4629 Ilya Tocar <ilya.tocar@intel.com>
4630 Andrey Turetskiy <andrey.turetskiy@intel.com>
4631 Ilya Verbin <ilya.verbin@intel.com>
4632 Kirill Yukhin <kirill.yukhin@intel.com>
4633 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4635 * config/i386/sse.md (VI48F_256_512): New.
4636 (avx2_permvar<mode>): Change to ...
4637 (<avx2_avx512f>_permvar<mode>): This.
4639 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4640 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4641 Sergey Lega <sergey.s.lega@intel.com>
4642 Anna Tikhonova <anna.tikhonova@intel.com>
4643 Ilya Tocar <ilya.tocar@intel.com>
4644 Andrey Turetskiy <andrey.turetskiy@intel.com>
4645 Ilya Verbin <ilya.verbin@intel.com>
4646 Kirill Yukhin <kirill.yukhin@intel.com>
4647 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4649 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
4650 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
4651 * config/i386/sse.md (VI4_AVX): New.
4652 (sf2simodelower): Ditto.
4653 (sse2_cvtps2dq): Change to ...
4654 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
4656 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4657 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4658 Sergey Lega <sergey.s.lega@intel.com>
4659 Anna Tikhonova <anna.tikhonova@intel.com>
4660 Ilya Tocar <ilya.tocar@intel.com>
4661 Andrey Turetskiy <andrey.turetskiy@intel.com>
4662 Ilya Verbin <ilya.verbin@intel.com>
4663 Kirill Yukhin <kirill.yukhin@intel.com>
4664 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4666 * config/i386/sse.md (V_512): New.
4668 (vcond<V_512:mode><VF_512:mode>): Ditto.
4669 (vcond<V_512:mode><VI_512:mode>): Ditto.
4670 (vcondu<V_512:mode><VI_512:mode>): Ditto.
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/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
4683 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
4685 (fma<mode>4): Removed condition.
4686 (fms<mode>4): Ditto.
4687 (fnma<mode>4): Ditto.
4688 (fnms<mode>4): Ditto.
4689 (fma4i_fmadd_<mode>): Ditto.
4690 (*fma_fmadd_<mode>): Ditto.
4691 (*fma_fmsub_<mode>): Ditto.
4692 (*fma_fnmadd_<mode>): Ditto.
4693 (*fma_fnmsub_<mode>): Ditto.
4694 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
4695 (*fma_fmaddsub_<mode>): Ditto.
4696 (*fma_fmsubadd_<mode>): Ditto.
4697 (*fmai_fmadd_<mode>): Ditto.
4698 (*fmai_fmsub_<mode>): Ditto.
4699 (*fmai_fnmadd_<mode>): Ditto.
4700 (*fmai_fnmsub_<mode>): Ditto.
4702 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4703 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4704 Sergey Lega <sergey.s.lega@intel.com>
4705 Anna Tikhonova <anna.tikhonova@intel.com>
4706 Ilya Tocar <ilya.tocar@intel.com>
4707 Andrey Turetskiy <andrey.turetskiy@intel.com>
4708 Ilya Verbin <ilya.verbin@intel.com>
4709 Kirill Yukhin <kirill.yukhin@intel.com>
4710 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4712 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
4713 (VI124_256): Changed to ...
4714 (VI124_256_48_512): This.
4715 (ssepackmode): Extended with wider modes.
4716 (<code><mode>3): Changed iterator.
4717 (*avx2_<code><mode>3): Ditto.
4718 (vec_pack_trunc_<mode>): Ditto.
4720 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4721 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4722 Sergey Lega <sergey.s.lega@intel.com>
4723 Anna Tikhonova <anna.tikhonova@intel.com>
4724 Ilya Tocar <ilya.tocar@intel.com>
4725 Andrey Turetskiy <andrey.turetskiy@intel.com>
4726 Ilya Verbin <ilya.verbin@intel.com>
4727 Kirill Yukhin <kirill.yukhin@intel.com>
4728 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4730 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
4731 (VI8F_256_512): Ditto.
4732 (abs<mode>2): Changed iterator.
4733 (avx2_perm<mode>): Changed to ...
4734 (<avx2_avx512f>_perm<mode>): This.
4735 (avx2_perm<mode>_1): Changed to ...
4736 (<avx2_avx512f>_perm<mode>_1): This.
4738 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4739 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4740 Sergey Lega <sergey.s.lega@intel.com>
4741 Anna Tikhonova <anna.tikhonova@intel.com>
4742 Ilya Tocar <ilya.tocar@intel.com>
4743 Andrey Turetskiy <andrey.turetskiy@intel.com>
4744 Ilya Verbin <ilya.verbin@intel.com>
4745 Kirill Yukhin <kirill.yukhin@intel.com>
4746 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4748 * config/i386/sse.md (VI48_AVX512F): New.
4749 (VI48_AVX2): Changed to ...
4750 (VI48_AVX2_48_AVX512F): This.
4751 (avx2_ashrv<mode>): Changed to ...
4752 (<avx2_avx512f>_ashrv<mode>): This.
4753 (avx2_<shift_insn>v<mode>): Changed to ...
4754 (<avx2_avx512f>_<shift_insn>v<mode>): This.
4756 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4757 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4758 Sergey Lega <sergey.s.lega@intel.com>
4759 Anna Tikhonova <anna.tikhonova@intel.com>
4760 Ilya Tocar <ilya.tocar@intel.com>
4761 Andrey Turetskiy <andrey.turetskiy@intel.com>
4762 Ilya Verbin <ilya.verbin@intel.com>
4763 Kirill Yukhin <kirill.yukhin@intel.com>
4764 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4766 * config/i386/sse.md (VI4_AVX512F): New.
4767 (VI8_AVX2_AVX512F): Ditto.
4768 (mul<mode>3): Extended with wider modes.
4769 (*<sse4_1_avx2>_mul<mode>3): Ditto.
4770 (mul<mode>3): Ditto.
4771 (vec_widen_<s>mult_odd_<mode>): Ditto.
4773 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4774 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4775 Sergey Lega <sergey.s.lega@intel.com>
4776 Anna Tikhonova <anna.tikhonova@intel.com>
4777 Ilya Tocar <ilya.tocar@intel.com>
4778 Andrey Turetskiy <andrey.turetskiy@intel.com>
4779 Ilya Verbin <ilya.verbin@intel.com>
4780 Kirill Yukhin <kirill.yukhin@intel.com>
4781 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4783 * config/i386/sse.md (VI2_AVX512F): New.
4784 (VI124_AVX512F): Ditto.
4785 (sseunpackmode): Extended with wider modes.
4786 (sseunpackfltmode): Ditto.
4787 (vec_unpacks_float_hi_<mode>): Ditto.
4788 (vec_unpacks_float_lo_<mode>): Ditto.
4789 (vec_unpacku_float_hi_<mode>): Ditto.
4790 (vec_unpacku_float_lo_<mode>): Ditto.
4791 (vec_unpacks_lo_<mode>): Ditto.
4792 (vec_unpacks_hi_<mode>): Ditto.
4793 (vec_unpacku_lo_<mode>): Ditto.
4794 (vec_unpacku_hi_<mode>): Ditto.
4796 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4797 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4798 Sergey Lega <sergey.s.lega@intel.com>
4799 Anna Tikhonova <anna.tikhonova@intel.com>
4800 Ilya Tocar <ilya.tocar@intel.com>
4801 Andrey Turetskiy <andrey.turetskiy@intel.com>
4802 Ilya Verbin <ilya.verbin@intel.com>
4803 Kirill Yukhin <kirill.yukhin@intel.com>
4804 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4806 * config/i386/i386.md (multdiv): New.
4807 (multdiv_mnemonic): Ditto.
4808 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
4809 (<sse>_vm<multdiv_mnemonic><mode>3): This.
4810 (<sse>_vmdiv<mode>3): Removed.
4812 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4813 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4814 Sergey Lega <sergey.s.lega@intel.com>
4815 Anna Tikhonova <anna.tikhonova@intel.com>
4816 Ilya Tocar <ilya.tocar@intel.com>
4817 Andrey Turetskiy <andrey.turetskiy@intel.com>
4818 Ilya Verbin <ilya.verbin@intel.com>
4819 Kirill Yukhin <kirill.yukhin@intel.com>
4820 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4822 * config/i386/sse.md (V): Extended with wider modes.
4824 (ssehalfvecmode): Ditto.
4826 (ssepackfltmode): Ditto.
4827 (avx_vec_concat<mode>): Ditto.
4828 (V_256_512): New iterator.
4829 (VF2_512_256): Ditto.
4830 (si2dfmode): New attribute.
4831 (si2dfmodelower): Ditto.
4833 (concat_tg_mode): Ditto.
4834 (floatv4siv4df2): Changed to ...
4835 (float<si2dfmodelower><mode>2): This.
4836 (avx_cvtps2pd256): Changed to ...
4837 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
4838 (vec_pack_trunc_v4df): Changed to ...
4839 (vec_pack_trunc_<mode>): This.
4840 (avx_vpermil<mode>): Changed to ...
4841 (<sse2_avx_avx512f>_vpermil<mode>): This.
4842 (<fixsuffix>fix_truncv8dfv8si2): New.
4843 (vec_pack_sfix_trunc_v8df): Ditto.
4844 (avx512f_rndscale<mode>): Ditto.
4845 (avx512f_roundpd512): Ditto.
4846 (vec_pack_ufix_trunc_<mode>): Updated iterator.
4848 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4849 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4850 Sergey Lega <sergey.s.lega@intel.com>
4851 Anna Tikhonova <anna.tikhonova@intel.com>
4852 Ilya Tocar <ilya.tocar@intel.com>
4853 Andrey Turetskiy <andrey.turetskiy@intel.com>
4854 Ilya Verbin <ilya.verbin@intel.com>
4855 Kirill Yukhin <kirill.yukhin@intel.com>
4856 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4858 * config/i386/i386.md (any_fix): New iterator.
4859 (fixsuffix): New attribute.
4860 * config/i386/sse.md (VF1): Extened with wider modes.
4864 (sseintvecmodelower): Ditto.
4865 (ssescalarmode): Ditto.
4866 (ssescalarnum): Ditto.
4868 (ssexmmmode): Ditto.
4869 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
4870 (<sse>_rcp<mode>2): Change iterator.
4871 (rsqrt<mode>2): Ditto.
4872 (<sse>_rsqrt<mode>2): Ditto.
4873 (avx2_vec_dup<mode>): Ditto.
4874 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
4875 (round<mode>2_sfix): Ditto.
4876 (avx2_pbroadcast<mode>): Ditto.
4877 (*andnot<mode>3): Handle XI mode.
4878 (*<code><mode>3): Ditto.
4879 (AVXTOSSEMODE): Removed.
4880 (avx_vpermil<mode>): Changed to ...
4881 (<sse2_avx_avx512f>_vpermil<mode>): This.
4883 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4884 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4885 Sergey Lega <sergey.s.lega@intel.com>
4886 Anna Tikhonova <anna.tikhonova@intel.com>
4887 Ilya Tocar <ilya.tocar@intel.com>
4888 Andrey Turetskiy <andrey.turetskiy@intel.com>
4889 Ilya Verbin <ilya.verbin@intel.com>
4890 Kirill Yukhin <kirill.yukhin@intel.com>
4891 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4893 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
4894 (<sse>_comi): Ditto.
4895 (<sse>_ucomi): Ditto.
4896 (sse_cvtss2siq_2): Ditto.
4897 (sse2_cvtsd2si): Ditto.
4898 (sse2_cvtsd2siq): Ditto.
4899 (sse2_cvttsd2si): Ditto.
4900 (sse2_cvttsd2siq): Ditto.
4901 (<shift_insn><mode>3): Ditto.
4902 (sse2_cvtsi2sdq): Update constraint and prefix.
4903 (sse_cvtsi2ss): Update prefix.
4904 (sse_cvtsi2ssq): Ditto.
4906 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4908 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
4909 ignore internal calls.
4911 2013-10-11 Richard Biener <rguenther@suse.de>
4913 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
4914 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
4915 TREE_OVERFLOW is set.
4917 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
4919 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
4921 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
4922 * gimple.h: Likewise.
4924 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
4926 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
4929 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
4931 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
4932 * configure: Regenerate.
4934 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4936 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
4937 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
4938 formatting fixes, use pp_colon instead of pp_character (..., ':'),
4939 similarly pp_right_paren.
4940 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
4941 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
4942 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
4943 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
4945 (omp_declare_simd_clauses_equal,
4946 omp_remove_redundant_declare_simd_attrs): New functions.
4947 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
4948 (walk_tree_1): Handle new OpenMP 4.0 clauses.
4949 * tree.h (OMP_LOOP_CHECK): Define.
4950 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
4951 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
4952 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
4953 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
4954 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
4955 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
4956 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
4957 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
4958 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
4959 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
4960 OMP_CLAUSE_SIMDLEN_EXPR): Define.
4961 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
4962 (omp_remove_redundant_declare_simd_attrs): New prototype.
4963 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
4964 GIMPLE_OMP_TEAMS): New codes.
4965 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
4966 * omp-low.c (struct omp_context): Add cancel_label and cancellable
4968 (target_nesting_level): New variable.
4969 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
4970 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
4971 for collapse > 1 static schedule unless ordered.
4972 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
4973 (determine_parallel_type): Adjust get_ws_args_for caller.
4974 (install_var_field): Handle mask & 4 for double indirection.
4975 (scan_sharing_clauses): Ignore shared clause on teams construct.
4976 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
4977 (create_omp_child_function): If inside target or declare target
4978 constructs, set "omp declare target" attribute on the child function.
4979 (find_combined_for): New function.
4980 (scan_omp_parallel): Handle combined loops.
4981 (scan_omp_target, scan_omp_teams): New functions.
4982 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
4983 restrictions and set ctx->cancellable for cancellable constructs.
4984 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
4985 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
4986 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
4987 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
4988 (omp_clause_aligned_alignment): New function.
4989 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
4990 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
4991 on teams constructs. Handle user defined reductions and new
4993 (lower_reduction_clauses): Don't set placeholder to address of ref
4994 if it has already the right type.
4995 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
4996 (expand_parallel_call): Use the new non-_start suffixed builtins,
4997 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
4998 and GOMP_parallel_end after the call.
4999 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
5000 (expand_omp_for_init_counts): Handle combined loops.
5001 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
5003 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
5004 end of cancellable loops.
5005 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
5006 Likewise. Handle collapse > 1 loops.
5007 (expand_omp_simd): Handle combined loops.
5008 (expand_omp_for): Add inner_stmt argument, adjust callers of
5009 expand_omp_for* functions, use expand_omp_for_static*chunk even
5010 for collapse > 1 unless ordered.
5011 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
5012 of cancellable sections.
5013 (expand_omp_single): Remove need_barrier variable, just rely on
5014 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
5015 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
5016 (expand_omp_atomic_load, expand_omp_atomic_store,
5017 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
5018 (expand_omp_target): New function.
5019 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
5020 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
5021 (build_omp_regions_1): Immediately close region for
5022 GF_OMP_TARGET_KIND_UPDATE.
5023 (maybe_add_implicit_barrier_cancel): New function.
5024 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
5026 (lower_omp_single): Likewise. Add clobber after the barrier.
5027 (lower_omp_taskgroup): New function.
5028 (lower_omp_for): Handle combined loops. Adjust
5029 lower_rec_input_clauses caller. Handle cancellation.
5030 (lower_depend_clauses): New function.
5031 (lower_omp_taskreg): Lower depend clauses. Adjust
5032 lower_rec_input_clauses caller. Add clobber after the call. Handle
5034 (lower_omp_target, lower_omp_teams): New functions.
5035 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
5036 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
5037 and GOMP_cancellation_point calls.
5038 (lower_omp): Fold stmts inside of target region.
5039 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
5040 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5041 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
5042 (BT_FN_VOID_OMPFN_PTR_UINT,
5043 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
5044 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
5045 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
5046 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
5047 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
5048 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
5049 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
5050 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
5051 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
5052 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
5053 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
5054 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
5055 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
5056 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
5057 BUILT_IN_GOMP_PARALLEL_END.
5058 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
5059 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5060 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
5061 GF_OMP_FOR_KIND_DISTRIBUTE.
5062 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
5063 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
5064 (dump_gimple_omp_return): Print lhs if it has any.
5065 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
5066 gimple_omp_atomic_seq_cst_p.
5067 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
5068 and GIMPLE_OMP_TEAMS.
5069 * langhooks.c (lhd_omp_mappable_type): New function.
5070 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
5071 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
5073 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
5074 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
5075 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
5076 (struct gimplify_omp_ctx): Add combined_loop field.
5077 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
5078 on stmts inside of target region.
5079 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
5080 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
5081 ORT_TARGET and ORT_TARGET_DATA.
5082 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
5084 (omp_notice_threadprivate_variable): Complain about threadprivate
5085 variables in target region.
5086 (omp_notice_variable): Complain about vars with non-mappable type
5087 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
5088 (omp_check_private): Ignore ORT_TARGET* regions.
5089 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
5090 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
5091 (find_combined_omp_for): New function.
5092 (gimplify_omp_for): Handle gimplification of combined loops.
5093 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
5095 (gimplify_omp_target_update): New function.
5096 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
5097 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
5098 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
5099 (gimplify_body): If fndecl has "omp declare target" attribute, add
5100 implicit ORT_TARGET context around it.
5101 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
5102 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
5103 * tree-nested.c (convert_nonlocal_reference_stmt,
5104 convert_local_reference_stmt, convert_gimple_call): Handle
5105 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5106 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
5107 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
5108 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
5109 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
5110 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
5111 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
5112 BUILT_IN_GOMP_LOOP_END_CANCEL,
5113 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
5114 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
5115 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
5116 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
5117 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
5118 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
5119 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
5120 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
5121 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
5122 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
5123 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
5124 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
5125 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
5126 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
5127 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
5128 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5129 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
5130 gimple_build_omp_teams): New functions.
5131 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
5132 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
5133 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
5134 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5135 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
5136 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
5137 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
5138 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
5139 GF_OMP_ATOMIC_SEQ_CST): New.
5140 (gimple_build_omp_taskgroup, gimple_build_omp_target,
5141 gimple_build_omp_teams): New prototypes.
5142 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
5143 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5144 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
5145 GIMPLE_OMP_SINGLE as end of range.
5146 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
5147 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
5148 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
5149 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
5150 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
5151 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
5152 gimple_omp_target_kind, gimple_omp_target_set_kind,
5153 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
5154 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
5155 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
5156 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
5157 gimple_omp_teams_set_clauses): New inlines.
5158 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
5159 and GIMPLE_OMP_TASKGROUP.
5160 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
5161 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
5162 enum omp_clause_proc_bind_kind): New.
5163 (union omp_clause_subcode): Add depend_kind, map_kind and
5164 proc_bind_kind fields.
5165 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
5166 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5167 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
5168 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
5169 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
5171 2013-10-10 Teresa Johnson <tejohnson@google.com>
5173 * predict.c (tree_estimate_probability): Add new parameter
5174 for estimate_bb_frequencies.
5175 (estimate_bb_frequencies): Add new parameter to force estimation.
5176 (rebuild_frequencies): When max frequency in function is small,
5177 recompute counts from frequencies.
5178 * predict.h (estimate_bb_frequencies): New parameter.
5180 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5182 * ipa-inline.c (ipa_inline): Fix leak of "order" when
5183 optimizations are disabled.
5185 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5187 * coverage.c (coverage_finish): Fix leak of da_file_name.
5189 2013-10-10 Jan Hubicka <jh@suse.cz>
5191 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
5192 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
5195 2013-10-10 Jakub Jelinek <jakub@redhat.com>
5198 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
5199 if any labels are in FALLTHRU_BB, use a special label emitted
5200 immediately after the asm goto insn rather than label_rtx
5202 (expand_asm_stmt): Adjust caller.
5203 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
5204 edge if the last insn in predecessor is a jump with single successor,
5205 but it isn't simplejump_p.
5207 2013-10-10 Richard Biener <rguenther@suse.de>
5209 PR tree-optimization/58656
5210 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
5212 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5214 * gimplify.c: Include expr.h and tm_p.h for targets with special
5215 va-arg padding requirements.
5217 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5219 * tree-flow.h: Move some prototypes to gimple.h.
5220 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
5221 * gimple.h: Relocate some prototypes from tree-flow.h
5222 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
5224 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
5225 (build_va_arg_indirect_ref): Relocate and make static.
5226 (std_gimplify_va_arg_expr): Relocate here.
5227 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
5228 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
5230 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5232 * doc/md.texi: Document the mnemonic attribute.
5234 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5237 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
5238 (set_attr_alternative x ...) when searching for user defined
5241 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5243 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
5245 2013-10-09 Easwaran Raman <eraman@google.com>
5247 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
5248 * cfgexpand.c (defer_stack_allocation): ...use here
5249 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
5251 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
5253 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
5254 (operand_equal_for_value_replacement): New function, extracted from
5255 value_replacement and enhanced to catch more cases.
5256 (value_replacement): Use operand_equal_for_value_replacement.
5258 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5260 * loop-doloop.c (doloop_modify, doloop_optimize): Use
5261 get_max_loop_iterations.
5263 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5265 * config/arm/aarch-common.c (arm_early_load_addr_dep):
5266 Place comment above function.
5268 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5270 * tree-flow.h: Remove all remaining prototypes, enums and structs that
5271 are not related to tree-cfg.c.
5272 * tree-ssa-address.h: New file. Relocate prototypes.
5273 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
5274 (addr_for_mem_ref): New. Combine call to get_address_description and
5275 return addr_for_mem_ref.
5276 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
5277 * tree-ssa-live.h: Adjust prototypes.
5278 * passes.c: Include tree-ssa-live.h.
5279 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
5280 * graphite.c (graphite_transform_loops): Make static.
5281 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5282 make_pass_graphite, pass_data_graphite_transforms,
5283 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
5284 * ipa-pure-const.c (warn_function_noreturn): Make static.
5285 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5286 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5287 Relocate from tree-cfg.c
5288 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
5290 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5291 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5292 Move to ipa-pure-const.c.
5293 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
5294 Relocate from tree-optimize.c.
5295 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
5296 make_pass_fixup_cfg): Move to tree-cfg.c.
5297 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
5298 Relocate some prototypes.
5299 * tree-data-ref.h (tree_check_data_deps) Add prototype.
5300 * tree-dump.c (dump_function_to_file): Remove prototype.
5301 Add tree-flow.h to the include file.
5302 * tree-dump.h: Remove prototype.
5303 * tree-parloops.h: New File. Add prototypes.
5304 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
5305 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
5306 from tree-ssa-loop.c.
5307 * tree-predcom.c (run_tree_predictive_commoning,
5308 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
5309 Relocate here from tree-ssa-loop.c.
5310 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
5311 ssa_name_values.release ().
5312 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
5313 (ssa_name_values): Relocate from tree-flow.h.
5314 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
5315 * tree-ssa-loop.c (run_tree_predictive_commoning,
5316 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
5317 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5318 make_pass_graphite, pass_data_graphite_transforms,
5319 make_pass_graphite_transforms, gate_tree_parallelize_loops,
5320 tree_parallelize_loops, pass_data_parallelize_loops,
5321 make_pass_parallelize_loops): Move to other files.
5322 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
5324 * tree.h: Remove prototypes from tree-address.c.
5326 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5328 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
5329 (struct int_tree_map): Move to tree-hasher.h
5330 (SCALE, LABEL, PERCENT): Move to gimple.h
5331 * tree-flow-inline.h: Delete. Move functions to other files.
5332 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
5333 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
5334 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
5335 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
5336 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
5337 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
5338 * tree-hasher.h: Don't include tree-flow.h.
5339 (struct int_tree_map): Relocate from tree-flow.h.
5340 * tree-sra.c (contains_view_convert_expr_p): Relocate from
5341 tree-flow-inline.h and make static.
5342 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
5344 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
5345 tree-flow-inline.h and make static.
5346 * tree.h (is_global_var, may_be_aliased): Relocate from
5348 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
5349 * value-prof.c: No longer include tree-flow-inline.h.
5350 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
5352 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5354 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
5355 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
5356 (enum move_pos): Move to tree-ssa-loop-im.h
5357 * cfgloop.h: Move some prototypes.
5358 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
5359 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
5360 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
5361 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
5362 (loop_containing_stmt): Relocate from tree-flow-inline.h.
5363 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
5364 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
5365 (enum move_pos): Relocate here.
5366 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
5368 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
5369 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
5370 make_pass_lim): Relocate here from tree-ssa-loop.c.
5371 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
5373 (loop_edge_to_cancel, unloop_loops): Make static.
5374 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5375 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
5376 (tree_complete_unroll, gate_tree_complete_unroll,
5377 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
5378 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5379 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
5380 * tree-ssa-loop-ivopts.c: Remove local prototypes.
5381 (stmt_invariant_in_loop_p): Remove unused function.
5382 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
5383 * tree-ssa-loop-manip.h: New file. Add prototypes.
5384 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
5385 (gcov_type_to_double_int): Move to cfgloop.h.
5386 (double_int_cmp, bound_index,
5387 estimate_numbers_of_iterations_loop): Make static.
5388 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
5389 (max_loop_iterations): Factor out get_max_loop_iterations.
5390 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
5392 * tree-ssa-loop-niter.h: New file. Add prototypes.
5393 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
5394 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
5395 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
5396 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
5397 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
5398 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
5399 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
5400 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
5401 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
5402 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
5403 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5404 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
5405 pass_data_complete_unroll, make_pass_complete_unroll,
5406 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5407 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
5408 tree-ssa-loop-ivcanon.c.
5409 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
5410 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
5411 tree-ssa-loop-prefetch.c.
5412 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
5414 (get_lsm_tmp_name): Relocate and add suffix parameter.
5415 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
5416 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
5417 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
5418 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
5419 (get_estimated_loop_iterations): Factor out accessor from
5420 estimated_loop_iterations in tree-ssa-loop-niter.c.
5421 (get_max_loop_iterations): Factor out accessor from
5422 _max_loop_iterations in tree-ssa-niter.c.
5423 * loop-unroll.c (decide_unroll_constant_iterations,
5424 decide_unroll_runtime_iterations, decide_peel_simple,
5425 decide_unroll_stupid): Use new get_* accessors.
5427 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
5429 PR tree-optimization/20318
5430 * doc/extend.texi (returns_nonnull): New function attribute.
5431 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
5433 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
5434 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
5436 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
5439 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
5440 false if both components are bitfields.
5442 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5444 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
5445 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
5446 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
5447 * config/aarch64/aarch64.h
5448 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
5449 * config/aarch64/aarch64-simd-builtins.def
5450 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
5452 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5454 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
5456 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5458 * config/aarch64/arm_neon.h (vdiv_f64): Added.
5460 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5462 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
5463 (vneg_f64): New intrinsic.
5464 (vneg_s8): Asm replaced with C.
5465 (vneg_s16): Likewise.
5466 (vneg_s32): Likewise.
5467 (vneg_s64): New intrinsic.
5468 (vnegq_f32): Asm replaced with C.
5469 (vnegq_f64): Likewise.
5470 (vnegq_s8): Likewise.
5471 (vnegq_s16): Likewise.
5472 (vnegq_s32): Likewise.
5473 (vnegq_s64): Likewise.
5475 2013-10-09 Renlin Li <Renlin.Li@arm.com>
5477 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
5479 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5481 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
5482 packed stack special handling.
5483 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
5484 back to fixed stack slots for FPRs saved due to stdarg.
5486 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5488 * config/s390/s390.c (s390_frame_info): Restructure function.
5490 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5492 * config/s390/s390.c (struct s390_frame_layout): New field
5494 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
5495 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
5496 regs_ever_clobbered to char*.
5497 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
5498 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
5499 (s390_register_info): Enable FPR save slots. Move/Copy some
5500 functionality into ...
5501 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
5502 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
5504 (s390_frame_info): Do gpr slot allocation here now. stdarg does
5505 not imply a stack frame.
5506 (s390_init_frame_layout): Remove variable clobbered_regs.
5507 (s390_update_register_info): Remove function.
5508 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
5509 cannot be used for register renaming.
5510 (s390_hard_regno_scratch_ok): New function.
5511 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
5512 (s390_initial_elimination_offset): Change offset calculation of
5513 the return address pointer.
5514 (save_gprs): Deal with only r6 being saved from the call-saved regs.
5515 (restore_gprs): Set frame related flag.
5516 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
5517 (s390_emit_prologue): Call s390_register_info instead of
5518 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
5519 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
5520 (s390_optimize_prologue): Call s390_optimize_register_info.
5521 Try to remove also FPR slot save/restore INSNs. Remove frame
5522 related flags from restore INSNs.
5524 2013-10-08 DJ Delorie <dj@redhat.com>
5526 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
5529 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
5530 avoid conflict with the MI use of %c.
5531 * config/rl78/rl78-real.md: change %c to %C throughout.
5532 * config/rl78/rl78-virt.md: Likewise.
5534 2013-10-08 Jan Hubicka <jh@suse.cz>
5536 * config/i386/i386.c (ix86_option_override_internal): Switch
5537 to SSE math for -ffast-math when target ISA supports SSE2.
5539 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
5541 * tree-flow.h: Remove some prototypes.
5542 * tree.h: Remove some protypes, add a couple.
5543 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
5544 using_eh_for_cleanups_p): Add interface routines for front ends.
5545 * tree-eh.h: New file. Add protoptyes.
5546 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
5547 (add_stmt_to_eh_lp_fn): Make static.
5548 (lower_try_finally): Use new using_eh_for_cleanups_p.
5549 * emit-rtl.c: Include tree-eh.h.
5550 * gimple.h: Include tree-eh.h.
5552 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
5554 PR tree-optimization/58480
5555 * tree-vrp.c (infer_nonnull_range): New function.
5556 (infer_value_range): Call infer_nonnull_range.
5558 2013-10-08 Dehao Chen <dehao@google.com>
5560 PR tree-optimization/58619
5561 * tree-inline.c (copy_phis_for_bb): Combine location data
5564 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5567 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
5568 RTX_FRAME_RELATED_P on INSN.
5570 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5572 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
5573 (altivec_expand_vec_perm_const): Call it.
5575 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5577 * config/rs6000/vector.md (mov<mode>): Emit permuted move
5578 sequences for LE VSX loads and stores at expand time.
5579 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
5581 * config/rs6000/rs6000.c (rs6000_const_vec): New.
5582 (rs6000_gen_le_vsx_permute): New.
5583 (rs6000_gen_le_vsx_load): New.
5584 (rs6000_gen_le_vsx_store): New.
5585 (rs6000_gen_le_vsx_move): New.
5586 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
5587 (*vsx_le_perm_load_v4si): New.
5588 (*vsx_le_perm_load_v8hi): New.
5589 (*vsx_le_perm_load_v16qi): New.
5590 (*vsx_le_perm_store_v2di): New.
5591 (*vsx_le_perm_store_v4si): New.
5592 (*vsx_le_perm_store_v8hi): New.
5593 (*vsx_le_perm_store_v16qi): New.
5594 (*vsx_xxpermdi2_le_<mode>): New.
5595 (*vsx_xxpermdi4_le_<mode>): New.
5596 (*vsx_xxpermdi8_le_V8HI): New.
5597 (*vsx_xxpermdi16_le_V16QI): New.
5598 (*vsx_lxvd2x2_le_<mode>): New.
5599 (*vsx_lxvd2x4_le_<mode>): New.
5600 (*vsx_lxvd2x8_le_V8HI): New.
5601 (*vsx_lxvd2x16_le_V16QI): New.
5602 (*vsx_stxvd2x2_le_<mode>): New.
5603 (*vsx_stxvd2x4_le_<mode>): New.
5604 (*vsx_stxvd2x8_le_V8HI): New.
5605 (*vsx_stxvd2x16_le_V16QI): New.
5607 2013-10-07 Renlin Li <Renlin.Li@arm.com>
5609 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
5611 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5613 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
5614 loop to work also for 31bit ABI.
5615 Save the stack pointer for frame_size > 0.
5617 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5619 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
5620 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
5621 constraint letters from expanders.
5622 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
5623 retry count to general_operand.
5624 ("tabort"): Give operand 0 a mode.
5625 ("tabort_1"): Add mode and constraint letter for operand 0.
5626 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
5628 2013-10-04 Jeff Law <law@redhat.com>
5630 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
5632 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
5634 (thread_across_edge): Here. Call it.
5636 2013-10-04 Cary Coutant <ccoutant@google.com>
5638 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
5639 discarding a location list expression (or a piece of one).
5641 2013-10-03 Jan Hubicka <jh@suse.cz>
5643 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
5644 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
5645 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
5647 2013-10-03 Jan Hubicka <jh@suse.cz>
5649 * config/i386/i386.c (ix86_option_override_internal): Do not enable
5650 accumulate-outgoing-args when producing unwind info.
5652 2013-10-03 Wei Mi <wmi@google.com>
5654 * lra-constraints.c (insert_move_for_subreg): New function
5655 extracted from simplify_operand_subreg.
5656 (simplify_operand_subreg): Add reload for paradoxical subreg.
5658 2013-10-03 Rong Xu <xur@google.com>
5660 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
5661 the candidate of builtin_expect such that we should fix the
5662 size/time estimation.
5663 (estimate_function_body_sizes): Do the acutally size/time fix-up
5666 2013-10-03 Rong Xu <xur@google.com>
5668 * predict.c (tree_predict_by_opcode): Get the probability
5669 for builtin_expect from param builtin_expect_probability.
5670 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
5671 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
5672 * doc/invoke.texi: Add documentation for builtin-expect-probability.
5674 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
5677 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
5678 * calls.c (alloca_call_p): Use get_callee_fndecl.
5679 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
5680 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
5682 (vrp_visit_stmt): Remove duplicated code.
5684 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
5686 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
5687 ceildf2, btruncdf2, instead of vsx_* name.
5689 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
5690 iterators to only do V2DF and V4SF here. Move the DF code to
5691 rs6000.md where it is combined with SF mode. Replace <VSv> with
5692 just 'v' since only vector operations are handled with these insns
5693 after moving the DF support to rs6000.md.
5694 (vsx_sub<mode>3): Likewise.
5695 (vsx_mul<mode>3): Likewise.
5696 (vsx_div<mode>3): Likewise.
5697 (vsx_fre<mode>2): Likewise.
5698 (vsx_neg<mode>2): Likewise.
5699 (vsx_abs<mode>2): Likewise.
5700 (vsx_nabs<mode>2): Likewise.
5701 (vsx_smax<mode>3): Likewise.
5702 (vsx_smin<mode>3): Likewise.
5703 (vsx_sqrt<mode>2): Likewise.
5704 (vsx_rsqrte<mode>2): Likewise.
5705 (vsx_fms<mode>4): Likewise.
5706 (vsx_nfma<mode>4): Likewise.
5707 (vsx_copysign<mode>3): Likewise.
5708 (vsx_btrunc<mode>2): Likewise.
5709 (vsx_floor<mode>2): Likewise.
5710 (vsx_ceil<mode>2): Likewise.
5711 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
5712 (vsx_sminsf3): Likewise.
5713 (vsx_fmadf4): Likewise.
5714 (vsx_fmsdf4): Likewise.
5715 (vsx_nfmadf4): Likewise.
5716 (vsx_nfmsdf4): Likewise.
5717 (vsx_cmpdf_internal1): Likewise.
5719 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
5720 simpler to select whether a target has SPE or traditional floating
5721 point support in iterators.
5722 (TARGET_DF_SPE): Likewise.
5723 (TARGET_SF_FPR): Likewise.
5724 (TARGET_DF_FPR): Likewise.
5725 (TARGET_SF_INSN): Macros to say whether floating point support
5726 exists for a given operation for expanders.
5727 (TARGET_DF_INSN): Likewise.
5729 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
5730 combining of SF/DF mode operations, using both traditional and VSX
5738 (abs<mode>2): Combine SF/DF modes using traditional floating point
5739 instructions. Add support for using the upper DF registers with
5740 VSX support, and SF registers with power8-vector support. Update
5741 expanders for operations supported by both the SPE and traditional
5742 floating point units.
5743 (abs<mode>2_fpr): Likewise.
5744 (nabs<mode>2): Likewise.
5745 (nabs<mode>2_fpr): Likewise.
5746 (neg<mode>2): Likewise.
5747 (neg<mode>2_fpr): Likewise.
5748 (add<mode>3): Likewise.
5749 (add<mode>3_fpr): Likewise.
5750 (sub<mode>3): Likewise.
5751 (sub<mode>3_fpr): Likewise.
5752 (mul<mode>3): Likewise.
5753 (mul<mode>3_fpr): Likewise.
5754 (div<mode>3): Likewise.
5755 (div<mode>3_fpr): Likewise.
5756 (sqrt<mode>3): Likewise.
5757 (sqrt<mode>3_fpr): Likewise.
5758 (fre<Fs>): Likewise.
5759 (rsqrt<mode>2): Likewise.
5760 (cmp<mode>_fpr): Likewise.
5761 (smax<mode>3): Likewise.
5762 (smin<mode>3): Likewise.
5763 (smax<mode>3_vsx): Likewise.
5764 (smin<mode>3_vsx): Likewise.
5765 (negsf2): Delete SF operations that are merged with DF.
5772 (fmasf4_fpr): Likewise.
5773 (fmssf4_fpr): Likewise.
5774 (nfmasf4_fpr): Likewise.
5775 (nfmssf4_fpr): Likewise.
5776 (sqrtsf2): Likewise.
5777 (rsqrtsf_internal1): Likewise.
5778 (smaxsf3): Likewise.
5779 (sminsf3): Likewise.
5780 (cmpsf_internal1): Likewise.
5781 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
5782 (negdf2): Delete DF operations that are merged with SF.
5784 (nabsdf2): Likewise.
5790 (rsqrtdf_internal1): Likewise.
5791 (fmadf4_fpr): Likewise.
5792 (fmsdf4_fpr): Likewise.
5793 (nfmadf4_fpr): Likewise.
5794 (nfmsdf4_fpr): Likewise.
5795 (sqrtdf2): Likewise.
5796 (smaxdf3): Likewise.
5797 (smindf3): Likewise.
5798 (cmpdf_internal1): Likewise.
5799 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
5800 (btrunc<mode>2): Delete separate expander, and combine with the
5801 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
5802 (btrunc<mode>2_fpr): Likewise.
5803 (ceil<mode>2): Likewise.
5804 (ceil<mode>2_fpr): Likewise.
5805 (floor<mode>2): Likewise.
5806 (floor<mode>2_fpr): Likewise.
5807 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
5808 Add support for using the upper registers with VSX and
5809 power8-vector. Move insns to be closer to the define_expands. On
5810 VSX systems, prefer the traditional form of FMA over the VSX
5811 version, since the traditional form allows the target not to
5812 overlap with the inputs.
5813 (fms<mode>4_fpr): Likewise.
5814 (nfma<mode>4_fpr): Likewise.
5815 (nfms<mode>4_fpr): Likewise.
5817 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5818 Richard Earnshaw <richard.earnshaw@arm.com>
5820 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
5821 (struct mult_cost_table): Likewise.
5822 (struct mem_cost_table): Likewise.
5823 (struct fp_cost_table): Likewise.
5824 (struct vector_cost_table): Likewise.
5825 (cpu_cost_table): Likewise.
5826 * config/arm/arm.opt (mold-rts-costs): New option.
5827 (mnew-generic-costs): Likewise.
5828 * config/arm/arm.c (generic_extra_costs): New table.
5829 (cortexa15_extra_costs): Likewise.
5830 (arm_slowmul_tune): Use NULL as new costs.
5831 (arm_fastmul_tune): Likewise.
5832 (arm_strongarm_tune): Likewise.
5833 (arm_xscale_tune): Likewise.
5834 (arm_9e_tune): Likewise.
5835 (arm_v6t2_tune): Likewise.
5836 (arm_cortex_a5_tune): Likewise.
5837 (arm_cortex_a9_tune): Likewise.
5838 (arm_v6m_tune): Likewise.
5839 (arm_fa726te_tune): Likewise.
5840 (arm_cortex_a15_tune): Use cortex15_extra_costs.
5841 (arm_cortex_tune): Use generict_extra_costs.
5842 (shifter_op_p): New function.
5843 (arm_unspec_cost): Likewise.
5844 (LIBCALL_COST): Define.
5845 (arm_new_rtx_costs): New function.
5846 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
5847 table is available. Use old costs otherwise unless mnew-generic-costs
5849 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
5850 (cpu_cost_table): Declare.
5852 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
5855 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
5856 (*subs_mul_imm_<mode>)
5857 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
5858 (*sub_<shift>_<mode>)
5859 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
5860 Remove k constraint.
5862 2013-10-03 Ian Bolton <ian.bolton@arm.com>
5864 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
5866 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
5868 2013-10-02 Teresa Johnson <tejohnson@google.com>
5870 * predict.c (probably_never_executed): New function.
5871 (probably_never_executed_bb_p): Invoke probably_never_executed.
5872 (probably_never_executed_edge_p): Ditto.
5873 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
5874 Treat profile insanities conservatively.
5876 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
5878 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
5880 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
5882 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
5883 use it. Use smaller increase for scratch. Don't increase reject
5884 for early clobber scratch.
5885 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
5886 setting eliminated regs except setting fp from hfp.
5887 (lra_eliminate): Check lra_insn_recog_data on NULL.
5889 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5892 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
5893 setting -mvsx-timode by default until the underlying problem is fixed.
5894 (RS6000_CPU, power7 defaults): Likewise.
5896 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5898 * config/x-linux (host-linux.o): Remove header dependencies.
5899 Use $(COMPILE) and $(POSTCOMPILE).
5900 * config/t-linux-android (linux-android.o): Ditto.
5902 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5904 * Makefile.in (expmed.o-warn): Remove.
5906 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5908 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
5909 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
5911 2013-10-02 Teresa Johnson <tejohnson@google.com>
5913 * dojump.c (do_jump_1): Divide probability between
5914 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
5916 2013-10-02 Tom Tromey <tromey@redhat.com>
5918 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
5920 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5922 * tree-flow.h: Remove some prototypes.
5923 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
5924 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
5925 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
5926 mark_virtual_phi_result_for_renaming): Relocate here.
5927 * tree-into-ssa.h: Add prototypes.
5928 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
5929 single_pred_before_succ_order.
5930 (blocks_in_phiopt_order): Rename and move to cfganal.c.
5931 (nonfreeing_call_p) Move to gimple.c.
5932 * cfganal.c (single_pred_before_succ_order): Move and renamed from
5934 * basic-block.h (single_pred_before_succ_order): Add prototype.
5935 * gimple.c (nonfreeing_call_p): Relocate here.
5936 * gimple.h: Add prototype.
5937 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
5938 * tree-ssa-dom.h: New file. Relocate prototypes here.
5939 * tree-ssa.h: Include tree-ssa-dom.h.
5941 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5943 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
5944 Use $(COMPILE) and $(POSTCOMPILE).
5946 * config/alpha/x-alpha (driver-alpha.o): Ditto.
5948 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5950 * tree-flow.h: Remove some prototypes.
5951 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
5952 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5953 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
5954 propagate_tree_value*): Move from here to...
5955 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
5956 propagate_tree_value*): Relocate here.
5957 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5958 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
5959 * gimple-fold.c: Remove gimple-fold.h from include list.
5960 * tree-vrp.c: Remove gimple-fold.h from include list.
5961 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
5962 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
5963 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
5964 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
5965 * sese.c: Add tree-ssa-propagate.h to include list.
5967 2013-10-02 Richard Biener <rguenther@suse.de>
5969 * tree-loop-distribution.c: Include tree-vectorizer.h for
5971 (enum partition_kind): Remove PKIND_REDUCTION.
5972 (struct partition_s): Remove has_writes member, add reduction_p member.
5973 (partition_alloc): Adjust.
5974 (partition_builtin_p): Likewise.
5975 (partition_has_writes): Remove.
5976 (partition_reduction_p): New function.
5977 (partition_merge_into): Likewise.
5978 (generate_code_for_partition): Commonize builtin partition
5980 (rdg_cannot_recompute_vertex_p): Remove.
5981 (already_processed_vertex_p): Likewise.
5982 (rdg_flag_vertex): Do not set has_writes.
5983 (classify_partition): Adjust.
5984 (rdg_build_partitions): Do not set has_writes, treat all
5985 partitions as useful.
5986 (distribute_loop): Record number of library calls generated. Adjust.
5987 (tree_loop_distribution): Report number of loops and library
5988 calls generated as opt-info.
5990 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5992 * tree-flow.h: Include new .h files. Move prototypes.
5993 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
5994 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
5995 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
5996 * tree-pretty-print.h: Add prototypes from tree-flow.h.
5997 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
5998 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
5999 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
6000 * tree.h (get_ref_base_and_extent): Move prototype out.
6001 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
6003 * gimple-low.h: New File. Add prototypes from tree-flow.h.
6004 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
6005 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
6006 * tree-scalar-evolution.c: Include tree.h.
6007 * sese.c: Include tree.h.
6008 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
6009 * dwarf2out.c: Include tree-dfa.h.
6010 * tree-chrec.c: Include tree.h.
6011 * tree-data-ref.c: Include tree.h.
6013 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
6015 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
6018 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6020 * config/t-sol2 (sol2-c.o): Remove header dependencies.
6021 Use $(COMPILE) and $(POSTCOMPILE).
6022 (sol2-cxx.o): Likewise.
6023 (sol2-stubs.o): Likewise.
6025 * config/x-solaris (host-solaris.o): Likewise.
6027 * config/sparc/t-sparc (sparc.o): Remove.
6028 (sparc-c.o): Remove header dependencies.
6029 Use $(COMPILE) and $(POSTCOMPILE).
6030 * config/sparc/x-sparc: Likewise.
6032 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
6034 * config/arc/arc-opts.h: Add 2013 to Copyright years.
6035 * config/arc/arc700.md: Likewise.
6036 * config/arc/arc-modes.def: Likewise.
6037 * config/arc/arc-simd.h: Likewise.
6038 * config/arc/t-arc-uClibc: Likewise.
6039 * config/arc/t-arc-newlib: Likewise.
6041 2013-10-02 Renlin Li <renlin.li@arm.com>
6043 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
6045 (aarch64_expand_epilogue): Likewise.
6047 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6048 Yufeng Zhang <yufeng.zhang@arm.com>
6050 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
6052 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
6053 'base_in' represent a conversion and legal_cast_p_1 holds; set
6054 'base_in' with the returned value from get_unwidened.
6056 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6058 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
6059 plus_constant is not used.
6061 2013-10-01 Wei Mi <wmi@google.com>
6063 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
6064 * config/i386/i386.md: Add define_peephole2 to
6065 break partial reg stall for cvtss2sd/cvtsd2ss.
6067 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6069 * config/arc/arc.c (pass_arc_ifcvt::clone):
6070 Update for ctxt_ -> m_ctxt change.
6072 2013-10-01 Jeff Law <law@redhat.com>
6074 * tree-ssa-threadupdate.c (struct redirection_data): Delete
6075 outgoing_edge and intermediate_edge fields. Instead store the path.
6076 (redirection_data::hash): Hash on the last edge's destination index.
6077 (redirection_data::equal): Check the entire thread path.
6078 (lookup_redirectio_data): Corresponding changes.
6079 (create_edge_and_update_destination_phis): Likewise.
6080 (thread_single_edge): Likewise.
6082 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6083 Diego Novillo <dnovillo@google.com>
6085 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
6086 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
6087 (vld32wh_insn, vld32wl_insn): Delete commented-out old
6088 versions of these patterns.
6090 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
6091 (__builtin_arc_aligned): Likewise.
6093 * config/arc/arc.md: Expand adc_0 comment stating the intended
6094 purpose and why it isn't ready.
6095 Replace commented out call_value_via_label_mixed with a
6096 plain comment about bl_s.
6098 * config/arc/arc.c (stdio.h): Don't include directly.
6099 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
6100 Remove if (1) condition.
6101 (arc_encode_section_info): Fix comment.
6103 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6105 * config/arc/arc.c (arc_conditional_register_usage):
6106 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
6107 Also set reg_alloc_order for DMA config regs.
6109 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6110 Jeremy Bennett <jeremy.bennett@embecosm.com>
6112 * doc/install.texi (--with-cpu): Mention ARC.
6113 (arc-*-elf32): New paragraph.
6114 (arc-linux-uclibc): Likewise.
6115 * doc/md.texi (Machine Constraints): Add ARC part.
6116 * doc/invoke.texi: (menu): Add ARC Options.
6117 (Machine Dependent Options) <ARC Options>: Add synopsis.
6118 (node ARC Options): Add.
6119 * doc/extend.texi (long_call / short_call attribute): Add ARC.
6120 (ARC Built-in Functions): New section defining
6121 generic ARC built-in functions.
6122 (ARC SIMD Built-in Functions): New section defining SIMD specific
6124 (Declaring Attributes of Functions): Extended
6125 description of short_call and long_call attributes for ARC and
6126 added index entries.
6128 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
6129 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
6130 Joern Rennecke <joern.rennecke@embecosm.com>
6131 Muhammad Khurram Riaz <khurram.riaz@arc.com>
6132 Brendan Kehoe <brendan@zen.org>
6133 Michael Eager <eager@eagercon.com>
6134 Simon Cook <simon.cook@embecosm.com>
6135 Jeremy Bennett <jeremy.bennett@embecosm.com>
6137 * config/arc, common/config/arc: New directories.
6139 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6140 Brendan Kehoe <brendan@zen.org>
6141 Simon Cook <simon.cook@embecosm.com>
6143 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
6145 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6147 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
6148 * tree-ssa-coalesce.h: New. Move prototype to here.
6149 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
6150 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
6151 (gimple_can_coalesce_p): Move to...
6152 * gimple.c (gimple_can_coalesce_p): Here.
6154 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6156 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
6157 (dump_decl_set): Move to gimple.c.
6158 * gimple.h: Don't include tree-ssa-operands.h.
6159 (dump_decl_set): Add prototype.
6160 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6161 Move to gimple-ssa.h.
6162 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
6163 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
6164 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
6165 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
6166 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
6168 (dump_decl_set): Relocate here.
6169 * gimple-ssa.h: New file.
6170 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6171 Relocate from gimple.h.
6172 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
6173 * tree-ssa-operands.c (swap_ssa_operands): Rename from
6174 swap_tree_operands and remove non-ssa path.
6175 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
6176 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
6178 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
6179 vect_is_simple_reduction_1): Use swap_ssa_operands.
6180 * tree-flow.h: Move various prototypes to tree-phinodes.h.
6181 (enum need_phi_state): Move to tree-into-ssa.c.
6182 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6183 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
6184 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
6185 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
6186 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6187 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6188 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6189 num_imm_uses): Move to ssa-iterators.h.
6190 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
6191 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
6193 (op_iter_done, op_iter_next_def, op_iter_next_tree,
6194 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
6195 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
6196 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
6197 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
6198 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
6199 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
6200 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
6201 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
6202 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
6203 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
6204 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
6205 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
6206 (set_phi_nodes): Move to tree-phinodes.h.
6207 * tree-ssa-operands.h (enum ssa_op_iter_type,
6208 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
6209 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
6210 (dump_decl_set): Remove prototype.
6211 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
6212 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
6213 (set_phi_nodes): Relocate from tree-flow-inline.h.
6214 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
6216 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
6217 include list. Temporarily add gimple.h to include list.
6218 * ssa-iterators.h: New file.
6219 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6220 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
6221 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
6222 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
6223 Relocate from tree-ssa-operands.h.
6224 (delink_imm_use, link_imm_use_to_list, link_imm_use,
6225 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6226 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6227 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6228 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
6229 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
6230 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
6231 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
6232 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
6233 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
6234 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
6235 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
6236 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
6237 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
6238 Relocate from tree-flow-inline.h.
6239 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
6241 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
6244 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
6245 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
6246 Insert '\t' to output template.
6247 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
6248 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
6249 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
6250 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
6252 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
6254 * doc/install.texi (Host/target specific installation notes for GCC):
6255 Put @anchor before @heading.
6256 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
6257 Use @email for email addresses.
6259 2013-10-01 Jeff Law <law@redhat.com>
6261 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
6262 a vec. Only delete the path if we create one without successfully
6263 registering a jump thread.
6264 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
6266 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
6267 (paths): New vector of jump threading paths.
6268 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
6269 (THREAD_PATH): New accessor macro for the entire thread path.
6270 (lookup_redirection_data): Get intermediate and final outgoing edge
6271 from the thread path.
6272 (create_edge_and_update_destination_phis): Copy the threading path.
6273 (ssa_fix_duplicate_block_edges): Get edges and block types from the
6274 jump threading path.
6275 (ssa_redirect_edges): Get edges and block types from the jump threading
6276 path. Free the path vector.
6277 (thread_block): Get edges from the jump threading path. Look at the
6278 entire path to see if we thread to a loop exit. If we cancel a jump
6279 thread request, then free the path vector.
6280 (thread_single_edge): Get edges and block types from the jump threading
6281 path. Free the path vector.
6282 (thread_through_loop_header): Get edges and block types from the jump
6283 threading path. Free the path vector.
6284 (mark_threaded_blocks): Iterate over the vector of paths and store
6285 the path on the appropriate edge. Get edges and block types from the
6286 jump threading path.
6287 (mark_threaded_blocks): Get edges and block types from the jump
6288 threading path. Free the path vector.
6289 (thread_through_all_blocks): Use the vector of paths rather than
6290 a vector of 3-edge sets.
6291 (register_jump_thread): Accept pointer to a path vector rather
6292 than the path vector itself. Store the path vector for later use.
6295 2013-10-01 Jakub Jelinek <jakub@redhat.com>
6296 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6299 * config/s390/s390.c (s390_split_branches): Modify check for table
6301 (s390_chunkify_start): Rearrange table jump insn check in order to
6302 deal with compare and branch insns correctly.
6304 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
6308 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
6309 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
6310 define_insn_and_split.
6311 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
6312 (shiftsi3_compare): New pattern.
6314 * config/arm/unspecs.md (UNSPEC_RRX): New.
6316 2013-10-01 Alan Modra <amodra@gmail.com>
6318 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
6319 casing inout operands.
6321 2013-10-01 Richard Biener <rguenther@suse.de>
6323 PR tree-optimization/58553
6324 * tree-loop-distribution.c (struct partition_s): Add niter member.
6325 (classify_partition): Populate niter member for the partition
6326 and properly identify whether the relevant store happens before
6327 or after the loop exit.
6328 (generate_memset_builtin): Use niter member from the partition.
6329 (generate_memcpy_builtin): Likewise.
6331 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6333 * vec.h (vec_prefix, vec): Prefix member names with "m_".
6334 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
6336 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6338 * basic-block.h (edge_list): Prefix member names with "m_".
6339 * context.h (context): Likewise.
6340 * domwalk.h (dom_walker): Likewise.
6341 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
6342 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
6343 * hash-table.h (hash_table): Likewise.
6344 * machmode.h (bit_field_mode_iterator): Likewise.
6345 * pass_manager.h (pass_list): Likewise.
6346 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
6347 * tree-pass.h (pass_data): Likewise.
6348 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
6349 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
6350 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
6351 * asan.c (pass_data_asan): Update accordingly.
6352 * cfganal.c (control_dependences::find_control_dependence): Likewise.
6353 (control_dependences::control_dependences): Likewise.
6354 (control_dependences::~control_dependences): Likewise.
6355 (control_dependences::~control_dependences): Likewise.
6356 (control_dependences::get_edges_dependent_on): Likewise.
6357 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
6358 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
6359 * context.c (gcc::context::context): Likewise.
6360 * cprop.c (pass_rtl_cprop::clone): Likewise.
6361 * domwalk.c (dom_walker::walk): Likewise.
6362 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
6363 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
6364 * mode-switching.c (pass_mode_switching::clone): Likewise.
6365 * passes.c (opt_pass::opt_pass): Likewise.
6366 (pass_manager::pass_manager): Likewise.
6367 * predict.c (pass_strip_predict_hints::clone): Likewise.
6368 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
6369 (pass_split_all_insns::clone): Likewise.
6370 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
6372 (bit_field_mode_iterator::next_mode): Likewise.
6373 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
6374 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
6375 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
6376 * tree-complex.c (pass_lower_complex::clone): Likewise.
6377 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
6378 * tree-object-size.c (pass_object_sizes::clone): Likewise.
6379 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
6380 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
6381 (pass_fold_builtins::clone): Likewise.
6382 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
6383 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
6384 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
6385 (pass_cd_dce::clone): Likewise.
6386 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
6387 (pass_phi_only_cprop::clone): Likewise.
6388 * tree-ssa-dse.c (pass_dse::clone): Likewise.
6389 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
6390 * tree-ssa-loop.c (pass_lim::clone): Likewise.
6391 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
6392 * tree-ssa-pre.c (pass_fre::clone): Likewise.
6393 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
6394 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
6395 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
6396 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
6397 * tree-vrp.c (pass_vrp::clone): Likewise.
6398 * tsan.c (pass_tsan::clone): Likewise.
6400 2013-09-30 Jakub Jelinek <jakub@redhat.com>
6403 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
6404 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
6407 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
6408 optimization, punt if sign_bit_p looked through any zero extension.
6410 2013-09-30 Teresa Johnson <tejohnson@google.com>
6412 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
6413 Update redirected out edge count in joiner case.
6414 (ssa_redirect_edges): Common the joiner and non-joiner cases
6415 so that joiner case gets profile updates.
6417 2013-09-30 Richard Biener <rguenther@suse.de>
6419 PR tree-optimization/58554
6420 * tree-loop-distribution.c (classify_partition): Require
6421 unconditionally executed stores for memcpy and memset recognition.
6422 (tree_loop_distribution): Calculate dominance info.
6424 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
6426 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
6427 (NO_PROFILE_COUNTERS): Likewise.
6428 (PROFILE_HOOK): Likewise.
6429 (FUNCTION_PROFILER): Likewise.
6430 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
6432 2013-09-30 Iain Sandoe <iain@codesourcery.com>
6434 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
6435 calls and defines in TARGET_MACHO conditional.
6436 (load_macho_picbase_di): Likewise.
6437 (reload_macho_picbase): Likewise.
6438 (reload_macho_picbase_si): Likewise.
6439 (reload_macho_picbase_di): Likewise.
6440 (nonlocal_goto_receiver): Likewise.
6442 2013-09-30 Nick Clifton <nickc@redhat.com>
6444 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
6445 that use the MSP430X ISA.
6446 (msp430_option_override): Scan -mmcu command line option for any
6447 MCU name that supports the MSP430X ISA.
6448 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
6449 -mmcu options which enable the MSP430X ISA.
6451 2013-09-30 Richard Biener <rguenther@suse.de>
6454 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
6455 before looking for setjmp-like calls.
6457 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6460 * config/darwin-protos.h (machopic_get_function_picbase): New.
6461 * config/darwin.c (machopic_get_function_picbase): New.
6462 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
6463 label for a new func. (load_macho_picbase_di): Likewise.
6464 (reload_macho_picbase): New expand.
6465 (reload_macho_picbase_si): New insn.
6466 (reload_macho_picbase_di): New insn.
6467 (nonlocal_goto_receiver): New define and split.
6468 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
6469 (unspecv enum): Add UNSPECV_NLGR.
6471 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6473 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
6474 that altivec registers are correctly sized on Darwin.
6476 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6478 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
6479 darwin-driver.o): Use COMPILE and POSTCOMPILE.
6480 * config/x-darwin (host-darwin.o): Likewise.
6481 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
6482 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
6483 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
6485 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
6487 * doc/invoke.texi: Fix usage of @tie{} command.
6489 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
6491 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
6493 2013-09-28 Jan Hubicka <jh@suse.cz>
6495 * config/i386/x86-tune.def: Add documentation for each of the options;
6498 2013-09-28 Jan Hubicka <jh@suse.cz>
6500 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
6502 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
6503 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
6504 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
6505 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
6506 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
6508 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
6510 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
6511 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
6512 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
6513 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
6514 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
6515 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
6516 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
6517 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
6518 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
6519 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
6520 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
6521 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
6522 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
6523 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
6524 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
6525 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
6526 gimple.h, godump.c, graphite-clast-to-gimple.c,
6527 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
6528 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
6529 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
6530 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6531 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
6532 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
6533 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
6534 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
6535 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
6536 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
6537 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
6538 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
6539 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
6540 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
6541 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
6542 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
6543 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
6544 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
6545 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
6546 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
6547 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
6548 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
6549 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
6550 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
6551 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6552 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6553 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
6554 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
6555 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
6556 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6557 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6558 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
6559 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
6560 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6561 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6562 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
6563 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
6564 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
6565 whitespace before "(".
6567 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
6569 * expr.h (extract_bit_field): Remove packedp parameter.
6570 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
6571 from forward declaration.
6572 (store_split_bit_field): Remove packedp arg from calls to
6573 extract_fixed_bit_field.
6574 (extract_bit_field_1): Remove packedp parameter and packedp
6575 argument from recursive calls and calls to extract_fixed_bit_field.
6576 (extract_bit_field): Remove packedp parameter and corresponding
6577 arg to extract_bit_field_1.
6578 (extract_fixed_bit_field): Remove packedp parameter. Remove code
6580 (extract_split_bit_field): Remove packedp arg from call to
6581 extract_fixed_bit_field.
6582 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
6583 (copy_blkmode_from_reg): Likewise.
6584 (copy_blkmode_to_reg): Likewise.
6585 (read_complex_part): Likewise.
6586 (store_field): Likewise.
6587 (expand_expr_real_1): Likewise.
6588 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
6589 to extract_bit_field.
6590 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
6591 call to extract_bit_field.
6592 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
6593 call to extract_bit_field.
6594 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
6595 and special packedp behavior from -fstrict-volatile-bitfields
6598 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6600 * lra-eliminations.c (init_elim_table): Guard value_p.
6602 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6604 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
6605 DFmode, DImode, and SFmode in the upper VSX registers based on the
6606 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
6607 if -mpower8-vector. Combine -mvsx-timode handling with the rest
6608 of the VSX register handling.
6610 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
6612 (zero_extendsidi2_lfiwzx): Add support for loading into the
6613 Altivec registers with -mpower8-vector. Use wu/wv constraints to
6614 only do VSX memory options on Altivec registers.
6615 (extendsidi2_lfiwax): Likewise.
6616 (extendsfdf2_fpr): Likewise.
6617 (mov<mode>_hardfloat, SF/SD modes): Likewise.
6618 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
6619 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
6620 (movdi_internal64): Likewise.
6622 2013-09-27 Xinliang David Li <davidxl@google.com>
6624 * opts.c (finish_options): Adjust parameters
6625 according to vect cost model.
6626 (common_handle_option): Set dynamic vect cost
6628 targhooks.c (default_add_stmt_cost): Compute stmt cost
6630 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
6631 Use helper function.
6632 * tree-vectorizer.h (unlimited_cost_model): New function.
6633 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
6634 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
6636 (vect_enhance_data_refs_alignment): Ditto.
6637 * flag-types.h: New enum.
6638 * common/config/i386/i386-common.c (ix86_option_init_struct):
6639 No need to initialize vect_cost_model flag.
6640 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
6643 2013-09-27 Diego Novillo <dnovillo@google.com>
6645 * gimple.h (enum ssa_mode): Remove.
6647 2013-09-27 Paulo Matos <pmatos@broadcom.com>
6649 * cfgloop.h (number_of_loops): Fix typo in check for null.
6651 2013-09-27 Jakub Jelinek <jakub@redhat.com>
6654 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
6655 are children of outermost saved_cfun's loop, and set it up to
6656 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
6657 if loop != loop0 and SESE region contains bbs that belong to loop0.
6659 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6661 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
6662 (binary_scale_code_p, get_base_term, get_index_term): New functions.
6663 (set_address_segment, set_address_base, set_address_index)
6664 (set_address_disp): Accept the argument unconditionally.
6665 (baseness): Remove must_be_base_p and must_be_index_p checks.
6666 (decompose_normal_address): Classify as much as possible in the
6669 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6671 * cse.c (count_reg_usage): Handle INT_LIST.
6672 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
6673 * reginfo.c (reg_scan_mark_refs): Likewise.
6674 * reload1.c (eliminate_regs_1): Likewise.
6676 2013-09-27 Iain Sandoe <iain@codesourcery.com>
6679 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
6682 2013-09-27 Richard Biener <rguenther@suse.de>
6684 PR tree-optimization/58459
6685 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
6686 restriction not propagating into loops.
6688 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
6690 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
6691 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
6692 * doc/tree-ssa.texi (Walking use-def chains): Delete.
6694 2013-09-26 Richard Biener <rguenther@suse.de>
6696 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
6698 2013-09-26 Richard Biener <rguenther@suse.de>
6700 * alias.h (component_uses_parent_alias_set): Rename to ...
6701 (component_uses_parent_alias_set_from): ... this.
6702 * alias.c (component_uses_parent_alias_set): Rename to ...
6703 (component_uses_parent_alias_set_from): ... this and return
6705 (reference_alias_ptr_type_1): Use the result from
6706 component_uses_parent_alias_set_from instead of stripping
6707 components one at a time.
6708 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
6710 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6712 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
6713 Move prototypes to...
6714 * tree-ssa-ter.h: New File. Move prototypes here.
6715 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
6716 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
6717 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
6718 from is_replaceable_p.
6719 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
6720 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
6721 (process_replaceable): Use ter_is_replaceable_p.
6722 (find_replaceable_in_bb): Use ter_is_replaceable_p.
6723 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
6724 newly refactored ssa_is_replaceable_p.
6725 * cfgexpand.c: Include tree-outof-ssa.h.
6726 * ssaexpand.h: Delete.
6728 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6730 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
6731 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
6733 (create_gimple_tmp): Delete.
6734 (get_expr_type, build_assign, build_type_cast): Move to...
6735 * gimple-builder.c: New File.
6736 (get_expr_type): Relocate from gimple.c.
6737 (build_assign, build_type_cast): Change to only create ssanames.
6738 * gimple.h: Move prototypes to...
6739 * gimple-builder.h: New File. Here.
6740 * tree-ssa.h: And here.
6741 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
6742 count_uses_and_derefs): Relocate from gimple.c.
6743 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
6744 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
6745 * tree-ssa-math-opts (execute_cse_reciprocals): Use
6746 gimple_replace_ssa_lhs.
6747 * asan.c: Include gimple-builder.h.
6748 * Makefile.in: Add gimple-builder.o.
6750 2013-09-26 Richard Biener <rguenther@suse.de>
6752 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
6753 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
6754 (loe_visit_block): Use gcc_checking_assert.
6755 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
6756 gimple_assign_ssa_name_copy_p.
6757 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
6760 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
6762 * config/rs6000/t-rs6000 (rs6000.o): Remove.
6763 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
6765 2013-09-26 Richard Biener <rguenther@suse.de>
6767 PR tree-optimization/58539
6768 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
6769 the fact that debug statements are not taking part in loop-closed
6772 2013-09-26 Nick Clifton <nickc@redhat.com>
6774 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
6775 time warning message.
6776 (msp430_print_operand_raw): Delete unused letter parameter.
6777 (TARGET_PRINT_OPERAND_ADDRESS): Define.
6778 (msp430_print_operand_address): New function.
6779 (msp430_print_operand): Move address printing code from here to
6781 * config/msp430/msp430.md (movsipsi2): Add comment in generated
6783 (zero_extendpsisi2): Likewise.
6784 (extendpsisi2): New pattern.
6785 (andneghi3): New pattern.
6787 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
6789 * config/aarch64/aarch64.opt (mlra): New option.
6790 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
6791 (TARGET_LRA_P): Define.
6793 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
6795 * expr.c (expand_assignment): Remove obsolete comment.
6797 2013-09-25 Jeff Law <law@redhat.com>
6799 * tree-flow.h (thread_through_all_blocks): Prototype moved into
6800 tree-ssa-threadupdate.h.
6801 (register_jump_thread): Similarly.
6802 * tree-ssa-threadupdate.h: New header file.
6803 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
6804 * tree-vrp.c: Likewise.
6805 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
6806 (thread_around_empty_blocks): Change type of path vector argument to
6807 an edge,type pair from just an edge. Initialize both elements when
6808 appending to a jump threading path. Tweak references to elements
6810 (thread_across_edge): Similarly. Release memory for the elements
6812 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
6813 (dump_jump_thread_path): New function broken out from
6814 register_jump_thread.
6815 (register_jump_thread): Use dump_jump_thread_path. Change type of
6816 path vector entries. Search the path for NULL edges and dump
6817 the path if one is found. Tweak the conversion of path to 3-edge
6818 form to use the block copy type information embedded in the path.
6820 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6822 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
6824 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6825 Vladimir Makarov <vmakarov@redhat.com>
6827 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
6828 from the least significant bit.
6829 (strip_address_mutations): Add bitfield operations handling.
6830 (must_be_index_p): Add shifting and rotate operations handling.
6831 (set_address_base): Use must_be_base_p predicate.
6832 (set_address_index): Use must_be_index_p predicate.
6834 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
6835 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6836 Sergey Lega <sergey.s.lega@intel.com>
6837 Anna Tikhonova <anna.tikhonova@intel.com>
6838 Ilya Tocar <ilya.tocar@intel.com>
6839 Andrey Turetskiy <andrey.turetskiy@intel.com>
6840 Ilya Verbin <ilya.verbin@intel.com>
6841 Kirill Yukhin <kirill.yukhin@intel.com>
6842 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6844 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
6846 (ix86_expand_vector_move_misalign): Support new unaligned load and
6847 stores and use new names.
6848 (CODE_FOR_sse2_storedqu): Rename to ...
6849 (CODE_FOR_sse2_storedquv16qi): ... this.
6850 (CODE_FOR_sse2_loaddqu): Rename to ...
6851 (CODE_FOR_sse2_loaddquv16qi): ... this.
6852 (CODE_FOR_avx_loaddqu256): Rename to ...
6853 (CODE_FOR_avx_loaddquv32qi): ... this.
6854 (CODE_FOR_avx_storedqu256): Rename to ...
6855 (CODE_FOR_avx_storedquv32qi): ... this.
6856 * config/i386/i386.md (fpint_logic): New.
6857 * config/i386/sse.md (VMOVE): Extend for AVX512.
6861 (VI_UNALIGNED_LOADSTORE): Ditto.
6862 (sse2_avx_avx512f): Ditto.
6863 (sse2_avx2): Extend for AVX512.
6864 (sse4_1_avx2): Ditto.
6865 (avx2_avx512f): New.
6866 (sse): Extend for AVX512.
6869 (avxsizesuffix): Ditto.
6870 (sseintvecmode): Ditto.
6872 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
6873 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
6874 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6875 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
6876 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6877 (<sse2_avx_avx512f>_storedqu<mode): ... this.
6878 (<sse>_movnt<mode>): Replace constraint "x" with "v".
6879 (STORENT_MODE): Extend for AVX512.
6880 (*absneg<mode>2): Replace constraint "x" with "v".
6881 (*mul<mode>3): Ditto.
6882 (*ieee_smin<mode>3): Ditto.
6883 (*ieee_smax<mode>3): Ditto.
6884 (avx_cmp<mode>3): Replace VF with VF_128_256.
6885 (*<sse>_maskcmp<mode>3_comm): Ditto.
6886 (<sse>_maskcmp<mode>3): Ditto.
6887 (<sse>_andnot<mode>3): Extend for AVX512.
6888 (<code><mode>3, anylogic): Replace VF with VF_128_256.
6889 (<code><mode>3, fpint_logic): New.
6890 (*<code><mode>3): Extend for AVX512.
6891 (avx512flogicsuff): New.
6892 (avx512f_<logic><mode>): Ditto.
6893 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
6895 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
6896 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
6897 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
6898 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
6899 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
6900 (xop_vpermil2<mode>3): Ditto.
6901 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
6902 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
6903 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
6904 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
6906 2013-09-25 Tom Tromey <tromey@redhat.com>
6908 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
6909 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
6910 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
6911 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
6912 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
6913 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
6914 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
6915 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
6916 (GRAPHITE_HTAB_H): Remove.
6918 2013-09-25 Tom Tromey <tromey@redhat.com>
6920 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
6922 2013-09-25 Tom Tromey <tromey@redhat.com>
6924 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
6926 2013-09-25 Tom Tromey <tromey@redhat.com>
6928 * config/i386/t-i386 (i386.o): Remove.
6929 (i386-c.o): Use COMPILE and POSTCOMPILE.
6931 2013-09-25 Tom Tromey <tromey@redhat.com>
6933 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
6935 2013-09-25 Tom Tromey <tromey@redhat.com>
6937 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
6938 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
6939 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
6940 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
6941 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
6942 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
6943 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
6944 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
6945 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
6946 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
6947 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
6948 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
6949 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
6950 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
6951 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
6952 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
6953 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
6954 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
6955 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
6956 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
6957 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
6958 (tree-ssa-pre.o, tree-ssa-sccvn.o)
6959 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
6960 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
6961 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
6962 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
6963 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
6964 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
6965 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
6966 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
6967 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
6968 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
6969 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
6970 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
6971 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
6972 (graphite-clast-to-gimple.o, graphite-dependences.o)
6973 (graphite-interchange.o, graphite-poly.o)
6974 (graphite-scop-detection.o, graphite-sese-to-poly.o)
6975 (graphite-optimize-isl.o, tree-vect-loop.o)
6976 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
6977 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
6978 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
6979 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
6980 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
6981 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
6982 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
6983 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
6984 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
6985 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
6986 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
6987 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
6988 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
6989 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
6990 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
6991 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
6992 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
6993 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
6994 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
6995 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
6996 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
6997 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
6998 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
6999 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
7000 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
7001 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
7002 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
7003 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
7004 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
7005 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
7006 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
7007 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
7008 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
7009 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
7010 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
7011 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
7012 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
7013 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
7014 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
7015 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
7016 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
7017 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
7018 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
7019 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
7020 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
7021 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
7022 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
7023 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
7024 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
7025 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
7026 (gcov-dump.o): Remove.
7027 (default-c.o): Use COMPILE and POSTCOMPILE.
7028 (CFLAGS-gcc.o): New variable.
7029 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
7031 2013-09-25 Tom Tromey <tromey@redhat.com>
7033 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
7034 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
7035 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
7036 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
7037 (c-family/c-pragma.o, c-family/c-pretty-print.o)
7038 (c-family/c-semantics.o, c-family/c-ada-spec.o)
7039 (c-family/array-notation-common.o, c-family/stub-objc.o)
7040 (c-family/c-ubsan.o): Remove.
7042 2013-09-25 Tom Tromey <tromey@redhat.com>
7044 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
7046 2013-09-25 Tom Tromey <tromey@redhat.com>
7048 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
7049 to add -DENABLE_SHARED_LIBGCC.
7050 (gcc.o): Don't use subshell.
7052 2013-09-25 Tom Tromey <tromey@redhat.com>
7054 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
7055 * configure.ac: Don't invoke AM_PROG_CC_C_O.
7056 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
7057 * configure, config.in: Rebuild.
7059 2013-09-25 Tom Tromey <tromey@redhat.com>
7061 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
7062 (COMPILE, POSTCOMPILE): New variables.
7063 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
7064 (DEPFILES): New variable.
7065 Include ".Po" files.
7066 * configure.ac: Add checks for dependency checking.
7067 * configure, aclocal.m4: Regenerate.
7069 2013-09-25 Tom Tromey <tromey@redhat.com>
7071 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
7072 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
7074 2013-09-25 Tom Tromey <tromey@redhat.com>
7076 * Makefile.in (generated_files): Add options.h,
7077 target-hooks-def.h, insn-opinit.h,
7078 common/common-target-hooks-def.h, pass-instances.def,
7079 c-family/c-target-hooks-def.h.
7081 2013-09-25 Jeff Law <law@redhat.com>
7083 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
7084 than foo[foo.length () - 1] to access last member in a vec.
7085 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
7087 2013-09-25 Richard Biener <rguenther@suse.de>
7090 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
7092 2013-09-25 Jan Hubicka <jh@suse.cz>
7094 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
7097 2013-09-25 Marek Polacek <polacek@redhat.com>
7100 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
7101 when determining the type name.
7103 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
7105 * config/sh/sh.md: Fix formatting.
7107 2013-09-24 Xinliang David Li <davidxl@google.com>
7109 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
7110 max peel iterations parameter.
7111 * param.def: New parameter.
7112 * doc/invoke.texi: Document New parameter.
7114 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
7116 * gimple-pretty-print.c: Various whitespace tweaks.
7117 * tree-core.h: Likewise.
7118 * tree-pretty-print.c: Likewise.
7119 * tree-ssa-alias.c: Likewise.
7120 * tree-ssa-copy.c: Likewise.
7121 * tree-ssanames.c: Likewise.
7122 * tree-ssanames.h: Likewise.
7123 * tree-vrp.c: Likewise.
7125 2013-09-24 Alan Modra <amodra@gmail.com>
7129 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
7130 for output operands that disallow regs. Don't use EXPAND_WRITE on
7133 2013-09-24 Richard Biener <rguenther@suse.de>
7136 * tree.c (reference_alias_ptr_type): Move ...
7137 * alias.c (reference_alias_ptr_type): ... here and implement
7138 in terms of the new reference_alias_ptr_type_1.
7139 (ref_all_alias_ptr_type_p): New helper.
7140 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
7141 use ref_all_alias_ptr_type_p.
7142 (get_deref_alias_set): Add flag_strict_aliasing check here.
7143 (reference_alias_ptr_type_1): New function, split out from ...
7144 (get_alias_set): ... here.
7145 (alias_ptr_types_compatible_p): New function.
7146 * alias.h (reference_alias_ptr_type): Declare.
7147 (alias_ptr_types_compatible_p): Likewise.
7148 * tree.h (reference_alias_ptr_type): Remove.
7149 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
7150 to compare MEM_REF alias types.
7152 2013-09-24 Richard Biener <rguenther@suse.de>
7154 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
7156 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7158 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
7159 reload helper function arrays into a single array reg_addr.
7160 (reload_fpr_gpr): Likewise.
7161 (reload_gpr_vsx): Likewise.
7162 (reload_vsx_gpr): Likewise.
7163 (struct rs6000_reg_addr): Likewise.
7164 (reg_addr): Likewise.
7165 (rs6000_debug_reg_global): Change rs6000_vector_reload,
7166 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
7167 (rs6000_init_hard_regno_mode_ok): Likewise.
7168 (rs6000_secondary_reload_direct_move): Likewise.
7169 (rs6000_secondary_reload): Likewise.
7171 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
7172 constraints: wu, ww, and wy. Repurpose wv constraint added during
7173 power8 changes. Put wg constraint in alphabetical order.
7175 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
7176 for future work to add ISA 2.07 VSX single precision support.
7177 (-mvsx-scalar-double): Change default from -1 to 1, update
7178 documentation comment.
7179 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
7180 (-mupper-regs-df): New debug switch to control whether DF values
7181 can go in the traditional Altivec registers.
7182 (-mupper-regs-sf): New debug switch to control whether SF values
7183 can go in the traditional Altivec registers.
7185 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
7187 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
7188 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
7189 Add new constraints, wu/ww/wy. Repurpose wv constraint.
7190 (rs6000_debug_legitimate_address_p): Print if we are running
7191 before, during, or after reload.
7192 (rs6000_secondary_reload): Add a comment.
7193 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
7195 * config/rs6000/constraints.md (wa constraint): Sort w<x>
7196 constraints. Update documentation string.
7197 (wd constraint): Likewise.
7198 (wf constraint): Likewise.
7199 (wg constraint): Likewise.
7200 (wn constraint): Likewise.
7201 (ws constraint): Likewise.
7202 (wt constraint): Likewise.
7203 (wx constraint): Likewise.
7204 (wz constraint): Likewise.
7205 (wu constraint): New constraint for ISA 2.07 SFmode scalar
7207 (ww constraint): Likewise.
7208 (wy constraint): Likewise.
7209 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
7210 in the previous submissions.
7211 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
7213 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
7215 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
7216 (REG_BR_PROB): Say that the probability is stored in an int_list.
7217 * reg-notes.def: Update commentary to mention INT_LIST.
7218 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
7219 (INT_LIST): New rtx.
7220 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
7221 * rtlanal.c (int_reg_note_p): New function.
7222 (alloc_reg_note): Assert that the note does not have an int argument.
7223 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
7224 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
7225 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
7226 rather than an INSN_LIST. Handle INT_LIST.
7227 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
7228 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
7229 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
7230 Manipulate them as ints rather than rtxes.
7231 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
7232 * regmove.c (copy_src_to_dest): Likewise.
7233 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
7235 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
7236 into the cases that need it.
7237 * config/arm/arm.c (arm_unwind_emit): Likewise.
7239 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
7240 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
7241 * loop-doloop.c (add_test, doloop_modify): Likewise.
7242 * loop-unswitch.c (compare_and_jump_seq): Likewise.
7243 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
7244 * predict.c (combine_predictions_for_insn): Likewise.
7245 * print-rtl.c (print_rtx): Handle INT_LIST.
7246 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
7247 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
7248 * config/arm/arm.c (emit_unlikely_jump): Likewise.
7249 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
7250 (ix86_split_fp_branch, predict_jump): Likewise.
7251 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
7252 * config/sh/sh.c (expand_cbranchsi4): Likewise.
7253 * config/spu/spu.c (ea_load_store_inline): Likewise.
7255 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
7256 value of a REG_BR_PROB note.
7257 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
7258 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
7259 * emit-rtl.c (try_split): Likewise.
7260 * predict.c (br_prob_note_reliable_p): Likewise.
7261 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
7262 * reorg.c (mostly_true_jump): Likewise.
7263 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
7264 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
7265 * config/i386/i386.c (ix86_print_operand): Likewise.
7266 * config/ia64/ia64.c (ia64_print_operand): Likewise.
7267 * config/mmix/mmix.c (mmix_print_operand): Likewise.
7268 * config/rs6000/rs6000.c (output_cbranch): Likewise.
7269 * config/s390/s390.c (s390_expand_tbegin): Likewise.
7270 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
7271 * config/sparc/sparc.c (output_cbranch): Likewise.
7272 * config/spu/spu.c (get_branch_target): Likewise.
7273 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
7274 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
7276 2013-09-23 Jan Hubicka <jh@suse.cz>
7278 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
7280 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
7281 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
7283 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
7284 (gimple_fold_call): Dump inconsistent devirtualizations; add
7285 sanity check for type based devirtualizations.
7286 * ipa-prop.c: Include ipa-utils.h
7287 (ipa_intraprocedural_devirtualization): Add sanity check.
7288 (try_make_edge_direct_virtual_call): Likewise.
7290 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
7292 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
7293 assignment statements.
7295 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
7297 * gimple-pretty-print.c (dump_ssaname_info): New function.
7298 (dump_gimple_phi): Call it.
7299 (pp_gimple_stmt_1): Likewise.
7300 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
7301 (range_info_def): Declare.
7302 * tree-pretty-print.c (pp_double_int): New function.
7303 (dump_generic_node): Call it.
7304 * tree-pretty-print.h (pp_double_int): Declare.
7305 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
7306 * tree-ssanames.h (range_info_def): New structure.
7307 (value_range_type): Move definition here.
7308 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
7310 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
7312 (set_range_info): New function.
7313 (get_range_info): Likewise.
7314 (duplicate_ssa_name_range_info): Likewise.
7315 (duplicate_ssa_name_fn): Check pointer type and call
7316 duplicate_ssa_name_range_info.
7317 * tree-ssa-copy.c (fini_copy_prop): Likewise.
7318 * tree-vrp.c (value_range_type): Remove definition, now in
7320 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
7321 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
7322 (SSA_NAME_RANGE_INFO): New macro.
7324 2013-09-23 Richard Biener <rguenther@suse.de>
7326 PR tree-optimization/58464
7327 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7328 (phi_trans_add): Change to add conditionally on being not present.
7329 (phi_translate_1): Remove recursion detection here.
7330 (phi_translate): Pre-seed the cache with NULL to catch
7331 recursion here in a more generic way.
7332 (bitmap_find_leader): Adjust comment.
7333 (get_representative_for): Dump value-numbers.
7334 (create_expression_by_pieces): Likewise.
7335 (insert_into_preds_of_block): Likewise.
7337 2013-09-23 Christian Bruel <christian.bruel@st.com>
7340 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
7341 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
7343 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
7346 2013-09-20 Renlin Li <renlin.li@arm.com>
7348 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
7350 (aarch64_expand_epilogue): Likewise.
7351 (aarch64_legitimize_reload_address): Likewise.
7353 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
7355 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
7356 NULL_TREE before pushing them onto the vector. Likewise for labels.
7358 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
7360 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
7361 comparison operators when -fno-trapping-math is in effect.
7362 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
7363 comparison operators in TFmode and assert that unsupported operators
7365 (ia64_print_operand): Likewise.
7367 2013-09-21 Jan Hubicka <jh@suse.cz>
7369 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
7370 (sse_typeless_stores): Enable for core
7371 (sse_load0_by_pxor): Likewise.
7372 (four_jump_limit): Disable for core.
7373 (pad_returns): Likewise.
7374 (avoid_vector_decode): Likewise.
7375 (fuse_cmp_and_branch): Enable for cores.
7376 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
7378 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7381 * config/pa/pa.c (pa_option_override): Disable auto increment and
7382 decrement instructions until reload is completed.
7384 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
7385 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
7386 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
7388 2013-09-20 DJ Delorie <dj@redhat.com>
7389 Nick Clifton <nickc@redhat.com>
7391 * config/rl78/rl78.c: Various whitespace and comment tweaks.
7392 (need_to_save): Save bank 0 on interrupts.
7393 (characterize_address): Strip far address wrappers.
7394 (rl78_as_legitimate_address): Likewise.
7395 (transcode_memory_rtx): Likewise.
7396 (rl78_peep_movhi_p): Disable this peephole after devirt.
7397 (rl78_propogate_register_origins): Forget all origins when a
7399 * config/rl78/rl78-virt.md: Various whitespace tweaks.
7400 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
7402 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
7403 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
7404 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
7406 * config/rl78/constraints.md: Various whitespace and paren tweaks.
7408 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7410 * config/pa/pa.md: In "scc" insn patterns, change output template to
7411 handle const0_rtx in reg_or_0_operand operands.
7413 2013-09-20 Martin Husemann <martin@NetBSD.org>
7416 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
7417 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
7419 2013-09-20 Richard Biener <rguenther@suse.de>
7422 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
7423 remembering SSA name version and block index.
7424 (new_scev_info_str): Adjust.
7425 (hash_scev_info): Likewise. Also hash the block index.
7426 (eq_scev_info): Adjust.
7427 (find_var_scev_info): Likewise.
7428 (struct instantiate_cache_entry): Remove.
7429 (struct instantiate_cache_type): Use a htab to map name, block
7431 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7432 (get_instantiated_value_entry): Likewise.
7433 (hash_idx_scev_info, eq_idx_scev_info): New functions.
7434 (instantiate_scev_name): Adjust.
7436 2013-09-20 Jeff Law <law@redhat.com>
7438 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
7440 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
7442 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
7443 Call aarch64_simd_expand_args to update op[argc].
7445 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7447 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
7450 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7452 * gengtype.c (file_rules): Added rule for *.cc files.
7453 (get_output_file_with_visibility): Give fatal message when no
7456 2013-09-20 Renlin Li <renlin.li@arm.com>
7458 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
7459 (aarch64_expand_epilogue): Likewise.
7460 (aarch64_legitimize_reload_address): Likewise.
7462 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
7465 * expr.c (expand_assignment): Remove misalignp code path.
7467 2013-09-20 Marek Polacek <polacek@redhat.com>
7470 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
7471 TYPE_PRECISION. Add asserts.
7473 2013-09-20 Richard Biener <rguenther@suse.de>
7475 PR tree-optimization/58453
7476 * tree-loop-distribution.c (distribute_loop): Apply the cost
7477 model for -ftree-loop-distribute-patterns, too.
7479 2013-09-20 Richard Biener <rguenther@suse.de>
7482 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
7483 make type comparison less strict.
7485 2013-09-20 Alan Modra <amodra@gmail.com>
7487 * configure: Regenerate.
7488 * aclocal.m4: Regenerate.
7490 2013-09-20 Marek Polacek <polacek@redhat.com>
7493 * doc/extend.texi: Document that attribute used is meant to be used
7494 on variables with static storage duration.
7496 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7498 PR tree-optimization/58472
7499 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
7500 simd_lane_access set inv_p = false.
7501 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
7502 the simduid magic VAR_DECL.
7504 2013-09-19 Jan Hubicka <jh@suse.cz>
7506 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
7508 2013-09-17 Jeff Law <law@redhat.com>
7510 * tree-ssa-dom.c (record_temporary_equivalences): New function
7511 split out of dom_opt_dom_walker::after_dom_children.
7512 (dom_opt_dom_walker::thread_across_edge): Move common code
7513 in here from dom_opt_dom_walker::after_dom_children.
7514 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
7516 2013-09-19 Jan Hubicka <jh@suse.cz>
7518 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
7519 (TARGET_GENERIC): Use PROCESOR_GENERIC
7520 (enum processor_type): Unify generic32 and 64.
7521 * i386.md (cpu): Likewise.
7522 * x86-tune.def (use_leave): Enable for generic32.
7523 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
7524 * athlon.md: Change generic64 to generic in all occurences.
7525 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
7526 (ix86_target_macros_internal): Likewise.
7527 * driver-i386.c (host_detect_local_cpu): Likewise.
7528 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
7530 (generic_memcpy, generic_memset, generic_cost): This one.
7531 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
7532 (m_GENERIC32, m_GENERIC64): Remove.
7533 (m_GENERIC): Turn into one flag.
7534 (processor_target): Unify generic tunnings.
7535 (ix86_option_override_internal): Replace generic32/64 by generic.
7536 (ix86_issue_rate): Likewise.
7537 (ix86_adjust_cost): Likewise.
7539 2013-09-19 Jan Hubicka <jh@suse.cz>
7541 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
7542 of speculative flag.
7544 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7546 * omp-low.c (expand_omp_sections): Always pass len - 1 to
7547 GOMP_sections_start, even if !exit_reachable.
7549 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7551 * lra-constraints.c (need_for_all_save_p): Use macro
7552 HARD_REGNO_CALL_PART_CLOBBERED.
7553 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
7554 set up pseudo conflict hard regs.
7556 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7559 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
7562 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7564 PR rtl-optimization/58438
7565 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
7566 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
7567 from previous subpasses.
7569 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
7571 * arm.c (arm_get_frame_offsets): Validate architecture supports
7572 LDRD/STRD before accepting the tuning preference.
7573 (arm_expand_prologue): Likewise.
7574 (arm_expand_epilogue): Likewise.
7576 2013-09-18 Richard Biener <rguenther@suse.de>
7578 PR tree-optimization/58417
7579 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
7580 have chrecs with symbols defined in the loop as operands.
7581 (chrec_fold_multiply): Likewise.
7582 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
7583 parameters before folding binary operations.
7584 (struct instantiate_cache_entry_hasher): Remove.
7585 (struct instantiate_cache_type): Use a pointer-map.
7586 (instantiate_cache_type::instantiate_cache_type): New function.
7587 (instantiate_cache_type::get): Likewise.
7588 (instantiate_cache_type::set): Likewise.
7589 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7590 (get_instantiated_value_entry): Likewise.
7591 (global_cache): New global.
7592 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
7593 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
7594 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
7595 (instantiate_scev_name): Adjust.
7596 (instantiate_scev): Construct global instead of local cache.
7597 (resolve_mixers): Likewise.
7599 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
7600 Paolo Carlini <paolo.carlini@oracle.com>
7603 * doc/implement-cxx.texi: Fix references to the C++ standards.
7605 2013-09-18 Jakub Jelinek <jakub@redhat.com>
7607 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
7608 * tree-vect-data-refs.c (vect_analyze_data_refs): For
7609 simd_lane_access drs, update also DR_ALIGNED_TO.
7611 2013-09-18 Marek Polacek <polacek@redhat.com>
7614 * doc/extend.texi: Document no_sanitize_undefined attribute.
7615 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
7616 no_sanitize_undefined attribute.
7618 2013-09-18 Nick Clifton <nickc@redhat.com>
7620 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
7621 (ASM_DECLARE_FUNCTION_NAME): Define.
7623 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
7625 * compare-elim.c (find_comparison_dom_walker): New class
7626 (find_comparisons_in_bb): Rename to
7627 find_comparison_dom_walker::before_dom_children
7628 (find_comparisons): Adjust
7629 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
7631 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
7632 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
7633 (init_walk_dominator_tree): Remove declaration.
7634 (fini_walk_dominator_tree): Remove declaration.
7635 * fwprop.c (single_def_use_dom_walker): New class
7636 (single_def_use_enter_block): Convert to
7637 single_def_use_dom_walker::before_dom_children.
7638 (single_def_use_leave_block): Convert to
7639 single_def_use_dom_walker::after_dom_children.
7640 (build_single_def_use_links): Adjust.
7641 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
7643 (find_candidates_in_block): Convert to
7644 find_candidates_dom_walker::before_dom_children.
7645 (execute_strength_reduction): Adjust.
7646 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
7647 (sese_dom_walker): New class.
7648 (sese_dom_walker::sese_dom_walker): New constructor.
7649 (sese_dom_walker::~sese_dom_walker): New destructor.
7650 (build_sese_conditions_before): Convert to
7651 sese_dom_walker::before_dom_children.
7652 (build_sese_conditions_after): Convert to
7653 sese_dom_walker::after_dom_children.
7654 (build_poly_scop): Adjust
7655 * tree-into-ssa.c (rewrite_dom_walker): New class
7656 (rewrite_enter_block): Convert to
7657 rewrite_dom_walker::before_dom_children.
7658 (rewrite_leave_block): Convert to
7659 rewrite_dom_walker::after_dom_children.
7660 (rewrite_update_dom_walker): New class.
7661 (rewrite_update_enter_block): Convert to
7662 rewrite_update_dom_walker::before_dom_children.
7663 (rewrite_update_leave_block): Convert to
7664 rewrite_update_dom_walker::after_dom_children.
7665 (rewrite_blocks, rewrite_into_ssa): Adjust.
7666 (mark_def_dom_walker): New class.
7667 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
7668 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
7669 (mark_def_sites_blocks): Convert to
7670 mark_def_dom_walker::before_dom_children.
7671 (mark_def_site_blocks): Remove.
7672 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
7673 (tree_ssa_dominator_optimize): Adjust.
7674 (dom_thread_across_edge): Convert to method
7675 dom_opt_dom_walker::thread_across_edge.
7676 (dom_opt_enter_block): Convert to member function
7677 dom_opt_dom_walker::before_dom_children.
7678 (dom_opt_leave_block): Convert to member function
7679 dom_opt_dom_walker::after_dom_children.
7680 * tree-ssa-dse.c (dse_dom_walker): New class.
7681 (dse_enter_block): Convert to member function
7682 dse_dom_walker::before_dom_children.
7683 (tree_ssa_dse): Adjust.
7684 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
7685 (determine_invariantness_stmt): Convert to method
7686 invariantness_dom_walker::before_dom_children.
7687 (determine_invariantness): Remove
7688 (move_computations_dom_walker): New class.
7689 (move_computations_stmt): Convert to method
7690 move_computations_dom_walker::before_dom_children.
7691 (move_computations, tree_ssa_lim): Adjust.
7692 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
7693 (nt_init_block): Convert to method
7694 notrappping_dom_walker::before_dom_children.
7695 (nt_fini_block): Convert to method
7696 method nontrapping_dom_walker::after_dom_children.
7697 (get_non_trapping): Adjust.
7698 * tree-ssa-pre.c (eliminate_dom_walker): New class.
7699 (eliminate_bb): Convert to method
7700 eliminate_dom_walker::before_dom_children.
7701 (eliminate_leave_block): Convert to method
7702 eliminate_dom_walker::after_dom_children.
7703 (eliminate): Adjust.
7704 * tree-ssa-strlen.c (strlen_dom_walker): New class.
7705 (strlen_enter_block): Convert to method
7706 strlen_dom_walker::before_dom_children.
7707 (strlen_leave_block): Convert to method
7708 method strlen_dom_walker::after_dom_children.
7709 (tree_ssa_strlen): Adjust.
7710 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
7711 (tree_ssa_uncprop): Adjust.
7712 (uncprop_leave_block): Convert to method
7713 uncprop_dom_walker::after_dom_children.
7714 (uncprop_leave_block): Convert to method
7715 uncprop_dom_walker::before_dom_children.
7717 2013-09-18 Bin Cheng <bin.cheng@arm.com>
7719 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
7720 branch in same basic block. Check both src and dest of the move insn.
7722 2013-09-17 Nick Clifton <nickc@redhat.com>
7724 * config/rl78/rl78-real.md (bf): New pattern.
7726 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
7727 (rl78_print_operand): Do not put a # before a %B.
7728 * config/rl78/rl78.opt: Tweak doc strings.
7730 2013-09-17 DJ Delorie <dj@redhat.com>
7732 * config/rl78/constraints.md (Wcv): Allow up to $r31.
7733 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
7734 (rl78_option_override): Likewise, if -mallregs.
7735 (is_virtual_register): Likewise.
7736 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
7737 (REGNO_OK_FOR_BASE_P): Likewise.
7738 * config/rl78/rl78.opt (-mallregs): New.
7740 2013-09-17 Nick Clifton <nickc@redhat.com>
7742 * config/rl78/rl78.c (need_to_save): Change return type to bool.
7743 For interrupt functions: save all call clobbered registers if the
7744 interrupt handler is not a leaf function.
7745 (rl78_expand_prologue): Always recompute the frame information.
7746 For interrupt functions: only select bank 0 if one of the bank 0
7747 registers is going to be psuhed.
7749 2013-09-17 DJ Delorie <dj@redhat.com>
7751 * config/rl78/constraints.md: For each W* constraint, rename to C*
7752 and create a W* constraint that checks for an optional ES: prefix
7754 * config/rl78/rl78.md (UNS_ES_ADDR): New.
7755 (es_addr): New. Used to wrap far addresses.
7756 * config/rl78/rl78-protos.h (rl78_es_addr): New.
7757 (rl78_es_base): New.
7758 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
7759 wrapped far addresses.
7760 (rl78_print_operand_1): Unwrap far addresses before processing.
7761 (rl78_lo16): Wrap far addresses in unspecs.
7762 (rl78_es_addr): New.
7763 (rl78_es_base): New.
7764 (insn_ok_now): Check for not-yet-wrapped far addresses.
7765 (transcode_memory_rtx): Properly re-wrap far addresses.
7767 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
7769 * config/sparc/t-rtems: Add leon3 multilibs.
7771 2013-09-17 Cong Hou <congh@google.com>
7773 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
7774 when checking the dot production pattern. The type of rhs operand
7775 of multiply is now checked correctly.
7777 2013-09-17 Jeff Law <law@redhat.com>
7779 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
7780 edge implied equivalences into successor phis.
7781 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
7782 here from tree-ssa-threadedge.c.
7783 (mark_threaded_blocks): When threading through a joiner, if both
7784 successors of the joiner's clone reach the same block, verify the
7785 PHI arguments are equal. If not, cancel the jump threading request.
7786 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
7787 tree-ssa-threadupdate.c
7788 (thread_across_edge): Don't check PHI argument equality when
7789 threading through joiner block here.
7791 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
7793 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
7794 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
7795 (warn_uninit, warn_uninitialized_vars,
7796 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7797 Move to tree-ssa-uninit.c.
7798 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
7799 (has_undefined_value_p): New. Pass dependant parts of
7800 ssa_undefined_value_p.
7801 (uninit_undefined_value_p): Use has_undefined_value_p.
7802 (warn_uninit, warn_uninitialized_vars,
7803 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7804 Move from tree-ssa.c.
7805 * tree-ssa.h: Adjust prototypes.
7807 2013-09-17 Jan Hubicka <jh@suse.cz>
7810 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
7811 * ipa-inline.c (can_inline_edge_p): Do not downgrade
7812 FUNCTION_NOT_OPTIMIZED.
7813 * ipa-inline-analysis.c (compute_inline_parameters): Function
7814 not optimized is not inlinable unless it is alwaysinline.
7815 (inline_analyze_function): Force calls in not optimized
7816 function not inlinable.
7818 2013-09-17 Jan Hubicka <jh@suse.cz>
7821 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
7823 * ipa.c (function_and_variable_visibility): Likewise.
7824 * ipa-profile.c (ipa_profile): Likewise.
7826 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
7829 * cgraph.c (cgraph_function_body_availability): Check for ifunc
7830 attribute, and don't inline the resolver in this case.
7832 2013-09-17 Teresa Johnson <tejohnson@google.com>
7834 * coverage.c (get_coverage_counts): Add missing newline.
7836 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7838 PR tree-optimization/58088
7839 * fold-const.c (mask_with_trailing_zeros): New function.
7840 (fold_binary_loc): Make sure we don't recurse infinitely
7841 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
7842 Use mask_with_trailing_zeros where appropriate.
7844 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
7846 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
7847 of 'prev' var to get better distance estimation.
7849 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
7851 * tree-inline.h (struct copy_body_data): Add transform_parameter.
7852 * tree-inline.c (is_parameter_of): New predicate.
7853 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
7854 a parameter has been remapped.
7855 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
7856 (optimize_inline_calls): Initialize transform_parameter.
7857 (copy_gimple_seq_and_replace_locals): Likewise.
7858 (tree_function_versioning): Likewise.
7859 (maybe_inline_call_in_expr): Likewise.
7861 2013-09-17 Nick Clifton <nickc@redhat.com>
7863 * config/msp430/msp430-protos.h: Add prototypes for new functions.
7864 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
7866 (is_attr_func): New function.
7867 (msp430_is_interrupt_func): New function.
7868 (is_naked_func): New function.
7869 (is_reentrant_func): New function.
7870 (is_critical_func): New function.
7871 (msp430_start_function): Add annotations for function attributes.
7872 (msp430_attr): New function.
7873 (msp430_attribute_table): New.
7874 (msp430_function_section): New function.
7875 (TARGET_ASM_FUNCTION_SECTION): Define.
7876 (msp430_builtin): New enum.
7877 (msp430_init_builtins): New function.
7878 (msp430_builtin_devl): New function.
7879 (msp430_expand_builtin): New function.
7880 (TARGET_INIT_BUILTINS): Define.
7881 (TARGET_EXPAND_BUILTINS): Define.
7882 (TARGET_BUILTIN_DECL): Define.
7883 (msp430_expand_prologue): Add support for naked, interrupt,
7884 critical and reentrant functions.
7885 (msp430_expand_epilogue): Likewise.
7886 (msp430_print_operand): Handle 'O' character.
7887 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
7889 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
7890 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
7891 (pushm): Use a 'n' rather than an 'i' constraint.
7892 (msp_return): Add generation of the interrupt return instruction.
7893 (disable_interrupts): New pattern.
7894 (enable_interrupts): New pattern.
7895 (push_intr_state): New pattern.
7896 (pop_intr_state): New pattern.
7897 (bic_SR): New pattern.
7898 (bis_SR): New pattern.
7899 * doc/extend.texi: Document MSP430 function attributes and builtin
7902 2013-09-17 Richard Biener <rguenther@suse.de>
7904 PR tree-optimization/58432
7905 * tree-loop-distribution.c (tree_loop_distribution): Also
7906 scan PHIs for outside loop uses and seed a partition from them.
7908 2013-09-17 Bin Cheng <bin.cheng@arm.com>
7910 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
7911 (restructure_reference): Call backtrace_base_for_ref.
7913 2013-09-17 Alan Modra <amodra@gmail.com>
7916 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
7919 2013-09-16 DJ Delorie <dj@redhat.com>
7921 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
7922 vregs location for RL78/G10.
7923 (rl78_expand_prologue): Avoid SEL on G10.
7924 (rl78_expand_epilogue): Likewise.
7925 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
7926 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
7927 __RL78_G10__ when appropriate.
7928 (ASM_SPEC): Pass -mg10 along to the assembler.
7929 * config/rl78/rl78.md (sel_rb): Disable for G10.
7930 * config/rl78/rl78.opt: Add -mg10 option.
7931 * config/rl78/t-rl78: Add -mg10 multilib.
7933 2013-09-16 Xinliang David Li <davidxl@google.com>
7935 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
7936 * omp-low.c (omp_max_vf): Ditto.
7937 (expand_omp_simd): Ditto.
7938 * tree-vectorizer.c (vectorize_loops): Ditto.
7939 (gate_vect_slp): Ditto.
7940 (gate_increase_alignment): Ditto.
7941 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
7942 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
7943 (gate_tree_vectorize): Name change.
7944 (tree_vectorize): Ditto.
7945 (pass_vectorize::gate): Call new function.
7946 (pass_vectorize::execute): Ditto.
7947 * opts.c: O3 default setting change.
7948 (finish_options): Check new flag.
7949 * doc/invoke.texi: Document new flags.
7950 * common.opt: New flags.
7952 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
7954 * doc/tm.texi.in (Cond Exec Macros): Remove node.
7955 (Condition Code): Don't reference it.
7956 * doc/tm.texi: Regenerate.
7958 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
7961 * lra-constraints.c (undo_optional_reloads): Consider all optional
7962 reload even if it did not get a hard reg.
7964 2013-09-16 Teresa Johnson <tejohnson@google.com>
7966 * dumpfile.c (dump_loc): Remove newline emission.
7967 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
7968 emission to dump_printf_loc calls where missing.
7969 (vect_mark_for_runtime_alias_test): Ditto.
7970 (vect_analyze_data_ref_dependence): Ditto.
7971 (vect_analyze_data_ref_dependences): Ditto.
7972 (vect_slp_analyze_data_ref_dependence): Ditto.
7973 (vect_slp_analyze_data_ref_dependences): Ditto.
7974 (vect_compute_data_ref_alignment): Ditto.
7975 (vect_update_misalignment_for_peel): Ditto.
7976 (vect_verify_datarefs_alignment): Ditto.
7977 (vector_alignment_reachable_p): Ditto.
7978 (vect_get_data_access_cost): Ditto.
7979 (vect_enhance_data_refs_alignment): Ditto.
7980 (vect_find_same_alignment_drs): Ditto.
7981 (vect_analyze_data_refs_alignment): Ditto.
7982 (vect_analyze_group_access): Ditto.
7983 (vect_analyze_data_ref_access): Ditto.
7984 (vect_analyze_data_ref_accesses): Ditto.
7985 (vect_prune_runtime_alias_test_list): Ditto.
7986 (vect_analyze_data_refs): Ditto.
7987 (vect_create_addr_base_for_vector_ref): Ditto.
7988 (vect_create_data_ref_ptr): Ditto.
7989 (vect_grouped_store_supported): Ditto.
7990 (vect_grouped_load_supported): Ditto.
7991 * value-prof.c (check_counter): Ditto.
7992 (check_ic_target): Ditto.
7993 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
7994 (vect_recog_widen_mult_pattern): Ditto.
7995 (vect_recog_widen_sum_pattern): Ditto.
7996 (vect_recog_over_widening_pattern): Ditto.
7997 (vect_recog_widen_shift_pattern): Ditto.
7998 (vect_recog_rotate_pattern): Ditto.
7999 (vect_recog_vector_vector_shift_pattern): Ditto.
8000 (vect_recog_divmod_pattern): Ditto.
8001 (vect_recog_mixed_size_cond_pattern): Ditto.
8002 (vect_recog_bool_pattern): Ditto.
8003 (vect_pattern_recog_1): Ditto.
8004 (vect_pattern_recog): Ditto.
8005 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
8006 (vect_is_simple_iv_evolution): Ditto.
8007 (vect_analyze_scalar_cycles_1): Ditto.
8008 (vect_get_loop_niters): Ditto.
8009 (vect_analyze_loop_1): Ditto.
8010 (vect_analyze_loop_form): Ditto.
8011 (vect_analyze_loop_operations): Ditto.
8012 (vect_analyze_loop_2): Ditto.
8013 (vect_analyze_loop): Ditto.
8014 (report_vect_op): Ditto.
8015 (vect_is_slp_reduction): Ditto.
8016 (vect_is_simple_reduction_1): Ditto.
8017 (vect_get_known_peeling_cost): Ditto.
8018 (vect_estimate_min_profitable_iters): Ditto.
8019 (vect_model_reduction_cost): Ditto.
8020 (vect_model_induction_cost): Ditto.
8021 (get_initial_def_for_induction): Ditto.
8022 (vect_create_epilog_for_reduction): Ditto.
8023 (vectorizable_reduction): Ditto.
8024 (vectorizable_induction): Ditto.
8025 (vectorizable_live_operation): Ditto.
8026 (vect_loop_kill_debug_uses): Ditto.
8027 (vect_transform_loop): Ditto.
8028 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
8029 (vect_stmt_relevant_p): Ditto.
8030 (process_use): Ditto.
8031 (vect_mark_stmts_to_be_vectorized): Ditto.
8032 (vect_model_simple_cost): Ditto.
8033 (vect_model_promotion_demotion_cost): Ditto.
8034 (vect_model_store_cost): Ditto.
8035 (vect_get_store_cost): Ditto.
8036 (vect_model_load_cost): Ditto.
8037 (vect_get_load_cost): Ditto.
8038 (vect_init_vector_1): Ditto.
8039 (vect_get_vec_def_for_operand): Ditto.
8040 (vect_finish_stmt_generation): Ditto.
8041 (vectorizable_call): Ditto.
8042 (vectorizable_conversion): Ditto.
8043 (vectorizable_assignment): Ditto.
8044 (vectorizable_shift): Ditto.
8045 (vectorizable_operation): Ditto.
8046 (vectorizable_store): Ditto.
8047 (vectorizable_load): Ditto.
8048 (vectorizable_condition): Ditto.
8049 (vect_analyze_stmt): Ditto.
8050 (vect_transform_stmt): Ditto.
8051 (vect_is_simple_use): Ditto.
8052 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
8053 (vect_can_advance_ivs_p): Ditto.
8054 (vect_update_ivs_after_vectorizer): Ditto.
8055 (vect_do_peeling_for_loop_bound): Ditto.
8056 (vect_gen_niters_for_prolog_loop): Ditto.
8057 (vect_update_inits_of_drs): Ditto.
8058 (vect_create_cond_for_alias_checks): Ditto.
8059 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
8060 (vect_build_slp_tree_1): Ditto.
8061 (vect_supported_load_permutation_p): Ditto.
8062 (vect_analyze_slp_instance): Ditto.
8063 (vect_analyze_slp): Ditto.
8064 (vect_make_slp_decision): Ditto.
8065 (vect_detect_hybrid_slp): Ditto.
8066 (vect_bb_vectorization_profitable_p): Ditto.
8067 (vect_slp_analyze_bb_1): Ditto.
8068 (vect_update_slp_costs_according_to_vf): Ditto.
8069 (vect_get_mask_element): Ditto.
8070 (vect_transform_slp_perm_load): Ditto.
8071 (vect_schedule_slp_instance): Ditto.
8072 (vect_schedule_slp): Ditto.
8073 (vect_slp_transform_bb): Ditto.
8074 * profile.c (read_profile_edge_counts): Ditto.
8075 (compute_branch_probabilities): Ditto.
8076 * coverage.c (get_coverage_counts): Ditto.
8078 2013-09-16 Diego Novillo <dnovillo@google.com>
8080 * tree-core.h: Add missing comment lines from refactoring of tree.h.
8082 2013-09-16 Jan Hubicka <jh@suse.cz>
8084 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
8085 abstract functions; for static functions check the presence of body.
8087 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8089 * config/aarch64/aarch64-simd-builtins.def (fma): New.
8090 * config/aarch64/aarch64-simd.md
8091 (aarch64_mla_elt<mode>): New.
8092 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8093 (aarch64_mls_elt<mode>): Likewise.
8094 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8095 (aarch64_fma4_elt<mode>): Likewise.
8096 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8097 (aarch64_fma4_elt_to_128v2df): Likewise.
8098 (aarch64_fma4_elt_to_64df): Likewise.
8099 (fnma<mode>4): Likewise.
8100 (aarch64_fnma4_elt<mode>): Likewise.
8101 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8102 (aarch64_fnma4_elt_to_128v2df): Likewise.
8103 (aarch64_fnma4_elt_to_64df): Likewise.
8104 * config/aarch64/iterators.md (VDQSF): New.
8105 * config/aarch64/arm_neon.h
8106 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
8107 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
8109 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8111 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
8112 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
8113 (aarch64_mul3_elt_to_128df): Likewise.
8114 (aarch64_mul3_elt_to_64v2df): Likewise.
8115 * config/aarch64/iterators.md (VEL): Also handle DFmode.
8117 (VMUL_CHANGE_NLANES) Likewise.
8120 * config/aarch64/arm_neon.h
8121 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
8123 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8125 * config/aarch64/arm_neon.h
8126 (vcvtx_high_f32_f64): Fix parameters.
8128 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8129 Uros Bizjak <ubizjak@gmail.com>
8131 * config/alpha.c: Include tree-ssanames.h.
8133 2013-09-16 Richard Biener <rguenther@suse.de>
8135 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
8136 (dot_rdg_1): Handle control_dd.
8137 (create_edge_for_control_dependence): New function.
8138 (create_rdg_edges): Add control dependences if asked for.
8139 (build_rdg): Likewise.
8140 (generate_loops_for_partition): If there are not necessary
8141 control stmts remove all their dependencies.
8142 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
8143 (distribute_loop): Pass on control dependences.
8144 (tree_loop_distribution): Compute control dependences and remove
8145 restriction on number of loop nodes.
8147 2013-09-16 Jakub Jelinek <jakub@redhat.com>
8149 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
8152 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
8154 * cse.c (try_const_anchors): Punt on CC modes.
8156 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8158 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
8160 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
8163 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
8166 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8169 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
8171 (darwin_objc1_section): Likewise.
8172 (darwin_file_end): Emit Image Info section when required.
8174 2013-09-14 Jan Hubicka <jh@suse.cz>
8176 * tree-into-ssa.c (gate_into_ssa): New.
8177 (pass_data_build_ssa): Use it.
8178 * cgraph.h (expand_thunk): Update prototype.
8179 * cgraphunit.c (analyze_function): Expand thunks early.
8180 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
8181 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
8182 set lowered flag; do not add new function.
8183 (assemble_thunks_and_aliases): Update.
8184 * tree-ssa.c (gate_init_datastructures): New gate.
8185 (pass_data_init_datastructures): Use it.
8187 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8190 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
8191 xmm register set described in the psABI.
8193 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
8195 * dwarf2out.c (should_emit_struct_debug): Add check
8196 for type_decl variable is not NULL.
8198 2013-09-13 Jacek Caban <jacek@codeweavers.com>
8200 * config.gcc: Use new winnt-c.c target hooks
8201 * config/t-winnt: New file
8202 * config/winnt-c.c: New file
8203 * doc/tm.texi.in: Document new hook
8204 * doc/tm.texi: Regenerated
8206 2013-09-13 Jan Hubicka <jh@suse.cz>
8209 * ipa-inline.c (check_callers): New function.
8210 (check_caller_edge): Remove.
8211 (want_inline_function_to_all_callers_p): Also permit alises that are
8213 (inline_to_all_callers): Terminate the walk when devirtualization
8214 introduce new calls.
8216 2013-09-13 Jan Hubicka <jh@suse.cz>
8218 * ipa-inline-analysis.c (struct growth_data): Add node.
8219 (do_estimate_growth_1): Fix detection of recursion.
8221 2013-09-13 Jakub Jelinek <jakub@redhat.com>
8223 PR tree-optimization/58392
8224 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
8225 to avoid shadowing of outer loop variable. If
8226 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
8227 replace_by_duplicate_decl simduid of loops that have it set and
8228 set dest_cfun->has_simduid_loops and/or
8229 dest_cfun->has_force_vect_loops.
8230 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
8231 instead of maybe_lookup_decl.
8232 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
8233 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
8234 call. Copy over force_vect and copy and remap simduid. Set
8235 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
8236 (copy_cfg_body): Remove blocks_to_copy argument. Use
8237 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
8238 caller. Don't set cfun->has_simduid_loops and/or
8239 cfun->has_force_vect_loops here.
8240 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
8242 (expand_call_inline, tree_function_versioning): Adjust copy_body
8245 2013-09-13 Martin Jambor <mjambor@suse.cz>
8248 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
8249 the assert if the edge was a speculative one.
8251 2013-09-13 Richard Biener <rguenther@suse.de>
8253 * tree-data-ref.h (known_dependences_p): Move here ...
8254 * tree-loop-distribution.c (known_dependences_p): ... from here.
8255 (dump_rdg_component, debug_rdg_component): Remove.
8257 (generate_loops_for_partition): Use gimple_uid instead of
8258 relying on matching stmt visit order.
8259 (rdg_build_partitions): Take starting stmt vector.
8260 (ldist_gen): Merge into ...
8261 (distribute_loop): ... this function. Do not compute starting
8263 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
8265 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8267 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
8268 Set type attribute correctly. Set predicable_short_it attribute.
8269 (cmpsi_shiftsi): Remove %? from output template.
8271 2013-09-13 Richard Biener <rguenther@suse.de>
8273 * tree-loop-distribution.c (struct rdg_component,
8274 rdg_defs_used_in_other_loops_p, free_rdg_components,
8275 rdg_build_components): Remove.
8276 (stmts_from_loop): Do not record virtual PHIs.
8277 (generate_loops_for_partition): Skip virtual PHIs.
8278 (build_rdg_partition_for_component): Rename to ...
8279 (build_rdg_partition_for_vertex): ... this and adjust.
8280 (rdg_build_partitions): Take a vector of starting vertices
8281 instead of components. Remove unnecessary leftover handling.
8282 (ldist_gen): Do not build components or record other stores.
8283 (distribute_loop): Do not distribute loops containing stmts
8286 2013-09-13 Christian Bruel <christian.bruel@st.com>
8289 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
8291 2013-09-13 Kai Tietz <ktietz@redhat.com>
8293 * config.gcc: Separate cases for mingw and cygwin targets,
8294 and add 64-bit cygwin target case.
8296 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
8297 dll-export inline-functions.
8298 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
8301 2013-09-13 Jeff Law <law@redhat.com>
8305 2013-09-06 Jeff Law <law@redhat.com>
8307 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8308 edge implied equivalences into successor phis.
8310 2013-09-12 DJ Delorie <dj@redhat.com>
8312 * config/rl78/rl78-virt.md: Change from | to \; for asm line
8315 2013-09-12 Brooks Moses <bmoses@google.com>
8318 * Makefile.in: Do not install driver binaries in $(target)/bin.
8320 2013-09-12 DJ Delorie <dj@redhat.com>
8322 * config/rl78/rl78.opt (mrelax): New.
8323 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
8324 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
8326 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
8328 (rl78_expand_epilogue): Likewise.
8330 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
8333 * lra-eliminations.c (remove_reg_equal_offset_note): New.
8334 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
8335 pointer elimination with using remove_reg_equal_offset_note.
8337 2013-09-12 DJ Delorie <dj@redhat.com>
8339 * config/msp430/: New port.
8340 * config.gcc (msp430): Added.
8341 * doc/invoke.texi: Document MSP430 options.
8342 * doc/install.texi: Document msp430-elf
8343 * doc/md.texi: Document msp430-elf
8344 * doc/contrib.texi: Document msp430-elf
8346 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
8349 2013-09-12 Martin Jambor <mjambor@suse.cz>
8352 * ipa-prop.c (remove_described_reference): Give up if the edge in the
8353 reference descriptor is NULL.
8354 (ipa_edge_removal_hook): If owning a reference descriptor, set its
8357 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8359 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
8360 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
8361 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
8362 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
8363 make_temp_ssa_name): move to tree-ssanames.h
8364 * tree-ssa-alias.h: Move prototype.
8365 * tree-ssa.h: Include tree-ssanames.h.
8366 * tree-ssanames.c (FREE_SSANAMES): Move to here.
8367 * tree-ssanames.h: New. Move items from tree-flow*.h
8368 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
8370 2013-09-12 Richard Biener <rguenther@suse.de>
8372 PR tree-optimization/58404
8373 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
8374 propagate non-invariant addresses into dereferences wrapped
8375 in component references.
8377 2013-09-12 Richard Biener <rguenther@suse.de>
8379 PR tree-optimization/58402
8380 * passes.def: Move pass_late_warn_uninitialized later.
8382 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8384 * tree-ssa.h: New. Move content from tree-flow.h and
8386 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
8387 Move prototypes belonging to tree-ssa.c.
8388 * tree-flow-inline.h (redirect_edge_var_map_def,
8389 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
8391 * gimple.h: Adjust prototypes.
8392 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
8394 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
8395 * tree.h: Move prototype to tree-ssa.h.
8396 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
8397 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
8398 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
8399 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
8400 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
8401 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
8402 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
8403 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
8404 graphite-blocking.c, graphite-clast-to-gimple.c,
8405 graphite-dependences.c, graphite-interchange.c,
8406 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
8407 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
8408 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
8409 ipa-reference.c, ipa-split.c, ipa-utils.c,
8410 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
8411 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
8412 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
8413 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
8414 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
8415 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
8416 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
8417 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
8418 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
8419 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
8420 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
8421 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
8422 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
8423 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
8424 value-prof.c, var-tracking.c,
8425 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
8427 2013-09-12 Richard Biener <rguenther@suse.de>
8429 PR tree-optimization/58396
8430 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
8431 (build_rdg): Take a loop-nest parameter, fix memleaks.
8432 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
8434 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8436 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
8439 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
8441 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
8442 The FMA instruction names should have a 'v' prefix.
8444 2013-09-12 Richard Biener <rguenther@suse.de>
8446 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
8447 (dot_rdg): Use popen instead of system in optional code.
8448 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
8449 (already_processed_vertex_p): Adjust.
8450 (has_anti_or_output_dependence, predecessor_has_mem_write,
8451 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
8452 rdg_flag_uses): Remove.
8453 (rdg_flag_vertex): Simplify.
8454 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
8456 (build_rdg_partition_for_component): Process the first vertex
8457 of a component only.
8458 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
8460 2013-09-12 Alan Modra <amodra@gmail.com>
8462 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
8464 2013-09-11 DJ Delorie <dj@redhat.com>
8465 Nick Clifton <nickc@redhat.com>
8467 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
8468 (rl78_stack_based_mem): New.
8469 * config/rl78/constraints.md (Iv08): New.
8478 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
8480 (movsi): Change from expand to insn-and-split.
8481 (ashrsi3): Clobber AX.
8485 * config/rl78/rl78.md (CC_REG): Fix.
8486 (addsi3): Allow memory and immediate operands.
8487 (addsi3_internal): Split into...
8488 (addsi3_internal_virt): ...new, and ...
8489 (addsi3_internal_real): ...new.
8491 (subsi3_internal_virt): New.
8492 (subsi3_internal_real): New.
8493 (mulsi3): Add memory operand.
8494 (mulsi3_rl78): Likewise.
8495 (mulsi3_g13): Likewise.
8496 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
8497 (cbranchqi4_real): Add more constraint options.
8498 (cbranchhi4_real): Expand pattern.
8499 (cbranchhi4_real_signed): New.
8500 (cbranchhi4_real_inverted): New.
8501 (cbranchsi4_real_lt): New.
8502 (cbranchsi4_real_ge): New.
8503 (cbranchsi4_real_signed): New.
8504 (cbranchsi4_real): New.
8506 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
8508 (lshrsi3_virt): Likewise.
8509 (ashlsi3_virt): Likewise.
8510 (cbranchqi4_virt_signed): New.
8511 (cbranchhi4_virt_signed): New.
8512 (cbranchsi4_virt): New.
8513 * config/rl78/rl78.c: Whitespace fixes throughout.
8514 (move_elim_pass): New.
8515 (pass_data_rl78_move_elim): New.
8516 (pass_rl78_move_elim): New.
8517 (make_pass_rl78_move_elim): New.
8518 (rl78_devirt_info): Run devirt earlier.
8519 (rl78_move_elim_info): New.
8520 (rl78_asm_file_start): Register it.
8521 (rl78_split_movsi): New.
8522 (rl78_as_legitimate_address): Allow virtual base registers when
8524 (rl78_addr_space_convert): Remove spurious debug stuff.
8525 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
8526 (rl78_print_operand): More cases for not printing '#'.
8527 (rl78_expand_compare): Remove most of the logic.
8528 (content_memory): New.
8529 (clear_content_memory): New.
8530 (get_content_index): New.
8531 (get_content_name): New.
8532 (display_content_memory): New.
8533 (update_content): New.
8534 (record_content): New.
8535 (already_contains): New.
8536 (insn_ok_now): Re-recog insns with virtual registers.
8537 (add_postponed_content_update): New.
8538 (process_postponed_content_update): New.
8539 (gen_and_emit_move): New.
8540 (transcode_memory_rtx): Record new location content.
8541 Use gen_and_emit_move.
8542 (force_into_acc): New.
8543 (move_to_acc): Use gen_and_emit_move.
8544 (move_from_acc): Likewise.
8545 (move_acc_to_reg): Likewise.
8546 (move_to_x): Likewise.
8547 (move_to_hl): Likewise.
8548 (move_to_de): Likewise.
8549 (rl78_alloc_physical_registers_op1): Record location content.
8550 (has_constraint): New.
8551 (rl78_alloc_physical_registers_op2): Record location content.
8553 (rl78_alloc_physical_registers_ro1): Likewise.
8554 (rl78_alloc_physical_registers_cmp): Likewise.
8555 (rl78_alloc_physical_registers_umul): Likewise.
8556 (rl78_alloc_address_registers_macax): New.
8557 (rl78_alloc_physical_registers): Initialize and set location
8558 content memory as needed.
8559 (rl78_reorg): Make sure split2 is called.
8560 (rl78_rtx_costs): New.
8562 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
8564 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
8565 for (not (neg ...)) and (neg (not ...)) cases.
8567 2013-09-11 Richard Biener <rguenther@suse.de>
8570 * passes.def: Split critical edges before late uninit warning passes.
8571 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
8573 2013-09-11 Jakub Jelinek <jakub@redhat.com>
8575 PR tree-optimization/58385
8576 * fold-const.c (build_range_check): If both low and high are NULL,
8577 use omit_one_operand_loc to preserve exp side-effects.
8579 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8581 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
8583 2013-09-11 Richard Biener <rguenther@suse.de>
8585 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
8586 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
8587 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
8588 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
8589 stmts_from_loop, known_dependences_p, build_empty_rdg,
8590 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
8591 * tree-loop-distribution.c: ... here.
8592 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
8593 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
8594 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
8595 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
8596 * tree-loop-distribution.c: ... here.
8597 * tree-loop-distribution.c: Include gimple-pretty-print.h.
8598 (struct partition_s): Add loops member.
8599 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
8600 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
8601 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
8603 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8604 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8605 Sergey Lega <sergey.s.lega@intel.com>
8606 Anna Tikhonova <anna.tikhonova@intel.com>
8607 Ilya Tocar <ilya.tocar@intel.com>
8608 Andrey Turetskiy <andrey.turetskiy@intel.com>
8609 Ilya Verbin <ilya.verbin@intel.com>
8610 Kirill Yukhin <kirill.yukhin@intel.com>
8611 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8613 * config/i386/constraints.md (k): New.
8615 * config/i386/i386.c (const regclass_map): Add new mask registers.
8616 (dbx_register_map): Ditto.
8617 (dbx64_register_map): Ditto.
8618 (svr4_dbx_register_map): Ditto.
8619 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
8621 (ix86_preferred_reload_class): Disable constants for mask registers.
8622 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
8623 (ix86_hard_regno_mode_ok): Support new mask registers.
8624 (x86_order_regs_for_local_alloc): Ditto.
8625 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
8626 (FIXED_REGISTERS): Add new mask registers.
8627 (CALL_USED_REGISTERS): Ditto.
8628 (REG_ALLOC_ORDER): Ditto.
8629 (VALID_MASK_REG_MODE): New.
8630 (FIRST_MASK_REG): Ditto.
8631 (LAST_MASK_REG): Ditto.
8632 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
8633 (MAYBE_MASK_CLASS_P): New.
8634 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
8635 (REG_CLASS_CONTENTS): Ditto.
8636 (MASK_REGNO_P): New.
8637 (ANY_MASK_REG_P): Ditto.
8638 (HI_REGISTER_NAMES): Add new mask registers.
8639 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
8640 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
8642 (attribute "type"): Add mskmov, msklog.
8643 (attribute "length_immediate"): Support them.
8644 (attribute "memory"): Ditto.
8645 (attribute "prefix_0f"): Ditto.
8646 (*movhi_internal): Support new mask registers.
8647 (*movqi_internal): Ditto.
8648 (define_split): Split out clobber pattern is a logic
8649 insn on mask registers.
8650 (*k<logic><mode>): New.
8651 (*andhi_1): Extend to support mask regs.
8652 (*andqi_1): Extend to support mask regs.
8654 (define_split): Split and-not to and and not if operands
8656 (*<code><mode>_1): Separate HI mode to new pattern...
8657 (*<code>hi_1): This.
8658 (*<code>qi_1): Extend to support mask regs.
8660 (kortestzhi): Ditto.
8661 (kortestchi): Ditto.
8663 (*one_cmpl<mode>2_1): Remove HImode and handle it...
8664 (*one_cmplhi2_1): ...Here, now with mask registers support.
8665 (*one_cmplqi2_1): Support new mask registers.
8666 (HI/QImode arithmetics splitter): Don't split if mask registers
8668 (HI/QImode not splitter): Ditto.
8669 * config/i386/predicated.md (mask_reg_operand): New.
8670 (general_reg_operand): Ditto.
8672 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8674 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
8675 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
8677 2013-09-10 Jeff Law <law@redhat.com>
8679 PR tree-optimization/58380
8680 * tree-ssa-threadupdate.c (thread_block): Recognize another case
8681 of threading through a buried loop header.
8683 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
8684 return value for single successor case.
8686 2013-09-10 Jan Hubicka <jh@suse.cz>
8688 * ipa-devirt.c (ipa_devirt): Enable with LTO.
8690 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
8693 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
8694 support conditional execution.
8695 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
8697 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
8699 * lra.c (lra): Clear lra_optional_reload_pseudos before every
8701 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
8702 Check destination too to check move insn.
8703 (undo_optional_reloads): Add check that the original peudo did not
8704 changed its allocation and the optional reload was inherited on last
8705 inheritance pass. Break loop after deciding to keep optional reload.
8706 (lra_undo_inheritance): Add check that inherited pseudo still in
8709 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
8711 * config/aarch64/aarch64.md (generic_sched): New.
8712 * config/aarch64/aarch64-generic.md (load): Make conditional
8713 on generic_sched attribute.
8714 (nonload): Likewise.
8716 2013-09-10 Jan Hubicka <jh@suse.cz>
8718 * lto-cgraph.c: Include ipa-utils.h.
8719 (compute_ltrans_boundary): Also add possible targets into the boundary.
8721 2013-09-10 Jan Hubicka <jh@suse.cz>
8723 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
8724 VAR_DECL of vtable rather than full expression.
8726 2013-09-10 Jan Hubicka <jh@suse.cz>
8727 Paolo Carlini <paolo.carlini@oracle.com>
8729 * cgraphunit.c (analyze_functions): Save input_location, set it
8730 to UNKNOWN_LOCATION and restore it at the end.
8732 2013-09-10 Martin Jambor <mjambor@suse.cz>
8734 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
8735 represented by a thunk.
8737 2013-09-10 Jeff Law <law@redhat.com>
8739 PR tree-optimization/58343
8740 * tree-ssa-threadupdate.c (thread_block): Identify and disable
8741 jump threading requests through loop headers buried in the middle
8742 of a jump threading path.
8744 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
8745 in return value/type.
8747 2013-09-10 Jakub Jelinek <jakub@redhat.com>
8749 PR rtl-optimization/58365
8750 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
8751 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
8754 2013-09-10 Richard Biener <rguenther@suse.de>
8756 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
8757 * tree-data-ref.c (create_rdg_vertices): Collect all data
8758 references, signal failure to the caller, use data-ref API.
8759 (build_rdg): Compute data references only once. Maintain lifetime
8760 of data references and data dependences from within RDG.
8761 (free_rdg): Free dependence relations.
8762 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
8763 inventing extra dependences.
8764 (distribute_loop): Update for RDG API changes.
8766 2013-09-10 Kai Tietz <ktietz@redhat.com>
8768 * doc/invoke.texi (fms-extensions): Document changed
8769 behavior for ms-abi targets.
8770 * config/i386/i386.c (ix86_option_override_internal):
8771 Set default value of option -fms-extension for ms-abi targets.
8773 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
8775 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
8777 2013-09-10 Alan Modra <amodra@gmail.com>
8780 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
8782 2013-09-10 Alan Modra <amodra@gmail.com>
8784 * config/rs6000/predicates.md (add_cint_operand): New.
8785 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
8786 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
8787 using add_cint_operand.
8788 (largetoc_high_plus_aix): Likewise.
8790 2013-09-09 Jakub Jelinek <jakub@redhat.com>
8792 PR tree-optimization/58364
8793 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
8794 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
8795 the current range can't be an unconditional true or false.
8797 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
8799 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
8801 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
8803 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
8805 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
8808 * doc/invoke.texi (-Wdelete-incomplete): Document it.
8810 2013-09-09 Ian Bolton <ian.bolton@arm.com>
8812 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
8813 NO_REGS for immediate that can't be moved directly into FP_REGS.
8815 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8817 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
8818 comparison with negated operand.
8819 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
8822 2013-09-09 Richard Biener <rguenther@suse.de>
8825 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
8826 of a subloop record all its block as affecting loop-closed SSA form.
8828 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8830 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
8831 of an rtx/bitpos pair.
8832 (store_fixed_bit_field): Update accordingly.
8834 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8836 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
8838 * builtins.c (expand_errno_check): Likewise.
8839 * dwarf2cfi.c (init_return_column_size): Likewise.
8840 * except.c (sjlj_mark_call_sites): Likewise.
8841 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
8842 * lra-constraints.c (emit_inc): Likewise.
8843 * ree.c (combine_set_extension): Likewise.
8844 * regmove.c (fixup_match_2): Likewise.
8845 * reload1.c (inc_for_reload): Likewise.
8847 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8849 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
8850 (force_to_mode, simplify_shift_const_1, simplify_comparison):
8851 Use gen_int_mode with the mode of the associated simplify_* call.
8852 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
8853 * expmed.c (expand_shift_1): Likewise.
8854 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8855 * loop-iv.c (iv_number_of_iterations): Likewise.
8856 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8857 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
8859 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8861 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
8862 of the associated expand_* call.
8863 (asan_emit_stack_protection): Likewise.
8864 * builtins.c (round_trampoline_addr): Likewise.
8865 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
8866 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
8867 (emit_store_flag): Likewise.
8868 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
8869 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
8871 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8872 * ifcvt.c (noce_try_store_flag_constants): Likewise.
8873 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8874 * modulo-sched.c (generate_prolog_epilog): Likewise.
8875 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
8876 (expand_ctz, expand_ffs, expand_unop): Likewise.
8878 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8880 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
8881 of the associated gen_rtx_* call.
8882 * caller-save.c (init_caller_save): Likewise.
8883 * combine.c (find_split_point, make_extraction): Likewise.
8884 (make_compound_operation): Likewise.
8885 * dwarf2out.c (mem_loc_descriptor): Likewise.
8886 * explow.c (plus_constant, probe_stack_range): Likewise.
8887 * expmed.c (expand_mult_const): Likewise.
8888 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
8889 * reload1.c (init_reload): Likewise.
8890 * valtrack.c (cleanup_auto_inc_dec): Likewise.
8891 * var-tracking.c (adjust_mems): Likewise.
8892 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
8893 rather than gen_rtx_fmt_ee.
8895 2013-09-09 Jan Hubicka <jh@suse.cz>
8898 * value-prof.c (gimple_ic): Copy also abnormal edges.
8900 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8902 * asan.c (asan_shadow_cst): Use gen_int_mode.
8904 2013-09-08 Jan Hubicka <jh@suse.cz>
8906 * ipa-profile.c: Add toplevel comment.
8907 (ipa_propagate_frequency_1): Be more conservative when profile is read.
8908 (contains_hot_call_p): New function.
8909 (ipa_propagate_frequency): Set frequencies based on counts when
8911 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
8912 profile; do not tamper with profile after inlining if it is read.
8914 2013-09-08 Jan Hubicka <jh@suse.cz>
8916 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
8919 2013-09-08 Jan Hubicka <jh@suse.cz>
8921 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
8924 2013-09-08 Jeff Law <law@redhat.com>
8927 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
8930 2013-09-08 Andi Kleen <ak@linux.intel.com>
8932 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
8934 2013-09-08 Jan Hubicka <jh@suse.cz>
8936 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
8938 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
8940 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
8941 for non-debug insns.
8942 * lra.c (new_insn_reg): Take the containing insn as a parameter.
8943 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
8944 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
8947 2013-09-08 Jan Hubicka <jh@suse.cz>
8949 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
8950 targets and devirtualize to BUILT_IN_UNREACHABLE.
8951 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
8952 * ipa.c (walk_polymorphic_call_targets): New function.
8953 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
8954 functions; use the new timevar.
8955 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
8956 was removed from the program.
8957 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
8958 not consider it in the walk when its vtable is dead.
8959 (possible_polymorphic_call_targets_1): Pass anonymous flag to
8961 (devirt_variable_node_removal_hook): New function.
8962 (possible_polymorphic_call_targets): Also register
8963 devirt_variable_node_removal_hook.
8964 (ipa_devirt): Do not do non-speculative devirtualization.
8965 (gate_ipa_devirt): One execute if devirtualizing speculatively.
8967 2013-09-08 Jan Hubicka <jh@suse.cz>
8969 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
8970 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
8971 varpool_remove_variable_insertion_hook): Declare.
8972 * varpool.c (varpool_node_hook_list): New structure.
8973 (first_varpool_node_removal_hook,
8974 first_varpool_variable_insertion_hook): New variables.
8975 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
8976 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
8977 varpool_remove_variable_insertion_hook,
8978 varpool_call_variable_insertion_hooks): New functions.
8979 (varpool_remove_node): Use it.
8981 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
8984 * diagnostic.c (diagnostic_build_prefix): When s.file is
8985 "<built-in>" don't output line and column numbers.
8987 2013-09-06 Jan Hubicka <jh@suse.cz>
8989 * cgraphunit.c (expand_thunk): Get body before touching arguments.
8990 * lto-streamer-out.c: Stream thunks, too.
8991 * lto-streamer-in.c (input_function): Pop cfun here
8992 (lto_read_body): Instead of here.
8994 2013-09-06 Caroline Tice <cmtice@google.com>
8996 * doc/install.texi: Add documentation for the --enable-vtable-verify
8997 and the --disable-libvtv configure options.
8999 2013-09-06 Jeff Law <law@redhat.com>
9001 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
9002 edge implied equivalences into successor phis.
9004 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
9006 * resource.c (mark_referenced_resources): Handle COND_EXEC.
9008 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
9010 * resource.c (mark_target_live_regs): Compute resources taking
9011 into account if a call is predicated or not.
9013 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9015 * toplev.c (output_stack_usage): Be prepared for suffixes created by
9016 the compiler in the function names.
9018 2013-09-06 Jan Hubicka <jh@suse.cz>
9021 * ipa-inline.c (has_caller_p): New function.
9022 (want_inline_function_to_all_callers_p): Use it.
9023 (sum_callers, inline_to_all_callers): Break out from ...
9024 (ipa_inline): ... here.
9026 2013-09-06 Jan Hubicka <jh@suse.cz>
9028 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
9029 only when AVX is enabled.
9031 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9033 * config/aarch64/aarch64.md
9034 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
9035 is fpsimd_<load/store>2.
9036 (load_pair<mode>): Likewise.
9037 (store_pair<mode>): Likewise.
9039 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9041 * config/arm/types.md (type): Add "mrs" type.
9042 * config/aarch64/aarch64.md
9043 (aarch64_load_tp_hard): Make type "mrs".
9045 (load_tp_hard): Make type "mrs".
9046 * config/arm/cortex-a15.md: Update with new attributes.
9047 * config/arm/cortex-a5.md: Update with new attributes.
9048 * config/arm/cortex-a53.md: Update with new attributes.
9049 * config/arm/cortex-a7.md: Update with new attributes.
9050 * config/arm/cortex-a8.md: Update with new attributes.
9051 * config/arm/cortex-a9.md: Update with new attributes.
9052 * config/arm/cortex-m4.md: Update with new attributes.
9053 * config/arm/cortex-r4.md: Update with new attributes.
9054 * config/arm/fa526.md: Update with new attributes.
9055 * config/arm/fa606te.md: Update with new attributes.
9056 * config/arm/fa626te.md: Update with new attributes.
9057 * config/arm/fa726te.md: Update with new attributes.
9059 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9061 * config/aarch64/aarch64.md
9062 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
9063 (*movtf_aarch64): Likewise.
9065 (thumb1_movdi_insn): Use "multiple" for type where more than one
9066 instruction is used for a move.
9067 (*arm32_movhf): Likewise.
9068 (*thumb_movdf_insn): Likewise.
9070 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9072 * config/arm/types.md (type): Rename fcpys to fmov.
9074 (*arm_movsi_vfp): Rename type fcpys as fmov.
9075 (*thumb2_movsi_vfp): Likewise
9076 (*movhf_vfp_neon): Likewise
9077 (*movhf_vfp): Likewise
9078 (*movsf_vfp): Likewise
9079 (*thumb2_movsf_vfp): Likewise
9080 (*movsfcc_vfp): Likewise
9081 (*thumb2_movsfcc_vfp): Likewise
9082 * config/aarch64/aarch64-simd.md
9083 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
9084 * config/aarch64/aarch64.md
9085 (*movsi_aarch64): Replace type mov_reg with fmovs.
9086 (*movdi_aarch64): Likewise
9087 (*movsf_aarch64): Likewise
9088 (*movdf_aarch64): Likewise
9090 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
9091 * config/arm/iwmmxt.md
9092 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
9093 * config/arm/arm1020e.md: Update with new attributes.
9094 * config/arm/cortex-a15-neon.md: Update with new attributes.
9095 * config/arm/cortex-a5.md: Update with new attributes.
9096 * config/arm/cortex-a53.md: Update with new attributes.
9097 * config/arm/cortex-a7.md: Update with new attributes.
9098 * config/arm/cortex-a8-neon.md: Update with new attributes.
9099 * config/arm/cortex-a9.md: Update with new attributes.
9100 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9101 * config/arm/cortex-r4f.md: Update with new attributes.
9102 * config/arm/marvell-pj4.md: Update with new attributes.
9103 * config/arm/vfp11.md: Update with new attributes.
9105 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9107 * config/aarch64/aarch64.md
9108 (*madd<mode>): Fix type attribute.
9109 (*maddsi_uxtw): Likewise.
9110 (*msub<mode>): Likewise.
9111 (*msubsi_uxtw): Likewise.
9112 (<su_optab>maddsidi4): Likewise.
9113 (<su_optab>msubsidi4): Likewise.
9115 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9117 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
9118 * config/arm/arm.md (core_cycles): Remove fdiv.
9119 * config/arm/vfp.md:
9120 (*sqrtsf2_vfp): Update for attribute changes.
9121 (*sqrtdf2_vfp): Likewise.
9122 * config/aarch64/aarch64.md:
9123 (sqrt<mode>2): Update for attribute changes.
9124 * config/arm/arm1020e.md: Update with new attributes.
9125 * config/arm/cortex-a15-neon.md: Update with new attributes.
9126 * config/arm/cortex-a5.md: Update with new attributes.
9127 * config/arm/cortex-a53.md: Update with new attributes.
9128 * config/arm/cortex-a7.md: Update with new attributes.
9129 * config/arm/cortex-a8-neon.md: Update with new attributes.
9130 * config/arm/cortex-a9.md: Update with new attributes.
9131 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9132 * config/arm/cortex-r4f.md: Update with new attributes.
9133 * config/arm/marvell-pj4.md: Update with new attributes.
9134 * config/arm/vfp11.md: Update with new attributes.
9136 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9138 * config/arm/types.md
9139 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
9140 * config/aarch64/aarch64.md
9141 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
9143 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
9144 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
9145 (float<GPI:mode><GPF:mode>2): Likewise.
9147 (*truncsisf2_vfp): Update with new attributes.
9148 (*truncsidf2_vfp): Likewise.
9149 (fixuns_truncsfsi2): Likewise.
9150 (fixuns_truncdfsi2): Likewise.
9151 (*floatsisf2_vfp): Likewise.
9152 (*floatsidf2_vfp): Likewise.
9153 (floatunssisf2): Likewise.
9154 (floatunssidf2): Likewise.
9155 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
9156 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
9157 * config/arm/arm1020e.md: Update with new attributes.
9158 * config/arm/cortex-a15-neon.md: Update with new attributes.
9159 * config/arm/cortex-a5.md: Update with new attributes.
9160 * config/arm/cortex-a53.md: Update with new attributes.
9161 * config/arm/cortex-a7.md: Update with new attributes.
9162 * config/arm/cortex-a8-neon.md: Update with new attributes.
9163 * config/arm/cortex-a9.md: Update with new attributes.
9164 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9165 * config/arm/cortex-r4f.md: Update with new attributes.
9166 * config/arm/marvell-pj4.md: Update with new attributes.
9167 * config/arm/vfp11.md: Update with new attributes.
9169 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9171 * config/aarch64/arm_neon.h
9172 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
9173 (vqtbx<1,2,3,4><q>_s8): Likewise.
9175 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9177 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
9178 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
9179 (add<mode>3): Add type attribute.
9180 (add<mode>3): Likewise.
9181 (usadd<mode>3): Likewise.
9182 (ssadd<mode>3): Likewise.
9183 (sub<mode>3): Likewise.
9184 (sub<mode>3): Likewise.
9185 (ussub<mode>3): Likewise.
9186 (sssub<mode>3): Likewise.
9187 (ssmulsa3): Likewise.
9188 (usmulusa3): Likewise.
9189 (arm_usatsihi): Likewise.
9191 (*movdi_vfp): Add types for all instructions.
9192 (*movdi_vfp_cortexa8): Likewise.
9193 (*movhf_vfp_neon): Likewise.
9194 (*movhf_vfp): Likewise.
9195 (*movdf_vfp): Likewise.
9196 (*thumb2_movdf_vfp): Likewise.
9197 (*thumb2_movdfcc_vfp): Likewise.
9198 * config/arm/arm.md: Add type attribute to all insn patterns.
9199 (*thumb1_adddi3): Add type attribute.
9200 (*arm_adddi3): Likewise.
9201 (*adddi_sesidi_di): Likewise.
9202 (*adddi_zesidi_di): Likewise.
9203 (*thumb1_addsi3): Likewise.
9204 (addsi3_compare0): Likewise.
9205 (*addsi3_compare0_scratch): Likewise.
9206 (*compare_negsi_si): Likewise.
9207 (cmpsi2_addneg): Likewise.
9208 (*addsi3_carryin_<optab>): Likewise.
9209 (*addsi3_carryin_alt2_<optab>): Likewise.
9210 (*addsi3_carryin_clobercc_<optab>): Likewise.
9211 (*subsi3_carryin): Likewise.
9212 (*subsi3_carryin_const): Likewise.
9213 (*subsi3_carryin_compare): Likewise.
9214 (*subsi3_carryin_compare_const): Likewise.
9215 (*arm_subdi3): Likewise.
9216 (*thumb_subdi3): Likewise.
9217 (*subdi_di_zesidi): Likewise.
9218 (*subdi_di_sesidi): Likewise.
9219 (*subdi_zesidi_di): Likewise.
9220 (*subdi_sesidi_di): Likewise.
9221 (*subdi_zesidi_ze): Likewise.
9222 (thumb1_subsi3_insn): Likewise.
9223 (*arm_subsi3_insn): Likewise.
9224 (*anddi3_insn): Likewise.
9225 (*anddi_zesidi_di): Likewise.
9226 (*anddi_sesdi_di): Likewise.
9227 (*ne_zeroextracts): Likewise.
9228 (*ne_zeroextracts): Likewise.
9229 (*ite_ne_zeroextr): Likewise.
9230 (*ite_ne_zeroextr): Likewise.
9231 (*anddi_notdi_di): Likewise.
9232 (*anddi_notzesidi): Likewise.
9233 (*anddi_notsesidi): Likewise.
9234 (andsi_notsi_si): Likewise.
9235 (thumb1_bicsi3): Likewise.
9236 (*iordi3_insn): Likewise.
9237 (*iordi_zesidi_di): Likewise.
9238 (*iordi_sesidi_di): Likewise.
9239 (*thumb1_iorsi3_insn): Likewise.
9240 (*xordi3_insn): Likewise.
9241 (*xordi_zesidi_di): Likewise.
9242 (*xordi_sesidi_di): Likewise.
9243 (*arm_xorsi3): Likewise.
9244 (*andsi_iorsi3_no): Likewise.
9245 (*smax_0): Likewise.
9246 (*smax_m1): Likewise.
9247 (*arm_smax_insn): Likewise.
9248 (*smin_0): Likewise.
9249 (*arm_smin_insn): Likewise.
9250 (*arm_umaxsi3): Likewise.
9251 (*arm_uminsi3): Likewise.
9252 (*minmax_arithsi): Likewise.
9253 (*minmax_arithsi_): Likewise.
9254 (*satsi_<SAT:code>): Likewise.
9255 (arm_ashldi3_1bit): Likewise.
9256 (arm_ashrdi3_1bit): Likewise.
9257 (arm_lshrdi3_1bit): Likewise.
9258 (*arm_negdi2): Likewise.
9259 (*thumb1_negdi2): Likewise.
9260 (*arm_negsi2): Likewise.
9261 (*thumb1_negsi2): Likewise.
9262 (*negdi_extendsid): Likewise.
9263 (*negdi_zero_extend): Likewise.
9264 (*arm_abssi2): Likewise.
9265 (*thumb1_abssi2): Likewise.
9266 (*arm_neg_abssi2): Likewise.
9267 (*thumb1_neg_abss): Likewise.
9268 (one_cmpldi2): Likewise.
9269 (extend<mode>di2): Likewise.
9270 (*compareqi_eq0): Likewise.
9271 (*arm_extendhisi2addsi): Likewise.
9272 (*arm_movdi): Likewise.
9273 (*thumb1_movdi_insn): Likewise.
9274 (*arm_movt): Likewise.
9275 (*thumb1_movsi_insn): Likewise.
9276 (pic_add_dot_plus_four): Likewise.
9277 (pic_add_dot_plus_eight): Likewise.
9278 (tls_load_dot_plus_eight): Likewise.
9279 (*thumb1_movhi_insn): Likewise.
9280 (*thumb1_movsf_insn): Likewise.
9281 (*movdf_soft_insn): Likewise.
9282 (*thumb_movdf_insn): Likewise.
9283 (cbranchsi4_insn): Likewise.
9284 (cbranchsi4_scratch): Likewise.
9285 (*negated_cbranchsi4): Likewise.
9286 (*tbit_cbranch): Likewise.
9287 (*tlobits_cbranch): Likewise.
9288 (*tstsi3_cbranch): Likewise.
9289 (*cbranchne_decr1): Likewise.
9290 (*addsi3_cbranch): Likewise.
9291 (*addsi3_cbranch_scratch): Likewise.
9292 (*arm_cmpdi_insn): Likewise.
9293 (*arm_cmpdi_unsig): Likewise.
9294 (*arm_cmpdi_zero): Likewise.
9295 (*thumb_cmpdi_zero): Likewise.
9296 (*deleted_compare): Likewise.
9297 (*mov_scc): Likewise.
9298 (*mov_negscc): Likewise.
9299 (*mov_notscc): Likewise.
9300 (*cstoresi_eq0_thumb1_insn): Likewise.
9301 (cstoresi_nltu_thumb1): Likewise.
9302 (cstoresi_ltu_thu): Likewise.
9303 (thumb1_addsi3_addgeu): Likewise.
9304 (*arm_jump): Likewise.
9305 (*thumb_jump): Likewise.
9306 (*check_arch2): Likewise.
9307 (arm_casesi_internal): Likewise.
9308 (thumb1_casesi_dispatch): Likewise.
9309 (*arm_indirect_jump): Likewise.
9310 (*thumb1_indirect_jump): Likewise.
9312 (*and_scc): Likewise.
9313 (*ior_scc): Likewise.
9314 (*compare_scc): Likewise.
9315 (*cond_move): Likewise.
9316 (*cond_arith): Likewise.
9317 (*cond_sub): Likewise.
9318 (*cmp_ite0): Likewise.
9319 (*cmp_ite1): Likewise.
9320 (*cmp_and): Likewise.
9321 (*cmp_ior): Likewise.
9322 (*ior_scc_scc): Likewise.
9323 (*ior_scc_scc_cmp): Likewise.
9324 (*and_scc_scc): Likewise.
9325 (*and_scc_scc_cmp): Likewise.
9326 (*and_scc_scc_nod): Likewise.
9327 (*negscc): Likewise.
9328 (movcond_addsi): Likewise.
9329 (movcond): Likewise.
9330 (*ifcompare_plus_move): Likewise.
9331 (*if_plus_move): Likewise.
9332 (*ifcompare_move_plus): Likewise.
9333 (*if_move_plus): Likewise.
9334 (*ifcompare_arith_arith): Likewise.
9335 (*if_arith_arith): Likewise.
9336 (*ifcompare_arith_move): Likewise.
9337 (*if_arith_move): Likewise.
9338 (*ifcompare_move_arith): Likewise.
9339 (*if_move_arith): Likewise.
9340 (*ifcompare_move_not): Likewise.
9341 (*if_move_not): Likewise.
9342 (*ifcompare_not_move): Likewise.
9343 (*if_not_move): Likewise.
9344 (*ifcompare_shift_move): Likewise.
9345 (*if_shift_move): Likewise.
9346 (*ifcompare_move_shift): Likewise.
9347 (*if_move_shift): Likewise.
9348 (*ifcompare_shift_shift): Likewise.
9349 (*ifcompare_not_arith): Likewise.
9350 (*ifcompare_arith_not): Likewise.
9351 (*if_arith_not): Likewise.
9352 (*ifcompare_neg_move): Likewise.
9353 (*if_neg_move): Likewise.
9354 (*ifcompare_move_neg): Likewise.
9355 (*if_move_neg): Likewise.
9356 (prologue_thumb1_interwork): Likewise.
9357 (*cond_move_not): Likewise.
9358 (*sign_extract_onebit): Likewise.
9359 (*not_signextract_onebit): Likewise.
9360 (stack_tie): Likewise.
9361 (align_4): Likewise.
9362 (align_8): Likewise.
9363 (consttable_end): Likewise.
9364 (consttable_1): Likewise.
9365 (consttable_2): Likewise.
9366 (consttable_4): Likewise.
9367 (consttable_8): Likewise.
9368 (consttable_16): Likewise.
9369 (*thumb1_tablejump): Likewise.
9370 (prefetch): Likewise.
9371 (force_register_use): Likewise.
9372 (thumb_eh_return): Likewise.
9373 (load_tp_hard): Likewise.
9374 (load_tp_soft): Likewise.
9375 (tlscall): Likewise.
9376 (*arm_movtas_ze): Likewise.
9377 (*arm_rev): Likewise.
9378 (*arm_revsh): Likewise.
9379 (*arm_rev16): Likewise.
9380 * config/arm/thumb2.md
9381 (*thumb2_smaxsi3): Likewise.
9382 (*thumb2_sminsi3): Likewise.
9383 (*thumb32_umaxsi3): Likewise.
9384 (*thumb2_uminsi3): Likewise.
9385 (*thumb2_negdi2): Likewise.
9386 (*thumb2_abssi2): Likewise.
9387 (*thumb2_neg_abss): Likewise.
9388 (*thumb2_movsi_insn): Likewise.
9389 (tls_load_dot_plus_four): Likewise.
9390 (*thumb2_movhi_insn): Likewise.
9391 (*thumb2_mov_scc): Likewise.
9392 (*thumb2_mov_negs): Likewise.
9393 (*thumb2_mov_negs): Likewise.
9394 (*thumb2_mov_nots): Likewise.
9395 (*thumb2_mov_nots): Likewise.
9396 (*thumb2_movsicc_): Likewise.
9397 (*thumb2_movsfcc_soft_insn): Likewise.
9398 (*thumb2_indirect_jump): Likewise.
9399 (*thumb2_and_scc): Likewise.
9400 (*thumb2_ior_scc): Likewise.
9401 (*thumb2_ior_scc_strict_it): Likewise.
9402 (*thumb2_cond_move): Likewise.
9403 (*thumb2_cond_arith): Likewise.
9404 (*thumb2_cond_ari): Likewise.
9405 (*thumb2_cond_sub): Likewise.
9406 (*thumb2_negscc): Likewise.
9407 (*thumb2_movcond): Likewise.
9408 (thumb2_casesi_internal): Likewise.
9409 (thumb2_casesi_internal_pic): Likewise.
9410 (*thumb2_alusi3_short): Likewise.
9411 (*thumb2_mov<mode>_shortim): Likewise.
9412 (*thumb2_addsi_short): Likewise.
9413 (*thumb2_subsi_short): Likewise.
9414 (thumb2_addsi3_compare0): Likewise.
9415 (*thumb2_cbz): Likewise.
9416 (*thumb2_cbnz): Likewise.
9417 (*thumb2_one_cmplsi2_short): Likewise.
9418 (*thumb2_negsi2_short): Likewise.
9419 (*orsi_notsi_si): Likewise.
9420 * config/arm/arm1020e.md: Update with new attributes.
9421 * config/arm/arm1026ejs.md: Update with new attributes.
9422 * config/arm/arm1136jfs.md: Update with new attributes.
9423 * config/arm/arm926ejs.md: Update with new attributes.
9424 * config/arm/cortex-a15.md: Update with new attributes.
9425 * config/arm/cortex-a5.md: Update with new attributes.
9426 * config/arm/cortex-a53.md: Update with new attributes.
9427 * config/arm/cortex-a7.md: Update with new attributes.
9428 * config/arm/cortex-a8.md: Update with new attributes.
9429 * config/arm/cortex-a9.md: Update with new attributes.
9430 * config/arm/cortex-m4.md: Update with new attributes.
9431 * config/arm/cortex-r4.md: Update with new attributes.
9432 * config/arm/fa526.md: Update with new attributes.
9433 * config/arm/fa606te.md: Update with new attributes.
9434 * config/arm/fa626te.md: Update with new attributes.
9435 * config/arm/fa726te.md: Update with new attributes.
9437 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9439 * config/aarch64/aarch64-simd.md
9440 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
9441 <vwx> iterator to ensure correct register choice.
9442 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
9443 (aarch64_sqdmull_n<mode>): Likewise.
9444 (aarch64_sqdmull2_n<mode>_internal): Likewise.
9445 * config/aarch64/arm_neon.h
9446 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
9447 (vml<as><q>_n_<su>16): Likewise.
9448 (vml<as>l_high_lane<q>_<su>16): Likewise.
9449 (vml<as>l_high_n_<su>16): Likewise.
9450 (vml<as>l_lane<q>_<su>16): Likewise.
9451 (vml<as>l_n_<su>16): Likewise.
9452 (vmul<q>_lane<q>_<su>16): Likewise.
9453 (vmul<q>_n_<su>16): Likewise.
9454 (vmull_lane<q>_<su>16): Likewise.
9455 (vmull_n_<su>16): Likewise.
9456 (vmull_high_lane<q>_<su>16): Likewise.
9457 (vmull_high_n_<su>16): Likewise.
9458 (vqrdmulh<q>_n_s16): Likewise.
9460 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
9462 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
9463 have the correct lane parameter.
9465 2013-09-06 Richard Biener <rguenther@suse.de>
9467 * cfganal.c (control_dependences::~control_dependences):
9468 Properly free all of the vector.
9470 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
9473 * config/i386/i386.c (ix86_conditional_register_usage):
9474 Proper initialize extended SSE registers.
9476 2013-09-06 Jan Hubicka <jh@suse.cz>
9478 PR tree-optimization/58311
9479 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
9481 2013-09-06 Jan Hubicka <jh@suse.cz>
9483 * Makefile.in (tree-sra.o): Update dependencies.
9484 * tree-sra.c: Include ipa-utils.h
9485 (scan_function): Use recursive_call_p.
9486 (has_caller_p): New function.
9487 (cgraph_for_node_and_aliases): Count also callers of aliases.
9489 2013-09-06 Jan Hubicka <jh@suse.cz>
9492 * cgraph.h (symtab_semantically_equivalent_p): Declare.
9493 * tree-tailcall.c: Include ipa-utils.h.
9494 (find_tail_calls): Use it.
9495 * ipa-pure-const.c (check_call): Likewise.
9496 * ipa-utils.c (recursive_call_p): New function.
9497 * ipa-utils.h (recursive_call_p): Dclare.
9498 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
9499 (symtab_semantically_equivalent_p): New function.
9500 * Makefile.in (tree-tailcall.o): Update dependencies.
9502 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9504 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
9507 2013-09-06 Richard Biener <rguenther@suse.de>
9509 * lto-streamer.h (lto_global_var_decls): Remove.
9510 * Makefile.in (OBJS): Remove lto-symtab.o.
9511 (lto-symtab.o): Remove.
9512 (GTFILES): Remove lto-symtab.c
9513 * lto-symtab.c: Move to lto/
9515 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9517 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
9518 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
9519 (UNSPEC_FPINT_RINT): New constant definitions.
9520 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
9521 definition with 2 attributes.
9522 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
9523 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
9526 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9528 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
9529 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
9530 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
9531 attribute to "z196_alone".
9532 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
9535 2013-09-06 Richard Biener <rguenther@suse.de>
9537 * basic-block.h (class control_dependences): New.
9538 * tree-ssa-dce.c (control_dependence_map): Remove.
9540 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
9541 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
9542 find_pdom, find_control_dependence, find_all_control_dependences):
9544 (mark_control_dependent_edges_necessary,
9545 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
9546 tree_dce_done, perform_tree_ssa_dce): Adjust.
9547 * cfganal.c (set_control_dependence_map_bit,
9548 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
9549 find_all_control_dependences): Move from tree-ssa-dce.c and
9550 implement as methods of control_dependences class.
9551 (control_dependences::control_dependences): New.
9552 (control_dependences::~control_dependences): Likewise.
9553 (control_dependences::get_edges_dependent_on): Likewise.
9554 (control_dependences::get_edge): Likewise.
9556 2013-09-04 Jan Hubicka <jh@suse.cz>
9558 * tree.c (types_same_for_odr): Drop overactive check.
9559 * ipa-devirt.c (hash_type_name): Likewise.
9561 2013-09-04 Jan Hubicka <jh@suse.cz>
9563 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
9564 (analyze_functions): ... here.
9566 2013-09-04 Jan Hubicka <jh@suse.cz>
9569 * cgraphunit.c (analyze_functions): Clear AUX fields
9570 after processing; initialize assembler name has.
9572 2013-09-05 Jeff Law <law@redhat.com>
9574 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
9575 from thread_around_empty_block. Record threading path into PATH.
9576 Recurse if threading through the initial block is successful.
9577 (thread_across_edge): Corresponding changes to slightly simplify.
9579 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9581 * config/aarch64/aarch64.md
9582 (type): Remove frecpe, frecps, frecpx.
9583 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
9584 fix to be a TARGET_SIMD instruction.
9585 (aarch64_frecps): Remove.
9586 * config/aarch64/aarch64-simd.md
9587 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
9588 (aarch64_frecps<mode>): Handle all float/vector of float modes.
9590 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9591 Sofiane Naci <sofiane.naci@arm.com>
9593 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
9594 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
9595 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
9596 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
9597 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
9598 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
9599 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
9600 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
9601 into "clz, "rbit". Rename "shift" to "shift_imm".
9602 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
9603 changes. Update for attribute changes all occurrences of arlo_* and
9605 * config/arm/arm-fixed.md: Update for attribute changes
9606 all occurrences of arlo_* types.
9607 * config/arm/thumb2.md: Update for attribute changes all occurrences
9609 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
9610 (cortexa7_older_only): Likewise.
9611 (cortexa7_younger): Likewise.
9612 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9613 (1020alu_shift_op): Likewise.
9614 (1020alu_shift_reg_op): Likewise.
9615 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9616 (alu_shift_op): Likewise.
9617 (alu_shift_reg_op): Likewise.
9618 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9619 (11_alu_shift_op): Likewise.
9620 (11_alu_shift_reg_op): Likewise.
9621 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9622 (9_alu_shift_reg_op): Likewise.
9623 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
9625 (cortex_a15_alu_shift): Likewise.
9626 (cortex_a15_alu_shift_reg): Likewise.
9627 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
9629 (cortex_a5_alu_shift): Likewise.
9630 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
9632 (cortex_a53_alu_shift): Likewise.
9633 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
9635 (cortex_a7_alu_reg): Likewise.
9636 (cortex_a7_alu_shift): Likewise.
9637 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
9639 (cortex_a8_alu_shift): Likewise.
9640 (cortex_a8_alu_shift_reg): Likewise.
9641 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9642 (cortex_a9_dp_shift): Likewise.
9643 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
9645 * config/arm/cortex-r4.md
9646 (cortex_r4_alu): Update for attribute changes.
9647 (cortex_r4_mov): Likewise.
9648 (cortex_r4_alu_shift_reg): Likewise.
9649 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9650 (526_alu_shift_op): Likewise.
9651 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9652 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9653 (626te_alu_shift_op): Likewise.
9654 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
9655 (726te_alu_shift_op): Likewise.
9656 (726te_alu_shift_reg_op): Likewise.
9657 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9658 (mp626_alu_shift_op): Likewise.
9659 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
9660 (pj4_alu_conds): Likewise.
9661 (pj4_shift): Likewise.
9662 (pj4_shift_conds): Likewise.
9663 (pj4_alu_shift): Likewise.
9664 (pj4_alu_shift_conds): Likewise.
9665 * config/aarch64/aarch64.md: Update for attribute change
9666 all occurrences of arlo_* and shift* types.
9668 2013-09-05 Mike Stump <mikestump@comcast.net>
9670 * tree.h: Move documentation for tree_function_decl to tree-core.h
9671 with the declaration.
9673 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
9676 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
9677 looking for widest mode.
9679 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
9681 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
9683 2013-09-05 Richard Biener <rguenther@suse.de>
9685 PR tree-optimization/58137
9686 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
9687 Do not create vectors of pointers.
9688 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
9689 types for the components of the vector initializer.
9690 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
9691 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
9693 2013-09-05 Martin Jambor <mjambor@suse.cz>
9695 * ipa-prop.c (remove_described_reference): Accept missing references,
9696 return false if that hppens, otherwise return true.
9697 (cgraph_node_for_jfunc): New function.
9698 (try_decrement_rdesc_refcount): Likewise.
9699 (try_make_edge_direct_simple_call): Use them.
9700 (ipa_edge_removal_hook): Remove references from rdescs.
9701 (ipa_edge_duplication_hook): Clone rdescs and their references
9702 when the new edge has the same caller as the old one.
9703 * cgraph.c (cgraph_resolve_speculation): Remove speculative
9704 reference before removing any edges.
9706 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
9708 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
9710 * thumb2.md (thumb2_storewb_parisi): New pattern.
9712 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
9714 * config/aarch64/aarch64-option-extensions.def: Add
9715 AARCH64_OPT_EXTENSION of 'crc'.
9716 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
9717 (AARCH64_ISA_CRC): Ditto.
9718 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
9719 description of the CRC extension.
9721 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
9723 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
9724 * config/rs6000/linux.h: Ditto.
9725 * alpha/linux.h: Ditto.
9726 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
9727 no_c99_libc_has_function.
9728 * config/c6x/uclinux-elf.h: Ditto.
9729 * config/lm32/uclinux-elf.h: Ditto.
9730 * config/m68k/uclinux.h: Ditto.
9731 * config/moxie/uclinux.h: Ditto.
9732 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
9733 (crisv32-*-linux*, cris-*-linux*): Ditto.
9734 * config/bfin/bfin.c: Include "tm_p.h".
9736 2013-09-05 Richard Biener <rguenther@suse.de>
9738 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
9739 check for a definition without a basic-block.
9741 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9742 Sofiane Naci <sofiane.naci@arm.com>
9744 * config/aarch64/aarch64.md
9745 (*movti_aarch64): Rename r_2_f and f_2_r.
9746 (*movsf_aarch64): Likewise.
9747 (*movdf_aarch64): Likewise.
9748 (*movtf_aarch64): Likewise.
9749 (aarch64_movdi_<mode>low): Likewise.
9750 (aarch64_movdi_<mode>high): Likewise.
9751 (aarch64_mov<mode>high_di): Likewise.
9752 (aarch64_mov<mode>low_di): Likewise.
9753 (aarch64_movtilow_tilow): Likewise.
9754 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
9755 values to config/arm/types.md
9756 (attribute "conds"): Update for attribute change.
9757 (anddi3_insn): Likewise.
9758 (iordi3_insn): Likewise.
9759 (xordi3_insn): Likewise.
9760 (one_cmpldi2): Likewise.
9761 * config/arm/types.md (type): Add Neon types.
9762 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
9763 use "type" attribute.
9764 (movmisalign<mode>_neon_store): Likewise.
9765 (movmisalign<mode>_neon_load): Likewise.
9766 (vec_set<mode>_internal): Likewise.
9767 (vec_setv2di_internal): Likewise.
9768 (vec_extract<mode>): Likewise.
9769 (vec_extractv2di): Likewise.
9770 (add<mode>3_neon): Likewise.
9771 (adddi3_neon): Likewise.
9772 (sub<mode>3_neon): Likewise.
9773 (subdi3_neon): Likewise.
9774 (mul<mode>3_neon): Likewise.
9775 (mul<mode>3add<mode>_neon): Likewise.
9776 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
9777 (fma<VCVTF:mode>4)): Likewise.
9778 (fma<VCVTF:mode>4_intrinsic): Likewise.
9779 (fmsub<VCVTF:mode>4)): Likewise.
9780 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
9781 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
9782 (ior<mode>3): Likewise.
9783 (and<mode>3): Likewise.
9784 (anddi3_neon): Likewise.
9785 (orn<mode>3_neon): Likewise.
9786 (orndi3_neon): Likewise.
9787 (bic<mode>3_neon): Likewise.
9788 (bicdi3_neon): Likewise.
9789 (xor<mode>3): Likewise.
9790 (one_cmpl<mode>2): Likewise.
9791 (abs<mode>2): Likewise.
9792 (neg<mode>2): Likewise.
9793 (umin<mode>3_neon): Likewise.
9794 (umax<mode>3_neon): Likewise.
9795 (smin<mode>3_neon): Likewise.
9796 (smax<mode>3_neon): Likewise.
9797 (vashl<mode>3): Likewise.
9798 (vashr<mode>3_imm): Likewise.
9799 (vlshr<mode>3_imm): Likewise.
9800 (ashl<mode>3_signed): Likewise.
9801 (ashl<mode>3_unsigned): Likewise.
9802 (neon_load_count): Likewise.
9803 (ashldi3_neon_noclobber): Likewise.
9804 (signed_shift_di3_neon): Likewise.
9805 (unsigned_shift_di3_neon): Likewise.
9806 (ashrdi3_neon_imm_noclobber): Likewise.
9807 (lshrdi3_neon_imm_noclobber): Likewise.
9808 (widen_ssum<mode>3): Likewise.
9809 (widen_usum<mode>3): Likewise.
9810 (quad_halves_<code>v4si): Likewise.
9811 (quad_halves_<code>v4sf): Likewise.
9812 (quad_halves_<code>v8hi): Likewise.
9813 (quad_halves_<code>v16qi): Likewise.
9814 (reduc_splus_v2di): Likewise.
9815 (neon_vpadd_internal<mode>): Likewise.
9816 (neon_vpsmin<mode>): Likewise.
9817 (neon_vpsmax<mode>): Likewise.
9818 (neon_vpumin<mode>): Likewise.
9819 (neon_vpumax<mode>): Likewise.
9820 (ss_add<mode>_neon): Likewise.
9821 (us_add<mode>_neon): Likewise.
9822 (ss_sub<mode>_neon): Likewise.
9823 (us_sub<mode>_neon): Likewise.
9824 (neon_vadd<mode>_unspec): Likewise.
9825 (neon_vaddl<mode>): Likewise.
9826 (neon_vaddw<mode>): Likewise.
9827 (neon_vhadd<mode>): Likewise.
9828 (neon_vqadd<mode>): Likewise.
9829 (neon_vaddhn<mode>): Likewise.
9830 (neon_vmul<mode>): Likewise.
9831 (neon_vmla<mode>): Likewise.
9832 (neon_vmlal<mode>): Likewise.
9833 (neon_vmls<mode>): Likewise.
9834 (neon_vmlsl<mode>): Likewise.
9835 (neon_vqdmulh<mode>): Likewise.
9836 (neon_vqdmlal<mode>): Likewise.
9837 (neon_vqdmlsl<mode>): Likewise.
9838 (neon_vmull<mode>): Likewise.
9839 (neon_vqdmull<mode>): Likewise.
9840 (neon_vsub<mode>_unspec): Likewise.
9841 (neon_vsubl<mode>): Likewise.
9842 (neon_vsubw<mode>): Likewise.
9843 (neon_vqsub<mode>): Likewise.
9844 (neon_vhsub<mode>): Likewise.
9845 (neon_vsubhn<mode>): Likewise.
9846 (neon_vceq<mode>): Likewise.
9847 (neon_vcge<mode>): Likewise.
9848 (neon_vcgeu<mode>): Likewise.
9849 (neon_vcgt<mode>): Likewise.
9850 (neon_vcgtu<mode>): Likewise.
9851 (neon_vcle<mode>): Likewise.
9852 (neon_vclt<mode>): Likewise.
9853 (neon_vcage<mode>): Likewise.
9854 (neon_vcagt<mode>): Likewise.
9855 (neon_vtst<mode>): Likewise.
9856 (neon_vabd<mode>): Likewise.
9857 (neon_vabdl<mode>): Likewise.
9858 (neon_vaba<mode>): Likewise.
9859 (neon_vabal<mode>): Likewise.
9860 (neon_vmax<mode>): Likewise.
9861 (neon_vmin<mode>): Likewise.
9862 (neon_vpaddl<mode>): Likewise.
9863 (neon_vpadal<mode>): Likewise.
9864 (neon_vpmax<mode>): Likewise.
9865 (neon_vpmin<mode>): Likewise.
9866 (neon_vrecps<mode>): Likewise.
9867 (neon_vrsqrts<mode>): Likewise.
9868 (neon_vqabs<mode>): Likewise.
9869 (neon_vqneg<mode>): Likewise.
9870 (neon_vcls<mode>): Likewise.
9871 (clz<mode>2): Likewise.
9872 (popcount<mode>2): Likewise.
9873 (neon_vrecpe): Likewise.
9874 (neon_vrsqrte): Likewise.
9875 (neon_vget_lane<mode>_sext_internal): Likewise.
9876 (neon_vget_lane<mode>_zext_internal): Likewise.
9877 (neon_vdup_n<mode>): Likewise.
9878 (neon_vdup_nv2di): Likewise.
9879 (neon_vdpu_lane<mode>_internal): Likewise.
9880 (neon_vswp<mode>): Likewise.
9881 (float<mode><V_cvtto>2): Likewise.
9882 (floatuns<mode><V_cvtto>2): Likewise.
9883 (fix_trunc<mode><V_cvtto>)2): Likewise
9884 (fixuns_trunc<mode><V_cvtto)2): Likewise.
9885 (neon_vcvt<mode>): Likewise.
9886 (neon_vcvtv4sfv4hf): Likewise.
9887 (neon_vcvtv4hfv4sf): Likewise.
9888 (neon_vcvt_n<mode>): Likewise.
9889 (neon_vmovn<mode>): Likewise.
9890 (neon_vqmovn<mode>): Likewise.
9891 (neon_vqmovun<mode>): Likewise.
9892 (neon_vmovl<mode>): Likewise.
9893 (neon_vmul_lane<mode>): Likewise.
9894 (neon_vmull_lane<mode>): Likewise.
9895 (neon_vqdmull_lane<mode>): Likewise.
9896 (neon_vqdmulh_lane<mode>): Likewise.
9897 (neon_vmla_lane<mode>): Likewise.
9898 (neon_vmlal_lane<mode>): Likewise.
9899 (neon_vqdmlal_lane<mode>): Likewise.
9900 (neon_vmls_lane<mode>): Likewise.
9901 (neon_vmlsl_lane<mode>): Likewise.
9902 (neon_vqdmlsl_lane<mode>): Likewise.
9903 (neon_vext<mode>): Likewise.
9904 (neon_vrev64<mode>): Likewise.
9905 (neon_vrev32<mode>): Likewise.
9906 (neon_vrev16<mode>): Likewise.
9907 (neon_vbsl<mode>_internal): Likewise.
9908 (neon_vshl<mode>): Likewise.
9909 (neon_vqshl<mode>): Likewise.
9910 (neon_vshr_n<mode>): Likewise.
9911 (neon_vshrn_n<mode>): Likewise.
9912 (neon_vqshrn_n<mode>): Likewise.
9913 (neon_vqshrun_n<mode>): Likewise.
9914 (neon_vshl_n<mode>): Likewise.
9915 (neon_vqshl_n<mode>): Likewise.
9916 (neon_vqshlu_n<mode>): Likewise.
9917 (neon_vshll_n<mode>): Likewise.
9918 (neon_vsra_n<mode>): Likewise.
9919 (neon_vsri_n<mode>): Likewise.
9920 (neon_vsli_n<mode>): Likewise.
9921 (neon_vtbl1v8qi): Likewise.
9922 (neon_vtbl2v8qi): Likewise.
9923 (neon_vtbl3v8qi): Likewise.
9924 (neon_vtbl4v8qi): Likewise.
9925 (neon_vtbx1v8qi): Likewise.
9926 (neon_vtbx2v8qi): Likewise.
9927 (neon_vtbx3v8qi): Likewise.
9928 (neon_vtbx4v8qi): Likewise.
9929 (neon_vtrn<mode>_internal): Likewise.
9930 (neon_vzip<mode>_internal): Likewise.
9931 (neon_vuzp<mode>_internal): Likewise.
9932 (neon_vld1<mode>): Likewise.
9933 (neon_vld1_lane<mode>): Likewise.
9934 (neon_vld1_dup<mode>): Likewise.
9935 (neon_vld1_dupv2di): Likewise.
9936 (neon_vst1<mode>): Likewise.
9937 (neon_vst1_lane<mode>): Likewise.
9938 (neon_vld2<mode>): Likewise.
9939 (neon_vld2_lane<mode>): Likewise.
9940 (neon_vld2_dup<mode>): Likewise.
9941 (neon_vst2<mode>): Likewise.
9942 (neon_vst2_lane<mode>): Likewise.
9943 (neon_vld3<mode>): Likewise.
9944 (neon_vld3qa<mode>): Likewise.
9945 (neon_vld3qb<mode>): Likewise.
9946 (neon_vld3_lane<mode>): Likewise.
9947 (neon_vld3_dup<mode>): Likewise.
9948 (neon_vst3<mode>): Likewise.
9949 (neon_vst3qa<mode>): Likewise.
9950 (neon_vst3qb<mode>): Likewise.
9951 (neon_vst3_lane<mode>): Likewise.
9952 (neon_vld4<mode>): Likewise.
9953 (neon_vld4qa<mode>): Likewise.
9954 (neon_vld4qb<mode>): Likewise.
9955 (neon_vld4_lane<mode>): Likewise.
9956 (neon_vld4_dup<mode>): Likewise.
9957 (neon_vst4<mode>): Likewise.
9958 (neon_vst4qa<mode>): Likewise.
9959 (neon_vst4qb<mode>): Likewise.
9960 (neon_vst4_lane<mode>): Likewise.
9961 (neon_vec_unpack<US>_lo_<mode>): Likewise.
9962 (neon_vec_unpack<US>_hi_<mode>): Likewise.
9963 (neon_vec_<US>mult_lo_<mode>): Likewise.
9964 (neon_vec_<US>mult_hi_<mode>): Likewise.
9965 (neon_vec_<US>shiftl_<mode>): Likewise.
9966 (neon_unpack<US>_<mode>): Likewise.
9967 (neon_vec_<US>mult_<mode>): Likewise.
9968 (vec_pack_trunc_<mode>): Likewise.
9969 (neon_vec_pack_trunk_<mode>): Likewise.
9970 (neon_vabd<mode>_2): Likewise.
9971 (neon_vabd<mode>_3): Likewise.
9972 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
9973 (thumb2_movsi_vfp): Likewise.
9974 (movdi_vfp): Likewise.
9975 (movdi_vfp_cortexa8): Likewise.
9976 (movhf_vfp_neon): Likewise.
9977 (movhf_vfp): Likewiwse.
9978 (movsf_vfp): Likewiwse.
9979 (thumb2_movsf_vfp): Likewiwse.
9980 (movdf_vfp): Likewise.
9981 (thumb2_movdf_vfp): Likewise.
9982 (movsfcc_vfp): Likewise.
9983 (thumb2_movsfcc_vfp): Likewise.
9984 (movdfcc_vfp): Likewise.
9985 (thumb2_movdfcc_vfp): Likewise.
9986 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
9987 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
9988 (v10_v2c): Likewise.
9989 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
9991 (cortex_a15_neon_int_2): Likewise.
9992 (cortex_a15_neon_int_3): Likewise.
9993 (cortex_a15_neon_int_4): Likewise.
9994 (cortex_a15_neon_int_5): Likewise.
9995 (cortex_a15_neon_vqneg_vqabs): Likewise.
9996 (cortex_a15_neon_vmov): Likewise.
9997 (cortex_a15_neon_vaba): Likewise.
9998 (cortex_a15_neon_vaba_qqq): Likewise.
9999 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10000 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10001 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
10002 scalar_64_32_long_scalar): Likewise.
10003 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10004 (cortex_a15_neon_mla_qqq_8_16): Likewise.
10005 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10006 lotype_qdd_64_32_long): Likewise.
10007 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10008 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10009 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
10010 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10011 (cortex_a15_neon_shift_1): Likewise.
10012 (cortex_a15_neon_shift_2): Likewise.
10013 (cortex_a15_neon_shift_3): Likewise.
10014 (cortex_a15_neon_vshl_ddd): Likewise.
10015 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10016 (cortex_a15_neon_vsra_vrsra): Likewise.
10017 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
10018 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
10019 (cortex_a15_neon_fp_vmul_ddd): Likewise.
10020 (cortex_a15_neon_fp_vmul_qqd): Likewise.
10021 (cortex_a15_neon_fp_vmla_ddd): Likewise.
10022 (cortex_a15_neon_fp_vmla_qqq): Likewise.
10023 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
10024 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
10025 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10026 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10027 (cortex_a15_neon_bp_simple): Likewise.
10028 (cortex_a15_neon_bp_2cycle): Likewise.
10029 (cortex_a15_neon_bp_3cycle): Likewise.
10030 (cortex_a15_neon_vld1_1_2_regs): Likewise.
10031 (cortex_a15_neon_vld1_3_4_regs): Likewise.
10032 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10033 (cortex_a15_neon_vld2_4_regs): Likewise.
10034 (cortex_a15_neon_vld3_vld4): Likewise.
10035 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10036 (cortex_a15_neon_vst1_3_4_regs): Likewise.
10037 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
10038 (cortex_a15_neon_vst3_vst4): Likewise.
10039 (cortex_a15_neon_vld1_vld2_lane): Likewise.
10040 (cortex_a15_neon_vld3_vld4_lane" 10
10041 (cortex_a15_neon_vst1_vst2_lane): Likewise.
10042 (cortex_a15_neon_vst3_vst4_lane): Likewise.
10043 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
10044 (cortex_a15_neon_ldm_2): Likewise.0
10045 (cortex_a15_neon_stm_2): Likewise.
10046 (cortex_a15_neon_mcr): Likewise.
10047 (cortex_a15_neon_mcr_2_mcrr): Likewise.
10048 (cortex_a15_neon_mrc): Likewise.
10049 (cortex_a15_neon_mrrc): Likewise.
10050 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
10052 (cortex_a15_alu_shift): Likewise.
10053 (cortex_a15_alu_shift_reg): Likewise.
10054 (cortex_a15_mult32): Likewise.
10055 (cortex_a15_mult64): Likewise.
10056 (cortex_a15_block): Likewise.
10057 (cortex_a15_branch): Likewise.
10058 (cortex_a15_load1): Likewise.
10059 (cortex_a15_load3): Likewise.
10060 (cortex_a15_store1): Likewise.
10061 (cortex_a15_store3): Likewise.
10062 (cortex_a15_call): Likewise.
10063 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
10064 (cortex_a5_f2r): Likewise.
10065 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
10067 (cortex_a53_f2r): Likewise.
10068 * config/arm/cortex-a7.md
10069 (cortex_a7_branch): Update for attribute change.
10070 (cortex_a7_call): Likewise.
10071 (cortex_a7_alu_imm): Likewise.
10072 (cortex_a7_alu_reg): Likewise.
10073 (cortex_a7_alu_shift): Likewise.
10074 (cortex_a7_mul): Likewise.
10075 (cortex_a7_load1): Likewise.
10076 (cortex_a7_store1): Likewise.
10077 (cortex_a7_load2): Likewise.
10078 (cortex_a7_store2): Likewise.
10079 (cortex_a7_load3): Likewise.
10080 (cortex_a7_store3): Likewise.
10081 (cortex_a7_load4): Likewise.
10082 (cortex_a7_store4): Likewise.
10083 (cortex_a7_fpalu): Likewise.
10084 (cortex_a7_fconst): Likewise.
10085 (cortex_a7_fpmuls): Likewise.
10086 (cortex_a7_neon_mul): Likewise.
10087 (cortex_a7_fpmacs): Likewise.
10088 (cortex_a7_neon_mla: Likewise.
10089 (cortex_a7_fpmuld: Likewise.
10090 (cortex_a7_fpmacd: Likewise.
10091 (cortex_a7_fpfmad: Likewise.
10092 (cortex_a7_fdivs: Likewise.
10093 (cortex_a7_fdivd: Likewise.
10094 (cortex_a7_r2f: Likewise.
10095 (cortex_a7_f2r: Likewise.
10096 (cortex_a7_f_flags: Likewise.
10097 (cortex_a7_f_loads: Likewise.
10098 (cortex_a7_f_loadd: Likewise.
10099 (cortex_a7_f_stores: Likewise.
10100 (cortex_a7_f_stored: Likewise.
10101 (cortex_a7_neon): Likewise.
10102 * config/arm/cortex-a8-neon.md
10103 (cortex_a8_neon_mrc): Update for attribute change.
10104 (cortex_a8_neon_mrrc): Likewise.
10105 (cortex_a8_neon_int_1): Likewise.
10106 (cortex_a8_neon_int_2): Likewise.
10107 (cortex_a8_neon_int_3): Likewise.
10108 (cortex_a8_neon_int_4): Likewise.
10109 (cortex_a8_neon_int_5): Likewise.
10110 (cortex_a8_neon_vqneg_vqabs): Likewise.
10111 (cortex_a8_neon_vmov): Likewise.
10112 (cortex_a8_neon_vaba): Likewise.
10113 (cortex_a8_neon_vaba_qqq): Likewise.
10114 (cortex_a8_neon_vsma): Likewise.
10115 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10116 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10117 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10118 long_scalar): Likewise.
10119 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10120 (cortex_a8_neon_mla_qqq_8_16): Likewise.
10121 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10122 long_scalar_qdd_64_32_long): Likewise.
10123 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10124 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10125 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
10126 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10127 (cortex_a8_neon_shift_1): Likewise.
10128 (cortex_a8_neon_shift_2): Likewise.
10129 (cortex_a8_neon_shift_3): Likewise.
10130 (cortex_a8_neon_vshl_ddd): Likewise.
10131 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10132 (cortex_a8_neon_vsra_vrsra): Likewise.
10133 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
10134 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
10135 (cortex_a8_neon_fp_vsum): Likewise.
10136 (cortex_a8_neon_fp_vmul_ddd): Likewise.
10137 (cortex_a8_neon_fp_vmul_qqd): Likewise.
10138 (cortex_a8_neon_fp_vmla_ddd): Likewise.
10139 (cortex_a8_neon_fp_vmla_qqq): Likewise.
10140 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
10141 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
10142 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10143 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10144 (cortex_a8_neon_bp_simple): Likewise.
10145 (cortex_a8_neon_bp_2cycle): Likewise.
10146 (cortex_a8_neon_bp_3cycle): Likewise.
10147 (cortex_a8_neon_ldr): Likewise.
10148 (cortex_a8_neon_str): Likewise.
10149 (cortex_a8_neon_vld1_1_2_regs): Likewise.
10150 (cortex_a8_neon_vld1_3_4_regs): Likewise.
10151 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10152 (cortex_a8_neon_vld2_4_regs): Likewise.
10153 (cortex_a8_neon_vld3_vld4): Likewise.
10154 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10155 (cortex_a8_neon_vst1_3_4_regs): Likewise.
10156 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
10157 (cortex_a8_neon_vst3_vst4): Likewise.
10158 (cortex_a8_neon_vld1_vld2_lane): Likewise.
10159 (cortex_a8_neon_vld3_vld4_lane): Likewise.
10160 (cortex_a8_neon_vst1_vst2_lane): Likewise.
10161 (cortex_a8_neon_vst3_vst4_lane): Likewise.
10162 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
10163 (cortex_a8_neon_mcr): Likewise.
10164 (cortex_a8_neon_mcr_2_mcrr): Likewise.
10165 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10166 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
10168 (ca9_neon_mrrc): Likewise.
10169 (cortex_a9_neon_int_1): Likewise.
10170 (cortex_a9_neon_int_2): Likewise.
10171 (cortex_a9_neon_int_3): Likewise.
10172 (cortex_a9_neon_int_4): Likewise.
10173 (cortex_a9_neon_int_5): Likewise.
10174 (cortex_a9_neon_vqneg_vqabs): Likewise.
10175 (cortex_a9_neon_vmov): Likewise.
10176 (cortex_a9_neon_vaba): Likewise.
10177 (cortex_a9_neon_vaba_qqq): Likewise.
10178 (cortex_a9_neon_vsma): Likewise.
10179 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10180 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10181 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10182 long_scalar): Likewise.
10183 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10184 (cortex_a9_neon_mla_qqq_8_16): Likewise.
10185 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10186 long_scalar_qdd_64_32_long): Likewise.
10187 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10188 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10189 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
10190 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10191 (cortex_a9_neon_shift_1): Likewise.
10192 (cortex_a9_neon_shift_2): Likewise.
10193 (cortex_a9_neon_shift_3): Likewise.
10194 (cortex_a9_neon_vshl_ddd): Likewise.
10195 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10196 (cortex_a9_neon_vsra_vrsra): Likewise.
10197 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
10198 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
10199 (cortex_a9_neon_fp_vsum): Likewise.
10200 (cortex_a9_neon_fp_vmul_ddd): Likewise.
10201 (cortex_a9_neon_fp_vmul_qqd): Likewise.
10202 (cortex_a9_neon_fp_vmla_ddd): Likewise.
10203 (cortex_a9_neon_fp_vmla_qqq): Likewise.
10204 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
10205 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
10206 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10207 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10208 (cortex_a9_neon_bp_simple): Likewise.
10209 (cortex_a9_neon_bp_2cycle): Likewise.
10210 (cortex_a9_neon_bp_3cycle): Likewise.
10211 (cortex_a9_neon_ldr): Likewise.
10212 (cortex_a9_neon_str): Likewise.
10213 (cortex_a9_neon_vld1_1_2_regs): Likewise.
10214 (cortex_a9_neon_vld1_3_4_regs): Likewise.
10215 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10216 (cortex_a9_neon_vld2_4_regs): Likewise.
10217 (cortex_a9_neon_vld3_vld4): Likewise.
10218 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10219 (cortex_a9_neon_vst1_3_4_regs): Likewise.
10220 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
10221 (cortex_a9_neon_vst3_vst4): Likewise.
10222 (cortex_a9_neon_vld1_vld2_lane): Likewise.
10223 (cortex_a9_neon_vld3_vld4_lane): Likewise.
10224 (cortex_a9_neon_vst1_vst2_lane): Likewise.
10225 (cortex_a9_neon_vst3_vst4_lane): Likewise.
10226 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
10227 (cortex_a9_neon_mcr): Likewise.
10228 (cortex_a9_neon_mcr_2_mcrr): Likewise.
10229 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
10230 (cortex_a9_fps): Likewise.
10231 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
10233 (cortex_m4_fmuls): Likewise.
10234 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
10236 (cortex_r4_mrc): Likewise.
10237 * config/arm/iterators.md: Update comment referring to neon_type.
10238 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
10239 (iwmmxt_movsi_insn): Likewise.
10240 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
10242 (pj4_core_to_vfp): Likewise.
10243 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
10245 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
10246 (vfp_fstore): Likewise.
10247 * doc/md.texi: Change references to neon_type to refer to type.
10249 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10251 * tree.h (DECL_BUILT_IN): Fix typo in comment.
10253 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
10255 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
10256 lglobl if not weak.
10258 2013-09-04 Easwaran Raman <eraman@google.com>
10260 PR middle-end/57370
10261 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
10262 (build_and_add_sum): Use it.
10263 (appears_later_in_bb): Simplify code.
10265 2013-09-04 Teresa Johnson <tejohnson@google.com>
10267 * dumpfile.c (dump_finish): Don't close stderr/stdout.
10269 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
10271 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
10273 2013-09-04 Jan Hubicka <jh@suse.cz>
10275 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
10276 * ipa-devirt.c: Include diganostic.h
10277 (odr_type_d): Add types and types_set.
10278 (hash_type_name): Work for types with vtables during LTO.
10279 (odr_hasher::remove): Fix comment; destroy types_set.
10280 (add_type_duplicate): New function,
10281 (get_odr_type): Use it.
10282 (dump_type_inheritance_graph): Dump type duplicates.
10283 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
10285 * tree.c (types_same_for_odr): Give exact answers on types with
10288 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10290 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
10291 explaining their differences.
10293 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
10295 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
10297 2013-09-03 Jeff Law <law@redhat.com>
10299 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
10300 when not threading through a joiner block. Pass joiner/no joiner
10301 state to register_jump_thread.
10302 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
10303 state from argument rather than implying on path length.
10304 Dump the entire jump thread path into debugging dump.
10305 * tree-flow.h (register_jump_thread): Update prototype.
10307 2013-08-29 Xinliang David Li <davidxl@google.com>
10309 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10310 Remove a trivial gcc_assert.
10312 2013-08-29 Xinliang David Li <davidxl@google.com>
10314 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
10315 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
10316 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10317 Delay base decl alignment adjustment.
10318 * tree-vectorizer.c (vect_destroy_datarefs): New function.
10319 * tree-vectorizer.h: New data structure.
10320 (set_dr_misalignment): New function.
10321 (dr_misalignment): Ditto.
10322 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
10323 (vectorizable_load): Ditto.
10324 (ensure_base_align): New function.
10325 (vectorize_loops): Add dbg_cnt support.
10326 (execute_vect_slp): Ditto.
10327 * dbgcnt.def: New debug counter.
10328 * Makefile: New dependency.
10330 2013-09-03 Meador Inge <meadori@codesourcery.com>
10334 2013-08-30 Meador Inge <meadori@codesourcery.com>
10336 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10338 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
10340 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
10341 function descriptor.
10343 2013-09-03 Richard Biener <rguenther@suse.de>
10345 * tree-affine.c (add_elt_to_tree): Fix association issue,
10346 avoid useless converts and make sure to always return a
10347 properly typed result.
10349 2013-09-03 Richard Biener <rguenther@suse.de>
10351 PR middle-end/57656
10352 * fold-const.c (negate_expr_p): Fix division case.
10353 (negate_expr): Likewise.
10355 2013-09-03 Richard Biener <rguenther@suse.de>
10358 * tree-streamer-out.c: Include tm.h.
10359 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
10361 2013-09-03 Jan Hubicka <jh@suse.cz>
10363 * tree-profile.c (tree_profiling): Cleanup CFG when done.
10365 2013-09-03 Alan Modra <amodra@gmail.com>
10367 * config.gcc (powerpc*-*-linux*): Add support for little-endian
10368 multilibs to big-endian target and vice versa.
10369 * config/rs6000/t-linux64: Use := assignment on all vars.
10370 (MULTILIB_EXTRA_OPTS): Remove fPIC.
10371 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
10372 * config/rs6000/t-linux64le: New file.
10373 * config/rs6000/t-linux64bele: New file.
10374 * config/rs6000/t-linux64lebe: New file.
10376 2013-09-02 Jan Hubicka <jh@suse.cz>
10378 * ipa-inline-transform.c (inline_transform): Do not
10379 optimize_inline_calls when not optimizing.
10381 2013-09-02 Jan Hubicka <jh@suse.cz>
10383 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
10384 duplicated nodes for assembler names.
10385 * symtab.c (symtab_unregister_node): Do not attempt to unlink
10386 hard registers from assembler name hash.
10388 2013-09-02 Jan Hubicka <jh@suse.cz>
10390 * ipa-split.c (execute_split_functions): Split externally visible
10391 functions called once.
10393 2013-09-02 Martin Jambor <mjambor@suse.cz>
10396 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
10397 linked list. When finding the correct duplicate, also consider also
10398 the caller in additon to its inlined_to node.
10400 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
10402 * config/aarch64/aarch64-simd-builtins.def
10403 (dup_lane_scalar): Remove.
10404 * config/aarch64/aarch64-simd.md
10405 (aarch64_simd_dup): Add 'w->w' alternative.
10406 (aarch64_dup_lane<mode>): Allow for VALL.
10407 (aarch64_dup_lane_scalar<mode>): Remove.
10408 (aarch64_dup_lane_<vswap_width_name><mode>): New.
10409 (aarch64_get_lane_signed<mode>): Add w->w altenative.
10410 (aarch64_get_lane_unsigned<mode>): Likewise.
10411 (aarch64_get_lane<mode>): Likewise.
10412 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
10413 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
10414 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
10415 (VCON): Change container of V2SF.
10416 (vswap_width_name): Likewise.
10417 * config/aarch64/arm_neon.h
10418 (__aarch64_vdup_lane_any): New.
10419 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10420 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
10421 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10423 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
10425 PR middle-end/56382
10426 * expr.c (emit_move_complex): Do not move complex FP values as parts if
10427 the source or the destination is a single hard register.
10429 2013-09-02 Richard Biener <rguenther@suse.de>
10431 PR middle-end/57511
10432 * tree-scalar-evolution.c (instantiate_scev_name): Allow
10435 2013-09-02 Richard Biener <rguenther@suse.de>
10437 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
10438 arithmetic to sizetype.
10440 2013-09-02 Bin Cheng <bin.cheng@arm.com>
10442 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
10443 Find auto-increment use both before and after candidate.
10445 2013-09-02 Marek Polacek <polacek@redhat.com>
10447 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
10449 2013-09-01 Jan Hubicka <jh@suse.cz>
10451 * Makefile.in: Add ipa-profile.o
10452 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
10453 * cgraph.c (struct cgraph_propagate_frequency_data,
10454 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
10455 ipa-profile.c; replace cgraph_ by ipa_ prefix.
10456 * cgraph.h (cgraph_propagate_frequency): Remove.
10457 * ipa-inline-analysis.c: Include ipa-utils.h;
10458 drop duplicated cfgloop.h.
10459 (inline_update_callee_summaries): Update.
10460 * ipa-profile.c: New file.
10461 * ipa-utils.h (ipa_propagate_frequency): Declare.
10462 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
10463 data-streamer.h, value-prof.h.
10464 (symtab_remove_unreachable_nodes): Update profile.
10465 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
10466 account_time_size, cmp_counts, dump_histogram,
10467 ipa_profile_generate_summary, ipa_profile_write_summary,
10468 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
10469 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
10470 Move to ipa-profile.c.
10472 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
10474 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
10476 2013-09-01 Jan Hubicka <jh@suse.cz>
10478 * common.opt (fdevirtualize-speculatively): New function.
10479 * invoke.texi (fdevirtualize-speculatively): Document.
10480 * ipa-devirt.c: Include ipa-inline.h
10481 (likely_target_p): New function.
10482 (ipa_devirt): New function.
10483 (gate_ipa_devirt): New function.
10484 (pass_data_ipa_devirt): New static var.
10485 (pass_ipa_devirt): Likewise.
10486 (make_pass_ipa_devirt): New function.
10487 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
10488 (common_handle_option): Disable devirtualization when
10489 value range profiling is available.
10490 * passes.def (pass_ipa_devirt): Add.
10491 * timever.def (TV_IPA_DEVIRT): New timevar.
10492 * tree-pass.h (make_pass_ipa_devirt):
10494 2013-09-01 Iain Sandoe <iain@codesourcery.com>
10496 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
10497 include sanitize(undefined).
10499 2013-08-31 Diego Novillo <dnovillo@google.com>
10501 * Makefile.in (TREE_CORE_H): Define.
10503 (GTFILES): Add tree-core.h.
10504 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
10506 * tree-core.h: New file.
10507 Move all data structures, enum, typedefs, global
10508 declarations and constants from ...
10509 * tree.h: ... here.
10511 2013-08-31 Jan Hubicka <jh@suse.cz>
10513 * bulitins.c (expand_builtin): Do not early exit for gcov
10514 instrumented functions.
10516 2013-08-31 Marek Polacek <polacek@redhat.com>
10518 * ubsan.c: Include tm_p.h.
10520 2013-08-31 Jan Hubicka <jh@suse.cz>
10522 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
10525 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
10526 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
10527 * tree-cfg.c (verify_gimple_label): ... here.
10528 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
10529 (ipa_merge_profiles): New function.
10530 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
10531 (lto_input_function_body): Likewise.
10532 * ipa-utils.h (ipa_merge_profiles): Declare.
10533 * lto-streamer.h (lto_input_function_body): Update prototype.
10534 (emit_label_in_global_context_p): Remove.
10535 * lto-symtab.c: Include ipa-utils.h
10536 (lto_cgraph_replace_node): Use ipa_merge_profiles.
10538 2013-08-31 Jan Hubicka <jh@suse.cz>
10540 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
10542 2013-08-31 Jan Hubicka <jh@suse.cz>
10544 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
10546 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
10548 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
10549 "cmp" RTX before signed_comparison_operator check to account
10550 for "code" changes.
10552 2013-08-30 Jan Hubicka <jh@suse.cz>
10554 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
10555 (detect_type_change_1): Rename to ...
10556 (detect_type_change): ... this one; early return on non-polymorphic
10558 (detect_type_change_ssa): Add comp_type parameter; update
10559 use of detect_type_change.
10560 (compute_complex_assign_jump_func): Add param_type parameter;
10561 update use of detect_type_change_ssa.
10562 (compute_complex_ancestor_jump_func): Likewise.
10563 (ipa_get_callee_param_type): New function.
10564 (ipa_compute_jump_functions_for_edge): Compute parameter type;
10565 update calls to the jump function computation functions.
10567 2013-08-30 Teresa Johnson <tejohnson@google.com>
10568 Steven Bosscher <steven@gcc.gnu.org>
10570 * cfgrtl.c (fixup_new_cold_bb): New routine.
10571 (commit_edge_insertions): Invoke fixup_partitions.
10572 (find_partition_fixes): New routine.
10573 (fixup_partitions): Ditto.
10574 (verify_hot_cold_block_grouping): Update comments.
10575 (rtl_verify_edges): Invoke find_partition_fixes.
10576 (rtl_verify_bb_pointers): Update comments.
10577 (rtl_verify_bb_layout): Ditto.
10578 * basic-block.h (probably_never_executed_edge_p): Declare.
10579 (fixup_partitions): Ditto.
10580 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
10581 * bb-reorder.c (sanitize_hot_paths): New function.
10582 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
10583 sanitize_hot_paths.
10584 * predict.c (probably_never_executed_edge_p): New routine.
10585 * cfg.c (check_bb_profile): Add partition insanity warnings.
10587 2013-08-30 Meador Inge <meadori@codesourcery.com>
10589 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10591 2013-08-30 Marek Polacek <polacek@redhat.com>
10593 * Makefile.in (ubsan.o): Add.
10594 (c-family/c-ubsan.o): Add.
10595 (builtins.o): Add ubsan.h dependency.
10596 * ubsan.h: New file.
10597 * ubsan.c: New file.
10598 * common.opt: Add -fsanitize=undefined option.
10599 (flag_sanitize): Add variable.
10600 (fsanitize=): Add option. Add Driver.
10601 (fsanitize=thread): Remove option.
10602 (fsanitize=address): Likewise.
10603 (static-libubsan): New option.
10604 * doc/invoke.texi: Document the new flag and -static-libubsan.
10605 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
10606 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
10607 * builtin-attrs.def (ATTR_COLD): Define.
10608 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10609 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
10610 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
10611 * flag-types.h (sanitize_code): New enum.
10612 * opts.c (common_handle_option): Parse command line arguments
10613 of -fsanitize=. Add -fsanitize=unreachable option.
10614 * varasm.c (get_variable_section): Adjust.
10615 (assemble_noswitch_variable): Likewise.
10616 (assemble_variable): Likewise.
10617 (output_constant_def_contents): Likewise.
10618 (categorize_decl_for_section): Likewise.
10619 (place_block_symbol): Likewise.
10620 (output_object_block): Likewise.
10621 * builtins.def: Likewise.
10622 * toplev.c (compile_file): Likewise.
10623 (process_options): Likewise.
10624 * cppbuiltin.c: Likewise.
10625 * tsan.c (tsan_pass): Likewise.
10626 (tsan_gate): Likewise.
10627 (tsan_gate_O0): Likewise.
10628 * cfgexpand.c (partition_stack_vars): Likewise.
10629 (expand_stack_vars): Likewise.
10630 (defer_stack_allocation): Likewise.
10631 (expand_used_vars): Likewise.
10632 * cfgcleanup.c (old_insns_match_p): Likewise.
10633 * asan.c (asan_finish_file): Likewise.
10634 (asan_instrument): Likewise.
10635 (gate_asan): Likewise.
10636 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
10637 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10638 (asan_global_struct): Use pointer_sized_int_node instead
10639 calling build_nonstandard_integer_type.
10640 (initialize_sanitizer_builtins): Likewise.
10641 (asan_finish_file): Likewise.
10642 * gcc.c: Document %{%:function(args):X}.
10643 (static_spec_functions): Add sanitize.
10644 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
10645 store funcval != NULL there.
10646 (do_spec_1): Adjust handle_spec_function caller.
10647 (handle_braces): Allow %:function(args) as condition.
10648 (sanitize_spec_function): New function.
10649 (ADD_STATIC_LIBUBSAN_LIBS): Define.
10650 (LIBUBSAN_SPEC): Likewise.
10651 (LIBUBSAN_EARLY_SPEC): Likewise.
10652 (SANITIZER_SPEC): Handle libubsan.
10653 (SANITIZER_EARLY_SPEC): Likewise.
10654 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
10655 instead of fsanitize=address.
10656 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
10657 instead of fsanitize=address*.
10658 * builtins.c: Include ubsan.h.
10659 (fold_builtin_0): Instrument __builtin_unreachable.
10660 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
10661 instead of flag_asan.
10662 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
10663 (pointer_sized_int_node): Define.
10664 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
10666 2013-08-30 Mike Stump <mikestump@comcast.net>
10668 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
10671 2013-08-29 Jan Hubicka <jh@suse.cz>
10673 * cgraph.c (cgraph_function_body_availability): Handle weakref
10675 * passes.def: Remove pass_fixup_cfg.
10676 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
10677 track when we need to remove functions.
10678 (gate_ipa_inline): Execute inlining always; add comment why.
10679 (pass_data_ipa_inline): Remove TODO_remove_functions.
10680 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
10681 do not produce summaries.
10682 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
10683 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
10684 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
10685 constant pool and vtable.
10687 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
10689 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
10690 New arm_neon.h's internal macros to specify 64-bit constants.
10691 Avoid using stdint.h's macros.
10693 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
10695 * recog.c (verify_changes): Verify that changes[i].old is non-zero
10696 before applying REG_P.
10698 2013-08-30 Jakub Jelinek <jakub@redhat.com>
10700 PR tree-optimization/58277
10701 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
10702 after seeing too many stmts with vdef in between dombb and current
10703 bb, invalidate everything.
10705 2013-08-30 Richard Biener <rguenther@suse.de>
10707 * fold-const.c (fold_single_bit_test): Fix overflow test.
10709 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
10711 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
10712 and which can live in a register, always retrieve the value on entry.
10713 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
10714 passed by invisible reference specially.
10715 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
10716 (vt_add_function_parameter): Correctly deal with a parameter passed by
10717 invisible reference.
10719 2013-08-30 Jan Hubicka <jh@suse.cz>
10721 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
10723 2013-08-30 Richard Biener <rguenther@suse.de>
10725 PR tree-optimization/58228
10726 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
10727 allow invariant loads in nested loop vectorization.
10729 2013-08-30 Richard Biener <rguenther@suse.de>
10731 PR tree-optimization/58223
10732 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
10733 (has_anti_or_output_dependence): ... this and adjust to also
10734 look for output dependences.
10735 (mark_nodes_having_upstream_mem_writes): Adjust.
10736 (rdg_flag_uses): Likewise.
10738 2013-08-30 Richard Biener <rguenther@suse.de>
10740 PR tree-optimization/58010
10741 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
10742 assert that we have a loop-closed PHI.
10744 2013-08-29 Jan Hubicka <jh@suse.cz>
10746 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
10747 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
10748 * lto-section-in.c (lto_free_function_in_decl_state): New function.
10749 (lto_free_function_in_decl_state_for_node): New function.
10751 2013-08-29 Xinliang David Li <davidxl@google.com>
10753 * loop-unroll.c (report_unroll_peel): Minor message change.
10754 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
10755 Emit alignment peeling message with default -fopt-info.
10756 (vect_loop_versioning): Emit loop version info message.
10757 * tree-vectorizer.c (vectorize_loops): Minor message change.
10758 (execute_vect_slp): Ditto.
10760 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
10762 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
10763 of the clone from the DECL_NAME of the original function.
10765 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
10767 * passes.c (register_pass): Add overload.
10768 * tree-pass.h (register_pass): Forward declare it. Add comment.
10770 2013-08-29 Jan Hubicka <jh@suse.cz>
10772 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
10773 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
10774 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
10775 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10776 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
10777 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
10778 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10779 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
10781 2013-08-29 Teresa Johnson <tejohnson@google.com>
10783 * dumpfile.c (dump_loc): Output column number.
10784 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
10785 * doc/invoke.texi: Document optall -fopt-info flag.
10786 * profile.c (read_profile_edge_counts): Use new dump framework.
10787 (compute_branch_probabilities): Ditto.
10788 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
10789 when pass not in any opt group.
10790 * pass_manager.h (pass_manager::get_pass_profile): New method.
10791 * value-prof.c (check_counter): Use new dump framework.
10792 (check_ic_target): Ditto.
10793 * coverage.c (get_coverage_counts): Ditto.
10794 (coverage_init): Setup new dump framework.
10796 2013-08-29 Richard Biener <rguenther@suse.de>
10798 PR tree-optimization/58246
10799 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
10800 handle the dominance check inside a basic-block.
10802 2013-08-29 Richard Biener <rguenther@suse.de>
10804 PR middle-end/57287
10805 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
10806 of default defs that appear in abnormal PHI nodes.
10808 2013-08-29 Richard Biener <rguenther@suse.de>
10810 PR tree-optimization/57685
10811 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
10812 single-use operands to avoid exponential complexity.
10814 2013-08-28 Dehao Chen <dehao@google.com>
10816 * ipa-inline.c (edge_badness): Fix integer underflow.
10818 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
10820 * gtm-builtins.def (_ITM_free): Declare leaf.
10822 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10825 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
10826 (*tls_local_dynamic_base_64_largepic): Likewise.
10827 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
10828 Remove predicate from call operand.
10829 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
10830 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
10832 2013-08-28 Jeff Law <law@redhat.com>
10834 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
10835 checks for the number of predecessors and successors allowed.
10836 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
10837 which require copying a joiner block if there is a request which
10838 is a subpath that requires no joiner block copying.
10840 2013-08-28 Jan Hubicka <jh@suse.cz>
10842 * lto-streamer-out.c (DFS_write_tree_body): Drop
10843 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
10844 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
10845 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
10846 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
10847 Do not read DECL_ERROR_ISSUED.
10848 (unpack_ts_decl_with_vis_value_fields): Do not read
10850 (lto_input_ts_binfo_tree_pointers): Do not read
10851 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
10852 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
10853 write DECL_ERROR_ISSUED..
10854 (pack_ts_decl_with_vis_value_fields): Do not write
10856 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
10857 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
10858 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
10859 * tree.h (tree_decl_common): Update comment.
10860 (DECL_ERROR_ISSUED): Remove.
10862 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10864 PR middle-end/58257
10865 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
10867 2013-08-28 Jan Hubicka <jh@suse.cz>
10869 * builtins.def (free): Declare leaf.
10871 2013-08-27 David Malcolm <dmalcolm@redhat.com>
10873 * gdbhooks.py: New.
10874 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
10875 * configure: Regenerate.
10877 2013-08-27 Martin Jambor <mjambor@suse.cz>
10879 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
10880 (ipa_ancestor_jf_data): Likewise.
10881 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
10882 (ipa_get_jf_pass_through_type_preserved): New function.
10883 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
10884 (ipa_get_jf_ancestor_type_preserved): New function.
10885 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
10886 (ipa_get_jf_ancestor_result): Likewise.
10887 (propagate_vals_accross_pass_through): Use
10888 ipa_get_jf_pass_through_result to do all the value mappings.
10889 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
10890 type_preserved flag.
10891 (ipa_set_jf_cst_copy): New function.
10892 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
10893 (ipa_set_jf_arith_pass_through): Likewise.
10894 (ipa_set_ancestor_jf): Likewise.
10895 (compute_complex_assign_jump_func): Set type_preserved instead of
10897 (ipa_compute_jump_functions_for_edge): Likewise.
10898 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
10899 (update_jump_functions_after_inlining): Update type_preserved.
10900 Explicitely create jump functions when combining one with pass_through.
10901 (ipa_write_jump_function): Stream the type_preserved flags.
10902 (ipa_read_jump_function): Likewise.
10904 2013-08-27 Jakub Jelinek <jakub@redhat.com>
10905 Aldy Hernandez <aldyh@redhat.com>
10907 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
10908 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
10909 * function.h (struct function): Add has_force_vect_loops and
10911 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
10912 * gimple.c (gimple_build_omp_critical): Add KIND argument and
10914 * gimple.def: Update CLAUSES comments.
10915 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
10916 (gimple_build_omp_for): Add argument to prototype.
10917 (gimple_omp_for_kind): New.
10918 (gimple_omp_for_set_kind): New.
10919 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
10920 GOVD_DATA_SHARE_CLASS.
10921 (enum omp_region_type): Add ORT_SIMD.
10922 (gimple_add_tmp_var): Handle ORT_SIMD.
10923 (gimplify_var_or_parm_decl): Same.
10924 (is_gimple_stmt): Same.
10925 (omp_firstprivatize_variable): Same.
10926 (omp_add_variable): Only use splay_tree_insert if lookup failed.
10927 (omp_notice_variable): Handle ORT_SIMD.
10928 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
10929 (omp_check_private): Handle ORT_SIMD.
10930 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10931 OMP_CLAUSE_SAFELEN.
10932 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
10933 Handle OMP_CLAUSE_LASTPRIVATE.
10934 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10935 OMP_CLAUSE_SAFELEN.
10936 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
10937 (gimplify_expr): Handle OMP_SIMD.
10938 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
10939 (expand_GOMP_SIMD_VF): New.
10940 (expand_GOMP_SIMD_LAST_LANE): New.
10941 * internal-fn.def (GOMP_SIMD_LANE): New.
10942 (GOMP_SIMD_VF): New.
10943 (GOMP_SIMD_LAST_LANE): New.
10944 * omp-low.c: Include target.h.
10945 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
10946 OMP_CLAUSE_SAFELEN.
10947 (check_omp_nesting_restrictions): Same.
10949 (lower_rec_simd_input_clauses): New.
10950 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
10952 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
10953 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
10954 (expand_omp_build_assign): New.
10955 (expand_omp_for_init_counts): New.
10956 (expand_omp_for_init_vars): New.
10957 (extract_omp_for_update_vars): New.
10958 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
10959 and rewrite accordingly.
10960 (expand_omp_simd): New.
10961 (expand_omp_for): Use expand_omp_simd.
10962 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
10963 (lower_omp_for): Do not lower the body.
10964 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
10965 in their own loops.
10966 * tree-flow.h (find_omp_clause): Remove prototype.
10967 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
10968 forcing vectorization of the loop, or if flag_tree_vectorize.
10969 (gate_tree_if_conversion): Similarly.
10970 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
10971 gimple_build_omp_for.
10972 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
10973 * tree-parloops (create_parallel_loop): Pass kind argument to
10974 gimple_build_omp_for.
10975 * tree-pretty-print.c (dump_omp_clause): Add cases for
10976 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
10977 OMP_CLAUSE__SIMDUID_.
10978 (dump_generic_node): Handle OMP_SIMD.
10979 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
10980 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
10981 unroll OMP_SIMD loops here.
10982 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
10983 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
10985 (vect_analyze_data_refs): Handle simd loops.
10986 * tree-vect-loop.c (vectorizable_live_operation): Handle
10988 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
10989 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
10990 (vectorizable_load): Same.
10991 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
10992 (struct simduid_to_vf): New.
10993 (simduid_to_vf::hash): New.
10994 (simduid_to-vf::equal): New.
10995 (struct simd_array_to_simduid): New.
10996 (simd_array_to_simduid::hash): New.
10997 (simd_array_to_simduid::equal): New.
10998 (adjust_simduid_builtins): New.
10999 (struct note_simd_array_uses_struct): New.
11000 (note_simd_array_uses_cb): New.
11001 (note_simd_array_uses): New.
11002 (vectorize_loops): Handle simd hints and adjust simd builtins
11004 * tree-vectorizer.h (struct _stmt_vec_info): Add
11005 simd_lane_access_p field.
11006 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
11007 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
11008 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
11009 (omp_clause_code_name): Same.
11010 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
11011 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
11012 * tree.def (OMP_SIMD): New entry.
11013 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
11014 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
11015 (OMP_CLAUSE_DECL): Adjust range for new clauses.
11016 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
11017 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
11018 (OMP_CLAUSE_LINEAR_STEP): New.
11019 (OMP_CLAUSE_SAFELEN_EXPR): New.
11020 (OMP_CLAUSE__SIMDUID__DECL): New.
11021 (find_omp_clause): New prototype.
11023 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
11025 * config/i386/driver-i386.c (host_detect_local_cpu): Update
11026 Haswell processor detection.
11028 2013-08-27 Christian Widmer <shadow@umbrox.de>
11031 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
11032 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
11033 AVX2 capable processors.
11035 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
11037 * config/aarch64/arm_neon.h: Replace all inline asm implementations
11038 of vget_low_* with implementations in terms of other intrinsics.
11040 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
11042 PR middle-end/57219
11043 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
11044 values to -1, 0 and 1.
11046 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
11048 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
11049 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
11050 (<optab><mode>3_insn): Remove.
11051 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
11052 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
11053 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
11054 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
11055 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
11056 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
11057 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
11058 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
11059 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
11060 (ror<mode>3_insn): Likewise.
11061 * config/aarch64/predicates.md (aarch64_simd_register): New.
11063 2013-08-27 Richard Biener <rguenther@suse.de>
11065 PR tree-optimization/57521
11066 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
11067 one edge is non-critical.
11068 (find_phi_replacement_condition): Make sure to use a non-critical
11069 edge. Cleanup and remove old bug workarounds.
11070 (bb_postdominates_preds): Remove.
11071 (if_convertible_loop_p_1): Do not compute post-dominators.
11072 (combine_blocks): Do not free post-dominators.
11073 (main_tree_if_conversion): Likewise.
11074 (pass_data_if_conversion): Add TODO_verify_ssa.
11076 2013-08-27 DJ Delorie <dj@redhat.com>
11078 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
11080 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
11082 * function.c (assign_parm_find_data_types): Set passed_mode and
11083 nominal_mode to the TYPE_MODE of nominal_type for the built
11084 pointer type in case of the struct-pass-by-reference.
11086 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11088 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
11089 (UINT16_TYPE): Change default to "unsigned int".
11091 * config/avr/avr.opt (mfract-convert-truncate): New option.
11092 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
11093 is set, round negative fractional integers according to n1169
11094 when converting to integer types.
11096 2013-08-26 Jan Hubicka <jh@suse.cz>
11098 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
11099 methods can not be called indirectly when their address is not taken.
11101 2013-08-26 Jan Hubicka <jh@suse.cz>
11103 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
11106 2013-08-26 Jan Hubicka <jh@suse.cz>
11108 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
11111 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11113 * reload.c (find_valid_class): Allow classes that do not include
11114 FIRST_PSEUDO_REGISTER - 1.
11116 2013-08-26 Jan Hubicka <jh@suse.cz>
11118 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
11119 fix edge count/frequency when speculation failed; fix type check
11120 for the direct call.
11122 2013-08-26 Jan Hubicka <jh@suse.cz>
11124 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
11126 2013-08-26 Jan Hubicka <jh@suse.cz>
11128 * ipa-inline-transform.c (inline_transform): Be ready for basic block
11129 to be changed by edge redirection.
11131 2013-08-26 Jan Hubicka <jh@suse.cz>
11133 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
11134 formating; add sanity check.
11135 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
11136 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
11137 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
11138 (resolve_noninline_speculation): Update callee keys, too.
11140 2013-08-26 Jan Hubicka <jh@suse.cz>
11142 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
11143 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
11145 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11147 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
11150 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
11152 * config/i386/i386.c (ix86_debug_options): Remove prototype.
11153 (x86_64_elf_select_section): Ditto.
11154 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
11156 (ix86_pass_by_reference): Ditto.
11157 (output_set_got): Ditto.
11158 (ix86_unary_operator_ok): Ditto.
11159 (ix86_expand_builtin): Ditto.
11161 2013-08-23 Jan Hubicka <jh@suse.cz>
11163 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
11165 2013-08-23 Jan Hubicka <jh@suse.cz>
11167 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
11168 (tree_decl_with_vis): Add FINAL field.
11170 2013-08-23 Jeff Law <law@redhat.com>
11172 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
11173 the debugging dump when the expression is fully redundant.
11175 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11177 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
11178 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
11179 * pretty-print.c (pp_formatted_text_data): Likewise.
11180 (pp_write_text_to_stream): Likewise.
11181 (pp_write_text_as_dot_label_to_stream): Likewise.
11182 (pp_append_r): Likewise.
11183 (pp_format): Likewise.
11184 (pp_flush): Likewise.
11185 (pp_clear_output_area): Likewise.
11186 (pp_append_text): Likewise.
11187 (pp_formatted_text): Likewise.
11188 (pp_remaining_character_count_for_line): Likewise.
11189 (pp_newline): Likewise.
11190 (pp_character): Likewise.
11191 (output_buffer::~output_buffer): Define.
11192 (pretty_printer::~pretty_printer): Destruct output buffer.
11193 * pretty-print.h (output_buffer::~output_buffer): Declare.
11194 (pretty_printer::~pretty_printer): Declare virtual.
11196 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
11199 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
11200 HOST_WIDE_INT_M1U): New macros.
11201 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
11202 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
11203 unsigned -1 for lshift.
11204 * cse.c (cse_insn): Likewise.
11205 * double-int.c (rshift_double, lshift_double): Likewise.
11206 * builtins.c (fold_builtin_bitop): Likewise.
11207 * combine.c (force_to_mode): Likewise.
11208 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
11209 * simplify-rtx.c (simplify_const_unary_operation,
11210 simplify_const_binary_operation): Likewise.
11211 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
11212 check_va_list_escapes): Likewise.
11213 * rtlanal.c (nonzero_bits1): Likewise.
11214 * expmed.c (expand_smod_pow2): Likewise.
11215 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
11217 2013-08-23 Jan Hubicka <jh@suse.cz>
11219 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
11220 as having address taken.
11222 2013-08-23 Jan Hubicka <jh@suse.cz>
11224 * ipa-utils.h (method_class_type): Declare.
11225 * ipa-devirt.c (method_class_type): Export.
11227 * cgraphunit.c (analyze_functions): Do basic devirtualization;
11228 do not walk base classes of anonymous types.
11230 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
11232 PR rtl-optimization/58220
11233 PR regression/58221
11234 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
11235 handle SEQUENCE insns properly.
11237 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11239 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
11241 (pp_newline_and_indent): Likewise.
11242 (pp_separate_with): Likewise.
11243 * pretty-print.c (pp_newline_and_flush): Define.
11244 (pp_newline_and_indent): Likewise.
11245 (pp_separate_with): Likewise.
11247 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11250 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
11251 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
11253 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
11255 * config/i386/predicates.md (ext_sse_reg_operand): New.
11256 * config/i386/i386.md (*movti_internal): Use
11257 predicate to determine if EVEX is needed.
11258 (*movsi_internal): Ditto.
11259 (*movdf_internal): Ditto.
11260 (*movsf_internal): Ditto.
11261 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
11263 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11265 PR tree-optimization/58209
11266 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
11267 (find_tail_calls): Give up for pointer result types if m is non-NULL.
11268 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
11269 emit POINTER_PLUS_EXPR.
11270 (create_tailcall_accumulator): For pointer result type accumulate in
11273 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
11275 * configure.ac: Add backslashes missing from the last change.
11276 * configure: Regenerate.
11278 2013-08-22 Jan Hubicka <jh@susue.cz>
11280 * ipa.c (function_and_variable_visibility): First remember function
11281 was global and then make it local.
11283 2013-08-22 Julian Brown <julian@codesourcery.com>
11285 * configure.ac: Add aarch64 to list of arches which use "nop" in
11287 * configure: Regenerate.
11289 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11291 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
11292 gnu_libc_has_function.
11293 * config/s390/tpf.h: Likewise.
11295 2013-08-22 Jan Hubicka <jh@susue.cz>
11297 * timevar.c (validate_phases): Add cast.
11299 2013-08-22 Jan Hubicka <jh@susue.cz>
11301 * timevar.c (validate_phases): Use size_t for memory.
11302 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
11304 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
11306 * pretty-print.h (output_buffer::output_buffer): Declare.
11307 (pretty_printer::pretty_printer): Likewise.
11308 (pp_construct): Remove.
11309 * pretty-print.c (output_buffer::output_buffer): Define.
11310 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
11311 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
11312 (print_gimple_expr): Likewise.
11313 (print_gimple_seq): Likewise.
11314 (gimple_dump_bb): Likewise.
11315 * sched-vis.c (dump_value_slim): Likewise.
11316 (dump_insn_slim): Likewise.
11317 (dump_rtl_slim): Likewise.
11318 (str_pattern_slim): Likewise.
11319 * tree-mudflap.c (mf_varname_tree): Likewise.
11320 * graph.c (print_graph_cfg): Likewise.
11321 (start_graph_dump): Likewise.
11322 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
11324 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
11325 pretty printer initialization.
11326 * coretypes.h (diagnostic_context): Remove superflous type alias
11328 (pretty_printer): Likewise. Declare directly as a class.
11329 (pretty_print_info): Remove declaration as class.
11330 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
11331 and pp_clear_output_area.
11332 (asan_add_global): Likewise.
11334 2013-08-22 Jan Hubicka <jh@suse.cz>
11336 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
11337 * ipa-utils.h (update_type_inheritance_graph): Declare.
11338 (possible_polymorphic_call_target_p): Declare.
11339 (possible_polymorphic_call_target_p): New.
11340 * ipa-devirt.c: Update toplevel comments.
11341 (cached_polymorphic_call_targets): Move up.
11342 (odr_type_d): Move ID down.
11343 (polymorphic_type_binfo_p): Update comment.
11344 (odr_hasher::remove): Likewise;
11345 (get_odr_type): Set anonymous_namespace.
11346 (dump_odr_type): Dump it.
11347 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
11348 (maybe_record_node): Record node in cached_polymorphic_call_targets.
11349 (record_binfo): Add comment.
11350 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
11352 (devirt_node_removal_hook): Do not iCE when cache is freed.
11353 (possible_polymorphic_call_target_p): New predicate.
11354 (update_type_inheritance_graph): New function.
11356 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
11357 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
11358 Sergey Lega <sergey.s.lega@intel.com>
11359 Anna Tikhonova <anna.tikhonova@intel.com>
11360 Ilya Tocar <ilya.tocar@intel.com>
11361 Andrey Turetskiy <andrey.turetskiy@intel.com>
11362 Ilya Verbin <ilya.verbin@intel.com>
11363 Kirill Yukhin <kirill.yukhin@intel.com>
11364 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11366 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
11367 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
11368 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
11369 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
11370 (OPTION_MASK_ISA_AVX2_UNSET): Update.
11371 (OPTION_MASK_ISA_AVX512F_UNSET): New.
11372 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
11373 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
11374 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
11375 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
11376 OPT_mavx512pf, OPT_mavx512er cases.
11377 * config/i386/constraints.md (v): New constraint.
11378 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
11379 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
11380 (bit_AVX512CD): New.
11381 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11382 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
11383 * config/i386/i386-c.c (ix86_target_macros_internal):
11384 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
11386 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
11387 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
11388 * config/i386/i386.c (regclass_map, dbx_register_map)
11389 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
11390 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
11391 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
11392 -mavx512pf options.
11393 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
11394 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
11395 -mavx512pf options. Fix formatting.
11396 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
11397 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
11398 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
11399 -mavx512cd, -mavx512pf options.
11400 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
11401 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
11402 (ix86_preferred_output_reload_class): Replace SSE_REGS with
11404 (ix86_hard_regno_mode_ok): Support 512-bit registers.
11405 (ix86_set_reg_reg_cost): Ditto.
11406 (x86_order_regs_for_local_alloc): Ditto.
11407 (MAX_VECT_LEN): Extend to 64-byte.
11408 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
11409 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
11410 (TARGET_AVX512ER, TARGET_AVX512CD): New.
11411 (BIGGEST_ALIGNMENT): Extend to 512-bits.
11412 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
11413 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
11414 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
11415 (SSE_REG_MODE_P): Support new modes.
11416 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
11417 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
11418 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
11419 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
11420 (REG_CLASS_CONTENTS): Add new registers.
11421 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
11422 (EXT_REX_SSE_REGNO_P): New.
11423 (HI_REGISTER_NAMES): Add new registers.
11424 * config/i386/i386.md: Define constants for new registers.
11425 (mode): Add new 512-bit modes.
11426 (prefix): Support evex prefix.
11427 (isa): Support avx512f, noavx512f, fma_avx512f.
11428 (ssemodesuffix): Add new 512-bit modes.
11430 (*movxi_internal_avx512f): Ditto.
11431 (*movdi_internal): Replace constraint "x" with the new constraint "v".
11433 (*movsi_internal): Likewise.
11434 (*movdf_internal): Likewise.
11435 (*movsf_internal): Likewise.
11436 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
11437 (<code><mode>3): Likewise.
11438 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
11440 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
11441 with the new constraint "v".
11442 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
11444 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
11445 with the new constraint "v".
11446 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
11447 (<sse2>_storedqu<avxsizesuffix>): Likewise.
11448 (*<plusminus_insn><mode>3): Likewise.
11449 (<sse>_vm<plusminus_insn><mode>3): Likewise.
11450 (*mul<mode>3): Likewise.
11451 (<sse>_vmmul<mode>3): Likewise.
11452 (<sse>_div<mode>3): Likewise.
11453 (<sse>_vmdiv<mode>3): Likewise.
11454 (<sse>_sqrt<mode>2): Likewise.
11455 (<sse>_vmsqrt<mode>2): Likewise.
11456 (*<code><mode>3_finite): Likewise.
11457 (*<code><mode>3) <smaxmin>: Likewise.
11458 (<sse>_vm<code><mode>3): Likewise.
11459 (*<code><mode>3) <any_logic>: Likewise.
11460 (*fma_fmadd_<mode>): Likewise.
11461 (*fma_fmsub_<mode>): Likewise.
11462 (*fma_fnmadd_<mode>): Likewise.
11463 (*fma_fnmsub_<mode>): Likewise.
11464 (*fma_fmaddsub_<mode>): Likewise.
11465 (*fma_fmsubadd_<mode>): Likewise.
11466 (*fmai_fmadd_<mode>): Likewise.
11467 (*fmai_fmsub_<mode>): Likewise.
11468 (*fmai_fnmadd_<mode>): Likewise.
11469 (*fmai_fnmsub_<mode>): Likewise.
11470 (sse_cvtsi2ss): Likewise.
11471 (sse_cvtsi2ssq): Likewise.
11472 (sse_cvtss2si): Likewise.
11473 (sse_cvtss2si_2): Likewise.
11474 (sse_cvtss2siq): Likewise.
11475 (sse_cvtss2siq_2): Likewise.
11476 (sse_cvttss2si): Likewise.
11477 (sse_cvtss2siq_2): Likewise.
11478 (float<sseintvecmodelower><mode>2): Likewise.
11479 (sse2_cvtsd2si_2): Likewise.
11480 (sse2_cvtsd2siq_2): Likewise.
11481 (*<plusminus_insn><mode>3): Likewise.
11482 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
11483 (*<sse4_1_avx2>_mul<mode>3): Likewise.
11484 (ashr<mode>3): Likewise.
11485 (<shift_insn><mode>3): Likewise.
11486 (avx2_<code><mode>3): Likewise.
11487 (*avx2_<code><mode>3): Likewise.
11488 (*andnot<mode>3): Likewise.
11489 (*<code><mode>3) <any_logic>: Likewise.
11490 (abs<mode>2): Likewise.
11491 (avx2_permvar<mode>): Likewise.
11492 (avx2_perm<mode>_1): Likewise.
11493 (*avx_vpermilp<mode>): Likewise.
11494 (avx_vpermilvar<mode>3): Likewise.
11495 (avx2_ashrv<mode>): Likewise.
11496 (avx2_<shift_insn>v<mode>): Likewise.
11497 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
11499 * doc/rtl.texi: Document XImode.
11501 2013-08-21 Jeff Law <law@redhat.com>
11503 * tree-flow.h (register_jump_thread): Pass vector of edges
11504 instead of each important edge.
11505 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
11506 thread path into a vector and pass that to register_jump_thread.
11507 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
11508 passed in edge vector to the current 3-edge form.
11511 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11513 * dce.c (fini_dce): Call df_analyze again just in case
11514 delete_unmarked_insns removed anything.
11516 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
11518 * reload.h (struct reg_equivs): Rename to ..
11519 (struct reg_equivs_s): .. this.
11521 2013-08-20 Martin Liska <marxin.liska@gmail.com>
11523 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
11525 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11527 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
11529 2013-08-21 Jeff Law <law@redhat.com>
11531 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
11532 simplify assignments too. If the RHS collapses to a singleton
11533 range, then return the value for the range.
11535 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
11537 * config/i386/sse.md (V16): Rename to...
11539 (mov<mode>): Update iterator name.
11540 (*mov<mode>_internal): Ditto.
11541 (push<mode>1): Ditto.
11542 (movmisalign<mode>): Ditto.
11544 2013-08-20 Jan Hubicka <jh@suse.cz>
11547 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
11548 entry for direct edges.
11549 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
11551 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11553 Revert my last two changes, r201865 and r201864:
11556 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11558 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11559 instances can own GC refs.
11561 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11562 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11563 (gcc::context::gt_pch_nx): Likewise.
11564 (gcc::context::gt_pch_nx): Likewise.
11565 * ggc.h (gt_ggc_mx <T>): New.
11566 (gt_pch_nx_with_op <T>): New.
11567 (gt_pch_nx <T>): New.
11568 * passes.c (opt_pass::gt_ggc_mx): New.
11569 (opt_pass::gt_pch_nx): New.
11570 (opt_pass::gt_pch_nx_with_op): New.
11571 (pass_manager::gt_ggc_mx): New.
11572 (pass_manager::gt_pch_nx): New.
11573 (pass_manager::gt_pch_nx_with_op): New.
11574 (pass_manager::operator new): Use
11575 ggc_internal_cleared_alloc_stat rather than xcalloc.
11576 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11577 (pass_manager::gt_ggc_mx): New.
11578 (pass_manager::gt_pch_nx): New.
11579 (pass_manager::gt_pch_nx_with_op): New.
11580 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11581 (opt_pass::operator new): New.
11582 (opt_pass::gt_ggc_mx): New.
11583 (opt_pass::gt_pch_nx): New.
11584 (opt_pass::gt_pch_nx_with_op): New.
11587 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11589 * Makefile.in (GTFILES): Add context.h.
11590 * context.c (gcc::context::operator new): New.
11591 (gcc::context::gt_ggc_mx): New.
11592 (gcc::context::gt_pch_nx): New.
11593 (gcc::context::gt_pch_nx): New.
11594 * context.h (gcc::context): Add GTY((user)) marking.
11595 (gcc::context::operator new): New.
11596 (gcc::context::gt_ggc_mx): New.
11597 (gcc::context::gt_pch_nx): New.
11598 (gcc::context::gt_pch_nx): New.
11599 (g): Add GTY marking.
11600 (gt_ggc_mx (gcc::context *)): New.
11601 (gt_pch_nx (gcc::context *)): New.
11602 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11603 void *cookie)): New.
11604 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11606 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11608 * dce.c (fini_dce): Call df_analyze again just in case
11609 delete_unmarked_insns removed anything.
11611 2013-08-20 Teresa Johnson <tejohnson@google.com>
11613 PR rtl-optimizations/57451
11614 * final.c (reemit_insn_block_notes): Prevent lexical blocks
11615 from crossing split section boundaries.
11617 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11619 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
11620 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
11621 with MULTLIB_DEFAULTS.
11623 2013-08-20 Nick Clifton <nickc@redhat.com>
11625 * target.def (narrow_volatile_bitfield): Note that the default
11626 value is false, not !TARGET_STRICT_ALIGN.
11627 * doc/tm.texi: Regenerate.
11629 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
11631 Fix LIB_SPEC for systems without libpthread.
11633 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
11634 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
11636 * config/i386/linux-common.h: Likewise.
11637 * config/mips/linux-common.h: Likewise.
11639 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11641 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
11644 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11646 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11647 instances can own GC refs.
11649 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11650 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11651 (gcc::context::gt_pch_nx): Likewise.
11652 (gcc::context::gt_pch_nx): Likewise.
11653 * ggc.h (gt_ggc_mx <T>): New.
11654 (gt_pch_nx_with_op <T>): New.
11655 (gt_pch_nx <T>): New.
11656 * passes.c (opt_pass::gt_ggc_mx): New.
11657 (opt_pass::gt_pch_nx): New.
11658 (opt_pass::gt_pch_nx_with_op): New.
11659 (pass_manager::gt_ggc_mx): New.
11660 (pass_manager::gt_pch_nx): New.
11661 (pass_manager::gt_pch_nx_with_op): New.
11662 (pass_manager::operator new): Use
11663 ggc_internal_cleared_alloc_stat rather than xcalloc.
11664 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11665 (pass_manager::gt_ggc_mx): New.
11666 (pass_manager::gt_pch_nx): New.
11667 (pass_manager::gt_pch_nx_with_op): New.
11668 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11669 (opt_pass::operator new): New.
11670 (opt_pass::gt_ggc_mx): New.
11671 (opt_pass::gt_pch_nx): New.
11672 (opt_pass::gt_pch_nx_with_op): New.
11674 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11676 * Makefile.in (GTFILES): Add context.h.
11677 * context.c (gcc::context::operator new): New.
11678 (gcc::context::gt_ggc_mx): New.
11679 (gcc::context::gt_pch_nx): New.
11680 (gcc::context::gt_pch_nx): New.
11681 * context.h (gcc::context): Add GTY((user)) marking.
11682 (gcc::context::operator new): New.
11683 (gcc::context::gt_ggc_mx): New.
11684 (gcc::context::gt_pch_nx): New.
11685 (gcc::context::gt_pch_nx): New.
11686 (g): Add GTY marking.
11687 (gt_ggc_mx (gcc::context *)): New.
11688 (gt_pch_nx (gcc::context *)): New.
11689 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11690 void *cookie)): New.
11691 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11693 2013-08-20 Alan Modra <amodra@gmail.com>
11696 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
11697 (rs6000_emit_epilogue): Likewise.
11699 2013-08-19 Dehao Chen <dehao@google.com>
11701 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
11703 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
11704 Jakub Jelinek <jakub@redhat.com>
11706 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
11707 (BUILT_IN_FABSD64): Likewise.
11708 (BUILT_IN_FABSD128): Likewise.
11709 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
11710 (fold_builtin_1): Likewise.
11711 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
11712 destination and source operands.
11713 (*abstd2_fpr): Likewise.
11714 (*nabstd2_fpr): Likewise.
11716 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
11718 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
11721 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
11723 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
11726 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
11728 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
11729 * builtins.c (default_libc_has_function): New.
11730 (gnu_libc_has_function): Ditto.
11731 (no_c99_libc_has_function): Ditto.
11732 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
11733 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
11734 (fold_builtin_sincos): Likewise.
11735 (fold_builtin_cexp): Likewise.
11736 * builtins.def (DEF_C94_BUILTIN): Likewise.
11737 (DEF_C99_BUILTIN): Likewise.
11738 (DEF_C99_C90RES_BUILTIN): Likewise.
11739 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
11740 definitions to using this define.
11741 * config/darwin-protos.h (darwin_libc_has_function): New.
11742 * config/darwin.c (darwin_libc_has_function): Ditto.
11743 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
11744 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
11745 * config/darwin.h: Ditto.
11746 * config/elfos.h: Ditto.
11747 * config/freebsd.h: Ditto.
11748 * config/i386/cygming.h: Ditto.
11749 * config/i386/djgpp.h: Ditto.
11750 * config/i386/i386-interix.h: Ditto.
11751 * config/microblaze/microblaze.h: Ditto.
11752 * config/mmix/mmix.h: Ditto.
11753 * config/gnu-user.h: Ditto.
11754 * config/ia64/hpux.h: Ditto.
11755 * config/pa/pa-hpux.h: Ditto.
11756 * config/pdp11/pdp11.h: Ditto.
11757 * config/picochip/picochip.h: Ditto.
11758 * config/linux.h: Ditto.
11759 * config/netbsd.h: Ditto.
11760 * config/openbsd.h: Ditto.
11761 * config/rs6000/aix43.h: Ditto.
11762 * config/rs6000/aix51.h: Ditto.
11763 * config/rs6000/aix52.h: Ditto.
11764 * config/rs6000/aix53.h: Ditto.
11765 * config/rs6000/aix61.h: Ditto.
11766 * config/rs6000/darwin.h: Ditto.
11767 * config/rs6000/linux.h: Ditto.
11768 * config/rs6000/linux64.h: Ditto.
11769 * config/s390/tpf.h: Ditto.
11770 * config/sol2-10.h: Ditto.
11771 * config/sol2.h: Ditto.
11772 * config/vms/vms.h: Ditto.
11773 * config/vxworks.h: Ditto.
11774 * config/linux-android.c (linux_android_libc_has_function):
11775 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
11776 * config/linux-protos.h (linux_android_libc_has_function):
11778 * config/i386/i386.c (ix86_libc_has_function): New.
11779 * config/i386/i386-protos.h
11780 (ix86_libc_has_function): New declaration.
11781 * config/i386/i386.md
11782 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
11783 ("isinf<mode>2): Likewise.
11784 * convert.c (convert_to_integer): Using new target hook
11785 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
11786 TARGET_C99_FUNCTIONS.
11787 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
11788 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
11789 * coretypes.h (function_class): New enum for different
11790 classes of functions.
11791 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
11792 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
11793 (TARGET_HAS_SINCOS): Likewise.
11794 (TARGET_LIBC_HAS_FUNCTION): New.
11795 * doc/tm.texi: Regenerated.
11796 * targhooks.h (default_libc_has_function): New declaration.
11797 (no_c99_libc_has_function): Ditto.
11798 (gnu_libc_has_function): Ditto.
11799 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
11800 and TARGET_HAS_SINCOS.
11802 2013-08-18 Jan Hubicka <jh@suse.cz>
11804 * Makeifle-in (ipa-devirt.o): New.
11805 (GTFILES): Add ipa-utils.h and ipa-devirt.c
11806 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
11807 (analyze_functions): Look into possible targets of polymorphic call.
11808 * dumpfile.c (dump_files): Add type-inheritance dump.
11809 * dumpfile.h (TDI_inheritance): New.
11810 * ipa-devirt.c: New file.
11811 * ipa-utils.h (odr_type_d): Forward declare.
11812 (odr_type): New type.
11813 (build_type_inheritance_graph): Declare.
11814 (possible_polymorphic_call_targets): Declare and introduce inline
11815 variant when only edge is pased.
11816 (dump_possible_polymorphic_call_targets): Likewise.
11817 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
11818 * tree.c (type_in_anonymous_namespace_p): Break out from ...
11819 (types_same_for_odr): ... here.
11820 * tree.h (type_in_anonymous_namespace_p): Declare.
11822 2013-08-18 Jakub Jelinek <jakub@redhat.com>
11824 PR tree-optimization/58006
11825 * tree-parloops.c (take_address_of): Don't ICE if get_name
11827 (eliminate_local_variables_stmt): Remove clobber stmts.
11829 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
11831 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
11832 error message is issued for an alias to undefined symbol.
11834 2013-08-18 Jan Hubicka <jh@suse.cz>
11836 * cgraph.c (cgraph_create_indirect_edge): Discover
11837 polymorphic calls and record basic info into indirect_info.
11838 * gimple-fold.c (gimple_fold_call): When doing BINFO based
11839 devirtualization, ignore objc function calls.
11840 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
11841 call with no parm index info.
11842 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
11843 * tree.c (virtual_method_call_p): New function.
11844 * tree.h (virtual_method_call_p): Declare.
11846 2013-08-16 Jan Hubicka <jh@suse.cz>
11848 PR middle-end/58179
11849 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
11851 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
11853 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
11856 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11858 * gengtype.c (type_for_name): Add special-case support for
11859 locating types within the "gcc::" namespace.
11860 (open_base_files): Emit a "using namespace gcc" directive.
11862 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11865 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
11866 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
11868 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
11869 array instead of each individual operand as a separate argument.
11870 (emit_fusion_gpr_load): Likewise.
11871 (expand_fusion_gpr_load): Add new function declaration.
11873 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
11874 signature to have the operands passed as an array, instead of as
11875 separate arguments. Allow ZERO_EXTEND to be in the memory
11876 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
11877 depend on the register live/dead flags when peepholes are run.
11878 (expand_fusion_gpr_load): New function to be called from the
11879 peephole2 pass, to change the register that addis sets to be the
11881 (emit_fusion_gpr_load): Change the calling signature to have the
11882 operands passed as an array, instead of as separate arguments.
11883 Allow ZERO_EXTEND to be in the memory address, and also
11884 SIGN_EXTEND if -mpower8-fusion-sign.
11886 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
11887 unspec enumeration.
11888 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
11889 adjust the register addis loads up in the peephole2 pass. Do not
11890 depend on the register live/dead state when the peephole pass is done.
11892 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11894 * gengtype.c (create_user_defined_type): Ensure that the kind
11895 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
11896 declaration is seen before the GTY((user)) marking.
11898 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
11901 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
11903 2013-08-16 Jan Hubicka <jh@suse.cz>
11905 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
11906 arugment expected_type.
11907 (gimple_fold_call): Use it.
11908 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
11909 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
11910 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
11911 (try_make_edge_direct_virtual_call): Likewise.
11912 * tree.c (obj_type_ref_class): New.
11913 * tree.h (obj_type_ref_class): Use it.
11915 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
11917 * sched-vis.c (rtl_slim_pp_initialized): Remove.
11918 (rtl_slim_pp): Likewise.
11919 (init_rtl_slim_pretty_print): Likewise.
11920 (dump_value_slim): Don't call it. Use local pretty printer.
11921 (dump_insn_slim): Likewise.
11922 (dump_rtl_slim): Likewise.
11923 (str_pattern_slim): Likewise.
11924 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
11927 2013-08-16 Jakub Jelinek <jakub@redhat.com>
11929 PR tree-optimization/58164
11930 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
11931 walk gimple_goto_dest of GIMPLE_GOTO.
11933 PR tree-optimization/58165
11934 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
11935 bi_call must be the last stmt in a bb, don't split_block, instead
11936 use fallthru edge from it and give up if there is none.
11937 Release conds vector when returning early.
11939 2013-08-14 Xinliang David Li <davidxl@google.com>
11941 * config/i386/i386.c (ix86_option_override_internal):
11942 Remove unused variable and field.
11944 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11947 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
11948 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
11949 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
11950 and Linux, correct BLKmode alignment when 128-bit alignment is
11951 required and compatibility flag is not set.
11952 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
11953 for zero-size arguments when compatibility flag is not set.
11955 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11957 PR tree-optimization/58145
11958 * tree-sra.c (build_ref_for_offset): If prev_base has
11959 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
11961 2013-08-14 Xinliang David Li <davidxl@google.com>
11963 * config/i386/i386.c (ix86_option_override_internal):
11964 Fix uninitialized variable error.
11966 2013-08-14 Xinliang David Li <davidxl@google.com>
11968 * config/i386/i386.opt: Define two new options.
11969 * config/i386/x86-tune.def: Add arch selector field in macros.
11970 * config/i386/i386.h: Adjust macro definition.
11971 * config/i386/i386.c (ix86_option_override_internal):
11973 (parse_mtune_ctrl_str): New function.
11974 (set_ix86_tune_features): New function.
11975 (ix86_function_specific_restore): Call the new helper function.
11977 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
11979 PR rtl-optimization/57662
11980 * sel-sched.c (code_motion_process_successors): When the current insn
11981 is removed after the recursive traversal, break from the loop.
11982 Add comments and debug printouts.
11984 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11985 Alexandre Oliva <aoliva@redhat.com>
11988 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
11989 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
11990 there also UNSPEC_PLTOFF.
11992 2013-08-14 Marek Polacek <polacek@redhat.com>
11994 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
11995 NUM_CONDITIONS bit positions.
11997 2013-08-13 Cary Coutant <ccoutant@google.com>
11999 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
12000 (attr_checksum): Hash vector contents instead of pointer.
12001 (attr_checksum_ordered): Likewise.
12003 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
12005 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
12006 when Pmode != word_mode. Add length_address attribute.
12007 (sse3_monitor_<mode>): Merge from sse3_monitor and
12008 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
12009 Pmode != word_mode. Update insn length attribute.
12010 * config/i386/i386.c (ix86_option_override_internal): Update
12011 ix86_gen_monitor selection for merged sse3_monitor insn.
12013 2013-08-13 Julian Brown <julian@codesourcery.com>
12015 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
12016 perform invalid legitimization on greater-than-word-size modes for
12017 TARGET_E500_DOUBLE.
12019 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
12021 * ira.c (setup_class_translate_array): Use aclass instead of cl
12022 for classes not fully covered by allocno classes.
12024 2013-08-13 Jakub Jelinek <jakub@redhat.com>
12026 PR tree-optimization/57661
12027 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
12028 * tree-inline.c (tree_function_versioning): Initialize it.
12029 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
12030 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
12031 that is not being copied.
12034 * asan.c (instrument_strlen_call): Fix typo in comment.
12035 Use char * type even for the lhs of POINTER_PLUS_EXPR.
12037 2013-08-13 Steve Ellcey <sellcey@mips.com>
12039 * config/mips/mips.md (prefetch): Use lw instead of ld on
12040 loongson in 32bit mode.
12042 2013-08-13 Nick Clifton <nickc@redhat.com>
12044 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
12046 2013-08-13 Jan Hubicka <jh@suse.cz>
12048 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
12049 introduced edge; fix typo in sanity check.
12050 (cgraph_resolve_speculation): Export; improve diagnostic.
12051 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
12052 speculation at type mismatch.
12053 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
12054 (cgraph_resolve_speculation): Declare.
12055 (symtab_can_be_discarded): New function.
12056 * value-prof.c (gimple_ic_transform): Remove actual transform code.
12057 * ipa-inline-transform.c (speculation_removed): New global var.
12058 (clone_inlined_nodes): See if speculation can be removed.
12059 (inline_call): If speculations was removed, we growths may not match.
12060 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
12061 (speculation_useful_p): New function.
12062 (resolve_noninline_speculation): New function.
12063 (inline_small_functions): Resolve useless speculations.
12064 * ipa-inline.h (speculation_useful_p): Declare
12065 * ipa.c (can_replace_by_local_alias): Simplify.
12066 (ipa_profile): Produce speculative calls in non-lto, too;
12067 add simple cost model; produce local aliases.
12069 2013-08-13 David Malcolm <dmalcolm@redhat.com>
12071 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
12074 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
12076 * config/i386/i386.c (ix86_function_versions): Use error + inform.
12078 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
12080 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
12081 iterator instead of X87MODEF.
12083 2013-08-12 Perez Read <netfirewall@gmail.com>
12086 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
12087 operand 0 for intel asm alternative.
12088 (*movabs<mode>_2): Ditto for operand 1.
12090 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
12092 * config/aarch64/arm_none.h
12093 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
12095 2013-08-12 Nick Clifton <nickc@redhat.com>
12097 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
12099 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
12101 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
12102 expand for QI/HImode operand to produce more effictive code for
12103 unsigned char(short) --> float(double) conversion.
12105 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
12107 * doc/invoke.texi: Mention that -ftls-model does not force the final
12110 2013-08-12 Marek Polacek <polacek@redhat.com>
12111 Marc Glisse <marc.glisse@inria.fr>
12113 PR tree-optimization/57980
12114 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
12115 when creating -1 constant.
12117 2013-08-10 Jan Hubicka <jh@suse.cz>
12119 Workaround binutils PR14342.
12120 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
12121 (gimple_init_edge_profiler): Likewise.
12122 (gimple_gen_ic_func_profiler): Likewise.
12124 2013-08-09 Jan Hubicka <jh@suse.cz>
12126 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
12128 2013-08-09 Xinliang David Li <davidxl@google.com>
12130 * config/i386/stringop.def: New file.
12131 * config/i386/stringop.opt: New file.
12132 * config/i386/i386-opts.h: Include stringopt.def.
12133 * config/i386/i386.opt: Include stringopt.opt.
12134 * config/i386/i386.c (ix86_option_override_internal):
12135 Override default size based stringop inline strategies with options.
12136 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
12139 2013-08-09 Jan Hubicka <jh@suse.cz>
12141 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
12143 2013-08-09 Jan Hubicka <jh@suse.cz>
12145 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
12147 (dump_cgraph_node): Dump profile-id.
12148 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
12149 and common_target_probability.
12150 * lto-cgraph.c (lto_output_edge): Stream common targets.
12151 (lto_output_node): Stream profile ids.
12152 (input_node): Stream profile ids.
12153 (input_edge): Stream common targets.
12154 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
12155 * ipa.c: Include value-prof.h
12156 (ipa_profile_generate_summary): Turn indirect call statement histograms
12157 into common targets.
12158 (ipa_profile): Turn common targets into speculative edges.
12160 2013-08-09 Jan Hubicka <jh@suse.cz>
12162 * cgraph.h (cgraph_node): Add profile_id.
12163 * value-prof.c (cgraph_node_map): Turn into pointer_map.
12164 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
12165 (del_node_map): Update.
12166 (find_func_by_funcdef_no): Replace by ...
12167 (find_func_by_profile_id): ... this one.
12168 (gimple_ic_transform): Do not remove useful histograms when
12169 speculation is not done; dump info when indirect call removal
12171 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
12172 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
12173 (__gcov_indirect_call_profiler_v2): .. this one.
12174 * profile.h (init_node_map): Update.
12175 * coverage.c (coverage_compute_profile_id): New function.
12176 * coverage.h (coverage_compute_profile_id): Declare.
12177 * tree-profile.c (init_ic_make_global_vars): Make
12178 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
12179 (gimple_init_edge_profiler): Update prototype of
12180 __gcov_indirect_call_profiler.
12181 (gimple_gen_ic_func_profiler): Simplify.
12182 (tree_profiling): Use init_node_map
12184 2013-08-09 Jan Hubicka <jh@suse.cz>
12186 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
12187 non-speculative refs.
12188 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
12189 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
12190 (cgraph_set_call_stmt): Likewise.
12191 (cgraph_create_edge_1): Fix release checking compilatoin;
12192 clear lto_stmt_uid.
12193 (cgraph_free_edge): Free indirect info.
12194 (cgraph_turn_edge_to_speculative): New function.
12195 (cgraph_speculative_call_info): New function.
12196 (cgraph_make_edge_direct): Return direct edge; handle speculation.
12197 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
12198 (dump_cgraph_node): Dump speculation.
12199 (verify_edge_count_and_frequency): Accept speculative edges.
12200 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
12201 (verify_cgraph_node): Handle speculation.
12202 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
12203 (cgraph_set_call_stmt): Update prototype.
12204 (cgraph_make_edge_direct): Update prototype.
12205 (cgraph_speculative_call_info): Declare.
12206 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
12207 to change; update call of ipa_find_references.
12208 * ipa-ref.c (ipa_record_reference): Fix return value; clear
12209 lto_stmt_uid and speculative flags.
12210 (ipa_dump_references): Dump speculation.
12211 (ipa_clone_references): Clone speculative flag.
12212 (ipa_clone_referring): Likewise.
12213 (ipa_clone_ref): New function.
12214 (ipa_find_reference): Look into lto_stmt_uids
12215 (ipa_clear_stmts_in_references): Do not clear speculative calls.
12216 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
12217 (ipa_find_reference): Update declaration.
12218 (ipa_clone_ref): Declare.
12219 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
12220 stream speculative flag.
12221 (lto_output_ref): Stream statements uids and speculation.
12222 (input_ref): Likewise.
12223 (input_edge): Stream speuclation.
12224 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
12225 (cgraph_set_call_stmt_including_clones): Handle speculation.
12226 * ipa-inline.c (heap_edge_removal_hook): New function.
12227 (inline_small_functions): Register it.
12228 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
12229 also initialize refs.
12230 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
12232 (try_make_edge_direct_simple_call): Likewise.
12233 (try_make_edge_direct_simple_call): Likewise.
12234 (update_indirect_edges_after_inlining): Likewise.
12235 (remove_described_reference): Look proper lto_stmt_uid.
12236 (propagate_controlled_uses): Likewise.
12237 (propagate_controlled_uses): Liekwise.
12238 * tree-inline.c (copy_bb): Copy speculative edges.
12239 (redirect_all_calls): New function.
12240 (copy_cfg_body): Do redirection after loop info is updated.
12241 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
12243 2013-08-09 Jan Hubicka <jh@suse.cz>
12245 * lto-streamer-out.c (output_function): Renumber PHIs.
12246 * lto-streamer-in.c (input_function): Likewise.
12248 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
12250 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
12251 (get_lane_unsigned): Likewise.
12252 (dup_lane_scalar): Likewise.
12253 (get_lane): enable for VALL.
12254 * config/aarch64/aarch64-simd.md
12255 (aarch64_dup_lane_scalar<mode>): Remove.
12256 (aarch64_get_lane_signed<mode>): Likewise.
12257 (aarch64_get_lane_unsigned<mode>): Likewise.
12258 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
12259 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
12260 (aarch64_get_lane<mode>): Enable for all vector modes.
12261 (aarch64_get_lanedi): Remove misleading constraints.
12262 * config/aarch64/arm_neon.h
12263 (__aarch64_vget_lane_any): Define.
12264 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
12265 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
12266 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
12267 * config/aarch64/iterators.md (VDQQH): New.
12268 (VDQQHS): Likewise.
12269 (vwcore): Likewise.
12271 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
12273 * configure.ac: Add GAS check for LEON instructions on SPARC.
12274 * configure: Regenerate.
12275 * config.in: Likewise.
12276 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
12278 * config/sparc/sparc.opt (LEON, LEON3): New masks.
12279 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
12281 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
12282 (AS_LEON_FLAG): New macro.
12283 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
12284 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
12285 Deal with LEON and LEON3 for the memory model.
12286 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
12287 (atomic_compare_and_swap<mode>_1): Likewise.
12288 (*atomic_compare_and_swap<mode>_1): Likewise.
12290 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12292 * config/arm/neon.md (vcond): Fix floating-point vector
12293 comparisons against 0.
12295 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
12297 * lra-constraints.c (emit_spill_move): Remove assert.
12298 (process_alt_operands): Add more debugging
12299 output. Increase reject for spilling into memory. Decrease
12300 reject for reloading scratch.
12301 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
12303 2013-08-08 Steve Ellcey <sellcey@mips.com>
12305 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
12306 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
12307 micromips incompatible. Add nan2008.
12308 (MULTILIB_DIRNAMES): Add nan2008.
12309 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12310 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
12311 and micromips incompatible. Add nan2008.
12312 (MULTILIB_DIRNAMES): Add nan2008.
12313 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12315 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
12317 PR rtl-optimization/58079
12318 * combine.c (combine_simplify_rtx): Avoid using SUBST if
12319 simplify_comparison has widened a comparison with an integer.
12321 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12323 * config/arm/neon.md (movmisalign<mode>): Disable when we
12324 don't allow unaligned accesses.
12325 (*movmisalign<mode>_neon_store): Likewise.
12326 (*movmisalign<mode>_neon_load): Likewise.
12327 (*movmisalign<mode>_neon_store): Likewise.
12328 (*movmisalign<mode>_neon_load): Likewise.
12330 2013-08-08 Jan Hubicka <jh@suse.cz>
12332 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
12333 (make_pass_rebuild_cgraph_edges): Also clear references.
12334 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
12335 * ipa-inline-transform.c (inline_transform): Remove all references
12337 * cgraphunit.c (expand_function): Remove all references after
12339 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
12340 (ipa_find_reference): Rewrite to iterator.
12341 (remove_stmt_references): Likewise.
12342 (ipa_clear_stmts_in_references): New function.
12343 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
12344 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
12346 * ipa-split.c (split_function): Remove references in split function.
12348 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12351 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
12352 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
12354 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12357 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
12358 suggested mode for the assignment isn't compatible with the
12359 registers required.
12361 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12364 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
12366 2013-08-07 Xinliang David Li <davidxl@google.com>
12368 * config/i386/i386.opt: New option -mtune-ctrl=.
12369 * config/i386/x86-tune.def: New file.
12370 * config/i386/i386.h: include x86-tune.def.
12371 * config/i386/i386.c (ix86_option_override_internal):
12372 Parsing -mtune-ctrl= option and set tune features.
12374 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
12377 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
12378 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
12381 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12383 * rtl.h (update_alignments): Declare.
12384 * final.c (grow_label_align): New function extracted from...
12385 (shorten_branches): ...here. Call it.
12386 (update_alignments): New function.
12387 * reorg.c (sibling_labels): New variable.
12388 (get_label_before): Add SIBLING parameter. If it is non-zero, push
12389 the new label along with it onto the sibling_labels vector.
12390 (fill_simple_delay_slots): Adjust call to get_label_before.
12391 (fill_slots_from_thread): Likewise.
12392 (relax_delay_slots): Likewise.
12393 (make_return_insns): Likewise.
12394 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
12396 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12398 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
12399 document its semantics.
12400 (diagnostic_report_diagnostic): Adjust accordingly.
12402 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12404 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
12405 (sparc_option_override): ...and port to new C++ pass API.
12406 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
12408 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
12410 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
12412 2013-08-06 Caroline Tice <cmtice@google.com>
12414 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
12415 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
12416 * tree-pass.h: Add pass_vtable_verify.
12417 * varasm.c (assemble_variable): Add code to properly set the comdat
12418 section and name for the .vtable_map_vars section.
12419 (assemble_vtyv_preinit_initializer): New function.
12420 (default_sectin_type_flags): Make sure .vtable_map_vars section has
12422 * output.h: Add function decl for assemble_vtv_preinit_initializer.
12423 * vtable-verify.c: New file.
12424 * vtable-verify.h: New file.
12425 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
12426 initialiation levels.
12427 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
12428 * passes.def: Insert pass_vtable_verify.
12429 * aclocal.m4: Reorder includes.
12430 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
12431 -fvtv-counts options.
12432 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
12433 as appropriate, if -fvtable-verify=... is used.
12434 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
12435 -fvtable-verify=... is used.
12436 * Makefile.in (OBJS): Add vtable-verify.o to list.
12437 (vtable-verify.o): Add new build rule.
12438 (GTFILES): Add vtable-verify.c to list.
12439 * common.opt (fvtable-verify=): New flag.
12440 (vtv_priority): Values for fvtable-verify= flag.
12441 (fvtv-counts): New flag.
12442 (fvtv-debug): New flag.
12443 * tree.h (save_vtable_map_decl): New extern function decl.
12445 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12447 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
12448 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
12449 (pass_data_rl78_devirt): ...new pass_data instance and...
12450 (make_pass_rl78_devirt): ...new function.
12451 (rl78_asm_file_start): Port pass registration to new C++ API.
12453 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12455 * coretypes.h (rtl_opt_pass): Add.
12456 (gcc::context): Add.
12457 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
12458 (epiphany_init): Port to new C++ pass API.
12459 (epiphany_optimize_mode_switching): Likewise.
12460 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
12461 (pass_manager::get_pass_mode_switching): New.
12462 (pass_manager::get_pass_peephole2): New.
12463 * mode-switching.c (pass_mode_switching): Add clone method.
12464 * recog.c (pass_peephole2): Add clone method.
12465 (pass_split_all_insns): Add clone method.
12467 2013-08-06 David Malcolm <dmalcolm@redhat.com>
12469 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
12470 (mips_option_override): ...here, porting to new C++ API for passes.
12472 2013-08-06 Jan Hubicka <jh@suse.cz>
12474 * cgraph.c (cgraph_get_body): New function based on lto.c
12476 * cgraph.h (cgraph_get_body): Declare.
12477 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
12479 * cgraphunit.c (expand_function): Get body prior expanding.
12480 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
12481 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
12483 * passes.c (do_per_function_toporder): Get body.
12484 * tree-inline.c (expand_call_inline): Get body prior inlining it.
12485 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
12487 2013-08-06 Martin Jambor <mjambor@suse.cz>
12490 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
12491 is not re-finalized. Rename second parameter to no_collect.
12493 2013-08-06 Martin Jambor <mjambor@suse.cz>
12495 PR middle-end/58041
12496 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
12497 MEM_REF has proper alignment information.
12499 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
12502 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
12504 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
12505 rtx (*) (rtx, ...) with insn_gen_fn.
12506 * genoutput.c (output_insn_data): Cast gen_? function pointers to
12507 insn_gen_fn::stored_funcptr. Add initializer braces.
12509 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12511 Rewrite how instances of passes are cloned to remove assumptions
12512 about their sizes (thus allowing pass subclasses to have
12513 additional data fields, albeit non-GC-managed ones at this point).
12515 * passes.c (make_pass_instance): Now that passes have clone
12516 methods, rewrite this function to eliminate XNEW and memcpy
12517 calls that used hardcoded sizes. Since this function no longer
12518 creates pass instances, rename it to...
12519 (add_pass_instance): ...this. Document the old way that passes were
12520 numbered and flagged, and rework this function to continue using it.
12521 (next_pass_1): Add an initial_pass argument for use by
12523 (position_pass): When adding multiple instances of a pass, use
12524 the pass's clone method, rather than relying on the XNEW/memcpy
12525 within the former make_pass_instance (now add_pass_instance).
12526 (pass_manager::pass_manager): When invoking next_pass_1, also supply
12527 the initial instance of the current pass within the pass manager.
12529 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12531 This is the automated part of the conversion of passes from C
12532 structs to C++ classes.
12534 Patch autogenerated by refactor_passes.py from
12535 https://github.com/davidmalcolm/gcc-refactoring-scripts
12536 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
12538 * asan.c (pass_asan): Convert from a global struct to a subclass of
12539 gimple_opt_pass along with...
12540 (pass_data_asan): ...new pass_data instance and...
12541 (make_pass_asan): ...new function.
12542 (pass_asan_O0): Convert from a global struct to a subclass of
12543 gimple_opt_pass along with...
12544 (pass_data_asan_O0): ...new pass_data instance and...
12545 (make_pass_asan_O0): ...new function.
12546 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
12547 subclass of rtl_opt_pass along with...
12548 (pass_data_inc_dec): ...new pass_data instance and...
12549 (make_pass_inc_dec): ...new function.
12550 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
12551 a subclass of rtl_opt_pass along with...
12552 (pass_data_reorder_blocks): ...new pass_data instance and...
12553 (make_pass_reorder_blocks): ...new function.
12554 (pass_duplicate_computed_gotos): Convert from a global struct to a
12555 subclass of rtl_opt_pass along with...
12556 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
12557 (make_pass_duplicate_computed_gotos): ...new function.
12558 (pass_partition_blocks): Convert from a global struct to a subclass of
12559 rtl_opt_pass along with...
12560 (pass_data_partition_blocks): ...new pass_data instance and...
12561 (make_pass_partition_blocks): ...new function.
12562 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
12563 struct to a subclass of rtl_opt_pass along with...
12564 (pass_data_branch_target_load_optimize1): ...new pass_data instance
12566 (make_pass_branch_target_load_optimize1): ...new function.
12567 (pass_branch_target_load_optimize2): Convert from a global struct to a
12568 subclass of rtl_opt_pass along with...
12569 (pass_data_branch_target_load_optimize2): ...new pass_data instance
12571 (make_pass_branch_target_load_optimize2): ...new function.
12572 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
12573 of rtl_opt_pass along with...
12574 (pass_data_jump): ...new pass_data instance and...
12575 (make_pass_jump): ...new function.
12576 (pass_jump2): Convert from a global struct to a subclass of
12577 rtl_opt_pass along with...
12578 (pass_data_jump2): ...new pass_data instance and...
12579 (make_pass_jump2): ...new function.
12580 * cfgexpand.c (pass_expand): Convert from a global struct to a
12581 subclass of rtl_opt_pass along with...
12582 (pass_data_expand): ...new pass_data instance and...
12583 (make_pass_expand): ...new function.
12584 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
12585 of rtl_opt_pass along with...
12586 (pass_data_free_cfg): ...new pass_data instance and...
12587 (make_pass_free_cfg): ...new function.
12588 (pass_into_cfg_layout_mode): Convert from a global struct to a
12589 subclass of rtl_opt_pass along with...
12590 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
12591 (make_pass_into_cfg_layout_mode): ...new function.
12592 (pass_outof_cfg_layout_mode): Convert from a global struct to a
12593 subclass of rtl_opt_pass along with...
12594 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
12595 (make_pass_outof_cfg_layout_mode): ...new function.
12596 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
12597 struct to a subclass of gimple_opt_pass along with...
12598 (pass_data_build_cgraph_edges): ...new pass_data instance and...
12599 (make_pass_build_cgraph_edges): ...new function.
12600 (pass_rebuild_cgraph_edges): Convert from a global struct to a
12601 subclass of gimple_opt_pass along with...
12602 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
12603 (make_pass_rebuild_cgraph_edges): ...new function.
12604 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
12605 subclass of gimple_opt_pass along with...
12606 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
12608 (make_pass_remove_cgraph_callee_edges): ...new function.
12609 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
12610 struct to a subclass of rtl_opt_pass along with...
12611 (pass_data_stack_adjustments): ...new pass_data instance and...
12612 (make_pass_stack_adjustments): ...new function.
12613 * combine.c (pass_combine): Convert from a global struct to a subclass
12614 of rtl_opt_pass along with...
12615 (pass_data_combine): ...new pass_data instance and...
12616 (make_pass_combine): ...new function.
12617 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
12618 global struct to a subclass of rtl_opt_pass along with...
12619 (pass_data_compare_elim_after_reload): ...new pass_data instance
12621 (make_pass_compare_elim_after_reload): ...new function.
12622 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
12623 of rtl_opt_pass along with...
12624 (pass_data_rtl_cprop): ...new pass_data instance and...
12625 (make_pass_rtl_cprop): ...new function.
12626 * cse.c (pass_cse): Convert from a global struct to a subclass of
12627 rtl_opt_pass along with...
12628 (pass_data_cse): ...new pass_data instance and...
12629 (make_pass_cse): ...new function.
12630 (pass_cse2): Convert from a global struct to a subclass of
12631 rtl_opt_pass along with...
12632 (pass_data_cse2): ...new pass_data instance and...
12633 (make_pass_cse2): ...new function.
12634 (pass_cse_after_global_opts): Convert from a global struct to a
12635 subclass of rtl_opt_pass along with...
12636 (pass_data_cse_after_global_opts): ...new pass_data instance and...
12637 (make_pass_cse_after_global_opts): ...new function.
12638 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
12639 of rtl_opt_pass along with...
12640 (pass_data_ud_rtl_dce): ...new pass_data instance and...
12641 (make_pass_ud_rtl_dce): ...new function.
12642 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
12643 rtl_opt_pass along with...
12644 (pass_data_fast_rtl_dce): ...new pass_data instance and...
12645 (make_pass_fast_rtl_dce): ...new function.
12646 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
12647 a subclass of rtl_opt_pass along with...
12648 (pass_data_df_initialize_opt): ...new pass_data instance and...
12649 (make_pass_df_initialize_opt): ...new function.
12650 (pass_df_initialize_no_opt): Convert from a global struct to a
12651 subclass of rtl_opt_pass along with...
12652 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
12653 (make_pass_df_initialize_no_opt): ...new function.
12654 (pass_df_finish): Convert from a global struct to a subclass of
12655 rtl_opt_pass along with...
12656 (pass_data_df_finish): ...new pass_data instance and...
12657 (make_pass_df_finish): ...new function.
12658 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
12659 rtl_opt_pass along with...
12660 (pass_data_rtl_dse1): ...new pass_data instance and...
12661 (make_pass_rtl_dse1): ...new function.
12662 (pass_rtl_dse2): Convert from a global struct to a subclass of
12663 rtl_opt_pass along with...
12664 (pass_data_rtl_dse2): ...new pass_data instance and...
12665 (make_pass_rtl_dse2): ...new function.
12666 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
12667 subclass of rtl_opt_pass along with...
12668 (pass_data_dwarf2_frame): ...new pass_data instance and...
12669 (make_pass_dwarf2_frame): ...new function.
12670 * except.c (pass_set_nothrow_function_flags): Convert from a global
12671 struct to a subclass of rtl_opt_pass along with...
12672 (pass_data_set_nothrow_function_flags): ...new pass_data instance
12674 (make_pass_set_nothrow_function_flags): ...new function.
12675 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
12676 subclass of rtl_opt_pass along with...
12677 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
12679 (make_pass_convert_to_eh_region_ranges): ...new function.
12680 * final.c (pass_compute_alignments): Convert from a global struct to a
12681 subclass of rtl_opt_pass along with...
12682 (pass_data_compute_alignments): ...new pass_data instance and...
12683 (make_pass_compute_alignments): ...new function.
12684 (pass_final): Convert from a global struct to a subclass of
12685 rtl_opt_pass along with...
12686 (pass_data_final): ...new pass_data instance and...
12687 (make_pass_final): ...new function.
12688 (pass_shorten_branches): Convert from a global struct to a subclass of
12689 rtl_opt_pass along with...
12690 (pass_data_shorten_branches): ...new pass_data instance and...
12691 (make_pass_shorten_branches): ...new function.
12692 (pass_clean_state): Convert from a global struct to a subclass of
12693 rtl_opt_pass along with...
12694 (pass_data_clean_state): ...new pass_data instance and...
12695 (make_pass_clean_state): ...new function.
12696 * function.c (pass_instantiate_virtual_regs): Convert from a global
12697 struct to a subclass of rtl_opt_pass along with...
12698 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
12699 (make_pass_instantiate_virtual_regs): ...new function.
12700 (pass_leaf_regs): Convert from a global struct to a subclass of
12701 rtl_opt_pass along with...
12702 (pass_data_leaf_regs): ...new pass_data instance and...
12703 (make_pass_leaf_regs): ...new function.
12704 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
12705 subclass of rtl_opt_pass along with...
12706 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
12708 (make_pass_thread_prologue_and_epilogue): ...new function.
12709 (pass_match_asm_constraints): Convert from a global struct to a
12710 subclass of rtl_opt_pass along with...
12711 (pass_data_match_asm_constraints): ...new pass_data instance and...
12712 (make_pass_match_asm_constraints): ...new function.
12713 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
12714 subclass of rtl_opt_pass along with...
12715 (pass_data_rtl_fwprop): ...new pass_data instance and...
12716 (make_pass_rtl_fwprop): ...new function.
12717 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
12718 rtl_opt_pass along with...
12719 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
12720 (make_pass_rtl_fwprop_addr): ...new function.
12721 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
12722 rtl_opt_pass along with...
12723 (pass_data_rtl_pre): ...new pass_data instance and...
12724 (make_pass_rtl_pre): ...new function.
12725 (pass_rtl_hoist): Convert from a global struct to a subclass of
12726 rtl_opt_pass along with...
12727 (pass_data_rtl_hoist): ...new pass_data instance and...
12728 (make_pass_rtl_hoist): ...new function.
12729 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
12730 subclass of gimple_opt_pass along with...
12731 (pass_data_lower_cf): ...new pass_data instance and...
12732 (make_pass_lower_cf): ...new function.
12733 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
12734 from a global struct to a subclass of gimple_opt_pass along with...
12735 (pass_data_strength_reduction): ...new pass_data instance and...
12736 (make_pass_strength_reduction): ...new function.
12737 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
12738 of rtl_opt_pass along with...
12739 (pass_data_rtl_ifcvt): ...new pass_data instance and...
12740 (make_pass_rtl_ifcvt): ...new function.
12741 (pass_if_after_combine): Convert from a global struct to a subclass of
12742 rtl_opt_pass along with...
12743 (pass_data_if_after_combine): ...new pass_data instance and...
12744 (make_pass_if_after_combine): ...new function.
12745 (pass_if_after_reload): Convert from a global struct to a subclass of
12746 rtl_opt_pass along with...
12747 (pass_data_if_after_reload): ...new pass_data instance and...
12748 (make_pass_if_after_reload): ...new function.
12749 * init-regs.c (pass_initialize_regs): Convert from a global struct to
12750 a subclass of rtl_opt_pass along with...
12751 (pass_data_initialize_regs): ...new pass_data instance and...
12752 (make_pass_initialize_regs): ...new function.
12753 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
12754 of ipa_opt_pass_d along with...
12755 (pass_data_ipa_cp): ...new pass_data instance and...
12756 (make_pass_ipa_cp): ...new function.
12757 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
12758 global struct to a subclass of gimple_opt_pass along with...
12759 (pass_data_inline_parameters): ...new pass_data instance and...
12760 (make_pass_inline_parameters): ...new function.
12761 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
12762 subclass of gimple_opt_pass along with...
12763 (pass_data_early_inline): ...new pass_data instance and...
12764 (make_pass_early_inline): ...new function.
12765 (pass_ipa_inline): Convert from a global struct to a subclass of
12766 ipa_opt_pass_d along with...
12767 (pass_data_ipa_inline): ...new pass_data instance and...
12768 (make_pass_ipa_inline): ...new function.
12769 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
12770 struct to a subclass of gimple_opt_pass along with...
12771 (pass_data_local_pure_const): ...new pass_data instance and...
12772 (make_pass_local_pure_const): ...new function.
12773 (pass_ipa_pure_const): Convert from a global struct to a subclass of
12774 ipa_opt_pass_d along with...
12775 (pass_data_ipa_pure_const): ...new pass_data instance and...
12776 (make_pass_ipa_pure_const): ...new function.
12777 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
12778 to a subclass of ipa_opt_pass_d along with...
12779 (pass_data_ipa_reference): ...new pass_data instance and...
12780 (make_pass_ipa_reference): ...new function.
12781 * ipa-split.c (pass_split_functions): Convert from a global struct to
12782 a subclass of gimple_opt_pass along with...
12783 (pass_data_split_functions): ...new pass_data instance and...
12784 (make_pass_split_functions): ...new function.
12785 (pass_feedback_split_functions): Convert from a global struct to a
12786 subclass of gimple_opt_pass along with...
12787 (pass_data_feedback_split_functions): ...new pass_data instance and...
12788 (make_pass_feedback_split_functions): ...new function.
12789 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
12790 global struct to a subclass of simple_ipa_opt_pass along with...
12791 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
12793 (make_pass_ipa_function_and_variable_visibility): ...new function.
12794 (pass_ipa_free_inline_summary): Convert from a global struct to a
12795 subclass of simple_ipa_opt_pass along with...
12796 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
12797 (make_pass_ipa_free_inline_summary): ...new function.
12798 (pass_ipa_whole_program_visibility): Convert from a global struct to a
12799 subclass of ipa_opt_pass_d along with...
12800 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
12802 (make_pass_ipa_whole_program_visibility): ...new function.
12803 (pass_ipa_profile): Convert from a global struct to a subclass of
12804 ipa_opt_pass_d along with...
12805 (pass_data_ipa_profile): ...new pass_data instance and...
12806 (make_pass_ipa_profile): ...new function.
12807 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
12808 ipa_opt_pass_d along with...
12809 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
12810 (make_pass_ipa_cdtor_merge): ...new function.
12811 * ira.c (pass_ira): Convert from a global struct to a subclass of
12812 rtl_opt_pass along with...
12813 (pass_data_ira): ...new pass_data instance and...
12814 (make_pass_ira): ...new function.
12815 (pass_reload): Convert from a global struct to a subclass of
12816 rtl_opt_pass along with...
12817 (pass_data_reload): ...new pass_data instance and...
12818 (make_pass_reload): ...new function.
12819 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
12820 subclass of rtl_opt_pass along with...
12821 (pass_data_cleanup_barriers): ...new pass_data instance and...
12822 (make_pass_cleanup_barriers): ...new function.
12823 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
12824 of rtl_opt_pass along with...
12825 (pass_data_loop2): ...new pass_data instance and...
12826 (make_pass_loop2): ...new function.
12827 (pass_rtl_loop_init): Convert from a global struct to a subclass of
12828 rtl_opt_pass along with...
12829 (pass_data_rtl_loop_init): ...new pass_data instance and...
12830 (make_pass_rtl_loop_init): ...new function.
12831 (pass_rtl_loop_done): Convert from a global struct to a subclass of
12832 rtl_opt_pass along with...
12833 (pass_data_rtl_loop_done): ...new pass_data instance and...
12834 (make_pass_rtl_loop_done): ...new function.
12835 (pass_rtl_move_loop_invariants): Convert from a global struct to a
12836 subclass of rtl_opt_pass along with...
12837 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
12838 (make_pass_rtl_move_loop_invariants): ...new function.
12839 (pass_rtl_unswitch): Convert from a global struct to a subclass of
12840 rtl_opt_pass along with...
12841 (pass_data_rtl_unswitch): ...new pass_data instance and...
12842 (make_pass_rtl_unswitch): ...new function.
12843 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
12844 subclass of rtl_opt_pass along with...
12845 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
12847 (make_pass_rtl_unroll_and_peel_loops): ...new function.
12848 (pass_rtl_doloop): Convert from a global struct to a subclass of
12849 rtl_opt_pass along with...
12850 (pass_data_rtl_doloop): ...new pass_data instance and...
12851 (make_pass_rtl_doloop): ...new function.
12852 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
12853 a subclass of rtl_opt_pass along with...
12854 (pass_data_lower_subreg): ...new pass_data instance and...
12855 (make_pass_lower_subreg): ...new function.
12856 (pass_lower_subreg2): Convert from a global struct to a subclass of
12857 rtl_opt_pass along with...
12858 (pass_data_lower_subreg2): ...new pass_data instance and...
12859 (make_pass_lower_subreg2): ...new function.
12860 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
12861 struct to a subclass of ipa_opt_pass_d along with...
12862 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
12863 (make_pass_ipa_lto_gimple_out): ...new function.
12864 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
12865 of ipa_opt_pass_d along with...
12866 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
12867 (make_pass_ipa_lto_finish_out): ...new function.
12868 * mode-switching.c (pass_mode_switching): Convert from a global struct
12869 to a subclass of rtl_opt_pass along with...
12870 (pass_data_mode_switching): ...new pass_data instance and...
12871 (make_pass_mode_switching): ...new function.
12872 * modulo-sched.c (pass_sms): Convert from a global struct to a
12873 subclass of rtl_opt_pass along with...
12874 (pass_data_sms): ...new pass_data instance and...
12875 (make_pass_sms): ...new function.
12876 * omp-low.c (pass_expand_omp): Convert from a global struct to a
12877 subclass of gimple_opt_pass along with...
12878 (pass_data_expand_omp): ...new pass_data instance and...
12879 (make_pass_expand_omp): ...new function.
12880 (pass_lower_omp): Convert from a global struct to a subclass of
12881 gimple_opt_pass along with...
12882 (pass_data_lower_omp): ...new pass_data instance and...
12883 (make_pass_lower_omp): ...new function.
12884 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
12885 of gimple_opt_pass along with...
12886 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
12887 (make_pass_diagnose_omp_blocks): ...new function.
12888 * passes.c (pass_early_local_passes): Convert from a global struct to
12889 a subclass of simple_ipa_opt_pass along with...
12890 (pass_data_early_local_passes): ...new pass_data instance and...
12891 (make_pass_early_local_passes): ...new function.
12892 (pass_all_early_optimizations): Convert from a global struct to a
12893 subclass of gimple_opt_pass along with...
12894 (pass_data_all_early_optimizations): ...new pass_data instance and...
12895 (make_pass_all_early_optimizations): ...new function.
12896 (pass_all_optimizations): Convert from a global struct to a subclass
12897 of gimple_opt_pass along with...
12898 (pass_data_all_optimizations): ...new pass_data instance and...
12899 (make_pass_all_optimizations): ...new function.
12900 (pass_all_optimizations_g): Convert from a global struct to a subclass
12901 of gimple_opt_pass along with...
12902 (pass_data_all_optimizations_g): ...new pass_data instance and...
12903 (make_pass_all_optimizations_g): ...new function.
12904 (pass_rest_of_compilation): Convert from a global struct to a subclass
12905 of rtl_opt_pass along with...
12906 (pass_data_rest_of_compilation): ...new pass_data instance and...
12907 (make_pass_rest_of_compilation): ...new function.
12908 (pass_postreload): Convert from a global struct to a subclass of
12909 rtl_opt_pass along with...
12910 (pass_data_postreload): ...new pass_data instance and...
12911 (make_pass_postreload): ...new function.
12912 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
12913 subclass of rtl_opt_pass along with...
12914 (pass_data_gcse2): ...new pass_data instance and...
12915 (make_pass_gcse2): ...new function.
12916 * postreload.c (pass_postreload_cse): Convert from a global struct to
12917 a subclass of rtl_opt_pass along with...
12918 (pass_data_postreload_cse): ...new pass_data instance and...
12919 (make_pass_postreload_cse): ...new function.
12920 * predict.c (pass_profile): Convert from a global struct to a subclass
12921 of gimple_opt_pass along with...
12922 (pass_data_profile): ...new pass_data instance and...
12923 (make_pass_profile): ...new function.
12924 (pass_strip_predict_hints): Convert from a global struct to a subclass
12925 of gimple_opt_pass along with...
12926 (pass_data_strip_predict_hints): ...new pass_data instance and...
12927 (make_pass_strip_predict_hints): ...new function.
12928 * recog.c (pass_peephole2): Convert from a global struct to a subclass
12929 of rtl_opt_pass along with...
12930 (pass_data_peephole2): ...new pass_data instance and...
12931 (make_pass_peephole2): ...new function.
12932 (pass_split_all_insns): Convert from a global struct to a subclass of
12933 rtl_opt_pass along with...
12934 (pass_data_split_all_insns): ...new pass_data instance and...
12935 (make_pass_split_all_insns): ...new function.
12936 (pass_split_after_reload): Convert from a global struct to a subclass
12937 of rtl_opt_pass along with...
12938 (pass_data_split_after_reload): ...new pass_data instance and...
12939 (make_pass_split_after_reload): ...new function.
12940 (pass_split_before_regstack): Convert from a global struct to a
12941 subclass of rtl_opt_pass along with...
12942 (pass_data_split_before_regstack): ...new pass_data instance and...
12943 (make_pass_split_before_regstack): ...new function.
12944 (pass_split_before_sched2): Convert from a global struct to a subclass
12945 of rtl_opt_pass along with...
12946 (pass_data_split_before_sched2): ...new pass_data instance and...
12947 (make_pass_split_before_sched2): ...new function.
12948 (pass_split_for_shorten_branches): Convert from a global struct to a
12949 subclass of rtl_opt_pass along with...
12950 (pass_data_split_for_shorten_branches): ...new pass_data instance
12952 (make_pass_split_for_shorten_branches): ...new function.
12953 * ree.c (pass_ree): Convert from a global struct to a subclass of
12954 rtl_opt_pass along with...
12955 (pass_data_ree): ...new pass_data instance and...
12956 (make_pass_ree): ...new function.
12957 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
12958 subclass of rtl_opt_pass along with...
12959 (pass_data_stack_regs): ...new pass_data instance and...
12960 (make_pass_stack_regs): ...new function.
12961 (pass_stack_regs_run): Convert from a global struct to a subclass of
12962 rtl_opt_pass along with...
12963 (pass_data_stack_regs_run): ...new pass_data instance and...
12964 (make_pass_stack_regs_run): ...new function.
12965 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
12966 subclass of rtl_opt_pass along with...
12967 (pass_data_cprop_hardreg): ...new pass_data instance and...
12968 (make_pass_cprop_hardreg): ...new function.
12969 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
12970 subclass of rtl_opt_pass along with...
12971 (pass_data_reginfo_init): ...new pass_data instance and...
12972 (make_pass_reginfo_init): ...new function.
12973 * regmove.c (pass_regmove): Convert from a global struct to a subclass
12974 of rtl_opt_pass along with...
12975 (pass_data_regmove): ...new pass_data instance and...
12976 (make_pass_regmove): ...new function.
12977 * regrename.c (pass_regrename): Convert from a global struct to a
12978 subclass of rtl_opt_pass along with...
12979 (pass_data_regrename): ...new pass_data instance and...
12980 (make_pass_regrename): ...new function.
12981 * reorg.c (pass_delay_slots): Convert from a global struct to a
12982 subclass of rtl_opt_pass along with...
12983 (pass_data_delay_slots): ...new pass_data instance and...
12984 (make_pass_delay_slots): ...new function.
12985 (pass_machine_reorg): Convert from a global struct to a subclass of
12986 rtl_opt_pass along with...
12987 (pass_data_machine_reorg): ...new pass_data instance and...
12988 (make_pass_machine_reorg): ...new function.
12989 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
12990 of rtl_opt_pass along with...
12991 (pass_data_sched): ...new pass_data instance and...
12992 (make_pass_sched): ...new function.
12993 (pass_sched2): Convert from a global struct to a subclass of
12994 rtl_opt_pass along with...
12995 (pass_data_sched2): ...new pass_data instance and...
12996 (make_pass_sched2): ...new function.
12997 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
12998 to a subclass of rtl_opt_pass along with...
12999 (pass_data_stack_ptr_mod): ...new pass_data instance and...
13000 (make_pass_stack_ptr_mod): ...new function.
13001 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
13002 to a subclass of rtl_opt_pass along with...
13003 (pass_data_rtl_store_motion): ...new pass_data instance and...
13004 (make_pass_rtl_store_motion): ...new function.
13005 * tracer.c (pass_tracer): Convert from a global struct to a subclass
13006 of gimple_opt_pass along with...
13007 (pass_data_tracer): ...new pass_data instance and...
13008 (make_pass_tracer): ...new function.
13009 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
13010 to a subclass of gimple_opt_pass along with...
13011 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
13012 (make_pass_diagnose_tm_blocks): ...new function.
13013 (pass_lower_tm): Convert from a global struct to a subclass of
13014 gimple_opt_pass along with...
13015 (pass_data_lower_tm): ...new pass_data instance and...
13016 (make_pass_lower_tm): ...new function.
13017 (pass_tm_init): Convert from a global struct to a subclass of
13018 gimple_opt_pass along with...
13019 (pass_data_tm_init): ...new pass_data instance and...
13020 (make_pass_tm_init): ...new function.
13021 (pass_tm_mark): Convert from a global struct to a subclass of
13022 gimple_opt_pass along with...
13023 (pass_data_tm_mark): ...new pass_data instance and...
13024 (make_pass_tm_mark): ...new function.
13025 (pass_tm_edges): Convert from a global struct to a subclass of
13026 gimple_opt_pass along with...
13027 (pass_data_tm_edges): ...new pass_data instance and...
13028 (make_pass_tm_edges): ...new function.
13029 (pass_tm_memopt): Convert from a global struct to a subclass of
13030 gimple_opt_pass along with...
13031 (pass_data_tm_memopt): ...new pass_data instance and...
13032 (make_pass_tm_memopt): ...new function.
13033 (pass_ipa_tm): Convert from a global struct to a subclass of
13034 simple_ipa_opt_pass along with...
13035 (pass_data_ipa_tm): ...new pass_data instance and...
13036 (make_pass_ipa_tm): ...new function.
13037 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
13038 subclass of gimple_opt_pass along with...
13039 (pass_data_call_cdce): ...new pass_data instance and...
13040 (make_pass_call_cdce): ...new function.
13041 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
13042 subclass of gimple_opt_pass along with...
13043 (pass_data_build_cfg): ...new pass_data instance and...
13044 (make_pass_build_cfg): ...new function.
13045 (pass_split_crit_edges): Convert from a global struct to a subclass of
13046 gimple_opt_pass along with...
13047 (pass_data_split_crit_edges): ...new pass_data instance and...
13048 (make_pass_split_crit_edges): ...new function.
13049 (pass_warn_function_return): Convert from a global struct to a
13050 subclass of gimple_opt_pass along with...
13051 (pass_data_warn_function_return): ...new pass_data instance and...
13052 (make_pass_warn_function_return): ...new function.
13053 (pass_warn_function_noreturn): Convert from a global struct to a
13054 subclass of gimple_opt_pass along with...
13055 (pass_data_warn_function_noreturn): ...new pass_data instance and...
13056 (make_pass_warn_function_noreturn): ...new function.
13057 (pass_warn_unused_result): Convert from a global struct to a subclass
13058 of gimple_opt_pass along with...
13059 (pass_data_warn_unused_result): ...new pass_data instance and...
13060 (make_pass_warn_unused_result): ...new function.
13061 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
13062 a subclass of gimple_opt_pass along with...
13063 (pass_data_merge_phi): ...new pass_data instance and...
13064 (make_pass_merge_phi): ...new function.
13065 * tree-complex.c (pass_lower_complex): Convert from a global struct to
13066 a subclass of gimple_opt_pass along with...
13067 (pass_data_lower_complex): ...new pass_data instance and...
13068 (make_pass_lower_complex): ...new function.
13069 (pass_lower_complex_O0): Convert from a global struct to a subclass of
13070 gimple_opt_pass along with...
13071 (pass_data_lower_complex_O0): ...new pass_data instance and...
13072 (make_pass_lower_complex_O0): ...new function.
13073 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
13074 subclass of gimple_opt_pass along with...
13075 (pass_data_lower_eh): ...new pass_data instance and...
13076 (make_pass_lower_eh): ...new function.
13077 (pass_refactor_eh): Convert from a global struct to a subclass of
13078 gimple_opt_pass along with...
13079 (pass_data_refactor_eh): ...new pass_data instance and...
13080 (make_pass_refactor_eh): ...new function.
13081 (pass_lower_resx): Convert from a global struct to a subclass of
13082 gimple_opt_pass along with...
13083 (pass_data_lower_resx): ...new pass_data instance and...
13084 (make_pass_lower_resx): ...new function.
13085 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
13086 of gimple_opt_pass along with...
13087 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
13088 (make_pass_lower_eh_dispatch): ...new function.
13089 (pass_cleanup_eh): Convert from a global struct to a subclass of
13090 gimple_opt_pass along with...
13091 (pass_data_cleanup_eh): ...new pass_data instance and...
13092 (make_pass_cleanup_eh): ...new function.
13093 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
13094 to a subclass of simple_ipa_opt_pass along with...
13095 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
13096 (make_pass_ipa_lower_emutls): ...new function.
13097 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
13098 a subclass of gimple_opt_pass along with...
13099 (pass_data_if_conversion): ...new pass_data instance and...
13100 (make_pass_if_conversion): ...new function.
13101 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
13102 subclass of gimple_opt_pass along with...
13103 (pass_data_build_ssa): ...new pass_data instance and...
13104 (make_pass_build_ssa): ...new function.
13105 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
13106 global struct to a subclass of gimple_opt_pass along with...
13107 (pass_data_loop_distribution): ...new pass_data instance and...
13108 (make_pass_loop_distribution): ...new function.
13109 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
13110 subclass of gimple_opt_pass along with...
13111 (pass_data_mudflap_1): ...new pass_data instance and...
13112 (make_pass_mudflap_1): ...new function.
13113 (pass_mudflap_2): Convert from a global struct to a subclass of
13114 gimple_opt_pass along with...
13115 (pass_data_mudflap_2): ...new pass_data instance and...
13116 (make_pass_mudflap_2): ...new function.
13117 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
13118 subclass of gimple_opt_pass along with...
13119 (pass_data_mudflap_1): ...new pass_data instance and...
13120 (make_pass_mudflap_1): ...new function.
13121 (pass_mudflap_2): Convert from a global struct to a subclass of
13122 gimple_opt_pass along with...
13123 (pass_data_mudflap_2): ...new pass_data instance and...
13124 (make_pass_mudflap_2): ...new function.
13125 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
13126 gimple_opt_pass along with...
13127 (pass_data_nrv): ...new pass_data instance and...
13128 (make_pass_nrv): ...new function.
13129 (pass_return_slot): Convert from a global struct to a subclass of
13130 gimple_opt_pass along with...
13131 (pass_data_return_slot): ...new pass_data instance and...
13132 (make_pass_return_slot): ...new function.
13133 * tree-object-size.c (pass_object_sizes): Convert from a global struct
13134 to a subclass of gimple_opt_pass along with...
13135 (pass_data_object_sizes): ...new pass_data instance and...
13136 (make_pass_object_sizes): ...new function.
13137 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
13138 global struct to a subclass of gimple_opt_pass along with...
13139 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
13141 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13142 (pass_fixup_cfg): Convert from a global struct to a subclass of
13143 gimple_opt_pass along with...
13144 (pass_data_fixup_cfg): ...new pass_data instance and...
13145 (make_pass_fixup_cfg): ...new function.
13146 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
13147 (make_pass_mudflap_1): ...new function.
13148 (pass_mudflap_2): Replace declaration with that of...
13149 (make_pass_mudflap_2): ...new function.
13150 (pass_asan): Replace declaration with that of...
13151 (make_pass_asan): ...new function.
13152 (pass_asan_O0): Replace declaration with that of...
13153 (make_pass_asan_O0): ...new function.
13154 (pass_tsan): Replace declaration with that of...
13155 (make_pass_tsan): ...new function.
13156 (pass_tsan_O0): Replace declaration with that of...
13157 (make_pass_tsan_O0): ...new function.
13158 (pass_lower_cf): Replace declaration with that of...
13159 (make_pass_lower_cf): ...new function.
13160 (pass_refactor_eh): Replace declaration with that of...
13161 (make_pass_refactor_eh): ...new function.
13162 (pass_lower_eh): Replace declaration with that of...
13163 (make_pass_lower_eh): ...new function.
13164 (pass_lower_eh_dispatch): Replace declaration with that of...
13165 (make_pass_lower_eh_dispatch): ...new function.
13166 (pass_lower_resx): Replace declaration with that of...
13167 (make_pass_lower_resx): ...new function.
13168 (pass_build_cfg): Replace declaration with that of...
13169 (make_pass_build_cfg): ...new function.
13170 (pass_early_tree_profile): Replace declaration with that of...
13171 (make_pass_early_tree_profile): ...new function.
13172 (pass_cleanup_eh): Replace declaration with that of...
13173 (make_pass_cleanup_eh): ...new function.
13174 (pass_sra): Replace declaration with that of...
13175 (make_pass_sra): ...new function.
13176 (pass_sra_early): Replace declaration with that of...
13177 (make_pass_sra_early): ...new function.
13178 (pass_early_ipa_sra): Replace declaration with that of...
13179 (make_pass_early_ipa_sra): ...new function.
13180 (pass_tail_recursion): Replace declaration with that of...
13181 (make_pass_tail_recursion): ...new function.
13182 (pass_tail_calls): Replace declaration with that of...
13183 (make_pass_tail_calls): ...new function.
13184 (pass_tree_loop): Replace declaration with that of...
13185 (make_pass_tree_loop): ...new function.
13186 (pass_tree_loop_init): Replace declaration with that of...
13187 (make_pass_tree_loop_init): ...new function.
13188 (pass_lim): Replace declaration with that of...
13189 (make_pass_lim): ...new function.
13190 (pass_tree_unswitch): Replace declaration with that of...
13191 (make_pass_tree_unswitch): ...new function.
13192 (pass_predcom): Replace declaration with that of...
13193 (make_pass_predcom): ...new function.
13194 (pass_iv_canon): Replace declaration with that of...
13195 (make_pass_iv_canon): ...new function.
13196 (pass_scev_cprop): Replace declaration with that of...
13197 (make_pass_scev_cprop): ...new function.
13198 (pass_empty_loop): Replace declaration with that of...
13199 (make_pass_empty_loop): ...new function.
13200 (pass_record_bounds): Replace declaration with that of...
13201 (make_pass_record_bounds): ...new function.
13202 (pass_graphite): Replace declaration with that of...
13203 (make_pass_graphite): ...new function.
13204 (pass_graphite_transforms): Replace declaration with that of...
13205 (make_pass_graphite_transforms): ...new function.
13206 (pass_if_conversion): Replace declaration with that of...
13207 (make_pass_if_conversion): ...new function.
13208 (pass_loop_distribution): Replace declaration with that of...
13209 (make_pass_loop_distribution): ...new function.
13210 (pass_vectorize): Replace declaration with that of...
13211 (make_pass_vectorize): ...new function.
13212 (pass_slp_vectorize): Replace declaration with that of...
13213 (make_pass_slp_vectorize): ...new function.
13214 (pass_complete_unroll): Replace declaration with that of...
13215 (make_pass_complete_unroll): ...new function.
13216 (pass_complete_unrolli): Replace declaration with that of...
13217 (make_pass_complete_unrolli): ...new function.
13218 (pass_parallelize_loops): Replace declaration with that of...
13219 (make_pass_parallelize_loops): ...new function.
13220 (pass_loop_prefetch): Replace declaration with that of...
13221 (make_pass_loop_prefetch): ...new function.
13222 (pass_iv_optimize): Replace declaration with that of...
13223 (make_pass_iv_optimize): ...new function.
13224 (pass_tree_loop_done): Replace declaration with that of...
13225 (make_pass_tree_loop_done): ...new function.
13226 (pass_ch): Replace declaration with that of...
13227 (make_pass_ch): ...new function.
13228 (pass_ccp): Replace declaration with that of...
13229 (make_pass_ccp): ...new function.
13230 (pass_phi_only_cprop): Replace declaration with that of...
13231 (make_pass_phi_only_cprop): ...new function.
13232 (pass_build_ssa): Replace declaration with that of...
13233 (make_pass_build_ssa): ...new function.
13234 (pass_build_alias): Replace declaration with that of...
13235 (make_pass_build_alias): ...new function.
13236 (pass_build_ealias): Replace declaration with that of...
13237 (make_pass_build_ealias): ...new function.
13238 (pass_dominator): Replace declaration with that of...
13239 (make_pass_dominator): ...new function.
13240 (pass_dce): Replace declaration with that of...
13241 (make_pass_dce): ...new function.
13242 (pass_dce_loop): Replace declaration with that of...
13243 (make_pass_dce_loop): ...new function.
13244 (pass_cd_dce): Replace declaration with that of...
13245 (make_pass_cd_dce): ...new function.
13246 (pass_call_cdce): Replace declaration with that of...
13247 (make_pass_call_cdce): ...new function.
13248 (pass_merge_phi): Replace declaration with that of...
13249 (make_pass_merge_phi): ...new function.
13250 (pass_split_crit_edges): Replace declaration with that of...
13251 (make_pass_split_crit_edges): ...new function.
13252 (pass_pre): Replace declaration with that of...
13253 (make_pass_pre): ...new function.
13254 (pass_profile): Replace declaration with that of...
13255 (make_pass_profile): ...new function.
13256 (pass_strip_predict_hints): Replace declaration with that of...
13257 (make_pass_strip_predict_hints): ...new function.
13258 (pass_lower_complex_O0): Replace declaration with that of...
13259 (make_pass_lower_complex_O0): ...new function.
13260 (pass_lower_complex): Replace declaration with that of...
13261 (make_pass_lower_complex): ...new function.
13262 (pass_lower_vector): Replace declaration with that of...
13263 (make_pass_lower_vector): ...new function.
13264 (pass_lower_vector_ssa): Replace declaration with that of...
13265 (make_pass_lower_vector_ssa): ...new function.
13266 (pass_lower_omp): Replace declaration with that of...
13267 (make_pass_lower_omp): ...new function.
13268 (pass_diagnose_omp_blocks): Replace declaration with that of...
13269 (make_pass_diagnose_omp_blocks): ...new function.
13270 (pass_expand_omp): Replace declaration with that of...
13271 (make_pass_expand_omp): ...new function.
13272 (pass_expand_omp_ssa): Replace declaration with that of...
13273 (make_pass_expand_omp_ssa): ...new function.
13274 (pass_object_sizes): Replace declaration with that of...
13275 (make_pass_object_sizes): ...new function.
13276 (pass_strlen): Replace declaration with that of...
13277 (make_pass_strlen): ...new function.
13278 (pass_fold_builtins): Replace declaration with that of...
13279 (make_pass_fold_builtins): ...new function.
13280 (pass_stdarg): Replace declaration with that of...
13281 (make_pass_stdarg): ...new function.
13282 (pass_early_warn_uninitialized): Replace declaration with that of...
13283 (make_pass_early_warn_uninitialized): ...new function.
13284 (pass_late_warn_uninitialized): Replace declaration with that of...
13285 (make_pass_late_warn_uninitialized): ...new function.
13286 (pass_cse_reciprocals): Replace declaration with that of...
13287 (make_pass_cse_reciprocals): ...new function.
13288 (pass_cse_sincos): Replace declaration with that of...
13289 (make_pass_cse_sincos): ...new function.
13290 (pass_optimize_bswap): Replace declaration with that of...
13291 (make_pass_optimize_bswap): ...new function.
13292 (pass_optimize_widening_mul): Replace declaration with that of...
13293 (make_pass_optimize_widening_mul): ...new function.
13294 (pass_warn_function_return): Replace declaration with that of...
13295 (make_pass_warn_function_return): ...new function.
13296 (pass_warn_function_noreturn): Replace declaration with that of...
13297 (make_pass_warn_function_noreturn): ...new function.
13298 (pass_cselim): Replace declaration with that of...
13299 (make_pass_cselim): ...new function.
13300 (pass_phiopt): Replace declaration with that of...
13301 (make_pass_phiopt): ...new function.
13302 (pass_forwprop): Replace declaration with that of...
13303 (make_pass_forwprop): ...new function.
13304 (pass_phiprop): Replace declaration with that of...
13305 (make_pass_phiprop): ...new function.
13306 (pass_tree_ifcombine): Replace declaration with that of...
13307 (make_pass_tree_ifcombine): ...new function.
13308 (pass_dse): Replace declaration with that of...
13309 (make_pass_dse): ...new function.
13310 (pass_nrv): Replace declaration with that of...
13311 (make_pass_nrv): ...new function.
13312 (pass_rename_ssa_copies): Replace declaration with that of...
13313 (make_pass_rename_ssa_copies): ...new function.
13314 (pass_sink_code): Replace declaration with that of...
13315 (make_pass_sink_code): ...new function.
13316 (pass_fre): Replace declaration with that of...
13317 (make_pass_fre): ...new function.
13318 (pass_check_data_deps): Replace declaration with that of...
13319 (make_pass_check_data_deps): ...new function.
13320 (pass_copy_prop): Replace declaration with that of...
13321 (make_pass_copy_prop): ...new function.
13322 (pass_vrp): Replace declaration with that of...
13323 (make_pass_vrp): ...new function.
13324 (pass_uncprop): Replace declaration with that of...
13325 (make_pass_uncprop): ...new function.
13326 (pass_return_slot): Replace declaration with that of...
13327 (make_pass_return_slot): ...new function.
13328 (pass_reassoc): Replace declaration with that of...
13329 (make_pass_reassoc): ...new function.
13330 (pass_rebuild_cgraph_edges): Replace declaration with that of...
13331 (make_pass_rebuild_cgraph_edges): ...new function.
13332 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
13333 (make_pass_remove_cgraph_callee_edges): ...new function.
13334 (pass_build_cgraph_edges): Replace declaration with that of...
13335 (make_pass_build_cgraph_edges): ...new function.
13336 (pass_local_pure_const): Replace declaration with that of...
13337 (make_pass_local_pure_const): ...new function.
13338 (pass_tracer): Replace declaration with that of...
13339 (make_pass_tracer): ...new function.
13340 (pass_warn_unused_result): Replace declaration with that of...
13341 (make_pass_warn_unused_result): ...new function.
13342 (pass_diagnose_tm_blocks): Replace declaration with that of...
13343 (make_pass_diagnose_tm_blocks): ...new function.
13344 (pass_lower_tm): Replace declaration with that of...
13345 (make_pass_lower_tm): ...new function.
13346 (pass_tm_init): Replace declaration with that of...
13347 (make_pass_tm_init): ...new function.
13348 (pass_tm_mark): Replace declaration with that of...
13349 (make_pass_tm_mark): ...new function.
13350 (pass_tm_memopt): Replace declaration with that of...
13351 (make_pass_tm_memopt): ...new function.
13352 (pass_tm_edges): Replace declaration with that of...
13353 (make_pass_tm_edges): ...new function.
13354 (pass_split_functions): Replace declaration with that of...
13355 (make_pass_split_functions): ...new function.
13356 (pass_feedback_split_functions): Replace declaration with that of...
13357 (make_pass_feedback_split_functions): ...new function.
13358 (pass_strength_reduction): Replace declaration with that of...
13359 (make_pass_strength_reduction): ...new function.
13360 (pass_ipa_lower_emutls): Replace declaration with that of...
13361 (make_pass_ipa_lower_emutls): ...new function.
13362 (pass_ipa_function_and_variable_visibility): Replace declaration with
13364 (make_pass_ipa_function_and_variable_visibility): ...new function.
13365 (pass_ipa_tree_profile): Replace declaration with that of...
13366 (make_pass_ipa_tree_profile): ...new function.
13367 (pass_early_local_passes): Replace declaration with that of...
13368 (make_pass_early_local_passes): ...new function.
13369 (pass_ipa_whole_program_visibility): Replace declaration with that
13371 (make_pass_ipa_whole_program_visibility): ...new function.
13372 (pass_ipa_lto_gimple_out): Replace declaration with that of...
13373 (make_pass_ipa_lto_gimple_out): ...new function.
13374 (pass_ipa_increase_alignment): Replace declaration with that of...
13375 (make_pass_ipa_increase_alignment): ...new function.
13376 (pass_ipa_inline): Replace declaration with that of...
13377 (make_pass_ipa_inline): ...new function.
13378 (pass_ipa_free_lang_data): Replace declaration with that of...
13379 (make_pass_ipa_free_lang_data): ...new function.
13380 (pass_ipa_free_inline_summary): Replace declaration with that of...
13381 (make_pass_ipa_free_inline_summary): ...new function.
13382 (pass_ipa_cp): Replace declaration with that of...
13383 (make_pass_ipa_cp): ...new function.
13384 (pass_ipa_reference): Replace declaration with that of...
13385 (make_pass_ipa_reference): ...new function.
13386 (pass_ipa_pure_const): Replace declaration with that of...
13387 (make_pass_ipa_pure_const): ...new function.
13388 (pass_ipa_pta): Replace declaration with that of...
13389 (make_pass_ipa_pta): ...new function.
13390 (pass_ipa_lto_finish_out): Replace declaration with that of...
13391 (make_pass_ipa_lto_finish_out): ...new function.
13392 (pass_ipa_tm): Replace declaration with that of...
13393 (make_pass_ipa_tm): ...new function.
13394 (pass_ipa_profile): Replace declaration with that of...
13395 (make_pass_ipa_profile): ...new function.
13396 (pass_ipa_cdtor_merge): Replace declaration with that of...
13397 (make_pass_ipa_cdtor_merge): ...new function.
13398 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
13400 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13401 (pass_init_datastructures): Replace declaration with that of...
13402 (make_pass_init_datastructures): ...new function.
13403 (pass_fixup_cfg): Replace declaration with that of...
13404 (make_pass_fixup_cfg): ...new function.
13405 (pass_expand): Replace declaration with that of...
13406 (make_pass_expand): ...new function.
13407 (pass_instantiate_virtual_regs): Replace declaration with that of...
13408 (make_pass_instantiate_virtual_regs): ...new function.
13409 (pass_rtl_fwprop): Replace declaration with that of...
13410 (make_pass_rtl_fwprop): ...new function.
13411 (pass_rtl_fwprop_addr): Replace declaration with that of...
13412 (make_pass_rtl_fwprop_addr): ...new function.
13413 (pass_jump): Replace declaration with that of...
13414 (make_pass_jump): ...new function.
13415 (pass_jump2): Replace declaration with that of...
13416 (make_pass_jump2): ...new function.
13417 (pass_lower_subreg): Replace declaration with that of...
13418 (make_pass_lower_subreg): ...new function.
13419 (pass_cse): Replace declaration with that of...
13420 (make_pass_cse): ...new function.
13421 (pass_fast_rtl_dce): Replace declaration with that of...
13422 (make_pass_fast_rtl_dce): ...new function.
13423 (pass_ud_rtl_dce): Replace declaration with that of...
13424 (make_pass_ud_rtl_dce): ...new function.
13425 (pass_rtl_dce): Replace declaration with that of...
13426 (make_pass_rtl_dce): ...new function.
13427 (pass_rtl_dse1): Replace declaration with that of...
13428 (make_pass_rtl_dse1): ...new function.
13429 (pass_rtl_dse2): Replace declaration with that of...
13430 (make_pass_rtl_dse2): ...new function.
13431 (pass_rtl_dse3): Replace declaration with that of...
13432 (make_pass_rtl_dse3): ...new function.
13433 (pass_rtl_cprop): Replace declaration with that of...
13434 (make_pass_rtl_cprop): ...new function.
13435 (pass_rtl_pre): Replace declaration with that of...
13436 (make_pass_rtl_pre): ...new function.
13437 (pass_rtl_hoist): Replace declaration with that of...
13438 (make_pass_rtl_hoist): ...new function.
13439 (pass_rtl_store_motion): Replace declaration with that of...
13440 (make_pass_rtl_store_motion): ...new function.
13441 (pass_cse_after_global_opts): Replace declaration with that of...
13442 (make_pass_cse_after_global_opts): ...new function.
13443 (pass_rtl_ifcvt): Replace declaration with that of...
13444 (make_pass_rtl_ifcvt): ...new function.
13445 (pass_into_cfg_layout_mode): Replace declaration with that of...
13446 (make_pass_into_cfg_layout_mode): ...new function.
13447 (pass_outof_cfg_layout_mode): Replace declaration with that of...
13448 (make_pass_outof_cfg_layout_mode): ...new function.
13449 (pass_loop2): Replace declaration with that of...
13450 (make_pass_loop2): ...new function.
13451 (pass_rtl_loop_init): Replace declaration with that of...
13452 (make_pass_rtl_loop_init): ...new function.
13453 (pass_rtl_move_loop_invariants): Replace declaration with that of...
13454 (make_pass_rtl_move_loop_invariants): ...new function.
13455 (pass_rtl_unswitch): Replace declaration with that of...
13456 (make_pass_rtl_unswitch): ...new function.
13457 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
13458 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13459 (pass_rtl_doloop): Replace declaration with that of...
13460 (make_pass_rtl_doloop): ...new function.
13461 (pass_rtl_loop_done): Replace declaration with that of...
13462 (make_pass_rtl_loop_done): ...new function.
13463 (pass_web): Replace declaration with that of...
13464 (make_pass_web): ...new function.
13465 (pass_cse2): Replace declaration with that of...
13466 (make_pass_cse2): ...new function.
13467 (pass_df_initialize_opt): Replace declaration with that of...
13468 (make_pass_df_initialize_opt): ...new function.
13469 (pass_df_initialize_no_opt): Replace declaration with that of...
13470 (make_pass_df_initialize_no_opt): ...new function.
13471 (pass_reginfo_init): Replace declaration with that of...
13472 (make_pass_reginfo_init): ...new function.
13473 (pass_inc_dec): Replace declaration with that of...
13474 (make_pass_inc_dec): ...new function.
13475 (pass_stack_ptr_mod): Replace declaration with that of...
13476 (make_pass_stack_ptr_mod): ...new function.
13477 (pass_initialize_regs): Replace declaration with that of...
13478 (make_pass_initialize_regs): ...new function.
13479 (pass_combine): Replace declaration with that of...
13480 (make_pass_combine): ...new function.
13481 (pass_if_after_combine): Replace declaration with that of...
13482 (make_pass_if_after_combine): ...new function.
13483 (pass_ree): Replace declaration with that of...
13484 (make_pass_ree): ...new function.
13485 (pass_partition_blocks): Replace declaration with that of...
13486 (make_pass_partition_blocks): ...new function.
13487 (pass_match_asm_constraints): Replace declaration with that of...
13488 (make_pass_match_asm_constraints): ...new function.
13489 (pass_regmove): Replace declaration with that of...
13490 (make_pass_regmove): ...new function.
13491 (pass_split_all_insns): Replace declaration with that of...
13492 (make_pass_split_all_insns): ...new function.
13493 (pass_fast_rtl_byte_dce): Replace declaration with that of...
13494 (make_pass_fast_rtl_byte_dce): ...new function.
13495 (pass_lower_subreg2): Replace declaration with that of...
13496 (make_pass_lower_subreg2): ...new function.
13497 (pass_mode_switching): Replace declaration with that of...
13498 (make_pass_mode_switching): ...new function.
13499 (pass_sms): Replace declaration with that of...
13500 (make_pass_sms): ...new function.
13501 (pass_sched): Replace declaration with that of...
13502 (make_pass_sched): ...new function.
13503 (pass_ira): Replace declaration with that of...
13504 (make_pass_ira): ...new function.
13505 (pass_reload): Replace declaration with that of...
13506 (make_pass_reload): ...new function.
13507 (pass_clean_state): Replace declaration with that of...
13508 (make_pass_clean_state): ...new function.
13509 (pass_branch_prob): Replace declaration with that of...
13510 (make_pass_branch_prob): ...new function.
13511 (pass_value_profile_transformations): Replace declaration with that
13513 (make_pass_value_profile_transformations): ...new function.
13514 (pass_postreload_cse): Replace declaration with that of...
13515 (make_pass_postreload_cse): ...new function.
13516 (pass_gcse2): Replace declaration with that of...
13517 (make_pass_gcse2): ...new function.
13518 (pass_split_after_reload): Replace declaration with that of...
13519 (make_pass_split_after_reload): ...new function.
13520 (pass_branch_target_load_optimize1): Replace declaration with that
13522 (make_pass_branch_target_load_optimize1): ...new function.
13523 (pass_thread_prologue_and_epilogue): Replace declaration with that
13525 (make_pass_thread_prologue_and_epilogue): ...new function.
13526 (pass_stack_adjustments): Replace declaration with that of...
13527 (make_pass_stack_adjustments): ...new function.
13528 (pass_peephole2): Replace declaration with that of...
13529 (make_pass_peephole2): ...new function.
13530 (pass_if_after_reload): Replace declaration with that of...
13531 (make_pass_if_after_reload): ...new function.
13532 (pass_regrename): Replace declaration with that of...
13533 (make_pass_regrename): ...new function.
13534 (pass_cprop_hardreg): Replace declaration with that of...
13535 (make_pass_cprop_hardreg): ...new function.
13536 (pass_reorder_blocks): Replace declaration with that of...
13537 (make_pass_reorder_blocks): ...new function.
13538 (pass_branch_target_load_optimize2): Replace declaration with that
13540 (make_pass_branch_target_load_optimize2): ...new function.
13541 (pass_leaf_regs): Replace declaration with that of...
13542 (make_pass_leaf_regs): ...new function.
13543 (pass_split_before_sched2): Replace declaration with that of...
13544 (make_pass_split_before_sched2): ...new function.
13545 (pass_compare_elim_after_reload): Replace declaration with that of...
13546 (make_pass_compare_elim_after_reload): ...new function.
13547 (pass_sched2): Replace declaration with that of...
13548 (make_pass_sched2): ...new function.
13549 (pass_stack_regs): Replace declaration with that of...
13550 (make_pass_stack_regs): ...new function.
13551 (pass_stack_regs_run): Replace declaration with that of...
13552 (make_pass_stack_regs_run): ...new function.
13553 (pass_df_finish): Replace declaration with that of...
13554 (make_pass_df_finish): ...new function.
13555 (pass_compute_alignments): Replace declaration with that of...
13556 (make_pass_compute_alignments): ...new function.
13557 (pass_duplicate_computed_gotos): Replace declaration with that of...
13558 (make_pass_duplicate_computed_gotos): ...new function.
13559 (pass_variable_tracking): Replace declaration with that of...
13560 (make_pass_variable_tracking): ...new function.
13561 (pass_free_cfg): Replace declaration with that of...
13562 (make_pass_free_cfg): ...new function.
13563 (pass_machine_reorg): Replace declaration with that of...
13564 (make_pass_machine_reorg): ...new function.
13565 (pass_cleanup_barriers): Replace declaration with that of...
13566 (make_pass_cleanup_barriers): ...new function.
13567 (pass_delay_slots): Replace declaration with that of...
13568 (make_pass_delay_slots): ...new function.
13569 (pass_split_for_shorten_branches): Replace declaration with that of...
13570 (make_pass_split_for_shorten_branches): ...new function.
13571 (pass_split_before_regstack): Replace declaration with that of...
13572 (make_pass_split_before_regstack): ...new function.
13573 (pass_convert_to_eh_region_ranges): Replace declaration with that
13575 (make_pass_convert_to_eh_region_ranges): ...new function.
13576 (pass_shorten_branches): Replace declaration with that of...
13577 (make_pass_shorten_branches): ...new function.
13578 (pass_set_nothrow_function_flags): Replace declaration with that of...
13579 (make_pass_set_nothrow_function_flags): ...new function.
13580 (pass_dwarf2_frame): Replace declaration with that of...
13581 (make_pass_dwarf2_frame): ...new function.
13582 (pass_final): Replace declaration with that of...
13583 (make_pass_final): ...new function.
13584 (pass_rtl_seqabstr): Replace declaration with that of...
13585 (make_pass_rtl_seqabstr): ...new function.
13586 (pass_release_ssa_names): Replace declaration with that of...
13587 (make_pass_release_ssa_names): ...new function.
13588 (pass_early_inline): Replace declaration with that of...
13589 (make_pass_early_inline): ...new function.
13590 (pass_inline_parameters): Replace declaration with that of...
13591 (make_pass_inline_parameters): ...new function.
13592 (pass_update_address_taken): Replace declaration with that of...
13593 (make_pass_update_address_taken): ...new function.
13594 (pass_convert_switch): Replace declaration with that of...
13595 (make_pass_convert_switch): ...new function.
13596 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
13597 to a subclass of simple_ipa_opt_pass along with...
13598 (pass_data_ipa_tree_profile): ...new pass_data instance and...
13599 (make_pass_ipa_tree_profile): ...new function.
13600 * tree-sra.c (pass_sra_early): Convert from a global struct to a
13601 subclass of gimple_opt_pass along with...
13602 (pass_data_sra_early): ...new pass_data instance and...
13603 (make_pass_sra_early): ...new function.
13604 (pass_sra): Convert from a global struct to a subclass of
13605 gimple_opt_pass along with...
13606 (pass_data_sra): ...new pass_data instance and...
13607 (make_pass_sra): ...new function.
13608 (pass_early_ipa_sra): Convert from a global struct to a subclass of
13609 gimple_opt_pass along with...
13610 (pass_data_early_ipa_sra): ...new pass_data instance and...
13611 (make_pass_early_ipa_sra): ...new function.
13612 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
13613 subclass of gimple_opt_pass along with...
13614 (pass_data_ccp): ...new pass_data instance and...
13615 (make_pass_ccp): ...new function.
13616 (pass_fold_builtins): Convert from a global struct to a subclass of
13617 gimple_opt_pass along with...
13618 (pass_data_fold_builtins): ...new pass_data instance and...
13619 (make_pass_fold_builtins): ...new function.
13620 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
13621 subclass of gimple_opt_pass along with...
13622 (pass_data_copy_prop): ...new pass_data instance and...
13623 (make_pass_copy_prop): ...new function.
13624 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
13625 global struct to a subclass of gimple_opt_pass along with...
13626 (pass_data_rename_ssa_copies): ...new pass_data instance and...
13627 (make_pass_rename_ssa_copies): ...new function.
13628 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
13629 subclass of gimple_opt_pass along with...
13630 (pass_data_dce): ...new pass_data instance and...
13631 (make_pass_dce): ...new function.
13632 (pass_dce_loop): Convert from a global struct to a subclass of
13633 gimple_opt_pass along with...
13634 (pass_data_dce_loop): ...new pass_data instance and...
13635 (make_pass_dce_loop): ...new function.
13636 (pass_cd_dce): Convert from a global struct to a subclass of
13637 gimple_opt_pass along with...
13638 (pass_data_cd_dce): ...new pass_data instance and...
13639 (make_pass_cd_dce): ...new function.
13640 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
13641 subclass of gimple_opt_pass along with...
13642 (pass_data_dominator): ...new pass_data instance and...
13643 (make_pass_dominator): ...new function.
13644 (pass_phi_only_cprop): Convert from a global struct to a subclass of
13645 gimple_opt_pass along with...
13646 (pass_data_phi_only_cprop): ...new pass_data instance and...
13647 (make_pass_phi_only_cprop): ...new function.
13648 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
13649 subclass of gimple_opt_pass along with...
13650 (pass_data_dse): ...new pass_data instance and...
13651 (make_pass_dse): ...new function.
13652 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
13653 a subclass of gimple_opt_pass along with...
13654 (pass_data_forwprop): ...new pass_data instance and...
13655 (make_pass_forwprop): ...new function.
13656 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
13657 struct to a subclass of gimple_opt_pass along with...
13658 (pass_data_tree_ifcombine): ...new pass_data instance and...
13659 (make_pass_tree_ifcombine): ...new function.
13660 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
13661 subclass of gimple_opt_pass along with...
13662 (pass_data_ch): ...new pass_data instance and...
13663 (make_pass_ch): ...new function.
13664 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
13665 subclass of gimple_opt_pass along with...
13666 (pass_data_tree_loop): ...new pass_data instance and...
13667 (make_pass_tree_loop): ...new function.
13668 (pass_tree_loop_init): Convert from a global struct to a subclass of
13669 gimple_opt_pass along with...
13670 (pass_data_tree_loop_init): ...new pass_data instance and...
13671 (make_pass_tree_loop_init): ...new function.
13672 (pass_lim): Convert from a global struct to a subclass of
13673 gimple_opt_pass along with...
13674 (pass_data_lim): ...new pass_data instance and...
13675 (make_pass_lim): ...new function.
13676 (pass_tree_unswitch): Convert from a global struct to a subclass of
13677 gimple_opt_pass along with...
13678 (pass_data_tree_unswitch): ...new pass_data instance and...
13679 (make_pass_tree_unswitch): ...new function.
13680 (pass_predcom): Convert from a global struct to a subclass of
13681 gimple_opt_pass along with...
13682 (pass_data_predcom): ...new pass_data instance and...
13683 (make_pass_predcom): ...new function.
13684 (pass_vectorize): Convert from a global struct to a subclass of
13685 gimple_opt_pass along with...
13686 (pass_data_vectorize): ...new pass_data instance and...
13687 (make_pass_vectorize): ...new function.
13688 (pass_graphite): Convert from a global struct to a subclass of
13689 gimple_opt_pass along with...
13690 (pass_data_graphite): ...new pass_data instance and...
13691 (make_pass_graphite): ...new function.
13692 (pass_graphite_transforms): Convert from a global struct to a subclass
13693 of gimple_opt_pass along with...
13694 (pass_data_graphite_transforms): ...new pass_data instance and...
13695 (make_pass_graphite_transforms): ...new function.
13696 (pass_check_data_deps): Convert from a global struct to a subclass of
13697 gimple_opt_pass along with...
13698 (pass_data_check_data_deps): ...new pass_data instance and...
13699 (make_pass_check_data_deps): ...new function.
13700 (pass_iv_canon): Convert from a global struct to a subclass of
13701 gimple_opt_pass along with...
13702 (pass_data_iv_canon): ...new pass_data instance and...
13703 (make_pass_iv_canon): ...new function.
13704 (pass_scev_cprop): Convert from a global struct to a subclass of
13705 gimple_opt_pass along with...
13706 (pass_data_scev_cprop): ...new pass_data instance and...
13707 (make_pass_scev_cprop): ...new function.
13708 (pass_record_bounds): Convert from a global struct to a subclass of
13709 gimple_opt_pass along with...
13710 (pass_data_record_bounds): ...new pass_data instance and...
13711 (make_pass_record_bounds): ...new function.
13712 (pass_complete_unroll): Convert from a global struct to a subclass of
13713 gimple_opt_pass along with...
13714 (pass_data_complete_unroll): ...new pass_data instance and...
13715 (make_pass_complete_unroll): ...new function.
13716 (pass_complete_unrolli): Convert from a global struct to a subclass of
13717 gimple_opt_pass along with...
13718 (pass_data_complete_unrolli): ...new pass_data instance and...
13719 (make_pass_complete_unrolli): ...new function.
13720 (pass_parallelize_loops): Convert from a global struct to a subclass
13721 of gimple_opt_pass along with...
13722 (pass_data_parallelize_loops): ...new pass_data instance and...
13723 (make_pass_parallelize_loops): ...new function.
13724 (pass_loop_prefetch): Convert from a global struct to a subclass of
13725 gimple_opt_pass along with...
13726 (pass_data_loop_prefetch): ...new pass_data instance and...
13727 (make_pass_loop_prefetch): ...new function.
13728 (pass_iv_optimize): Convert from a global struct to a subclass of
13729 gimple_opt_pass along with...
13730 (pass_data_iv_optimize): ...new pass_data instance and...
13731 (make_pass_iv_optimize): ...new function.
13732 (pass_tree_loop_done): Convert from a global struct to a subclass of
13733 gimple_opt_pass along with...
13734 (pass_data_tree_loop_done): ...new pass_data instance and...
13735 (make_pass_tree_loop_done): ...new function.
13736 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
13737 struct to a subclass of gimple_opt_pass along with...
13738 (pass_data_cse_reciprocals): ...new pass_data instance and...
13739 (make_pass_cse_reciprocals): ...new function.
13740 (pass_cse_sincos): Convert from a global struct to a subclass of
13741 gimple_opt_pass along with...
13742 (pass_data_cse_sincos): ...new pass_data instance and...
13743 (make_pass_cse_sincos): ...new function.
13744 (pass_optimize_bswap): Convert from a global struct to a subclass of
13745 gimple_opt_pass along with...
13746 (pass_data_optimize_bswap): ...new pass_data instance and...
13747 (make_pass_optimize_bswap): ...new function.
13748 (pass_optimize_widening_mul): Convert from a global struct to a
13749 subclass of gimple_opt_pass along with...
13750 (pass_data_optimize_widening_mul): ...new pass_data instance and...
13751 (make_pass_optimize_widening_mul): ...new function.
13752 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
13753 subclass of gimple_opt_pass along with...
13754 (pass_data_phiopt): ...new pass_data instance and...
13755 (make_pass_phiopt): ...new function.
13756 (pass_cselim): Convert from a global struct to a subclass of
13757 gimple_opt_pass along with...
13758 (pass_data_cselim): ...new pass_data instance and...
13759 (make_pass_cselim): ...new function.
13760 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
13761 subclass of gimple_opt_pass along with...
13762 (pass_data_phiprop): ...new pass_data instance and...
13763 (make_pass_phiprop): ...new function.
13764 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
13765 subclass of gimple_opt_pass along with...
13766 (pass_data_pre): ...new pass_data instance and...
13767 (make_pass_pre): ...new function.
13768 (pass_fre): Convert from a global struct to a subclass of
13769 gimple_opt_pass along with...
13770 (pass_data_fre): ...new pass_data instance and...
13771 (make_pass_fre): ...new function.
13772 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
13773 subclass of gimple_opt_pass along with...
13774 (pass_data_reassoc): ...new pass_data instance and...
13775 (make_pass_reassoc): ...new function.
13776 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
13777 subclass of gimple_opt_pass along with...
13778 (pass_data_sink_code): ...new pass_data instance and...
13779 (make_pass_sink_code): ...new function.
13780 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
13781 subclass of gimple_opt_pass along with...
13782 (pass_data_strlen): ...new pass_data instance and...
13783 (make_pass_strlen): ...new function.
13784 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
13785 struct to a subclass of gimple_opt_pass along with...
13786 (pass_data_build_alias): ...new pass_data instance and...
13787 (make_pass_build_alias): ...new function.
13788 (pass_build_ealias): Convert from a global struct to a subclass of
13789 gimple_opt_pass along with...
13790 (pass_data_build_ealias): ...new pass_data instance and...
13791 (make_pass_build_ealias): ...new function.
13792 (pass_ipa_pta): Convert from a global struct to a subclass of
13793 simple_ipa_opt_pass along with...
13794 (pass_data_ipa_pta): ...new pass_data instance and...
13795 (make_pass_ipa_pta): ...new function.
13796 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
13797 subclass of gimple_opt_pass along with...
13798 (pass_data_uncprop): ...new pass_data instance and...
13799 (make_pass_uncprop): ...new function.
13800 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
13801 global struct to a subclass of gimple_opt_pass along with...
13802 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
13803 (make_pass_late_warn_uninitialized): ...new function.
13804 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
13805 to a subclass of gimple_opt_pass along with...
13806 (pass_data_init_datastructures): ...new pass_data instance and...
13807 (make_pass_init_datastructures): ...new function.
13808 (pass_early_warn_uninitialized): Convert from a global struct to a
13809 subclass of gimple_opt_pass along with...
13810 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
13811 (make_pass_early_warn_uninitialized): ...new function.
13812 (pass_update_address_taken): Convert from a global struct to a
13813 subclass of gimple_opt_pass along with...
13814 (pass_data_update_address_taken): ...new pass_data instance and...
13815 (make_pass_update_address_taken): ...new function.
13816 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
13817 struct to a subclass of gimple_opt_pass along with...
13818 (pass_data_release_ssa_names): ...new pass_data instance and...
13819 (make_pass_release_ssa_names): ...new function.
13820 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
13821 subclass of gimple_opt_pass along with...
13822 (pass_data_stdarg): ...new pass_data instance and...
13823 (make_pass_stdarg): ...new function.
13824 * tree-switch-conversion.c (pass_convert_switch): Convert from a
13825 global struct to a subclass of gimple_opt_pass along with...
13826 (pass_data_convert_switch): ...new pass_data instance and...
13827 (make_pass_convert_switch): ...new function.
13828 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
13829 to a subclass of gimple_opt_pass along with...
13830 (pass_data_tail_recursion): ...new pass_data instance and...
13831 (make_pass_tail_recursion): ...new function.
13832 (pass_tail_calls): Convert from a global struct to a subclass of
13833 gimple_opt_pass along with...
13834 (pass_data_tail_calls): ...new pass_data instance and...
13835 (make_pass_tail_calls): ...new function.
13836 * tree-vect-generic.c (pass_lower_vector): Convert from a global
13837 struct to a subclass of gimple_opt_pass along with...
13838 (pass_data_lower_vector): ...new pass_data instance and...
13839 (make_pass_lower_vector): ...new function.
13840 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
13841 gimple_opt_pass along with...
13842 (pass_data_lower_vector_ssa): ...new pass_data instance and...
13843 (make_pass_lower_vector_ssa): ...new function.
13844 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
13845 to a subclass of gimple_opt_pass along with...
13846 (pass_data_slp_vectorize): ...new pass_data instance and...
13847 (make_pass_slp_vectorize): ...new function.
13848 (pass_ipa_increase_alignment): Convert from a global struct to a
13849 subclass of simple_ipa_opt_pass along with...
13850 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
13851 (make_pass_ipa_increase_alignment): ...new function.
13852 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
13853 gimple_opt_pass along with...
13854 (pass_data_vrp): ...new pass_data instance and...
13855 (make_pass_vrp): ...new function.
13856 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
13857 subclass of simple_ipa_opt_pass along with...
13858 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
13859 (make_pass_ipa_free_lang_data): ...new function.
13860 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
13861 gimple_opt_pass along with...
13862 (pass_data_tsan): ...new pass_data instance and...
13863 (make_pass_tsan): ...new function.
13864 (pass_tsan_O0): Convert from a global struct to a subclass of
13865 gimple_opt_pass along with...
13866 (pass_data_tsan_O0): ...new pass_data instance and...
13867 (make_pass_tsan_O0): ...new function.
13868 * var-tracking.c (pass_variable_tracking): Convert from a global
13869 struct to a subclass of rtl_opt_pass along with...
13870 (pass_data_variable_tracking): ...new pass_data instance and...
13871 (make_pass_variable_tracking): ...new function.
13872 * web.c (pass_web): Convert from a global struct to a subclass of
13873 rtl_opt_pass along with...
13874 (pass_data_web): ...new pass_data instance and...
13875 (make_pass_web): ...new function.
13876 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
13877 declaration with that of...
13878 (make_pass_mode_switch_use): ...new function.
13879 (pass_resolve_sw_modes): Replace declaration with that of...
13880 (make_pass_resolve_sw_modes): ...new function.
13881 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
13882 from a global struct to a subclass of rtl_opt_pass along with...
13883 (pass_data_mode_switch_use): ...new pass_data instance and...
13884 (make_pass_mode_switch_use): ...new function.
13885 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
13886 from a global struct to a subclass of rtl_opt_pass along with...
13887 (pass_data_resolve_sw_modes): ...new pass_data instance and...
13888 (make_pass_resolve_sw_modes): ...new function.
13889 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
13890 struct to a subclass of rtl_opt_pass along with...
13891 (pass_data_insert_vzeroupper): ...new pass_data instance and...
13892 (make_pass_insert_vzeroupper): ...new function.
13893 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
13894 global struct to a subclass of rtl_opt_pass along with...
13895 (pass_data_work_around_errata): ...new pass_data instance and...
13896 (make_pass_work_around_errata): ...new function.
13897 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
13898 struct to a subclass of rtl_opt_pass along with...
13899 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
13900 (make_pass_mips_machine_reorg2): ...new function.
13902 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13904 * passes.c (pass_manager::operator new): New.
13906 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13908 Handwritten part of conversion of passes to C++ classes.
13910 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
13911 (toplev.o): Add dep on PASS_MANAGER_H.
13912 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
13913 of early local pases to reflect this moving from a global to a
13914 member of gcc::pass_manager.
13915 (cgraph_add_new_function): Likewise.
13916 * lto-cgraph.c (lto_output_node): Update for conversion of
13917 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
13918 * passes.c (opt_pass::clone): New.
13919 (opt_pass::gate): New.
13920 (opt_pass::execute): New.
13921 (opt_pass::opt_pass): New.
13922 (pass_manager::execute_early_local_passes): New.
13923 (pass_manager::execute_pass_mode_switching): new.
13924 (finish_optimization_passes): Convert to...
13925 (pass_manager::finish_optimization_passes): ...this.
13926 (finish_optimization_passes): Update for conversion of passes to
13928 (register_dump_files_1): Use has_gate since we cannot portably
13929 check a vtable entry against NULL.
13930 (dump_one_pass): Likewise.
13931 (ipa_write_summaries_2): Likewise.
13932 (ipa_write_optimization_summaries_1): Likewise.
13933 (ipa_read_summaries_1): Likewise.
13934 (ipa_read_optimization_summaries_1): Likewise.
13935 (execute_ipa_stmt_fixups): Likewise.
13936 (pass_manager::pass_manager): Rewrite pass-creation, invoking
13937 pass-creation functions rather than wiring up globals, and
13938 storing the results in fields of pass_manager generated using
13939 pass-instances.def.
13940 (pass_manager::dump_profile_report): Update for conversion of
13941 passes to C++ classes.
13942 (pass_manager::execute_ipa_summary_passes): Likewise.
13943 (execute_one_ipa_transform_pass): Likewise.
13944 (execute_one_pass): Use has_gate and has_execute since we cannot
13945 portably check a vtable entry against NULL.
13946 * pass_manager.h (pass_manager::finish_optimization_passes): New.
13947 (pass_manager): Use pass-instances.def to add fields for the
13948 various pass instances.
13949 * toplev.c (finalize): Update for move of
13950 finish_optimization_passes to a method of gcc::pass_manager.
13951 * toplev.h (finish_optimization_passes): Move to method of class
13953 * tree-pass.h (struct pass_data): New.
13954 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
13955 (opt_pass::gate): Convert to virtual function.
13956 (opt_pass::~opt_pass): New.
13957 (opt_pass::clone): New.
13958 (opt_pass::execute): Convert to virtual function.
13959 (opt_pass::opt_pass): New.
13960 (opt_pass::ctxt_): new.
13961 (gimple_opt_pass): Convert to subclass of opt_pass.
13962 (gimple_opt_pass::gimple_opt_pass): New.
13963 (rtl_opt_pass): Convert to subclass of opt_pass.
13964 (rtl_opt_pass::rtl_opt_pass): New.
13965 (ipa_opt_pass_d): Convert to subclass of opt_pass.
13966 (ipa_opt_pass_d::ipa_opt_pass_d): New.
13967 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
13968 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
13969 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
13970 invocation of pass_mode_switching to reflect this moving from a
13971 global to a member of gcc::pass_manager.
13972 (ix86_option_override): Rework how pass_insert_vzeroupper is
13973 added to the pass_manager to reflect autogenerated changes.
13974 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
13976 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
13978 PR rtl-optimization/57708
13979 * recog.c (peep2_find_free_register): Validate all regs in a
13982 2013-08-05 Jan Hubicka <jh@suse.cz>
13985 * cgraph.c (verify_cgraph_node): Accept local flags from other
13987 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
13988 (function_and_variable_visibility): Likewise.
13989 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
13991 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13993 * graph.c (init_graph_slim_pretty_print): Remove.
13994 (print_graph_cfg): Do not call it. Use local pretty printer.
13995 (start_graph_dump): Likewise.
13997 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13999 * gimple-pretty-print.c (buffer): Remove.
14000 (initialized): Likewise.
14001 (maybe_init_pretty_print): Likewise.
14002 (print_gimple_stmt): Do not call it. Use non-static local
14003 pretty_printer variable.
14004 (print_gimple_expr): Likewise.
14005 (print_gimple_seq): Likewise.
14006 (gimple_dump_bb): Likewise.
14008 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14010 * asan.c (asan_pp): Remove.
14011 (asan_pp_initialized): Likewise.
14012 (asan_pp_initialize): Likewise.
14013 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
14014 (asan_emit_stack_protection): Tidy. Use local pretty printer.
14015 (asan_add_global): Likewise.
14017 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
14019 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
14020 * diagnostic.h (diagnostic_flush_buffer): Adjust.
14021 * pretty-print.c (pp_formatted_text_data): Likewise.
14022 (pp_indent): Rename from pp_base_indent.
14023 (pp_format): Rename from pp_base_format.
14024 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
14025 (pp_format_verbatim): Rename from pp_base_format_verbatim.
14026 (pp_flush): Rename from pp_base_flush.
14027 (pp_set_line_maximum_length): Rename from
14028 pp_base_set_line_maximum_length.
14029 (pp_clear_output_area): Rename from pp_base_clear_output_area.
14030 (pp_set_prefix): Rename from pp_base_set_prefix.
14031 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
14032 (pp_emit_prefix): Rename from pp_base_emit_prefix.
14033 (pp_append_text): Rename from pp_base_append_text.
14034 (pp_formatted_text): Rename from pp_base_formatted_text.
14035 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
14036 (pp_remaining_character_count_for_line): Rename from
14037 pp_base_remaining_character_count_for_line.
14038 (pp_newline): Rename from pp_base_newline.
14039 (pp_character): Rename from pp_base_character.
14040 (pp_string): Rename from pp_base_string.
14041 (pp_maybe_space): Rename from pp_base_maybe_space.
14042 * asan.c (asan_pp_string): Adjust.
14043 (asan_emit_stack_protection): Likewise.
14044 (asan_add_global): Likewise.
14045 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
14046 * tree-mudflap.c (mf_varname_tree): Likewise.
14047 * tree-pretty-print.c (pp_tree_identifier): Rename from
14048 pp_base_tree_identifier.
14049 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
14050 Declare as function.
14052 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14054 * pretty-print.h (pp_bar_bar): New.
14055 (pp_ampersand_ampersand): Likewise.
14056 (pp_less_equal): Likewise.
14057 (pp_greater_equal): Likewise.
14058 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
14059 printer functions instead of pp_string or operators and punctuators.
14060 (dump_gimple_call): Likewise.
14061 (dump_gimple_omp_for): Likewise.
14062 (dump_gimple_transaction): Likewise.
14063 (dump_gimple_phi): Likewise.
14064 (pp_gimple_stmt_1): Likewise.
14065 * sched-vis.c (print_insn): Likewise.
14066 * tree-mudflap.c (mf_varname_tree): Likewise.
14067 * tree-pretty-print.c (dump_block_node): Likewise.
14068 (dump_generic_node): Likewise.
14070 2013-08-02 Jan Hubicka <jh@suse.cz>
14072 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
14074 * lto-streamer-out.c (tree_is_indexable): Results decls and
14075 parm decls are not indexable.
14076 (DFS_write_tree_body): Do not follow args and results.
14077 (hash_tree): Likewise.
14078 (output_functions): Rearrange so struct function is needed
14079 only when real body is output; be able to also ouptut abstract
14080 functions; output DECL_ARGUMENTS and DECL_RESULT.
14081 (lto_output): When not in WPA, ale store abstract functions.
14082 (write_symbol): Do not care about RESULT_DECL.
14083 (output_symbol_p): Handle correctly sbtract decls.
14084 * lto-streamer-in.c (input_function): Rearrange so struct
14085 function can be NULL at entry; allow streaming of
14086 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
14087 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
14088 sanity check during LTO.
14089 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
14090 RESULT_DECl and DECL_ARGUMENTS.
14091 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
14094 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14096 * pretty-print.h (pp_underscore): New.
14098 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
14099 printer functions instead of pp_character.
14100 (dump_binary_rhs): Likewise.
14101 (dump_ternary_rhs): Likewise.
14102 (dump_gimple_call_args): Likewise.
14103 (pp_points_to_solution): Likewise.
14104 (dump_gimple_call): Likewise.
14105 (dump_gimple_switch): Likewise.
14106 (dump_gimple_cond): Likewise.
14107 (dump_gimple_bind): Likewise.
14108 (dump_gimple_try): Likewise.
14109 (dump_gimple_omp_for): Likewise.
14110 (dump_gimple_omp_continue): Likewise.
14111 (dump_gimple_omp_single): Likewise.
14112 (dump_gimple_omp_sections): Likewise.
14113 (dump_gimple_omp_block): Likewise.
14114 (dump_gimple_omp_critical): Likewise.
14115 (dump_gimple_transaction): Likewise.
14116 (dump_gimple_asm): Likewise.
14117 (dump_gimple_phi): Likewise.
14118 (dump_gimple_omp_parallel): Likewise.
14119 (dump_gimple_omp_task): Likewise.
14120 (dump_gimple_omp_atomic_load): Likewise.
14121 (dump_gimple_omp_atomic_store): Likewise.
14122 (dump_gimple_mem_ops): Likewise.
14123 (pp_gimple_stmt_1): Likewise.
14124 (pp_cfg_jump): Likewise.
14125 (dump_implicit_edges): Likewise.
14126 (gimple_dump_bb_for_graph): Likewise.
14127 * graph.c (draw_cfg_node): Likewise.
14128 * langhooks.c (lhd_print_error_function): Likewise.
14129 * sched-vis.c (print_exp): Likewise.
14130 (print_value): Likewise.
14131 (print_pattern): Likewise.
14132 (print_insn): Likewise.
14133 (rtl_dump_bb_for_graph): Likewise.
14134 * tree-pretty-print.c (dump_function_declaration): Likewise.
14135 (dump_array_domain): Likewise.
14136 (dump_omp_clause): Likewise.
14137 (dump_location): Likewise.
14138 (dump_generic_node): Likewise.
14139 (print_struct_decl): Likewise.
14140 * diagnostic.c (diagnostic_show_locus): Use pp_space.
14142 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14144 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
14145 candidate table when replacing a candidate statement.
14146 (replace_rhs_if_not_dup): Likewise.
14147 (replace_one_candidate): Likewise.
14149 2013-08-02 Jan Hubicka <jh@suse.cz>
14150 Martin Liska <marxin.liska@gmail.com>
14152 * cgraphunit.c (add_new_function): Fix logic when adding from
14154 (assemble_thunk): Rename to ...
14155 (expand_thunk); .. this one; export; get it working with
14156 general functions; make produced gimple valid.
14157 * cgraph.h (expand_thunk): Declare.
14159 2013-08-02 Jan Hubicka <jh@suse.cz>
14161 * ipa-cp.c (gather_context_independent_values): Use
14162 ipa_get_param_move_cost.
14163 (get_replacement_map): Remove PARAM; move parameter folding
14165 (create_specialized_node): Update.
14166 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
14167 assert that we have gimple body; update move_cost.
14168 (count_formal_params): Assert that we have gimple body.
14169 (ipa_dump_param): New function.
14170 (ipa_alloc_node_params): Break out from ...
14171 (ipa_initialize_node_params): ... here.
14172 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
14173 (ipa_write_node_info): Stream move costs.
14174 (ipa_read_node_info): Read move costs.
14175 (ipa_update_after_lto_read): Do not recompute node params.
14176 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
14177 (ipa_get_param): Check we are not in WPA.
14178 (ipa_get_param_move_cost): New.
14179 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
14180 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
14181 parm numbers to be present.
14183 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14185 PR rtl-optimization/58048
14186 * lra-constraints.c (process_alt_operands): Don't check asm
14187 operand on register.
14189 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
14191 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
14192 the implied StoreLoad barrier for atomic operations if before.
14194 2013-08-02 Jan Hubicka <jh@suse.cz>
14195 Martin Liska <marxin.liska@gmail.com>
14197 * cgraph.c (cgraph_function_body_availability): Do not check
14199 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
14200 symtab_node_availability): Declare.
14201 * ipa.c (can_replace_by_local_alias): New.
14202 (function_and_variable_visibility): Use it.
14203 * symtab.c (symtab_for_node_and_aliases,
14204 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
14206 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14208 PR rtl-optimization/57963
14209 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
14210 (lra_constraints): Use them.
14212 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
14214 * config/arm/types.md (define_attr "type"): Add "load_acq"
14216 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
14218 (cortex_a53_store1): Likewise.
14220 2013-08-01 Jan Hubicka <jh@suse.cz>
14222 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
14223 partitions are not needed.
14225 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
14227 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
14228 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
14229 MAYBE_NON_Q_CLASS_P where appropriate.
14231 2013-08-01 Jan Hubicka <jh@suse.cz>
14233 * cgraph.h (release_function_body): Declare.
14234 * tree.c (free_lang_data_in_decl): Free, parameters and return values
14235 of unused delcarations.
14237 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14239 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
14240 RTL form when subtracting a constant.
14242 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14244 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
14245 Generate canonical plus rtx with negated immediate instead of minus
14247 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
14249 2013-08-01 Jan Hubicka <jh@suse.cz>
14251 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
14252 (cgraph_release_function_body): Likewise.
14253 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
14254 * cgraph.h (cgrpah_node): Rename abstract_and_needed
14255 to used_as_abstract_origin.
14256 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
14257 symbols used as abstract origins.
14258 * cgraphunit.c (analyze_functions): Update.
14259 * ipa.c (symtab_remove_unreachable_nodes): Recompute
14260 used_as_abstract_origin.
14261 * tree-inline.c (tree_function_versioning): Update
14262 used_as_abstract_origin; be ready for DECL_RESULT and
14263 DECL_ARGUMENTS to be NULL.
14265 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
14266 for abstract functions.
14267 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
14270 2013-08-01 Jan Hubicka <jh@suse.cz>
14272 * profile.c (compute_value_histograms): Fix thinko.
14274 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14276 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
14277 aarch-common-protos.h to extra_headers.
14278 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14279 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
14280 * config/aarch64/t-aarch64 (aarch-common.o): Define.
14282 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14284 * config/aarch64/aarch64.md (define_attr "type"): Delete.
14285 Include "../arm/types.md". Define "type" attribute for all patterns.
14286 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
14289 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
14291 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
14292 to support power8 load fusion.
14293 (fusion_gpr_mem_load): Likewise.
14295 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
14297 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
14298 declarations for power8 load fusion.
14299 (emit_fusion_gpr_load): Likewise.
14301 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
14302 tuning for power8, turn on fusion mode by default. Turn on sign
14303 extending fusion mode if normal fusion mode is on, and we are at
14305 (fusion_gpr_load_p): New function, return true if we can fuse an
14306 addis instruction with a dependent load to a GPR.
14307 (emit_fusion_gpr_load): Emit the instructions for power8 load
14310 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
14311 (VSX load fusion peepholes): New peepholes to fuse together an
14312 addi instruction with a VSX load instruction.
14314 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
14315 peepholes to fuse an addis instruction with a load to a GPR base
14316 register. If we are supporting sign extending fusions, convert
14317 sign extending loads to zero extending loads and add an explicit
14320 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14322 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
14323 aarch-common-protos.h to extra_headers.
14324 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14325 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
14326 (arm_early_store_addr_dep): Likewise.
14327 (arm_no_early_alu_shift_dep): Likewise.
14328 (arm_no_early_alu_shift_value_dep): Likewise.
14329 (arm_no_early_mul_dep): Likewise.
14330 (arm_no_early_store_addr_dep): Likewise.
14331 (arm_mac_accumulator_is_mul_result): Likewise.
14332 (arm_mac_accumulator_is_result): Likewise.
14333 * config/arm/aarch-common.c: ... here. New file.
14334 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
14336 (arm_early_store_addr_dep): Likewise.
14337 (arm_no_early_alu_shift_dep): Likewise.
14338 (arm_no_early_alu_shift_value_dep): Likewise.
14339 (arm_no_early_mul_dep): Likewise.
14340 (arm_no_early_store_addr_dep): Likewise.
14341 (arm_mac_accumulator_is_mul_result): Likewise.
14342 (arm_mac_accumulator_is_result): Likewise.
14343 * config/arm/aarch-common-protos.h: ... here. New file.
14344 * config/arm/t-arm (aarch-common.o): Define.
14346 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14348 * config/arm/arm.md: Include new file "types.md".
14349 (define_attr "type"): Move from here to ...
14350 (define_attr "mul32"): Likewise.
14351 (define_attr "mul64"): Likewise.
14352 * config/arm/types.md: ... here. New file.
14354 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
14356 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
14357 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
14359 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
14361 * gen-pass-instances.awk: Fix offset of substr().
14363 2013-07-31 David Malcolm <dmalcolm@redhat.com>
14365 * Makefile.in (pass-instances.def): New.
14366 (passes.o): Replace dependency on passes.def with one on
14369 * gen-pass-instances.awk: New.
14371 * passes.c (pass_manager::pass_manager): Use pass-instances.def
14372 rather than passes.def, updating local definition of NEXT_PASS
14373 macro to add an extra NUM parameter (currently unused).
14375 2013-07-30 David Malcolm <dmalcolm@redhat.com>
14377 * Makefile.in (PASS_MANAGER_H): New.
14378 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
14379 (passes.o): Likewise.
14380 (statistics.o): Likewise.
14381 (cgraphunit.o): Likewise.
14382 (context.o): Depend on PASS_MANAGER_H.
14384 * pass_manager.h: New.
14386 * cgraphunit.c (cgraph_add_new_function): Update for moves
14387 of globals to fields of pass_manager.
14388 (analyze_function): Likewise.
14389 (expand_function): Likewise.
14390 (ipa_passes): Likewise.
14391 (compile): Likewise.
14393 * context.c (context::context): New.
14394 * context.h (context::context): New.
14395 (context::get_passes): New.
14396 (context::passes_): New.
14398 * lto-cgraph.c (input_node): Update for moves of globals to
14399 fields of pass_manager.
14401 * passes.c (all_passes): Remove, in favor of a field of the
14402 same name within the new class pass_manager.
14403 (all_small_ipa_passes): Likewise.
14404 (all_lowering_passes): Likewise.
14405 (all_regular_ipa_passes): Likewise.
14406 (all_late_ipa_passes): Likewise.
14407 (all_lto_gen_passes): Likewise.
14408 (passes_by_id): Likewise.
14409 (passes_by_id_size): Likewise.
14410 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14411 the new class pass_manager.
14412 (set_pass_for_id): Convert to...
14413 (pass_manager::set_pass_for_id): ...method.
14414 (get_pass_for_id): Convert to...
14415 (pass_manager::get_pass_for_id): ...method.
14416 (register_one_dump_file): Move body of implementation into...
14417 (pass_manager::register_one_dump_file): ...here.
14418 (register_dump_files_1): Convert to...
14419 (pass_manager::register_dump_files_1): ...method.
14420 (register_dump_files): Convert to...
14421 (pass_manager::register_dump_files): ...method.
14422 (create_pass_tab): Update for moves of globals to fields of
14424 (dump_passes): Move body of implementation into...
14425 (pass_manager::dump_passes): ...here.
14426 (register_pass): Move body of implementation into...
14427 (pass_manager::register_pass): ...here.
14428 (init_optimization_passes): Convert into...
14429 (pass_manager::pass_manager): ...constructor for new
14430 pass_manager class, and initialize the pass_lists array.
14431 (check_profile_consistency): Update for moves of globals to
14432 fields of pass_manager.
14433 (dump_profile_report): Move body of implementation into...
14434 (pass_manager::dump_profile_report): ...here.
14435 (ipa_write_summaries_1): Update for moves of pass lists from
14436 being globals to fields of pass_manager.
14437 (ipa_write_optimization_summaries): Likewise.
14438 (ipa_read_summaries): Likewise.
14439 (ipa_read_optimization_summaries): Likewise.
14440 (execute_all_ipa_stmt_fixups): Likewise.
14442 * statistics.c (statistics_fini): Update for moves of globals to
14443 fields of pass_manager.
14445 * toplev.c (general_init): Replace call to
14446 init_optimization_passes with construction of the pass_manager
14449 * tree-pass.h (all_passes): Remove, in favor of a field of the
14450 same name within the new class pass_manager.
14451 (all_small_ipa_passes): Likewise.
14452 (all_lowering_passes): Likewise.
14453 (all_regular_ipa_passes): Likewise.
14454 (all_lto_gen_passes): Likewise.
14455 (all_late_ipa_passes): Likewise.
14456 (passes_by_id): Likewise.
14457 (passes_by_id_size): Likewise.
14458 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14459 the new class pass_manager.
14460 (get_pass_for_id): Remove.
14462 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14464 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
14467 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14469 * arm.md (mulhi3): New expand pattern.
14471 2013-07-30 Jan Hubicka <jh@suse.cz>
14472 Martin Liska <marxin.liska@gmail.com>
14474 * profile.c (compute_value_histograms): Do not ICE when
14475 there is mismatch only on some counters.
14477 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14479 PR rtl-optimization/57637
14480 * function.c (move_insn_for_shrink_wrap): Also check the
14481 GEN set of the LIVE problem for the liveness analysis
14482 if it exists, otherwise give up.
14484 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14486 PR tree-optimization/57993
14487 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
14488 replaced statement in the candidate table.
14489 (phi_add_costs): Return infinite cost when the hidden basis does
14490 not dominate all phis on which the candidate is dependent.
14491 (replace_one_candidate): Record replaced statement in the
14494 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
14496 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
14497 (ashlv2si3): New expander.
14498 (*ashlv2si3_i): New define_insn_and_split.
14499 * predicates.md (float_operation): Allow patterns with three
14500 basic sub-patterns.
14502 PR rtl-optimization/58021
14503 * mode-switching.c (create_pre_exit): Always split off preceding
14504 insns if we are not at the basic block head.
14506 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
14508 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
14509 (UCLIBC_DYNAMIC_LINKER): New macro.
14510 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
14512 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
14513 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
14515 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
14516 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
14517 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
14518 for SF and DF modes. Use ieee_quad_format for TF mode.
14519 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
14520 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
14521 (mips_option_override): Handle `-mnan=legacy'.
14522 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
14523 `-mabs=2008' and `-mnan=2008'.
14524 (OPTION_DEFAULT_SPECS): Add "nan" default.
14525 (ASM_SPEC): Handle `-mnan='.
14526 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
14527 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
14528 comment accordingly.
14529 (neg<mode>2): Likewise.
14530 * config/mips/mips.opt (mabs, mnan): New options.
14531 * doc/install.texi (Configuration): Document `--with-nan=' option.
14532 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
14534 (MIPS Options): Document them.
14535 * config.gcc <mips*-*-*>: Handle `--with-nan='.
14536 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
14537 * configure: Regenerate.
14538 * config.in: Regenerate.
14540 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14542 * config/i386/i386.md (float post-reload splitters): Do not check
14543 for subregs of SSE registers.
14545 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14546 H.J. Lu <hongjiu.lu@intel.com>
14550 * config/i386/i386.md (post-reload splitter
14551 to avoid partial SSE reg dependency stalls): New pattern.
14553 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
14555 * config/s390/s390.md ("movcc"): Swap load and store instructions.
14557 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
14559 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14560 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
14562 2013-07-26 Cary Coutant <ccoutant@google.com>
14564 * dwarf2out.c (die_checksum_ordered): Don't include template
14565 instantiations in signature.
14566 (is_template_parameter): New function.
14567 (is_template_instantiation): New function.
14568 (generate_skeleton_bottom_up): Don't include template instantiations
14570 (generate_skeleton): Likewise.
14571 (break_out_comdat_types): Move recursive call to break out nested
14573 (prune_unused_types_mark_generic_parms_dies): Call
14574 is_template_parameter.
14576 2013-07-26 Ian Bolton <ian.bolton@arm.com>
14578 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
14579 uses vector registers.
14580 * config/aarch64/iterators.md: Add attributes rtn and vas.
14582 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14583 Richard Earnshaw <richard.earnshaw@arm.com>
14585 * combine.c (simplify_comparison): Re-canonicalize operands
14587 * config/arm/arm.md (movcond_addsi): New splitter.
14589 2013-07-25 Sterling Augustine <saugustine@google.com>
14591 * dwarf2out.c (size_of_pubnames): Move code to...
14592 (include_pubname_in_output): ...here. New.
14593 (want_pubnames): Rearrange.
14594 (output_pubnames): Call include_pubname_in_output. Move assertion.
14596 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14598 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
14600 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14603 * doc/extend.texi: Remove obsolete builtins. Fix
14604 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
14606 2013-07-25 Jan Hubicka <jh@suse.cz>
14608 * cgraph.c (release_function_body): Break out from ...
14609 (cgraph_release_function_body): ... this one; also release DECL_RESULT
14610 and DECL_ARGUMENTS.
14611 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
14612 old_tree in the map.
14613 (create_specialized_node): Update.
14614 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
14616 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
14617 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
14618 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
14619 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
14622 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14624 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
14625 addsi3_carryin_alt2_<optab>): Correct output template.
14627 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14629 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
14630 Adjust for arm_restrict_it.
14631 Remove trailing whitespace.
14633 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
14635 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
14636 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
14638 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
14640 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
14642 PR rtl-optimization/57960
14643 * lra-constraints.c (process_alt_operands): Use the right mode
14644 when checking strict_low.
14646 2013-07-25 Jan Hubicka <jh@suse.cz>
14648 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
14649 * cgraph.c (cgraph_remove_node): Do not release function body
14650 when in cgraph streaming.
14651 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
14652 in other partitions are not considered reachable; fix handling of
14655 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14657 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
14659 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14664 * config/arm/arm.md ("*sibcall_insn): Replace use of
14665 Ss with US. Adjust output for v5 and v4t.
14666 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
14668 * config/arm/constraints.md ("Ss"): Rename to US.
14670 2013-07-25 Terry Guo <terry.guo@arm.com>
14672 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
14673 shift_add/shift_sub0/shift_sub1 RTXs.
14675 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
14676 Anton Blanchard <anton@au1.ibm.com>
14678 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
14679 (altivec_vpks<VI_char>ss): Likewise.
14680 (altivec_vpks<VI_char>us): Likewise.
14681 (altivec_vpku<VI_char>us): Likewise.
14682 (altivec_vpku<VI_char>um): Likewise.
14684 2013-07-24 David Malcolm <dmalcolm@redhat.com>
14686 Introduce context class.
14688 * Makefile.in (CONTEXT_H): New.
14689 (OBJS): Add context.o.
14690 (toplev.o): Add CONTEXT_H to dependencies.
14693 * toplev.c (general_init): Create the singleton gcc::context instance.
14699 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
14701 PR rtl-optimization/57968
14702 * mode-switching.c (create_pre_exit): Allow instructions that
14703 don't set a return register to need a non-exit mode.
14705 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14706 Anton Blanchard <anton@au1.ibm.com>
14708 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
14709 operands to vperm for little endian.
14710 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
14711 of lvsl to create the control mask for a vperm for little endian.
14713 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14714 Anton Blanchard <anton@au1.ibm.com>
14716 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
14717 two operands for little-endian.
14719 2013-07-23 Steve Ellcey <sellcey@mips.com>
14721 * config/mips/mips.c (mips_case_values_threshold): New.
14722 (TARGET_CASE_VALUES_THRESHOLD): Define.
14724 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14725 Anton Blanchard <anton@au1.ibm.com>
14727 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
14728 selection of field for vector splat in little endian mode.
14730 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
14732 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
14733 expanders to rs6000.md.
14734 (ior<mode>3): Likewise.
14735 (and<mode>3): Likewise.
14736 (one_cmpl<mode>2): Likewise.
14737 (nor<mode>3): Likewise.
14738 (andc<mode>3): Likewise.
14739 (eqv<mode>3): Likewise.
14740 (nand<mode>3): Likewise.
14741 (orc<mode>3): Likewise.
14743 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
14746 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
14747 to split multi-word logical operations.
14748 (rs6000_split_logical_di): Likewise.
14749 (rs6000_split_logical): Likewise.
14751 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
14752 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
14753 and allow TImode operations in 32-bit.
14754 (vsx_and<mode>3_64bit): Likewise.
14755 (vsx_ior<mode>3_32bit): Likewise.
14756 (vsx_ior<mode>3_64bit): Likewise.
14757 (vsx_xor<mode>3_32bit): Likewise.
14758 (vsx_xor<mode>3_64bit): Likewise.
14759 (vsx_one_cmpl<mode>2_32bit): Likewise.
14760 (vsx_one_cmpl<mode>2_64bit): Likewise.
14761 (vsx_nor<mode>3_32bit): Likewise.
14762 (vsx_nor<mode>3_64bit): Likewise.
14763 (vsx_andc<mode>3_32bit): Likewise.
14764 (vsx_andc<mode>3_64bit): Likewise.
14765 (vsx_eqv<mode>3_32bit): Likewise.
14766 (vsx_eqv<mode>3_64bit): Likewise.
14767 (vsx_nand<mode>3_32bit): Likewise.
14768 (vsx_nand<mode>3_64bit): Likewise.
14769 (vsx_orc<mode>3_32bit): Likewise.
14770 (vsx_orc<mode>3_64bit): Likewise.
14772 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
14773 logical types in GPRs.
14775 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
14776 logical insns to rs6000.md, and allow TImode operations in
14778 (altivec_ior<mode>3): Likewise.
14779 (altivec_xor<mode>3): Likewise.
14780 (altivec_one_cmpl<mode>2): Likewise.
14781 (altivec_nor<mode>3): Likewise.
14782 (altivec_andc<mode>3): Likewise.
14784 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
14785 attributes for moving the 128-bit logical operations into
14787 (BOOL_REGS_OUTPUT): Likewise.
14788 (BOOL_REGS_OP1): Likewise.
14789 (BOOL_REGS_OP2): Likewise.
14790 (BOOL_REGS_UNARY): Likewise.
14791 (BOOL_REGS_AND_CR0): Likewise.
14792 (one_cmpl<mode>2): Add support for DI logical operations on
14793 32-bit, splitting the operations to 32-bit.
14794 (anddi3): Likewise.
14795 (iordi3): Likewise.
14796 (xordi3): Likewise.
14797 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
14798 changes to combine the 32/64-bit code, allow logical operations on
14799 TI mode in 32-bit, and to use similar match_operator patterns like
14800 scalar mode uses. Combine the Altivec and VSX code for logical
14801 operations, and move it here.
14802 (ior<mode>3, 128-bit types): Likewise.
14803 (xor<mode>3, 128-bit types): Likewise.
14804 (one_cmpl<mode>3, 128-bit types): Likewise.
14805 (nor<mode>3, 128-bit types): Likewise.
14806 (andc<mode>3, 128-bit types): Likewise.
14807 (eqv<mode>3, 128-bit types): Likewise.
14808 (nand<mode>3, 128-bit types): Likewise.
14809 (orc<mode>3, 128-bit types): Likewise.
14810 (and<mode>3_internal): Likewise.
14811 (bool<mode>3_internal): Likewise.
14812 (boolc<mode>3_internal1): Likewise.
14813 (boolc<mode>3_internal2): Likewise.
14814 (boolcc<mode>3_internal1): Likewise.
14815 (boolcc<mode>3_internal2): Likewise.
14816 (eqv<mode>3_internal1): Likewise.
14817 (eqv<mode>3_internal2): Likewise.
14818 (one_cmpl1<mode>3_internal): Likewise.
14820 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14822 * config/microblaze/microblaze.c (microblaze_expand_prologue):
14823 Rename flag_stack_usage to flag_stack_usage_info.
14825 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14827 * config/microblaze/sync.md: New file.
14828 * config/microblaze/microblaze.md: Include sync.md
14829 * config/microblaze/microblaze.c: Add print_operand 'y'.
14830 * config/microblaze/constraints.md: Add memory_contraint
14831 'Q' which is a single register.
14833 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
14835 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
14837 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14839 * reload.c (find_reloads): Exit loop once we find this operand
14840 cannot be reloaded somehow for this alternative.
14842 * reload.c (find_reloads): Exit loop once we find a hard register.
14844 * rtlanal.c (computed_jump_p): Exit loop once we find label
14847 * i386.c (ix86_pad_returns): Exit loop after setting replace.
14849 * cfgloopmanip.c (remove_path): Exit loop after setting
14852 * gensupport.c (subst_dup): Avoid loop if code is not
14853 MATCH_DUP nor MATCH_OP_DUP.
14855 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
14857 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
14859 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14861 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
14862 true for SP_REGNUM if mode == ptr_mode.
14863 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
14864 with value R0_REGNUM + 31.
14866 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14868 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
14869 pad pointer-typed argument downward.
14871 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14873 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
14874 and __ILP32__ when the ILP32 model is in use.
14876 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14878 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
14879 (aarch64_load_symref_appropriately): In the case of
14880 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
14881 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
14882 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
14883 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
14884 if 'mode' doesn't equal to 'ptr_mode'.
14885 (aarch64_output_mi_thunk): Add an assertion on the alignment of
14886 'vcall_offset'; change to call aarch64_emit_move differently depending
14887 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
14888 to calculate the upper bound of 'vcall_offset'.
14889 (aarch64_cannot_force_const_mem): Change to also return true if
14891 (aarch64_legitimize_reload_address): In the case of large
14892 displacements, add new local variable 'xmode' and an assertion
14893 based on it; change to use 'xmode' to generate the new rtx and
14895 (aarch64_asm_trampoline_template): Change to generate the template
14896 differently depending on TARGET_ILP32 or not; change to use
14897 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
14898 (aarch64_trampoline_size): Removed.
14899 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
14900 and replace immediate literals with it. Change to use 'ptr_mode'
14901 instead of 'DImode' and call convert_memory_address if the mode
14902 of 'fnaddr' doesn't equal to 'ptr_mode'.
14903 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
14905 (aarch64_elf_asm_destructor): Likewise.
14906 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
14907 on TARGET_ILP32 instead of aarch64_trampoline_size.
14908 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
14909 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
14910 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
14911 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
14912 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
14913 (storewb_pair<GPI:mode>_<P:mode>): ... this.
14914 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
14915 depending on the value of 'mode'.
14916 (add_losym_<mode>): New.
14917 (ldr_got_small_<mode>): New, based on ldr_got_small.
14918 (ldr_got_small): Remove.
14919 (ldr_got_small_sidi): New.
14920 * config/aarch64/iterators.md (P): New.
14921 (PTR): Change to 'ptr_mode' in the condition.
14923 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14925 * config.gcc (aarch64*-*-*): Support --with-abi.
14926 (aarch64*-*-elf): Support --with-multilib-list.
14927 (aarch64*-*-linux*): Likewise.
14928 (supported_defaults): Add abi to aarch64*-*-*.
14929 * configure.ac: Mention AArch64 for --with-multilib-list.
14930 * configure: Re-generated.
14931 * config/aarch64/biarchilp32.h: New file.
14932 * config/aarch64/biarchlp64.h: New file.
14933 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
14935 (MULTILIB_DEFAULTS): Ditto.
14936 (DRIVER_SELF_SPECS): Ditto.
14937 (ASM_SPEC): Update to also substitute -mabi.
14938 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
14939 file whose name depends on -mabi= and -mbig-endian.
14940 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
14942 (POINTER_SIZE): New define.
14943 (POINTERS_EXTEND_UNSIGNED): Ditto.
14944 (enum aarch64_abi_type): New enumeration tag.
14945 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
14946 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
14947 (TARGET_ILP32): New define.
14948 * config/aarch64/aarch64.opt (mabi): New.
14949 (aarch64_abi): New.
14950 (ilp32, lp64): New values for -mabi.
14951 * config/aarch64/t-aarch64 (comma): New define.
14952 (MULTILIB_OPTIONS): Ditto.
14953 (MULTILIB_DIRNAMES): Ditto.
14954 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
14955 * doc/invoke.texi: Document -mabi for AArch64.
14957 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
14959 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
14961 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14962 Anton Blanchard <anton@au1.ibm.com>
14964 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
14965 endianness when selecting field to splat.
14967 2013-07-22 Eric Christopher <echristo@gmail.com>
14969 * dwarf2out.c (die_odr_checksum): New function to use
14970 CHECKSUM_ macros and ULEB128 for DIE tag.
14971 (generate_type_signature): Use.
14973 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
14975 * config.gcc (sparc*-*-*): Accept leon3 processor.
14976 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
14977 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
14978 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
14979 * config/sparc/sparc.opt (enum processor_type): Add leon3.
14980 (mfix-ut699): Adjust comment.
14981 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
14982 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
14983 (CPP_CPU_SPEC): Likewise.
14984 (ASM_CPU_SPEC): Likewise.
14985 * config/sparc/sparc.c (leon3_cost): New constant.
14986 (sparc_option_override): Add leon3 support.
14987 (mem_ref): New function.
14988 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
14989 (sparc_do_work_around_errata): Look into the instruction in the delay
14990 slot and adjust accordingly. Add fix for the data cache nullify issues
14991 of the UT699. Change insertion position for the NOP.
14992 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
14993 (leon3_load): New reservation.
14994 (leon_store): Bump latency to 2.
14995 (grfpu): New automaton.
14996 (grfpu_alu): New unit.
14997 (grfpu_ds): Likewise.
14998 (leon_fp_alu): Adjust.
14999 (leon_fp_mult): Delete.
15000 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
15001 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
15002 * config/sparc/sparc.md (cpu): Add leon3.
15003 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
15004 (swapsi): Likewise.
15005 (atomic_test_and_set): Likewise.
15006 (ldstub): Likewise.
15008 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
15010 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
15011 default for R5900 targets.
15012 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
15013 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
15014 * config/mips/mips.c (mips_option_override): Report an error for
15015 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
15016 for -march=r5900 -mhard-float.
15018 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15020 * df-problems.c (can_move_insns_across): Exit loop once we
15021 find a non-fixed, non-global register.
15023 * ipa-pure-const.c (propagate_nothrow): Exit loop after
15026 * omega.c (omega_eliminate_red): Break after setting red_found.
15027 (omega_problem_has_red_equations): Similarly after setting found.
15028 (omega_query_variable): Similarly after setting coupled.
15030 2013-07-22 Marek Polacek <polacek@redhat.com>
15032 * gimplify.c: Don't include gimple.h twice.
15034 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15036 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
15037 instead of TARGET_THUMB1.
15038 (Pz): New constraint.
15039 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
15041 (compare_negsi_si): Likewise.
15042 (compare_addsi2_op0): Likewise.
15043 (compare_addsi2_op1): Likewise.
15044 (addsi3_carryin_<optab>): Likewise.
15045 (addsi3_carryin_alt2_<optab>): Likewise.
15046 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
15047 for arm_restrict_it.
15048 (subsi3_carryin): Likewise.
15049 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
15050 (minmax_arithsi): Disable for arm_restrict_it.
15051 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
15052 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
15053 (satsi_<SAT:code>_shift): Likewise.
15054 (arm_shiftsi3): Add alternative for 16-bit encoding.
15055 (arm32_movhf): Disable for arm_restrict_it.
15056 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
15057 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
15059 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
15061 * config/arm/arm.md (attribute "insn"): Delete.
15062 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
15063 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
15064 (not_shiftsi): Update for attribute change.
15065 (not_shiftsi_compare0): Likewise.
15066 (not_shiftsi_compare0_scratch): Likewise.
15067 (arm_one_cmplsi2): Likewise.
15068 (thumb1_one_cmplsi2): Likewise.
15069 (notsi_compare0): Likewise.
15070 (notsi_compare0_scratch): Likewise.
15071 (thumb1_movdi_insn): Likewise.
15072 (arm_movsi_insn): Likewise.
15073 (movhi_insn_arch4): Likewise.
15074 (movhi_bytes): Likewise.
15075 (arm_movqi_insn): Likewise.
15076 (thumb1_movqi_insn): Likewise.
15077 (arm32_movhf): Likewise.
15078 (thumb1_movhf): Likewise.
15079 (arm_movsf_soft_insn): Likewise.
15080 (thumb1_movsf_insn): Likewise.
15081 (thumb_movdf_insn): Likewise.
15082 (movsicc_insn): Likewise.
15083 (movsfcc_soft_insn): Likewise.
15084 (and_scc): Likewise.
15085 (cond_move): Likewise.
15086 (if_move_not): Likewise.
15087 (if_not_move): Likewise.
15088 (if_shift_move): Likewise.
15089 (if_move_shift): Likewise.
15090 (if_shift_shift): Likewise.
15091 (if_not_arith): Likewise.
15092 (if_arith_not): Likewise.
15093 (cond_move_not): Likewise.
15094 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
15095 (neon_mov<mode>): Likewise.
15096 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
15097 (thumb2_movsi_vfp): Likewise.
15098 (movsf_vfp): Likewise.
15099 (thumb2_movsf_vfp): Likewise.
15100 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15102 (cortexa7_older_only): Likewise.
15103 (cortexa7_younger): Likewise.
15104 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
15105 (1020alu_shift_op): Likewise.
15106 (1020alu_shift_reg_op): Likewise.
15107 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
15108 (alu_shift_op): Likewise.
15109 (alu_shift_reg_op): Likewise.
15110 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
15111 (11_alu_shift_op): Likewise.
15112 (11_alu_shift_reg_op): Likewise.
15113 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
15114 (9_alu_shift_reg_op): Likewise.
15115 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15117 (cortex_a15_alu_shift): Likewise.
15118 (cortex_a15_alu_shift_reg): Likewise.
15119 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
15120 (cortex_a5_alu_shift): Likewise.
15121 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
15123 (cortex_a53_alu_shift): Likewise.
15124 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15126 (cortex_a7_alu_reg): Likewise.
15127 (cortex_a7_alu_shift): Likewise.
15128 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15129 (cortex_a8_alu_shift): Likewise.
15130 (cortex_a8_alu_shift_reg): Likewise.
15131 (cortex_a8_mov): Likewise.
15132 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
15133 (cortex_a9_dp_shift): Likewise.
15134 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
15135 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
15136 (cortex_r4_mov): Likewise.
15137 (cortex_r4_alu_shift): Likewise.
15138 (cortex_r4_alu_shift_reg): Likewise.
15139 * config/arm/fa526.md (526_alu_op): Update for attribute change.
15140 (526_alu_shift_op): Likewise.
15141 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
15142 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
15143 (626te_alu_shift_op): Likewise.
15144 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
15145 (726te_alu_op): Likewise.
15146 (726te_alu_shift_op): Likewise.
15147 (726te_alu_shift_reg_op): Likewise.
15148 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
15149 (mp626_alu_shift_op): Likewise.
15150 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
15151 (pj4_alu_e1_conds): Likewise.
15152 (pj4_alu): Likewise.
15153 (pj4_alu_conds): Likewise.
15154 (pj4_shift): Likewise.
15155 (pj4_shift_conds): Likewise.
15156 (pj4_alu_shift): Likewise.
15157 (pj4_alu_shift_conds): Likewise.
15159 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15161 * config/arm/predicates.md (shiftable_operator_strict_it):
15163 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
15164 Disable cond_exec version for arm_restrict_it.
15165 (thumb2_smaxsi3): Convert to generate cond_exec.
15166 (thumb2_sminsi3): Likewise.
15167 (thumb32_umaxsi3): Likewise.
15168 (thumb2_uminsi3): Likewise.
15169 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
15170 (thumb2_neg_abssi2): Likewise.
15171 (thumb2_mov_scc): Add alternative for 16-bit encoding.
15172 (thumb2_movsicc_insn): Adjust alternatives.
15173 (thumb2_mov_negscc): Disable for arm_restrict_it.
15174 (thumb2_mov_negscc_strict_it): New pattern.
15175 (thumb2_mov_notscc_strict_it): New pattern.
15176 (thumb2_mov_notscc): Disable for arm_restrict_it.
15177 (thumb2_ior_scc): Likewise.
15178 (thumb2_ior_scc_strict_it): New pattern.
15179 (thumb2_cond_move): Adjust for arm_restrict_it.
15180 (thumb2_cond_arith): Disable for arm_restrict_it.
15181 (thumb2_cond_arith_strict_it): New pattern.
15182 (thumb2_cond_sub): Adjust for arm_restrict_it.
15183 (thumb2_movcond): Likewise.
15184 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
15185 (thumb2_zero_extendhisi2_v6): Likewise.
15186 (thumb2_zero_extendqisi2_v6): Likewise.
15187 (orsi_notsi_si): Likewise.
15188 (orsi_not_shiftsi_si): Likewise.
15190 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
15192 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
15193 instruction sequence is 1 byte shorter.
15195 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
15197 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
15198 it is not needed after split.
15200 2013-07-20 Iain Sandoe <iain@codesourcery.com>
15203 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
15204 second label for nonlocal goto receivers. Don't output pic base labels
15205 unless we're producing PIC; mark that action unreachable().
15206 (ix86_save_reg): If the function contains a nonlocal label, save the
15208 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
15209 * config/darwin.c (emitted_pic_label_num): New GTY.
15210 (update_pic_label_number_if_needed): New.
15211 (machopic_output_function_base_name): Adjust for nonlocal receiver
15213 (machopic_should_output_picbase_label): New.
15214 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
15215 (nonlocal_goto_receiver): New insn and split.
15217 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
15219 * config/aarch64/aarch64-builtins.c
15220 (aarch64_fold_builtin): Fold abs in all modes.
15221 * config/aarch64/aarch64-simd-builtins.def
15222 (abs): Enable for all modes.
15223 * config/aarch64/arm_neon.h
15224 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
15225 (vabs_f64): Add missing intrinsic.
15227 2013-07-19 Ian Bolton <ian.bolton@arm.com>
15229 * config/aarch64/arm_neon.h (vabs_s64): New function
15231 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
15234 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
15235 * config/avr/avr.md (adjust_len): Add `round'.
15236 * config/avr/avr-protos.h (avr_out_round): New prototype.
15237 (avr_out_plus): Add `out_label' argument.
15238 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
15239 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
15240 Handle the case where `insn' is just a pattern.
15241 (avr_out_bitop): Handle the case where `insn' is just a pattern.
15242 (avr_out_round): New function.
15243 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
15245 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
15247 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15248 Add check for flag_stack_usage to handle -fstack-usage support
15250 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
15252 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
15253 interaction for new Power8 flags and VSX.
15255 2013-07-18 Sriraman Tallam <tmsriram@google.com>
15257 PR middle-end/57698
15258 * tree-inline.c (expand_call_inline): Emit errors during
15259 early_inlining only if optimization is not turned on.
15261 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15265 * passes.c (init_optimization_passes): Move the construction of
15266 the pass hierarchy into a new passes.def file.
15268 * Makefile.in (passes.o): Add dependency on passes.def.
15270 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15272 * passes.c (init_optimization_passes): Introduce macros for
15273 constructing the tree of passes (INSERT_PASSES_AFTER,
15274 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
15275 TERMINATE_PASS_LIST).
15277 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
15278 Wei Mi <wmi@google.com>
15280 PR rtl-optimization/57878
15281 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
15283 (reload_pseudo_compare_func): Check nregs first for reload
15286 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15288 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
15290 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
15292 * read-rtl.c (validate_const_int): Once an invalid character is
15293 seen, quit the loop.
15295 * gengtype.c (write_roots): Similarly once we find the "deletable"
15296 or "if_marked" option.
15298 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15300 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
15301 "xtab" and "sat". Move value "clz" from here to ...
15302 (attriubte "type"): ... here.
15303 (satsi_<SAT:code>): Delete "insn" attribute.
15304 (satsi_<SAT:code>_shift): Likewise.
15305 (arm_zero_extendqisi2addsi): Likewise.
15306 (arm_extendqisi2addsi): Likewise.
15307 (clzsi2): Update for attribute changes.
15308 (rbitsi2): Likewise.
15309 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
15311 (arm_usatsihi): Likewise.
15312 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15314 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15316 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
15317 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
15318 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
15319 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
15320 in alphabetical order.
15321 (attribute "core_cycles"): Update for attribute changes.
15322 (arm_addsi3): Likewise.
15323 (addsi3_compare0): Likewise.
15324 (addsi3_compare0_scratch): Likewise.
15325 (addsi3_compare_op1): Likewise.
15326 (addsi3_compare_op2): Likewise.
15327 (compare_addsi2_op0): Likewise.
15328 (compare_addsi2_op1): Likewise.
15329 (addsi3_carryin_shift_<optab>): Likewise.
15330 (subsi3_carryin_shift): Likewise.
15331 (rsbsi3_carryin_shift): Likewise.
15332 (arm_subsi3_insn): Likewise.
15333 (subsi3_compare0): Likewise.
15334 (subsi3_compare): Likewise.
15335 (arm_andsi3_insn): Likewise.
15336 (thumb1_andsi3_insn): Likewise.
15337 (andsi3_compare0): Likewise.
15338 (andsi3_compare0_scratch): Likewise.
15339 (zeroextractsi_compare0_scratch
15340 (andsi_not_shiftsi_si): Likewise.
15341 (iorsi3_insn): Likewise.
15342 (iorsi3_compare0): Likewise.
15343 (iorsi3_compare0_scratch): Likewise.
15344 (arm_xorsi3): Likewise.
15345 (thumb1_xorsi3_insn): Likewise.
15346 (xorsi3_compare0): Likewise.
15347 (xorsi3_compare0_scratch): Likewise.
15348 (satsi_<SAT:code>_shift): Likewise.
15350 (arm_shiftsi3): Likewise.
15351 (shiftsi3_compare0): Likewise.
15352 (not_shiftsi): Likewise.
15353 (not_shiftsi_compare0): Likewise.
15354 (not_shiftsi_compare0_scratch): Likewise.
15355 (arm_one_cmplsi2): Likewise.
15356 (thumb_one_complsi2): Likewise.
15357 (notsi_compare0): Likewise.
15358 (notsi_compare0_scratch): Likewise.
15359 (thumb1_zero_extendhisi2): Likewise.
15360 (arm_zero_extendhisi2): Likewise.
15361 (arm_zero_extendhisi2_v6): Likewise.
15362 (arm_zero_extendhisi2addsi): Likewise.
15363 (thumb1_zero_extendqisi2): Likewise.
15364 (thumb1_zero_extendqisi2_v6): Likewise.
15365 (arm_zero_extendqisi2): Likewise.
15366 (arm_zero_extendqisi2_v6): Likewise.
15367 (arm_zero_extendqisi2addsi): Likewise.
15368 (thumb1_extendhisi2): Likewise.
15369 (arm_extendhisi2): Likewise.
15370 (arm_extendhisi2_v6): Likewise.
15371 (arm_extendqisi): Likewise.
15372 (arm_extendqisi_v6): Likewise.
15373 (arm_extendqisi2addsi): Likewise.
15374 (thumb1_extendqisi2): Likewise.
15375 (thumb1_movdi_insn): Likewise.
15376 (arm_movsi_insn): Likewise.
15377 (movsi_compare0): Likewise.
15378 (movhi_insn_arch4): Likewise.
15379 (movhi_bytes): Likewise.
15380 (arm_movqi_insn): Likewise.
15381 (thumb1_movqi_insn): Likewise.
15382 (arm32_movhf): Likewise.
15383 (thumb1_movhf): Likewise.
15384 (arm_movsf_soft_insn): Likewise.
15385 (thumb1_movsf_insn): Likewise.
15386 (movdf_soft_insn): Likewise.
15387 (thumb_movdf_insn): Likewise.
15388 (arm_cmpsi_insn): Likewise.
15389 (cmpsi_shiftsi): Likewise.
15390 (cmpsi_shiftsi_swp): Likewise.
15391 (arm_cmpsi_negshiftsi_si): Likewise.
15392 (movsicc_insn): Likewise.
15393 (movsfcc_soft_insn): Likewise.
15394 (arith_shiftsi): Likewise.
15395 (arith_shiftsi_compare0
15396 (arith_shiftsi_compare0_scratch
15397 (sub_shiftsi): Likewise.
15398 (sub_shiftsi_compare0
15399 (sub_shiftsi_compare0_scratch
15400 (and_scc): Likewise.
15401 (cond_move): Likewise.
15402 (if_plus_move): Likewise.
15403 (if_move_plus): Likewise.
15404 (if_move_not): Likewise.
15405 (if_not_move): Likewise.
15406 (if_shift_move): Likewise.
15407 (if_move_shift): Likewise.
15408 (if_shift_shift): Likewise.
15409 (if_not_arith): Likewise.
15410 (if_arith_not): Likewise.
15411 (cond_move_not): Likewise.
15412 (thumb1_ashlsi3): Set type attribute.
15413 (thumb1_ashrsi3): Likewise.
15414 (thumb1_lshrsi3): Likewise.
15415 (thumb1_rotrsi3): Likewise.
15416 (shiftsi3_compare0_scratch): Likewise.
15417 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
15418 (neon_mov<mode>): Likewise.
15419 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
15421 (thumb2_movsi_insn): Likewise.
15422 (thumb2_cmpsi_neg_shiftsi): Likewise.
15423 (thumb2_extendqisi_v6): Likewise.
15424 (thumb2_zero_extendhisi2_v6): Likewise.
15425 (thumb2_zero_extendqisi2_v6): Likewise.
15426 (thumb2_shiftsi3_short): Likewise.
15427 (thumb2_addsi3_compare0_scratch): Likewise.
15428 (orsi_not_shiftsi_si): Likewise.
15429 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
15430 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
15432 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
15433 (1020alu_shift_op): Likewise.
15434 (1020alu_shift_reg_op): Likewise.
15435 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
15436 (alu_shift_op): Likewise.
15437 (alu_shift_reg_op): Likewise.
15438 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
15439 (11_alu_shift_op): Likewise.
15440 (11_alu_shift_reg_op): Likewise.
15441 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
15442 (9_alu_shift_reg_op): Likewise.
15443 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15445 (cortex_a15_alu_shift): Likewise.
15446 (cortex_a15_alu_shift_reg): Likewise.
15447 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
15449 (cortex_a5_alu_shift): Likewise.
15450 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
15452 (cortex_a53_alu_shift): Likewise.
15453 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15455 (cortex_a7_alu_reg): Likewise.
15456 (cortex_a7_alu_shift): Likewise.
15457 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
15459 (cortex_a8_alu_shift): Likewise.
15460 (cortex_a8_alu_shift_reg): Likewise.
15461 (cortex_a8_mov): Likewise.
15462 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
15463 (cortex_a9_dp_shift): Likewise.
15464 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
15466 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
15468 (cortex_r4_mov): Likewise.
15469 (cortex_r4_alu_shift): Likewise.
15470 (cortex_r4_alu_shift_reg): Likewise.
15471 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
15472 (526_alu_shift_op): Likewise.
15473 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
15474 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
15475 (626te_alu_shift_op): Likewise.
15476 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
15477 (726te_alu_op): Likewise.
15478 (726te_alu_shift_op): Likewise.
15479 (726te_alu_shift_reg_op): Likewise.
15480 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
15481 (mp626_alu_shift_op): Likewise.
15482 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
15483 (pj4_alu_e1_conds): Likewise.
15484 (pj4_alu): Likewise.
15485 (pj4_alu_conds): Likewise.
15486 (pj4_shift): Likewise.
15487 (pj4_shift_conds): Likewise.
15488 (pj4_alu_shift): Likewise.
15489 (pj4_alu_shift_conds): Likewise.
15490 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15492 (cortexa7_older_only): Likewise.
15493 (cortexa7_younger): Likewise.
15495 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15497 * ipa-pure-const.c (generate_summary): Rename to...
15498 (pure_const_generate_summary): ... this.
15500 2013-07-17 Iain Sandoe <iain@codesourcery.com>
15502 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
15504 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
15507 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
15510 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15512 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
15513 enabled without -march=zEC12.
15514 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
15517 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15519 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
15520 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
15521 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
15522 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
15523 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
15524 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
15525 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
15527 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
15528 and ISA_HAS_NMADD3_NMSUB3.
15529 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
15530 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
15531 (nmsub4<mode>, nmsub3<mode>): Likewise.
15532 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
15534 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15536 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
15537 TARGET_MIPS5400 checking.
15539 2013-07-16 Jakub Jelinek <jakub@redhat.com>
15540 Peter Bergner <bergner@vnet.ibm.com>
15542 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
15543 registers in the comment.
15544 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
15545 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
15546 rather than FIRST_PSEUDO_REGISTERS.
15548 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
15550 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
15551 enable extra ISA flags with TARGET_HTM.
15553 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15555 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
15558 2013-07-15 Cong Hou <congh@google.com>
15560 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
15561 in compare function for sorting.
15563 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
15565 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
15566 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
15567 * config/rs6000/rs6000.opt: Add -mhtm option.
15568 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
15569 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
15570 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
15571 __HTM__ if the HTM instructions are available.
15572 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
15573 htm_spr_reg_operand): New define_predicates.
15574 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
15575 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
15577 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
15578 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
15579 HTM builtin functions.
15580 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
15581 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
15582 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
15583 (rs6000_builtin_mask_calculate): Likewise.
15584 (rs6000_option_override_internal): Likewise.
15585 (bdesc_htm): Add new HTM builtin support.
15586 (htm_spr_num): New function.
15587 (htm_spr_regno): Likewise.
15588 (rs6000_htm_spr_icode): Likewise.
15589 (htm_expand_builtin): Likewise.
15590 (htm_init_builtins): Likewise.
15591 (rs6000_expand_builtin): Add support for HTM builtin functions.
15592 (rs6000_init_builtins): Likewise.
15593 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
15595 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
15596 (TARGET_HTM, MASK_HTM): Define macros.
15597 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
15598 (FIXED_REGISTERS): Likewise.
15599 (CALL_USED_REGISTERS): Likewise.
15600 (CALL_REALLY_USED_REGISTERS): Likewise.
15601 (REG_ALLOC_ORDER): Likewise.
15602 (enum reg_class): Likewise.
15603 (REG_CLASS_NAMES): Likewise.
15604 (REG_CLASS_CONTENTS): Likewise.
15605 (REGISTER_NAMES): Likewise.
15606 (ADDITIONAL_REGISTER_NAMES): Likewise.
15607 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
15608 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
15609 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
15610 * config/rs6000/htm.md: New file.
15611 * config/rs6000/htmintrin.h: New file.
15612 * config/rs6000/htmxlintrin.h: New file.
15614 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
15616 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15617 Define SYMBOL_TINY_GOT, update comment.
15618 * config/aarch64/aarch64.c
15619 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
15620 (aarch64_expand_mov_immediate): Likewise.
15621 (aarch64_print_operand): Likewise.
15622 (aarch64_classify_symbol): Likewise.
15623 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
15624 (ldr_got_tiny): Define.
15626 2013-07-13 Tobias Grosser <tobias@grosser.es>
15628 PR tree-optimization/54094
15629 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
15630 scheduling dimension for the parallelism check from the polyhedral
15631 information in the AST.
15632 * graphite-dependences.c (carries_deps): Do not assume the schedule is
15635 2013-07-13 Jason Merrill <jason@redhat.com>
15637 * print-tree.c (debug_vec_tree): Use debug_raw.
15638 (debug_raw (vec<tree, va_gc> &)): New.
15639 (debug_raw (vec<tree, va_gc> *)): New.
15640 * tree.h: Declare them.
15642 2013-07-13 Bin Cheng <bin.cheng@arm.com>
15644 * ifcvt.c (ifcvt_after_combine): New static variable.
15645 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
15647 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
15648 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
15649 rest_of_handle_if_after_reload): Pass new argument for if_convert.
15651 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
15653 * config/mips/mips.c (mips_expand_call): Remove empty statement.
15655 2013-07-12 Michael Matz <matz@suse.de>
15657 PR middle-end/55771
15658 * convert.c (convert_to_real): Reject non-float inner types.
15660 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
15662 * config/aarch64/aarch64-protos.h
15663 (aarch64_simd_immediate_valid_for_move): Remove.
15664 * config/aarch64/aarch64.c (simd_immediate_info): New member.
15665 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
15667 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
15669 2013-07-11 Steve Ellcey <sellcey@mips.com>
15671 * config/mips/mips.c (mips_conditional_register_usage): Do not
15672 use t[0-7] registers in MIPS16 mode when optimizing for size.
15674 2013-07-11 Sriraman Tallam <tmsriram@google.com>
15676 * config/i386/i386.c (dispatch_function_versions): Fix array
15677 indexing of function_version_info to match actual_versions.
15679 2013-07-11 Teresa Johnson <tejohnson@google.com>
15681 * vec.h (struct va_gc): Move release out-of-line.
15682 (va_gc::release): Call ggc_free on released vec.
15684 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15686 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
15687 Require GOT register as additional operand in UNSPEC.
15688 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
15689 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15690 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15691 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
15692 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
15693 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
15694 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
15696 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
15699 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
15700 name seen by assembler/linker rather if available.
15702 2013-07-11 Andreas Schwab <schwab@suse.de>
15704 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
15706 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
15708 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
15710 2013-07-10 Joseph Myers <joseph@codesourcery.com>
15712 * doc/tm.texi.in: Move hook documentation to ....
15713 * target.def: ... here.
15715 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
15716 text on @hook line.
15717 * doc/tm.texi: Regenerate.
15719 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
15722 * doc/invoke.texi: Document -Wconditionally-supported.
15724 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15727 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
15730 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15733 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
15734 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
15735 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
15736 Remove duplicate devices.
15737 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
15738 * config/avr/t-multilib: Regenerate.
15739 * config/avr/avr-tables.opt: Regenerate.
15740 * doc/avr-mmcu.texi: Regenerate.
15742 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15745 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
15747 2013-07-10 Graham Stott <graham.stott@btinternet.com>
15749 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
15750 the cost of MULT when optimizing for size.
15752 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
15754 * config/cr16/cr16-protos.h: Don't include target.h.
15756 2013-07-09 Joseph Myers <joseph@codesourcery.com>
15758 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
15759 adjust register size for TDmode and TFmode for VSX registers.
15761 2013-07-08 Kai Tietz <ktietz@redhat.com>
15764 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
15765 hook_bool_const_tree_true.
15767 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15769 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
15770 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
15771 * config/s390/s390.md: Define FPR*_REGNUM constants.
15772 Fix FPR2_REGNUM constant (18 -> 17).
15773 ("*trunc<BFP:mode><DFP_ALL:mode>2")
15774 ("*trunc<DFP_ALL:mode><BFP:mode>2")
15775 ("trunc<BFP:mode><DFP_ALL:mode>2")
15776 ("trunc<DFP_ALL:mode><BFP:mode>2")
15777 ("*extend<BFP:mode><DFP_ALL:mode>2")
15778 ("*extend<DFP_ALL:mode><BFP:mode>2")
15779 ("extend<BFP:mode><DFP_ALL:mode>2")
15780 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
15783 2013-07-08 Graham Stott <graham.stott@btinternet.com>
15785 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
15787 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15789 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
15790 and cfun_fpr_bit_p to cfun_fpr_save_p.
15791 (s390_frame_area, s390_register_info, s390_frame_info)
15792 (s390_emit_prologue, s390_emit_epilogue)
15793 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
15795 * config/s390/s390.h: Define *_REGNUM macros for floating point
15798 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
15800 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
15802 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
15804 PR rtl-optimization/57786
15805 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
15806 and break out of the loop when it is set to false.
15808 2013-07-08 Jakub Jelinek <jakub@redhat.com>
15811 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
15812 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
15813 (const_int 63)) 0)).
15814 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
15815 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
15816 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
15818 PR rtl-optimization/57829
15819 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
15820 mask bits outside of mode are just sign-extension from mode to HWI.
15822 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
15824 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
15825 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
15826 adjust_address instead of change_address to keep info about alignment.
15827 (emit_strmov): Remove.
15828 (emit_memmov): New function.
15829 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
15830 (expand_movmem_epilogue): Likewise and return updated rtx for
15832 (expand_constant_movmem_prologue): Likewise and return updated rtx for
15833 destination and source.
15834 (decide_alignment): Refactor, handle vector_loop.
15835 (ix86_expand_movmem): Likewise.
15836 (ix86_expand_setmem): Likewise.
15837 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
15839 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
15841 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
15842 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
15844 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15846 * config/i386/sse.md (sse_movlhps): Change alternative 3
15847 of operand 2 to "m".
15849 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15852 * config/i386/sse.md (iptr): New mode attribute.
15853 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
15854 (<sse>_vm<plusminus_insn><mode>3): Ditto.
15855 (<sse>_vmmul<mode>3): Ditto.
15856 (<sse>_vmdiv<mode>3): Ditto.
15857 (sse_vmrcpv4sf2): Ditto.
15858 (<sse>_vmsqrt<mode>2): Ditto.
15859 (sse_vmrsqrtv4sf2): Ditto.
15860 (<sse>_vm<code><mode>3): Ditto.
15861 (avx_vmcmp<mode>3): Ditto.
15862 (<sse>_vmmaskcmp<mode>3): Ditto.
15863 (<sse>_comi): Ditto.
15864 (<sse>_ucomi): Ditto.
15865 (*xop_vmfrcz_<mode>): Ditto.
15866 (*fmai_fmadd_<mode>): Ditto.
15867 (*fmai_fmsub_<mode>): Ditto.
15868 (*fmai_fnmadd_<mode>): Ditto.
15869 (*fmai_fnmsub_<mode>): Ditto.
15870 (*fma4i_vmfmadd_<mode>): Ditto.
15871 (*fma4i_vmfmsub_<mode>): Ditto.
15872 (*fma4i_vmfnmadd_<mode>): Ditto.
15873 (*fma4i_vmfnmsub_<mode>): Ditto.
15874 (*xop_vmfrcz_<mode>): Ditto.
15875 (sse_cvtps2pi): Ditto.
15876 (sse_cvttps2pi): Ditto.
15877 (sse_cvtss2si): Ditto.
15878 (sse_cvtss2si_2): Ditto.
15879 (sse_cvtss2siq_2): Ditto.
15880 (sse_cvttss2si): Ditto.
15881 (sse_cvttss2siq): Ditto.
15882 (sse_cvtsd2si): Ditto.
15883 (sse_cvtsd2si_2): Ditto.
15884 (sse_cvtsd2siq_2): Ditto.
15885 (sse_cvttsd2si): Ditto.
15886 (sse_cvttsd2siq): Ditto.
15887 (sse_cvtsd2ss): Ditto.
15888 (sse_cvtss2sd): Ditto.
15889 (avx2_pbroadcast<mode>): Ditto.
15890 (avx2_pbroadcast<mode>_1): Ditto.
15891 (*avx_vperm_broadcast_v4sf): Ditto.
15893 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
15894 (sse_movlhps): Ditto.
15895 (sse_storehps): Ditto.
15896 (sse_loadhps): Ditto.
15897 (sse_storelps): Ditto.
15898 (sse_loadlps): Ditto.
15899 (*vec_concatv4sf): Ditto.
15900 (*vec_interleave_highv2df): Ditto.
15901 (*vec_interleave_lowv2df): Ditto.
15902 (*vec_extractv2df_1_sse): Ditto.
15903 (*vec_extractv2df_0_sse): Ditto.
15904 (sse2_storelpd): Ditto.
15905 (sse2_loadlpd): Ditto.
15906 (sse2_movsd): Ditto.
15907 (*vec_concatv4si): Ditto.
15908 (vec_concatv2di): Ditto.
15910 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
15911 for Intel asm dialect.
15912 (mmx_punpcklwd): Ditto.
15913 (mmx_punpckldq): Ditto.
15915 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
15916 for intel assembler dialect.
15918 2013-07-06 Jakub Jelinek <jakub@redhat.com>
15921 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
15922 for BUILT_IN_C{LZ,LRSB}*.
15923 * tree.h (CASE_INT_FN): Add FN##IMAX case.
15924 * tree-vrp.c (extract_range_basic): Handle
15925 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
15926 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
15927 fall thru to code calling set_value*.
15928 * builtins.c (expand_builtin): Remove *IMAX cases.
15929 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
15930 if width is bigger than 2*HWI.
15932 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
15934 PR rtl-optimization/55342
15935 * lra-int.h (lra_subreg_reload_pseudos): New.
15936 * lra.c: Add undoing optional reloads to the block diagram.
15937 (lra_subreg_reload_pseudos): New.
15938 (lra_optional_reload_pseudos): Change comments.
15939 (lra): Init and clear lra_subreg_reload_pseudos. Clear
15940 lra_optional_reload_pseudos after undo transformations.
15941 * lra-assigns.c (pseudo_prefix_title): New.
15942 (lra_setup_reg_renumber): Use it.
15943 (spill_for): Ditto. Check subreg reload pseudos too.
15944 (assign_by_spills): Consider subreg reload pseudos too.
15945 * lra-constraints.c (simplify_operand_subreg): Use
15946 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
15947 (curr_insn_transform): Recognize and do optional reloads.
15948 (undo_optional_reloads): New.
15949 (lra_undo_inheritance): Call undo_optional_reloads.
15951 2013-07-05 Thomas Quinot <quinot@adacore.com>
15953 * tree-complex.c (expand_complex_operations_1): Fix typo.
15955 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
15957 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
15958 (tune_params): New member 'const vec_costs'.
15959 * config/aarch64/aarch64.c (generic_vector_cost): New.
15960 (generic_tunings): New member 'generic_vector_cost'.
15961 (aarch64_builtin_vectorization_cost): New.
15962 (aarch64_add_stmt_cost): New.
15963 (TARGET_VECTORIZE_ADD_STMT_COST): New.
15964 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
15966 2013-07-03 Jakub Jelinek <jakub@redhat.com>
15969 * config/i386/predicates.md (vsib_address_operand): Disallow
15970 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
15972 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
15974 PR middle-end/55030
15975 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
15976 expand_builtin_setjmp_receiver.
15977 (expand_label): Adjust, call expand_builtin_setjmp_receiver
15978 with NULL for the label parameter.
15979 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
15980 the frame-pointer. Adjust comments.
15981 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
15982 only if LABEL is non-NULL.
15984 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
15986 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
15987 (ARM_ABI_AAPCS64): Ditto.
15989 (ARM_DEFAULT_ABI): Ditto.
15991 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
15993 * config/aarch64/aarch64-builtins.c
15994 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
15995 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
15997 * config/aarch64/aarch64-simd.md
15998 (aarch64_ld1<VALL:mode>): New.
15999 (aarch64_st1<VALL:mode>): Likewise.
16000 * config/aarch64/arm_neon.h
16001 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
16003 2013-07-02 Sriraman Tallam <tmsriram@google.com>
16005 * config/i386/i386.c (gate_insert_vzeroupper): Check if
16007 (ix86_option_override_internal):Turn on all -mavx target flags by
16008 default as they are dependent on AVX anyway.
16010 2013-07-02 Cary Coutant <ccoutant@google.com>
16012 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
16013 deterministic hash.
16014 (loc_checksum_ordered): Likewise.
16015 (hash_loc_operands): Remove inline keyword.
16017 2013-07-02 Jakub Jelinek <jakub@redhat.com>
16019 PR tree-optimization/57741
16020 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
16021 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
16022 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
16023 Allow REAL_CST step_exprs if flag_associative_math.
16024 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
16026 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16028 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
16030 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16032 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
16034 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16036 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
16038 (iorsi3_insn): Likewise.
16039 (arm_xorsi3): Likewise.
16041 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
16043 * arm.md (attribute "wtype"): Delete. Move attribute values from here
16045 (attribute "type"): ... here, and prefix with "wmmx_".
16046 (attribute "core_cycles"): Update for attribute changes.
16047 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
16048 (tbcstv4hi): Likewise.
16049 (tbcstv2si): Likewise.
16050 (iwmmxt_iordi3): Likewise.
16051 (iwmmxt_xordi3): Likewise.
16052 (iwmmxt_anddi3): Likewise.
16053 (iwmmxt_nanddi3): Likewise.
16054 (iwmmxt_arm_movdi): Likewise.
16055 (iwmmxt_movsi_insn): Likewise.
16056 (mov<mode>_internal): Likewise.
16057 (and<mode>3_iwmmxt): Likewise.
16058 (ior<mode>3_iwmmxt): Likewise.
16059 (xor<mode>3_iwmmxt): Likewise.
16060 (add<mode>3_iwmmxt): Likewise.
16061 (ssaddv8qi3): Likewise.
16062 (ssaddv4hi3): Likewise.
16063 (ssaddv2si3): Likewise.
16064 (usaddv8qi3): Likewise.
16065 (usaddv4hi3): Likewise.
16066 (usaddv2si3): Likewise.
16067 (sub<mode>3_iwmmxt): Likewise.
16068 (sssubv8qi3): Likewise.
16069 (sssubv4hi3): Likewise.
16070 (sssubv2si3): Likewise.
16071 (ussubv8qi3): Likewise.
16072 (ussubv4hi3): Likewise.
16073 (ussubv2si3): Likewise.
16074 (mulv4hi3_iwmmxt): Likewise.
16075 (smulv4hi3_highpart): Likewise.
16076 (umulv4hi3_highpart): Likewise.
16077 (iwmmxt_wmacs): Likewise.
16078 (iwmmxt_wmacsz): Likewise.
16079 (iwmmxt_wmacu): Likewise.
16080 (iwmmxt_wmacuz): Likewise.
16081 (iwmmxt_clrdi): Likewise.
16082 (iwmmxt_clrv8qi): Likewise.
16083 (iwmmxt_clr4hi): Likewise.
16084 (iwmmxt_clr2si): Likewise.
16085 (iwmmxt_uavgrndv8qi3): Likewise.
16086 (iwmmxt_uavgrndv4hi3): Likewise.
16087 (iwmmxt_uavgv8qi3): Likewise.
16088 (iwmmxt_uavgv4hi3): Likewise.
16089 (iwmmxt_tinsrb): Likewise.
16090 (iwmmxt_tinsrh): Likewise.
16091 (iwmmxt_tinsrw): Likewise.
16092 (iwmmxt_textrmub): Likewise.
16093 (iwmmxt_textrmsb): Likewise.
16094 (iwmmxt_textrmuh): Likewise.
16095 (iwmmxt_textrmsh): Likewise.
16096 (iwmmxt_textrmw): Likewise.
16097 (iwmxxt_wshufh): Likewise.
16098 (eqv8qi3): Likewise.
16099 (eqv4hi3): Likewise.
16100 (eqv2si3): Likewise.
16101 (gtuv8qi3): Likewise.
16102 (gtuv4hi3): Likewise.
16103 (gtuv2si3): Likewise.
16104 (gtv8qi3): Likewise.
16105 (gtv4hi3): Likewise.
16106 (gtv2si3): Likewise.
16107 (smax<mode>3_iwmmxt): Likewise.
16108 (umax<mode>3_iwmmxt): Likewise.
16109 (smin<mode>3_iwmmxt): Likewise.
16110 (umin<mode>3_iwmmxt): Likewise.
16111 (iwmmxt_wpackhss): Likewise.
16112 (iwmmxt_wpackwss): Likewise.
16113 (iwmmxt_wpackdss): Likewise.
16114 (iwmmxt_wpackhus): Likewise.
16115 (iwmmxt_wpackwus): Likewise.
16116 (iwmmxt_wpackdus): Likewise.
16117 (iwmmxt_wunpckihb): Likewise.
16118 (iwmmxt_wunpckihh): Likewise.
16119 (iwmmxt_wunpckihw): Likewise.
16120 (iwmmxt_wunpckilb): Likewise.
16121 (iwmmxt_wunpckilh): Likewise.
16122 (iwmmxt_wunpckilw): Likewise.
16123 (iwmmxt_wunpckehub): Likewise.
16124 (iwmmxt_wunpckehuh): Likewise.
16125 (iwmmxt_wunpckehuw): Likewise.
16126 (iwmmxt_wunpckehsb): Likewise.
16127 (iwmmxt_wunpckehsh): Likewise.
16128 (iwmmxt_wunpckehsw): Likewise.
16129 (iwmmxt_wunpckelub): Likewise.
16130 (iwmmxt_wunpckeluh): Likewise.
16131 (iwmmxt_wunpckeluw): Likewise.
16132 (iwmmxt_wunpckelsb): Likewise.
16133 (iwmmxt_wunpckelsh): Likewise.
16134 (iwmmxt_wunpckelsw): Likewise.
16135 (ror<mode>3): Likewise.
16136 (ashr<mode>3_iwmmxt): Likewise.
16137 (lshr<mode>3_iwmmxt): Likewise.
16138 (ashl<mode>3_iwmmxt): Likewise.
16139 (ror<mode>3_di): Likewise.
16140 (ashr<mode>3_di): Likewise.
16141 (lshr<mode>3_di): Likewise.
16142 (ashl<mode>3_di): Likewise.
16143 (iwmmxt_wmadds): Likewise.
16144 (iwmmxt_wmaddu): Likewise.
16145 (iwmmxt_tmia): Likewise.
16146 (iwmmxt_tmiaph): Likewise.
16147 (iwmmxt_tmiabb): Likewise.
16148 (iwmmxt_tmiatb): Likewise.
16149 (iwmmxt_tmiabt): Likewise.
16150 (iwmmxt_tmiatt): Likewise.
16151 (iwmmxt_tmovmskb): Likewise.
16152 (iwmmxt_tmovmskh): Likewise.
16153 (iwmmxt_tmovmskw): Likewise.
16154 (iwmmxt_waccb): Likewise.
16155 (iwmmxt_wacch): Likewise.
16156 (iwmmxt_waccw): Likewise.
16157 (iwmmxt_waligni): Likewise.
16158 (iwmmxt_walignr): Likewise.
16159 (iwmmxt_walignr0): Likewise.
16160 (iwmmxt_walignr1): Likewise.
16161 (iwmmxt_walignr2): Likewise.
16162 (iwmmxt_walignr3): Likewise.
16163 (iwmmxt_wsadb): Likewise.
16164 (iwmmxt_wsadh): Likewise.
16165 (iwmmxt_wsadbz): Likewise.
16166 (iwmmxt_wsadhz): Likewise.
16167 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
16168 (iwmmxt_wabsdiffb): Likewise.
16169 (iwmmxt_wabsdiffh): Likewise.
16170 (iwmmxt_wabsdiffw): Likewise.
16171 (iwmmxt_waddsubhx): Likewise
16172 (iwmmxt_wsubaddhx): Likewise.
16173 (addc<mode>3): Likewise.
16174 (iwmmxt_avg4): Likewise.
16175 (iwmmxt_avg4r): Likewise.
16176 (iwmmxt_wmaddsx): Likewise.
16177 (iwmmxt_wmaddux): Likewise.
16178 (iwmmxt_wmaddsn): Likewise.
16179 (iwmmxt_wmaddun): Likewise.
16180 (iwmmxt_wmulwsm): Likewise.
16181 (iwmmxt_wmulwum): Likewise.
16182 (iwmmxt_wmulsmr): Likewise.
16183 (iwmmxt_wmulumr): Likewise.
16184 (iwmmxt_wmulwsmr): Likewise.
16185 (iwmmxt_wmulwumr): Likewise.
16186 (iwmmxt_wmulwl): Likewise.
16187 (iwmmxt_wqmulm): Likewise.
16188 (iwmmxt_wqmulwm): Likewise.
16189 (iwmmxt_wqmulmr): Likewise.
16190 (iwmmxt_wqmulwmr): Likewise.
16191 (iwmmxt_waddbhusm): Likewise.
16192 (iwmmxt_waddbhusl): Likewise.
16193 (iwmmxt_wqmiabb): Likewise.
16194 (iwmmxt_wqmiabt): Likewise.
16195 (iwmmxt_wqmiatb): Likewise.
16196 (iwmmxt_wqmiatt): Likewise.
16197 (iwmmxt_wqmiabbn): Likewise.
16198 (iwmmxt_wqmiabtn): Likewise.
16199 (iwmmxt_wqmiatbn): Likewise.
16200 (iwmmxt_wqmiattn): Likewise.
16201 (iwmmxt_wmiabb): Likewise.
16202 (iwmmxt_wmiabt): Likewise.
16203 (iwmmxt_wmiatb): Likewise.
16204 (iwmmxt_wmiatt): Likewise.
16205 (iwmmxt_wmiabbn): Likewise.
16206 (iwmmxt_wmiabtn): Likewise.
16207 (iwmmxt_wmiatbn): Likewise.
16208 (iwmmxt_wmiattn): Likewise.
16209 (iwmmxt_wmiawbb): Likewise.
16210 (iwmmxt_wmiawbt): Likewise.
16211 (iwmmxt_wmiawtb): Likewise.
16212 (iwmmxt_wmiawtt): Likewise.
16213 (iwmmxt_wmiawbbn): Likewise.
16214 (iwmmxt_wmiawbtn): Likewise.
16215 (iwmmxt_wmiawtbn): Likewise.
16216 (iwmmxt_wmiawttn): Likewise.
16217 (iwmmxt_wmerge): Likewise.
16218 (iwmmxt_tandc<mode>3): Likewise.
16219 (iwmmxt_torc<mode>3): Likewise.
16220 (iwmmxt_torvsc<mode>3): Likewise.
16221 (iwmmxt_textrc<mode>3): Likewise.
16222 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
16223 (wmmxt_pack): Likewise.
16224 (wmmxt_mult_c1): Likewise.
16225 (wmmxt_mult_c2): Likewise.
16226 (wmmxt_alu_c1): Likewise.
16227 (wmmxt_alu_c2): Likewise.
16228 (wmmxt_alu_c3): Likewise.
16229 (wmmxt_transfer_c1): Likewise.
16230 (wmmxt_transfer_c2): Likewise.
16231 (wmmxt_transfer_c3): Likewise.
16232 (marvell_f_iwmmxt_wstr): Likewise.
16233 (marvell_f_iwmmxt_wldr): Likewise.
16235 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
16237 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
16239 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16242 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16243 * lra-constraints.c (need_for_split_p): Check call used hard regs
16244 living through calls.
16246 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
16247 call used regs for call insn.
16249 2013-06-28 Jakub Jelinek <jakub@redhat.com>
16252 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
16253 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
16255 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16257 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
16259 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16261 * lra-constraints.c (need_for_split_p): Check call used hard regs
16262 living through calls.
16264 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
16267 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
16268 to tie with any other modes. Eliminate Altivec vector mode tests,
16269 since these are a subset of ALTIVEC or VSX vector modes. Simplify
16270 code, to return 0 if testing MODE2 for a condition, if we've
16271 already tested MODE1 for the same condition.
16273 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16275 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
16278 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16280 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
16281 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
16283 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16285 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
16287 (aarch64_symbolic_constant_p): Remove.
16288 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
16289 static. Fix line length and white space.
16290 (aarch64_symbolic_constant_p): Remove.
16291 * config/aarch64/predicates.md (aarch64_valid_symref):
16292 Use aarch64_classify_symbol_expression.
16294 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16296 * config/arm/constraints.md (Ts): New constraint.
16297 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
16299 (compare_scc): Use "Ts" constraint for operand 0.
16300 (ior_scc_scc): Likewise.
16301 (and_scc_scc): Likewise.
16302 (and_scc_scc_nodom): Likewise.
16303 (ior_scc_scc_cmp): Likewise for operand 7.
16304 (and_scc_scc_cmp): Likewise.
16305 * config/arm/thumb2.md (thumb2_movsi_insn):
16306 Add alternatives for 16-bit encodings.
16307 (thumb2_movhi_insn): Likewise.
16308 (thumb2_movsicc_insn): Likewise.
16309 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
16310 (thumb2_negscc): Use "Ts" constraint.
16311 Move mvn instruction outside cond_exec block.
16312 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
16313 for 16-bit encodings.
16315 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16317 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
16319 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
16320 (mulsi3subsi): Likewise.
16321 (mulsidi3adddi): Likewise.
16322 (mulsidi3_v6): Likewise.
16323 (umulsidi3_v6): Likewise.
16324 (umulsidi3adddi_v6): Likewise.
16325 (smulsi3_highpart_v6): Likewise.
16326 (umulsi3_highpart_v6): Likewise.
16327 (mulhisi3tb): Likewise.
16328 (mulhisi3bt): Likewise.
16329 (mulhisi3tt): Likewise.
16330 (maddhisi4): Likewise.
16331 (maddhisi4tb): Likewise.
16332 (maddhisi4tt): Likewise.
16333 (maddhidi4): Likewise.
16334 (maddhidi4tb): Likewise.
16335 (maddhidi4tt): Likewise.
16336 (zeroextractsi_compare0_scratch): Likewise.
16337 (insv_zero): Likewise.
16338 (insv_t2): Likewise.
16339 (anddi_notzesidi_di): Likewise.
16340 (anddi_notsesidi_di): Likewise.
16341 (andsi_notsi_si): Likewise.
16342 (iordi_zesidi_di): Likewise.
16343 (xordi_zesidi_di): Likewise.
16344 (andsi_iorsi3_notsi): Likewise.
16345 (smax_0): Likewise.
16346 (smax_m1): Likewise.
16347 (smin_0): Likewise.
16348 (not_shiftsi): Likewise.
16349 (unaligned_loadsi): Likewise.
16350 (unaligned_loadhis): Likewise.
16351 (unaligned_loadhiu): Likewise.
16352 (unaligned_storesi): Likewise.
16353 (unaligned_storehi): Likewise.
16354 (extv_reg): Likewise.
16355 (extzv_t2): Likewise.
16356 (divsi3): Likewise.
16357 (udivsi3): Likewise.
16358 (arm_zero_extendhisi2addsi): Likewise.
16359 (arm_zero_extendqisi2addsi): Likewise.
16360 (compareqi_eq0): Likewise.
16361 (arm_extendhisi2_v6): Likewise.
16362 (arm_extendqisi2addsi): Likewise.
16363 (arm_movt): Likewise.
16364 (thumb2_ldrd): Likewise.
16365 (thumb2_ldrd_base): Likewise.
16366 (thumb2_ldrd_base_neg): Likewise.
16367 (thumb2_strd): Likewise.
16368 (thumb2_strd_base): Likewise.
16369 (thumb2_strd_base_neg): Likewise.
16370 (arm_negsi2): Add alternative for 16-bit encoding.
16371 (arm_one_cmplsi2): Likewise.
16373 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16375 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
16376 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
16377 (movdfcc): Likewise.
16378 * config/arm/vfp.md (*thumb2_movsf_vfp):
16379 Disable predication for arm_restrict_it.
16380 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
16381 (*thumb2_movdfcc_vfp): Likewise.
16382 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
16383 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
16384 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
16385 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
16386 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
16387 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
16388 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
16389 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
16390 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
16391 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
16392 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
16393 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
16394 Disable predication for arm_restrict_it.
16396 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
16398 * config/i386/bmiintrin.h (_bextr_u32): New.
16399 (_bextr_u64): Ditto.
16401 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
16403 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
16404 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
16405 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
16406 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
16407 * config/mips/n32-elf.h: ...this new file.
16409 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
16412 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
16413 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
16415 2013-06-27 Catherine Moore <clm@codesourcery.com>
16417 * config/mips/mips-tables.opt: Regenerate.
16418 * config/mips/mips-cpus.def: Add m14ke and m14kec.
16419 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
16420 * doc/invoke.texi: Add -m14kc.
16422 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16425 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
16426 constraints of operand 1 and 2.
16429 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
16430 to match RTL canonicalization. Swap predicates and
16431 constraints of operand 1 and 2.
16433 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
16435 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
16436 Process OP_INOUT regs for splitting too.
16438 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16440 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
16441 decl before the loop, initialize to NULL.
16442 (vectorizable_load): Initialize ptr_incr to NULL.
16444 2013-06-27 Martin Jambor <mjambor@suse.cz>
16447 * ipa-ref.h (ipa_maybe_record_reference): Declare.
16448 * ipa-ref.c (ipa_maybe_record_reference): New function.
16449 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
16450 * ipa-cp.c (create_specialized_node): Record potential references from
16452 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
16454 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16456 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
16457 parameter 'mode' of type 'enum machine_mode mode'; change to pass
16458 'mode' to force_reg.
16459 (aarch64_add_offset): Update calls to aarch64_force_temporary.
16460 (aarch64_expand_mov_immediate): Likewise.
16462 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16464 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
16465 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
16467 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16469 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
16470 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16471 (struct machine_function): Add tbegin_p.
16472 (s390_canonicalize_comparison): Fold CC mode compares to
16473 conditional jump if possible.
16474 (s390_emit_jump): Return the emitted jump.
16475 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
16476 Handle CCRAWmode compares.
16477 (s390_option_override): Default to -mhtm if available.
16478 (s390_reg_clobbered_rtx): Handle floating point regs as well.
16479 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
16480 FPRs instead of df_regs_ever_live_p.
16481 (s390_optimize_nonescaping_tx): New function.
16482 (s390_init_frame_layout): Extend clobbered_regs array to cover
16484 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
16485 (s390_expand_tbegin): New function.
16486 (enum s390_builtin): New enum definition.
16487 (code_for_builtin): New array definition.
16488 (s390_init_builtins): New function.
16489 (s390_expand_builtin): New function.
16490 (TARGET_INIT_BUILTINS): Define.
16491 (TARGET_EXPAND_BUILTIN): Define.
16492 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
16493 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
16494 (s390_alc_comparison): Likewise.
16495 * config/s390/s390-modes.def: Add CCRAWmode.
16496 * config/s390/s390.h (processor_flags): Add PF_TX.
16497 (TARGET_CPU_HTM): Define macro.
16498 (TARGET_HTM): Define macro.
16499 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
16500 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
16501 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16502 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
16503 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
16504 (TBEGIN_MASK, TBEGINC_MASK): New constants.
16505 ("*cc_to_int"): Move up.
16506 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
16507 constants other than 0.
16508 ("*ccraw_to_int"): New insn and splitter definition.
16509 ("tbegin", "tbegin_nofloat", "tbegin_retry")
16510 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
16511 ("tx_assist"): New expander.
16512 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
16513 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
16514 * config/s390/s390.opt: Add -mhtm option.
16515 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
16516 * config/s390/htmxlintrin.h: New file.
16517 * config/s390/htmintrin.h: New file.
16518 * config/s390/s390intrin.h: New file.
16519 * doc/extend.texi: Document htm builtins.
16520 * config.gcc: Add the new header files to extra_headers.
16522 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
16524 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
16525 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
16527 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
16528 Pat Haugen <pthaugen@us.ibm.com>
16529 Peter Bergner <bergner@vnet.ibm.com>
16531 * config/rs6000/power8.md: New.
16532 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
16533 setting for power8 entry.
16534 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
16535 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
16536 test for Power4/Power5 only.
16537 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
16539 (force_new_group): Adjust comment.
16540 * config/rs6000/rs6000.md: Include power8.md.
16542 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
16544 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
16545 * config/arm/arm-protos.h (arm_max_conditional_execute): New
16547 (tune_params): Update comment.
16548 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
16549 (arm_max_conditional_execute): New function.
16550 (thumb2_final_prescan_insn): Use max_insn_skipped and
16551 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
16553 2013-06-25 Jakub Jelinek <jakub@redhat.com>
16555 PR tree-optimization/57705
16556 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
16557 SSA_NAME step, provided that it is not defined inside the loop.
16558 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
16559 (get_initial_def_for_induction): Handle SSA_NAME IV step.
16561 2013-06-25 Martin Jambor <mjambor@suse.cz>
16563 PR middle-end/57670
16564 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
16565 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
16567 (ipa_note_param_call): Initialize member_ptr flag.
16568 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
16569 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
16570 (ipa_write_indirect_edge_info): Stream member_ptr flag.
16571 (ipa_read_indirect_edge_info): Likewise.
16573 2013-06-25 Richard Biener <rguenther@suse.de>
16575 PR middle-end/56977
16576 * passes.c (init_optimization_passes): Move pass_fold_builtins
16577 and pass_dce earlier with -Og.
16579 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
16581 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
16582 <BIT_FIELD_REF>: Remove trailing TAB.
16583 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
16586 2013-06-24 Martin Jambor <mjambor@suse.cz>
16588 PR tree-optimization/57358
16589 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
16590 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
16591 (ipa_analyze_params_uses): Generate pessimistic info when true.
16593 2013-06-24 Martin Jambor <mjambor@suse.cz>
16595 PR tree-optimization/57539
16596 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
16597 global.inlined_to of the new node to it. All callers changed.
16598 * ipa-inline-transform.c (clone_inlined_nodes): New variable
16599 inlining_into, pass it to cgraph_clone_node.
16600 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
16601 ipa_free_edge_args_substructures.
16602 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
16603 rdesc linked list. Do not assert rdesc edges have inlined caller.
16604 Assert we have found an rdesc in the rdesc list.
16606 2013-06-24 Richard Biener <rguenther@suse.de>
16608 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
16609 (pointer_set_lookup): Declare.
16610 (class pointer_map): New template class implementing a
16611 generic pointer to T map.
16612 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
16613 pointer_map<T>::contains, pointer_map<T>::insert,
16614 pointer_map<T>::traverse): New functions.
16615 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
16616 (pointer_set_lookup): New function.
16617 (pointer_set_contains): Use pointer_set_lookup.
16618 (pointer_set_insert): Likewise.
16619 (insert_aux): Remove.
16620 (struct pointer_map_t): Embed a pointer_set_t.
16621 (pointer_map_create): Adjust.
16622 (pointer_map_destroy): Likewise.
16623 (pointer_map_contains): Likewise.
16624 (pointer_map_insert): Likewise.
16625 (pointer_map_traverse): Likewise.
16626 * tree-streamer.h (struct streamer_tree_cache_d): Use a
16627 pointer_map<unsigned> instead of a pointer_map_t.
16628 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
16629 (streamer_tree_cache_lookup): Likewise.
16630 (streamer_tree_cache_create): Likewise.
16631 (streamer_tree_cache_delete): Likewise.
16632 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
16633 pointer_map<unsigned> instead of a pointer_map_t.
16634 (lto_init_tree_ref_encoder): Adjust.
16635 (lto_destroy_tree_ref_encoder): Likewise.
16636 * lto-section-out.c (lto_output_decl_index): Likewise.
16637 (lto_record_function_out_decl_state): Likewise.
16638 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
16640 2013-06-24 Richard Biener <rguenther@suse.de>
16642 PR tree-optimization/57488
16643 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
16645 2013-06-24 Alan Modra <amodra@gmail.com>
16647 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
16648 (gen_easy_altivec_constant): Likewise.
16649 * config/rs6000/predicates.md (easy_vector_constant_add_self,
16650 easy_vector_constant_msb): Likewise.
16652 2013-06-23 Jakub Jelinek <jakub@redhat.com>
16655 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
16656 add missing return true.
16658 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
16661 * config/sh/predicates.md (general_extend_operand): Invoke
16662 general_movsrc_operand for memory operands.
16663 (general_movsrc_operand): Allow reg+reg addressing, do not use
16664 general_operand for memory operands.
16666 2013-06-23 Sriraman Tallam <tmsriram@google.com>
16668 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
16669 when current target options does not apply.
16670 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
16671 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
16672 * config/i386/bmiintrin.h: Pass appropriate target
16673 attributes to header.
16674 * config/i386/mmintrin.h: Ditto.
16675 * config/i386/nmmintrin.h: Ditto.
16676 * config/i386/avx2intrin.h: Ditto.
16677 * config/i386/fxsrintrin.h: Ditto.
16678 * config/i386/tbmintrin.h: Ditto.
16679 * config/i386/xsaveintrin.h: Ditto.
16680 * config/i386/f16cintrin.h: Ditto.
16681 * config/i386/xtestintrin.h: Ditto.
16682 * config/i386/xsaveoptintrin.h: Ditto.
16683 * config/i386/bmi2intrin.h: Ditto.
16684 * config/i386/lzcntintrin.h: Ditto.
16685 * config/i386/smmintrin.h: Ditto.
16686 * config/i386/wmmintrin.h: Ditto.
16687 * config/i386/x86intrin.h: Remove all header include guards.
16688 * config/i386/prfchwintrin.h: Ditto.
16689 * config/i386/pmmintrin.h: Ditto.
16690 * config/i386/tmmintrin.h: Ditto.
16691 * config/i386/xmmintrin.h: Ditto.
16692 * config/i386/popcntintrin.h: Ditto.
16693 * config/i386/rdseedintrin.h: Ditto.
16694 * config/i386/ammintrin.h: Ditto.
16695 * config/i386/emmintrin.h: Ditto.
16696 * config/i386/immintrin.h: Remove all header include guards.
16697 * config/i386/fma4intrin.h: Ditto.
16698 * config/i386/lwpintrin.h: Ditto.
16699 * config/i386/xopintrin.h: Ditto.
16700 * config/i386/ia32intrin.h: Ditto.
16701 * config/i386/avxintrin.h: Ditto.
16702 * config/i386/rtmintrin.h: Ditto.
16703 * config/i386/fmaintrin.h: Ditto.
16704 * config/i386/mm3dnow.h: Ditto.
16706 2013-06-22 Sriraman Tallam <tmsriram@google.com>
16708 * common/config/i386/i386-common.c: Handle LZCNT.
16710 2013-06-22 Andi Kleen <ak@linux.intel.com>
16712 * doc/extend.texi: Use __atomic_store_n instead of
16713 __atomic_store in HLE example.
16715 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
16717 * config/sh/sh.c: Remove <cstdlib> workaround.
16719 2013-06-21 Andi Kleen <ak@linux.intel.com>
16721 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
16723 2013-06-21 Andi Kleen <ak@linux.intel.com>
16725 * doc/extend.texi: Document that __atomic_clear and
16726 __atomic_test_and_set should only be used with bool.
16728 2013-06-20 Jan Hubicka <jh@suse.cz>
16730 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
16731 types_same_for_odr.
16732 * tree.c (decls_same_for_odr): New function.
16733 (same_for_edr): New function.
16734 (types_same_for_odr): New function.
16735 (get_binfo_at_offset): Use it.
16736 * tree.h (types_same_for_odr): Declare.
16738 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
16739 Jason Merrill <jason@redhat.com>
16741 * system.h: Include <cstdlib> as well as <stdlib.h>.
16743 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
16746 * config/i386/i386.c (construct_container): Report error if
16747 long double is used with disabled x87 float returns.
16749 2013-06-20 Jan Hubicka <jh@suse.cz>
16751 * lto-cgraph.c (input_symtab): Do not set cgraph state.
16753 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
16755 PR rtl-optimization/57425
16756 PR rtl-optimization/57569
16757 * alias.c (write_dependence_p): Remove parameters mem_mode and
16758 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
16759 Changed all callers.
16760 (canon_anti_dependence): Get comments and semantics in sync.
16761 Add parameter mem_canonicalized. Changed all callers.
16762 * rtl.h (canon_anti_dependence): Update prototype.
16764 2013-06-20 Richard Biener <rguenther@suse.de>
16766 * data-streamer-in.c (streamer_read_uhwi): Optimize single
16767 byte case, inline streamer_read_uchar and defer section
16770 2013-06-20 Richard Biener <rguenther@suse.de>
16772 PR tree-optimization/57584
16773 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
16774 SSA names into the expanded expression that take part in
16775 abnormal coalescing.
16777 2013-06-19 Sharad Singhai <singhai@google.com>
16779 * gcov.c (print_usage): Handle new option.
16780 (process_args): Ditto.
16781 (get_gcov_intermediate_filename): New function.
16782 (output_intermediate_file): New function.
16783 (output_gcov_file): New function
16784 (generate_results): Handle new option.
16785 (release_function): Relase demangled name.
16786 (read_graph_file): Handle demangled name.
16787 (output_lines): Ditto.
16788 * doc/gcov.texi: Document gcov intermediate format.
16790 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
16793 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
16794 (lra_emit_add): Use the functions. Add comment about Y as an
16797 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
16800 * collect2.c (collect_atexit): New.
16801 (collect_exit): Delete.
16802 (main): Register collect_atexit with atexit.
16803 (collect_wait): Change collect_exit to exit.
16805 * collect2.h (collect_exit): Delete.
16806 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
16808 2013-06-19 Wei Mi <wmi@google.com>
16810 PR rtl-optimization/57518
16811 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
16812 if regno is used in paradoxical subreg.
16813 (update_equiv_regs): Check pdx_subregs[regno] before
16814 set a reg to be equivalent with a mem.
16816 2013-06-19 Matthias Klose <doko@ubuntu.com>
16819 * file-find.h (find_a_file): Add a mode parameter.
16820 * file-find.c (find_a_file): Likewise.
16821 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
16822 with X_OK for the executables.
16823 * collect2.c (main): Call find_a_file with X_OK.
16825 2013-06-19 Steve Ellcey <sellcey@mips.com>
16828 * config/mips/mips.md (casesi_internal_mips16_<mode>):
16829 Use NEXT_INSN instead of next_real_insn.
16831 2013-06-19 Jan Hubicka <jh@suse.cz>
16833 * cgraph.h (const_value_known_p): Replace by ...
16834 (ctor_for_folding): .. this one.
16835 * cgraphunit.c (process_function_and_variable_attributes): Use it.
16836 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
16837 * expr.c (expand_expr_real_1): Likewise.
16838 (string_constant): Likewise.
16839 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
16840 * ipa.c (process_references): Likewise.
16841 (symtab_remove_unreachable_nodes): Likewise.
16842 * ipa-inline-analysis.c (param_change_prob): Likewise.
16843 * gimple-fold.c (canonicalize_constructor_val): Likewise.
16844 (get_base_constructor): Likwise.
16845 * varpool.c (varpool_remove_node): Likewise.
16846 (varpool_remove_initializer): LIkewise.
16847 (dump_varpool_node): LIkwise.
16848 (const_value_known_p): Rewrite to ...
16849 (ctor_for_folding): ... this one.
16851 2013-06-19 Jakub Jelinek <jakub@redhat.com>
16854 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
16855 PERSONALITY in $PATH derived prefixes.
16857 2013-06-19 Jeff Law <law@redhat.com>
16859 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
16862 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
16863 (simplify_bitwise_binary): Use it to simpify certain binary ops on
16866 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
16868 * config/arm/vfp.md: Move VFP instruction classification documentation
16870 * config/arm/arm.md: ... here. Update instruction classification
16873 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
16875 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
16876 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
16877 pattern. Use more efficient sequences on ARMv5 and Thumb2.
16879 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
16882 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16883 with NEXT_INSN. Use tablejump_p to check for jump table data
16886 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
16889 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
16890 that now in C++ the value is correct per the C++ standards.
16892 2013-06-19 Richard Biener <rguenther@suse.de>
16894 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
16895 for global context.
16897 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16900 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16903 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16904 with next_active_insn.
16906 2013-06-18 Sriraman Tallam <tmsriram@google.com>
16908 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
16909 functions are inlined during failures to flag an error.
16910 * tree-inline.c (expand_call_inline): Allow the error to be flagged
16911 in early inline pass.
16913 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
16915 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
16918 2013-06-18 Julian Brown <julian@codesourcery.com>
16920 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
16921 Permit virtual register pre-reload if !strict.
16922 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
16924 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
16926 * config/arm/neon.md (movmisalign<mode>): Use
16927 neon_perm_struct_or_reg_operand instead of
16928 neon_struct_or_register_operand.
16929 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
16930 neon_permissive_struct_operand instead of neon_struct_operand.
16931 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
16932 neon_vector_mem_operand.
16933 * config/arm/predicates.md (neon_struct_operand): Adjust call to
16934 neon_vector_mem_operand.
16935 (neon_permissive_struct_operand): New.
16936 (neon_struct_or_register_operand): Rename to...
16937 (neon_perm_struct_or_reg_operand): This. Adjust call to
16938 neon_vector_mem_operand.
16940 2013-06-18 Richard Biener <rguenther@suse.de>
16942 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
16943 * lto-streamer.h: Include pointer-set.h.
16944 (struct lto_decl_slot): Remove.
16945 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
16946 Remove next_index entry.
16947 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16948 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16949 (lto_init_tree_ref_encoder): Adjust.
16950 (lto_destroy_tree_ref_encoder): Likewise.
16951 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16952 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16953 (lto_output_decl_index): Adjust.
16954 (lto_new_out_decl_state): Likewise.
16955 (lto_record_function_out_decl_state): Likewise.
16956 * lto-streamer-out.c (copy_function): Likewise.
16958 2013-06-18 Richard Biener <rguenther@suse.de>
16960 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
16961 * cgraphunit.c: Include cfgloop.h.
16962 (init_lowered_empty_function): Initialize the loop tree.
16963 (assemble_thunk): Insert new BBs into loops.
16965 2013-06-18 Richard Biener <rguenther@suse.de>
16967 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
16968 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
16969 the map from cache entry to cache index optional.
16970 (streamer_tree_cache_replace_tree): Adjust accordingly.
16971 (streamer_tree_cache_append): Likewise.
16972 (streamer_tree_cache_delete): Likewise.
16973 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
16974 streamer cache map from cache entry to cache index.
16975 * lto-streamer-out.c (create_output_block): Adjust.
16977 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
16979 * config/arm/arm.md (attribute "insn"): Move multiplication and
16980 division attributes to...
16981 (attribute "type"): ... here. Remove mult.
16982 (attribute "mul32"): New attribute.
16983 (attribute "mul64"): Add umaal.
16984 (*arm_mulsi3): Update attributes.
16985 (*arm_mulsi3_v6): Likewise.
16986 (*thumb_mulsi3): Likewise.
16987 (*thumb_mulsi3_v6): Likewise.
16988 (*mulsi3_compare0): Likewise.
16989 (*mulsi3_compare0_v6): Likewise.
16990 (*mulsi_compare0_scratch): Likewise.
16991 (*mulsi_compare0_scratch_v6): Likewise.
16992 (*mulsi3addsi): Likewise.
16993 (*mulsi3addsi_v6): Likewise.
16994 (*mulsi3addsi_compare0): Likewise.
16995 (*mulsi3addsi_compare0_v6): Likewise.
16996 (*mulsi3addsi_compare0_scratch): Likewise.
16997 (*mulsi3addsi_compare0_scratch_v6): Likewise.
16998 (*mulsi3subsi): Likewise.
16999 (*mulsidi3adddi): Likewise.
17000 (*mulsi3addsi_v6): Likewise.
17001 (*mulsidi3adddi_v6): Likewise.
17002 (*mulsidi3_nov6): Likewise.
17003 (*mulsidi3_v6): Likewise.
17004 (*umulsidi3_nov6): Likewise.
17005 (*umulsidi3_v6): Likewise.
17006 (*umulsidi3adddi): Likewise.
17007 (*umulsidi3adddi_v6): Likewise.
17008 (*smulsi3_highpart_nov6): Likewise.
17009 (*smulsi3_highpart_v6): Likewise.
17010 (*umulsi3_highpart_nov6): Likewise.
17011 (*umulsi3_highpart_v6): Likewise.
17012 (mulhisi3): Likewise.
17013 (*mulhisi3tb): Likewise.
17014 (*mulhisi3bt): Likewise.
17015 (*mulhisi3tt): Likewise.
17016 (maddhisi4): Likewise.
17017 (*maddhisi4tb): Likewise.
17018 (*maddhisi4tt): Likewise.
17019 (maddhidi4): Likewise.
17020 (*maddhidi4tb): Likewise.
17021 (*maddhidi4tt): Likewise.
17022 (divsi3): Likewise.
17023 (udivsi3): Likewise.
17024 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
17025 (thumb2_mulsi_short_compare0): Likewise.
17026 (thumb2_mulsi_short_compare0_scratch): Likewise.
17027 * config/arm/arm1020e.md (1020mult1): Update attribute change.
17028 (1020mult2): Likewise.
17029 (1020mult3): Likewise.
17030 (1020mult4): Likewise.
17031 (1020mult5): Likewise.
17032 (1020mult6): Likewise.
17033 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
17035 (cortex_a15_mult64): Likewise.
17036 (cortex_a15_sdiv): Likewise.
17037 (cortex_a15_udiv): Likewise.
17038 * config/arm/arm1026ejs.md (mult1): Update attribute change.
17044 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
17045 (pj4_ir_div): Likewise.
17046 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
17047 (11_mult2): Likewise.
17048 (11_mult3): Likewise.
17049 (11_mult4): Likewise.
17050 (11_mult5): Likewise.
17051 (11_mult6): Likewise.
17052 (11_mult7): Likewise.
17053 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
17054 (cortex_a8_mla): Likewise.
17055 (cortex_a8_mull): Likewise.
17056 (cortex_a8_smulwy): Likewise.
17057 (cortex_a8_smlald): Likewise.
17058 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
17059 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
17060 (cortex_r4_mul_3): Likewise.
17061 (cortex_r4_mla_4): Likewise.
17062 (cortex_r4_mla_3): Likewise.
17063 (cortex_r4_smlald): Likewise.
17064 (cortex_r4_mull): Likewise.
17065 (cortex_r4_sdiv): Likewise.
17066 (cortex_r4_udiv): Likewise.
17067 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
17068 (cortex_a7_idiv): Likewise.
17069 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
17070 (9_mult2): Likewise.
17071 (9_mult3): Likewise.
17072 (9_mult4): Likewise.
17073 (9_mult5): Likewise.
17074 (9_mult6): Likewise.
17075 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
17076 (cortex_a53_sdiv): Likewise.
17077 (cortex_a53_udiv): Likewise.
17078 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
17079 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
17080 (mp626_mult2): Likewise.
17081 (mp626_mult3): Likewise.
17082 (mp626_mult4): Likewise.
17083 * config/arm/fa526.md (526_mult1): Update attribute change.
17084 (526_mult2): Likewise.
17085 * config/arm/arm-generic.md (mult): Update attribute change.
17086 (mult_ldsched_strongarm): Likewise.
17087 (mult_ldsched): Likewise.
17088 (multi_cycle): Likewise.
17089 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
17090 * config/arm/fa606te.md (606te_mult1): Update attribute change.
17091 (606te_mult2): Likewise.
17092 (606te_mult3): Likewise.
17093 (606te_mult4): Likewise.
17094 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
17095 (cortex_a9_mac16): Likewise.
17096 (cortex_a9_multiply): Likewise.
17097 (cortex_a9_mac): Likewise.
17098 (cortex_a9_multiply_long): Likewise.
17099 * config/arm/fa626te.md (626te_mult1): Update attribute change.
17100 (626te_mult2): Likewise.
17101 (626te_mult3): Likewise.
17102 (626te_mult4): Likewise.
17104 2013-06-18 Richard Biener <rguenther@suse.de>
17107 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
17109 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17112 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17113 with next_active_insn.
17115 2013-06-18 Alan Modra <amodra@gmail.com>
17117 * config/rs6000/rs6000.h (enum data_align): New.
17118 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
17119 (DATA_ABI_ALIGNMENT): Define.
17120 (CONSTANT_ALIGNMENT): Correct comment.
17121 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
17122 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
17124 2013-06-17 David Malcolm <dmalcolm@redhat.com>
17126 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
17127 ATTRIBUTE_UNUSED marking.
17129 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
17131 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
17132 alternative and update.
17133 (aarch64_dup_lanedi): Delete.
17134 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
17135 * config/aarch64/aarch64-simd-builtins.def: Update.
17137 2013-06-17 Richard Biener <rguenther@suse.de>
17139 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
17140 (lto_input_scc): Declare.
17141 (lto_input_tree_1): Likewise.
17142 (struct lto_stats_d): Add num_tree_bodies_output and
17143 num_pickle_refs_output.
17144 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
17145 (lto_read_tree_1): Split out from ...
17146 (lto_read_tree): ... this.
17147 (lto_input_scc): New function.
17148 (lto_input_tree_1): Split out from ...
17149 (lto_input_tree): ... this. Handle LTO_tree_scc.
17150 (lto_data_in_create): Create the streamer cache without hashes.
17151 * lto-streamer-out.c (create_output_block): Create the streamer
17152 cache with hashes when not doing WPA.
17153 (lto_write_tree_1): Split out from ...
17154 (lto_write_tree): ... this.
17155 (get_symbol_initial_value): New function.
17156 (lto_output_tree_1): Split out from ...
17157 (lto_output_tree): ... this. Write trees as series of SCCs
17158 using a DFS walk via DFS_write_tree.
17159 (struct sccs, struct scc_entry): New types.
17160 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
17161 (DFS_write_tree_body): New function.
17162 (DFS_write_tree): Likewise.
17163 (hash_tree): Likewise.
17164 (scc_entry_compare): Likewise.
17165 (hash_scc): Likewise.
17166 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
17167 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
17168 TREE_CHAIN as regular reference.
17169 (streamer_read_integer_cst): Remove.
17170 (streamer_get_pickled_tree): Adjust.
17171 * tree-streamer-out.c (streamer_write_chain): Disable streaming
17172 of DECL_EXTERNALs in BLOCK_VARS for now.
17173 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
17175 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
17176 Add hash value argument and record that if hashes are recorded
17178 (streamer_tree_cache_insert_1): Adjust.
17179 (streamer_tree_cache_insert): Likewise.
17180 (streamer_tree_cache_insert_at): Rename to ...
17181 (streamer_tree_cache_replace_tree): ... this and adjust.
17182 (streamer_tree_cache_append): Adjust.
17183 (record_common_node): Likewise.
17184 (streamer_tree_cache_create): Add argument whether to
17185 record hash values together with trees.
17186 (streamer_tree_cache_delete): Adjust.
17187 * tree-streamer.h (struct streamer_tree_cache_d): Add
17189 (streamer_read_integer_cst): Remove.
17190 (streamer_tree_cache_insert): Adjust.
17191 (streamer_tree_cache_append): Likewise.
17192 (streamer_tree_cache_insert_at): Rename to ...
17193 (streamer_tree_cache_replace_tree): ... this and adjust.
17194 (streamer_tree_cache_create): Add argument whether to record hashes.
17195 (streamer_tree_cache_get): Rename to ...
17196 (streamer_tree_cache_get_tree): ... this.
17197 (streamer_tree_cache_get_hash): New function.
17198 * tree.c (cache_integer_cst): New function.
17199 * tree.h (cache_integer_cst): Declare.
17200 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
17201 * lto-symtab.c (lto_varpool_replace_node): Only release
17202 DECL_INITIAL of non-prevailing decls.
17203 * varpool.c (varpool_remove_initializer): Do not release
17204 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
17206 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17208 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
17209 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
17210 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
17211 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
17212 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
17213 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
17214 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
17215 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
17216 instead of TARGET_64BIT.
17217 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
17218 Require ISA_HAS_<D>DIV.
17220 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
17222 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
17223 (mips*-*-linux*): Move default with_llsc setting to where other
17225 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
17227 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
17228 Likewise. Remove default with_tune setting. Move default float
17229 setting to its own block. Handle with_llsc in the same block as above.
17231 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
17233 PR rtl-optimization/57425
17234 PR rtl-optimization/57569
17235 * alias.c (write_dependence_p): Add new parameters mem_mode,
17236 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
17237 Changed all callers.
17238 (canon_anti_dependence): New function.
17239 * cse.c (check_dependence): Use canon_anti_dependence.
17240 * cselib.c (cselib_invalidate_mem): Likewise.
17241 * rtl.h (canon_anti_dependence): Declare.
17243 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17245 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
17246 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
17247 ".set mips3" for 64-bit targets.
17249 2013-06-15 Dehao Chen <dehao@google.com>
17251 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
17252 * gimple-low.c (gimple_check_call_matching_types): Likewise.
17253 (gimple_check_call_args): Likewise.
17254 * value-prof.c (check_ic_target): Likewise.
17255 * ipa-inline.c (early_inliner): Likewise.
17256 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
17257 * cgraph.c (cgraph_create_edge_1): Likewise.
17258 (cgraph_make_edge_direct): Likewise.
17260 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
17263 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
17264 rs6000_output_move_128bit to handle emitting quad memory
17265 operations. Set attribute length to 8 bytes.
17267 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
17269 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
17271 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
17272 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
17273 (aarch64_<su>mlsl<mode>): Likewise.
17275 2013-06-14 Mike Stump <mikestump@comcast.net>
17277 * Makefile.in (TARGET_H): Add insn-codes.h.
17279 2013-06-14 Alan Modra <amodra@gmail.com>
17281 PR middle-end/57134
17282 PR middle-end/57586
17283 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
17284 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
17285 bitfield expansion when EXPAND_MEMORY.
17286 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
17288 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
17290 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
17291 test for clearing quad memory on 32-bit later.
17293 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17295 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
17296 (fold_negate_expr): Likewise.
17297 (fold_real_zero_addition_p): Handle vectors.
17298 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
17300 2013-06-14 Alan Modra <amodra@gmail.com>
17302 * varasm.c (force_const_mem): Revert 2013-06-07 change.
17304 2013-06-13 Jan Hubicka <jh@suse.cz>
17306 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17307 Local comdats are not externally visible.
17308 * symtab.c (dump_symtab_base): Dump externally visible.
17309 (verify_symtab_base): Verify back links in the symtab hash.
17311 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17313 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
17314 CONVERT_EXPR as equal nodes.
17316 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17318 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
17320 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17322 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
17323 Generalize to complex and vector.
17324 * tree.c (build_all_ones_cst): New function.
17325 * tree.h (build_all_ones_cst): Declare it.
17327 2013-06-13 Alan Modra <amodra@gmail.com>
17329 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
17330 * config/rs6000/rs6000.md (signbittf2): New insn.
17331 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
17332 (abstf2_internal, cmptf_internal2): Likewise.
17333 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
17335 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
17336 Pat Haugen <pthaugen@us.ibm.com>
17337 Peter Bergner <bergner@vnet.ibm.com>
17339 * config/rs6000/rs6000.c (emit_load_locked): Add support for
17340 power8 byte, half-word, and quad-word atomic instructions.
17341 (emit_store_conditional): Likewise.
17342 (rs6000_expand_atomic_compare_and_swap): Likewise.
17343 (rs6000_expand_atomic_op): Likewise.
17345 * config/rs6000/sync.md (larx): Add new modes for power8.
17347 (AINT): New mode iterator to include TImode as well as normal
17348 integer modes on power8.
17349 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
17350 that VSX registers are not considered. Use AINT mode iterator
17351 instead of INT1 to allow inclusion of quad word atomic operations
17353 (load_locked<mode>): Likewise.
17354 (store_conditional<mode>): Likewise.
17355 (atomic_compare_and_swap<mode>): Likewise.
17356 (atomic_exchange<mode>): Likewise.
17357 (atomic_nand<mode>): Likewise.
17358 (atomic_fetch_<fetchop_name><mode>): Likewise.
17359 (atomic_nand_fetch<mode>): Likewise.
17360 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
17362 (ATOMIC): On power8, add QImode, HImode modes.
17363 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
17364 modes that promote to SImode.
17365 (load_lockedti): Convert TImode arguments to PTImode, so that we
17366 get a guaranteed even/odd register pair.
17367 (load_lockedpti): Likewise.
17368 (store_conditionalti): Likewise.
17369 (store_conditionalpti): Likewise.
17371 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
17372 atomic load/store instructions.
17375 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
17377 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
17379 (insn_count): New attribute, with most cases extracted from...
17380 (length): ...here. Redefine most cases in terms of insn_count.
17381 (single_insn): Delete.
17382 (can_delay): Use insn_count to check for single instructions.
17383 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
17384 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
17385 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
17386 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
17387 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
17388 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
17389 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
17390 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
17391 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
17392 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
17393 rather than "length".
17394 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
17395 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
17396 Use "insn_count" rather than "length".
17397 * config/mips/mips-dsp.md
17398 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
17399 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
17402 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
17404 PR tree-optimization/57361
17405 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
17407 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
17409 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
17411 (aarch64_simd_combine<mode>): New instruction expansion.
17412 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
17413 function prototype.
17414 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
17415 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
17417 2013-06-12 Jan Hubicka <jh@suse.cz>
17419 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
17420 decl has when in streaming stage.
17421 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
17422 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
17424 2013-06-12 Roland Stigge <stigge@antcom.de>
17427 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
17429 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17431 PR tree-optimization/57537
17432 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
17433 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
17435 2013-06-12 Richard Biener <rguenther@suse.de>
17437 * data-streamer.h (streamer_write_char_stream): CSE
17438 obs->current_pointer.
17439 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
17440 streamer_write_char_stream manually and optimize the resulting loop.
17441 (streamer_write_hwi_stream): Likewise.
17443 2013-06-12 Jan Hubicka <jh@suse.cz>
17445 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
17446 * cgraph.h (varpool_create_empty_node): Declare.
17447 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
17449 * symtab.c (symtab_unregister_node): Be lax about missin entries
17451 (symtab_get_node): Update comment.
17452 * varpool.c (varpool_create_empty_node): Break out from ...
17453 (varpool_node_for_decl): ... here.
17454 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
17456 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
17458 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
17459 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
17460 part. Use straight-line flow at the end.
17461 <COMPONENT_REF>: Remove superfluous else.
17462 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
17464 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17467 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
17468 target hook even for !TREE_PUBLIC decls. If no resolution info
17469 is available, return false for common and external decls.
17471 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17473 * config/rl78/constraints.md (U): New constraint.
17474 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
17477 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17480 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
17481 to allow returning address to AT_PLATFORM name.
17483 2013-06-11 Jan Hubicka <jh@suse.cz>
17485 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
17486 * cgraph.h (symtab_node_base): Add weakref flag.
17487 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
17488 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
17489 (output_weakrefs): Use weakref flag.
17490 * fold-const.c (simple_operand_p): Handle WEAK.
17491 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
17492 * ipa.c (varpool_externally_visible_p): Drop weakref.
17493 (function_and_variable_visibility): Update comment; fix weakref
17494 sanity checks; do not clear DECL_WEAK on them.
17495 * lto-cgraph.c (lto_output_node): update.
17496 (lto_output_varpool_node): Update.
17497 (input_overwrite_node): Update.
17498 (input_node): Update.
17499 (input_varpool_node): Update.
17500 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
17501 (lto_symtab_merge_symbols): Add sanity check.
17502 (lto_symtab_prevailing_decl): Do not special case weakrefs.
17503 * passes.c (rest_of_decl_compilation): Set static flag, too.
17504 * symtab.c (dump_symtab_base): Dump weakref.
17505 (verify_symtab_base): Sanity check weakrefs.
17506 (symtab_make_decl_local): Remove duplicated code.
17507 (symtab_alias_ultimate_target): Simplify.
17508 * varpool.c (varpool_create_variable_alias): Set weakref flag.
17510 2013-06-11 Tom de Vries <tom@codesourcery.com>
17512 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
17513 sequence_vect == NULL.
17515 2013-06-11 DJ Delorie <dj@redhat.com>
17517 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
17518 (rl78_unwind_word_mode): New.
17520 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17522 * final.c (debug_prefix_maps): Make static.
17524 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17526 * function.c (initial_trampoline): Remove stray copy.
17528 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
17530 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
17532 2013-06-11 Martin Jambor <mjambor@suse.cz>
17534 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
17535 within bounds at the beginning of the function.
17537 2013-06-11 Alan Modra <amodra@gmail.com>
17539 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
17541 (get_named_section): Don't NULL !DECL_P decl.
17543 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
17545 * doc/invoke.texi (core-avx2): Document.
17547 (atom): Updated with MOVBE.
17549 2013-06-11 Richard Biener <rguenther@suse.de>
17551 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
17553 2013-06-11 Anton Blanchard <anton@samba.org>
17555 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
17556 correct shift value in little-endian mode.
17558 2013-06-11 Jakub Jelinek <jakub@redhat.com>
17561 * varasm.c (get_variable_align): Move #endif to the right place.
17563 2013-06-10 Cary Coutant <ccoutant@google.com>
17565 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
17566 for hash so that hash table traversal order is deterministic.
17568 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
17569 Pat Haugen <pthaugen@us.ibm.com>
17570 Peter Bergner <bergner@vnet.ibm.com>
17572 * config/rs6000/vector.md (GPR move splitter): Do not split moves
17573 of vectors in GPRS if they are direct moves or quad word load or
17576 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
17578 (direct_move_p): Likewise.
17579 (quad_load_store_p): Likewise.
17581 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
17582 classes into bins based on the physical register type.
17583 (reg_class_to_reg_type): Likewise.
17584 (IS_STD_REG_TYPE): Likewise.
17585 (IS_FP_VECT_REG_TYPE): Likewise.
17586 (reload_fpr_gpr): Arrays to determine what insn to use if we can
17587 use direct move instructions.
17588 (reload_gpr_vsx): Likewise.
17589 (reload_vsx_gpr): Likewise.
17590 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
17591 information that is a simplification of register classes. Also
17592 precalculate direct move reload helpers.
17593 (direct_move_p): New function to return true if the operation can
17594 be done as a direct move instruciton.
17595 (quad_load_store_p): New function to return true if the operation
17596 is a quad memory operation.
17597 (rs6000_legitimize_address): If quad memory, only allow register
17598 indirect for TImode addresses.
17599 (rs6000_legitimate_address_p): Likewise.
17600 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
17601 (rs6000_reload_register_type): Likewise.
17602 (register_to_reg_type): Return register type.
17603 (rs6000_secondary_reload_simple_move): New helper function for
17604 secondary reload and secondary memory needed to identify anything
17605 that is a simple move, and does not need reloading.
17606 (rs6000_secondary_reload_direct_move): New helper function for
17607 secondary reload to identify cases that can be done with several
17608 instructions via the direct move instructions.
17609 (rs6000_secondary_reload_move): New helper function for secondary
17610 reload to identify moves between register types that can be done.
17611 (rs6000_secondary_reload): Add support for quad memory operations
17612 and for direct move.
17613 (rs6000_secondary_memory_needed): Likewise.
17614 (rs6000_debug_secondary_memory_needed): Change argument names.
17615 (rs6000_output_move_128bit): New function to return the move to
17616 use for 128-bit moves, including knowing about the various
17617 limitations of quad memory operations.
17619 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
17620 memory operations. call rs6000_output_move_128bit for the actual
17621 instruciton(s) to generate.
17622 (vsx_movti_64bit): Likewise.
17624 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
17625 (UNSPEC_P8V_MTVSRWZ): Likewise.
17626 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
17627 (UNSPEC_P8V_MTVSRD): Likewise.
17628 (UNSPEC_P8V_XXPERMDI): Likewise.
17629 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
17630 (UNSPEC_FUSION_GPR): Likewise.
17631 (FMOVE128_GPR): New iterator for direct move.
17632 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
17633 (f32_sv): Likewise.
17634 (f32_dm): Likewise.
17635 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
17636 loads and direct move instructions.
17637 (zero_extendsidi2_lfiwzx): Likewise.
17638 (extendsidi2_lfiwax): Likewise.
17639 (extendsidi2_nocell): Likewise.
17640 (floatsi<mode>2_lfiwax): Likewise.
17641 (lfiwax): Likewise.
17642 (floatunssi<mode>2_lfiwzx): Likewise.
17643 (lfiwzx): Likewise.
17644 (fix_trunc<mode>_stfiwx): Likewise.
17645 (fixuns_trunc<mode>_stfiwx): Likewise.
17646 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
17647 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
17648 (parity<mode>2_cmpb): Set length/type attr.
17649 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
17650 for 'mr.' to fast_compare.
17651 (bpermd_<mode>): Change type attr to popcnt.
17652 (p8_fmrgow_<mode>): New insns for power8 direct move support.
17653 (p8_mtvsrwz_1): Likewise.
17654 (p8_mtvsrwz_2): Likewise.
17655 (reload_fpr_from_gpr<mode>): Likewise.
17656 (p8_mtvsrd_1): Likewise.
17657 (p8_mtvsrd_2): Likewise.
17658 (p8_xxpermdi_<mode>): Likewise.
17659 (reload_vsx_from_gpr<mode>): Likewise.
17660 (reload_vsx_from_gprsf): Likewise.
17661 (p8_mfvsrd_3_<mode>): LIkewise.
17662 (reload_gpr_from_vsx<mode>): Likewise.
17663 (reload_gpr_from_vsxsf): Likewise.
17664 (p8_mfvsrd_4_disf): Likewise.
17665 (multi-word GPR splits): Do not split direct moves or quad memory
17668 2013-06-10 David Malcolm <dmalcolm@redhat.com>
17670 * tree-into-ssa.c (interesting_blocks): Make static.
17672 2013-06-10 Jakub Jelinek <jakub@redhat.com>
17675 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
17676 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
17677 Use DATA_ABI_ALIGNMENT for that case instead if defined.
17678 (get_variable_align): New function.
17679 (get_variable_section, emit_bss, emit_common,
17680 assemble_variable_contents, place_block_symbol): Use
17681 get_variable_align instead of DECL_ALIGN.
17682 (assemble_noswitch_variable): Add align argument, use it
17683 instead of DECL_ALIGN.
17684 (assemble_variable): Adjust caller. Use get_variable_align
17685 instead of DECL_ALIGN.
17686 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
17688 (DATA_ABI_ALIGNMENT): Define.
17689 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
17690 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
17691 opt is false, only return the psABI mandated alignment increase.
17692 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
17693 (DATA_ABI_ALIGNMENT): ... this.
17694 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
17695 (DATA_ABI_ALIGNMENT): ... this.
17696 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
17697 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
17698 (DATA_ABI_ALIGNMENT): ... this.
17699 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
17700 * doc/tm.texi: Regenerated.
17702 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
17704 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
17705 cmp_code to construct REG_EQUAL note.
17707 2013-06-09 Jakub Jelinek <jakub@redhat.com>
17710 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
17711 that operands[2] doesn't overlap with operands[0].
17713 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
17714 Jan Hubicka <jh@suse.cz>
17716 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
17717 hack to mark symbols as used.
17719 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
17721 PR rtl-optimization/57559
17722 * lra-constraints.c (process_alt_operands): Don't discourage
17723 memory with known offset for offsetable memory constraint.
17724 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
17726 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
17728 * varasm.c (struct oc_local_state): Reorder fields.
17729 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
17730 and adjust accordingly.
17731 (output_constructor): Reorder initialization code and adjust call to
17732 output_constructor_bitfield.
17734 2013-06-07 Jan Hubicka <jh@suse.cz>
17736 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
17738 2013-06-07 David Malcolm <dmalcolm@redhat.com>
17740 * tree-object-size.c (unknown): Make const.
17742 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17744 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
17745 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
17746 for last alternative in the cpu_facility attribute.
17748 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17751 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
17752 (xordi3): Change operand 2 constraint to arm_xordi_operand.
17753 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
17754 * config/arm/constraints.md (Dg): New constraint.
17755 * config/arm/neon.md (xordi3_neon): Remove.
17756 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
17757 * config/arm/predicates.md (arm_xordi_operand): New predicate.
17759 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17761 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
17762 Clean up alternatives.
17764 2013-06-07 Alan Modra <amodra@gmail.com>
17766 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
17769 2013-06-07 Alan Modra <amodra@gmail.com>
17771 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
17773 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17775 * config/arm/constraints.md (Df): New constraint.
17776 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
17777 Correct length attribute for last two alternatives.
17779 2013-06-07 Alan Modra <amodra@gmail.com>
17781 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
17782 override user -mfp-in-toc.
17783 (offsettable_ok_by_alignment): Consider just the current access
17784 rather than the whole object, unless BLKmode. Handle
17785 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
17786 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
17787 for -mcmodel=medium.
17788 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
17789 override user -mfp-in-toc or -msum-in-toc. Default to
17790 -mno-fp-in-toc for -mcmodel=medium.
17792 2013-06-06 DJ Delorie <dj@redhat.com>
17794 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
17795 TARGET_VALID_POINTER_MODE.
17797 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
17798 Pat Haugen <pthaugen@us.ibm.com>
17799 Peter Bergner <bergner@vnet.ibm.com>
17801 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
17802 Document new power8 builtins.
17804 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
17805 condition code register, to allow 128-bit logical operations to be
17806 done in the VSX or GPR registers.
17807 (nor<mode>3): Use the canonical form for nor.
17808 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
17809 vclz*, and vpopcnt* vector instructions.
17810 (nand<mode>3): Likewise.
17811 (orc<mode>3): Likewise.
17812 (clz<mode>2): LIkewise.
17813 (popcount<mode>2): Likewise.
17815 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
17816 that only the GPRs are recognized.
17818 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
17819 support for new power8 builtins.
17821 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
17823 (xscvdpspn): Likewise.
17829 (vpopcnt): Likewise.
17830 (vpopcntb): Likewise.
17831 (vpopcnth): Likewise.
17832 (vpopcntw): Likewise.
17833 (vpopcntd): Likewise.
17835 (vmrgew): Likewise.
17836 (vmrgow): Likewise.
17838 (eqv_v16qi3): Likewise.
17839 (eqv_v8hi3): Likewise.
17840 (eqv_v4si3): Likewise.
17841 (eqv_v2di3): Likewise.
17842 (eqv_v4sf3): Likewise.
17843 (eqv_v2df3): Likewise.
17845 (nand_v16qi3): Likewise.
17846 (nand_v8hi3): Likewise.
17847 (nand_v4si3): Likewise.
17848 (nand_v2di3): Likewise.
17849 (nand_v4sf3): Likewise.
17850 (nand_v2df3): Likewise.
17852 (orc_v16qi3): Likewise.
17853 (orc_v8hi3): Likewise.
17854 (orc_v4si3): Likewise.
17855 (orc_v2di3): Likewise.
17856 (orc_v4sf3): Likewise.
17857 (orc_v2df3): Likewise.
17859 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
17860 allow power8 quad mode in 64-bit.
17861 (rs6000_builtin_vectorized_function): Add support to vectorize
17862 ISA 2.07 count leading zeros, population count builtins.
17863 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
17864 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
17865 (builtin_function_type): Add vgbbd builtin function which takes an
17867 (altivec_expand_vec_perm_const): Add support for new power8 merge
17870 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
17871 that does not include TImdoe for use with 32-bit.
17872 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
17874 (UNSPEC_VSX_CVDPSPN): Likewise.
17875 (vsx_xscvdpspn): Likewise.
17876 (vsx_xscvspdpn): Likewise.
17877 (vsx_xscvdpspn_scalar): Likewise.
17878 (vsx_xscvspdpn_directmove): Likewise.
17879 (vsx_and<mode>3): Split logical operations into 32-bit and
17880 64-bit. Add support to do logical operations on TImode as well as
17881 VSX vector types. Allow logical operations to be done in either
17882 VSX registers or in general purpose registers in 64-bit mode. Add
17883 splitters if GPRs were used. For AND, add clobber of CCmode to
17884 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
17886 (vsx_and<mode>3_32bit): Likewise.
17887 (vsx_and<mode>3_64bit): Likewise.
17888 (vsx_ior<mode>3): Likewise.
17889 (vsx_ior<mode>3_32bit): Likewise.
17890 (vsx_ior<mode>3_64bit): Likewise.
17891 (vsx_xor<mode>3): Likewise.
17892 (vsx_xor<mode>3_32bit): Likewise.
17893 (vsx_xor<mode>3_64bit): Likewise.
17894 (vsx_one_cmpl<mode>2): Likewise.
17895 (vsx_one_cmpl<mode>2_32bit): Likewise.
17896 (vsx_one_cmpl<mode>2_64bit): Likewise.
17897 (vsx_nor<mode>3): Likewise.
17898 (vsx_nor<mode>3_32bit): Likewise.
17899 (vsx_nor<mode>3_64bit): Likewise.
17900 (vsx_andc<mode>3): Likewise.
17901 (vsx_andc<mode>3_32bit): Likewise.
17902 (vsx_andc<mode>3_64bit): Likewise.
17903 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
17904 and xxlorc instructions.
17905 (vsx_eqv<mode>3_64bit): Likewise.
17906 (vsx_nand<mode>3_32bit): Likewise.
17907 (vsx_nand<mode>3_64bit): Likewise.
17908 (vsx_orc<mode>3_32bit): Likewise.
17909 (vsx_orc<mode>3_64bit): Likewise.
17911 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
17913 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
17915 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
17916 (p8_vmrgow): Likewise.
17917 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
17918 GPRs to be split under VSX.
17919 (p8v_clz<mode>2): Add power8 count leading zero support.
17920 (p8v_popcount<mode>2): Add power8 population count support.
17921 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
17924 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
17927 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
17929 (vec_nand): Likewise.
17930 (vec_vclz): Likewise.
17931 (vec_vclzb): Likewise.
17932 (vec_vclzd): Likewise.
17933 (vec_vclzh): Likewise.
17934 (vec_vclzw): Likewise.
17935 (vec_vgbbd): Likewise.
17936 (vec_vmrgew): Likewise.
17937 (vec_vmrgow): Likewise.
17938 (vec_vpopcnt): Likewise.
17939 (vec_vpopcntb): Likewise.
17940 (vec_vpopcntd): Likewise.
17941 (vec_vpopcnth): Likewise.
17942 (vec_vpopcntw): Likewise.
17944 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17946 PR rtl-optimization/57468
17947 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
17950 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17952 PR rtl-optimization/57459
17953 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
17954 type when setting live regs.
17956 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17958 * config/s390/s390.opt (mlra): New option.
17959 * config/s390/s390.c (s390_decompose_address): Check displacement
17960 for all registers for LRA.
17961 (s390_secondary_reload): Don't used secondary reloads for LRA.
17962 (s390_lra_p): New function.
17963 (TARGET_LRA_P): Define.
17964 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
17965 of attribute cpu_facility to zarch for the last alternative.
17966 (*cmpmem_short): Ditto.
17968 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
17970 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
17971 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
17972 (arm_expand_prologue): Likewise.
17974 2013-06-06 Teresa Johnson <tejohnson@google.com>
17977 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
17978 as this is now done by redirect_edge_and_branch_force.
17979 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
17980 barriers, and fix interaction with splitting.
17981 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
17982 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
17983 reflect changes made in the routine.
17984 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
17985 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
17986 since this is called in cfglayout mode, and replace partition fixup
17987 with assert as that is now done by force_nonfallthru_and_redirect.
17988 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
17989 already be marked with region crossing note.
17990 (insert_section_boundary_note): Make non-static, gate on flag
17991 has_bb_partition, rewrite to also check for multiple partitions.
17992 (rest_of_handle_reorder_blocks): Remove call to
17993 insert_section_boundary_note, now done later during free_cfg.
17994 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
17995 * bb-reorder.h (insert_section_boundary_note): Declare.
17996 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
17997 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
17998 invoke insert_section_boundary_note.
17999 (try_redirect_by_replacing_jump): Remove unnecessary
18000 check for region crossing note.
18001 (fixup_partition_crossing): New function.
18002 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
18003 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
18004 in non-cfglayout mode.
18005 (force_nonfallthru_and_redirect): Fixup partition boundaries,
18006 remove old code that tried to do this. Emit barrier correctly
18007 when we are in cfglayout mode.
18008 (last_bb_in_partition): New function.
18009 (rtl_split_edge): Correctly fixup partition boundaries.
18010 (commit_one_edge_insertion): Remove old code that tried to
18011 fixup region crossing edge since this is now handled in
18012 split_block, and set up insertion point correctly since
18013 block may now end in a jump.
18014 (verify_hot_cold_block_grouping): Guard against checking when not in
18015 linearized RTL mode.
18016 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
18018 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
18019 rtl_verify_flow_info, so not called in cfglayout mode.
18020 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
18021 (fixup_reorder_chain): Remove old code that attempted to fixup region
18022 crossing note as this is now handled in force_nonfallthru_and_redirect.
18023 (duplicate_insn_chain): Don't duplicate switch section notes.
18024 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
18026 * basic-block.h (emit_barrier_after_bb): Declare.
18028 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18030 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
18031 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
18032 arm_usatsihi): Adjust alternatives for arm_restrict_it.
18034 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18036 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
18038 * config/arm/ldmstm.md: Regenerate.
18040 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18042 * config/arm/sync.md (atomic_loaddi_1):
18043 Disable predication for arm_restrict_it.
18044 (arm_load_exclusive<mode>): Likewise.
18045 (arm_load_exclusivesi): Likewise.
18046 (arm_load_exclusivedi): Likewise.
18047 (arm_load_acquire_exclusive<mode>): Likewise.
18048 (arm_load_acquire_exclusivesi): Likewise.
18049 (arm_load_acquire_exclusivedi): Likewise.
18050 (arm_store_exclusive<mode>): Likewise.
18051 (arm_store_exclusive<mode>): Likewise.
18052 (arm_store_release_exclusivedi): Likewise.
18053 (arm_store_release_exclusive<mode>): Likewise.
18055 2013-06-06 Richard Biener <rguenther@suse.de>
18057 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
18059 (lto_tag_is_tree_code_p): Adjust.
18060 (lto_tag_is_gimple_code_p): Likewise.
18061 (lto_gimple_code_to_tag): Likewise.
18062 (lto_tag_to_gimple_code): Likewise.
18063 (lto_tree_code_to_tag): Likewise.
18064 (lto_tag_to_tree_code): Likewise.
18065 * data-streamer.h (streamer_write_hwi_in_range): Use
18066 uhwi streaming to stream the normalized range.
18067 (streamer_read_hwi_in_range): Likewise.
18069 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18071 * config/arm/arm.md (enabled_for_depr_it): New attribute.
18072 (predicable_short_it): Likewise.
18073 (predicated): Likewise.
18074 (enabled): Handle above.
18075 (define_cond_exec): Set predicated attribute to yes.
18077 2013-06-05 Mike Stump <mikestump@comcast.net>
18079 * gdbinit.in (__FUNCTION__): Add.
18081 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
18083 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
18084 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
18086 2013-06-05 Jan Hubicka <jh@suse.cz>
18088 * varasm.c (mark_decl_referenced): Revert the removal until targets
18091 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
18093 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
18094 instead of mark_decl_referenced.
18096 2013-06-05 Jan Hubicka <jh@suse.cz>
18098 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
18099 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
18100 and symtab_used_from_object_file_p.
18101 (cgraph_make_node_local_1): Clear forced_by_abi.
18102 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
18103 * cgraph.h (symtab_node_base): Add forced_by_abi.
18104 (decide_is_variable_needed): Remove.
18105 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
18106 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
18107 (decide_is_symbol_needed): ... this one; handle symbols in general;
18108 always analyze virtuals; honnor forced_by_abi.
18109 (cgraph_finalize_function): Update.
18110 (varpool_finalize_decl): Update.
18111 (symbol_defined_and_needed): Remove.
18112 (analyze_functions): Update.
18113 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18114 output_refs, input_overwrite_node): Handle forced_by_abi.
18115 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
18116 (address_taken_from_non_vtable_p): ... this one.
18117 (comdat_can_be_unshared_p_1): New function.
18118 (cgraph_comdat_can_be_unshared_p): Rename to ...
18119 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
18120 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
18121 (function_and_variable_visibility): Clear forced_by_abi as needed.
18122 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
18123 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
18124 * symtab.c (dump_symtab_base): Dump forced_by_abi.
18125 * varpool.c (decide_is_variable_needed): Remove.
18127 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18129 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
18130 (arm_option_override): Override arm_restrict_it where appropriate.
18131 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
18132 * config/arm/arm.opt (mrestrict-it): New command-line option.
18133 * doc/invoke.texi: Document -mrestrict-it.
18135 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18137 * tsan.c (tsan_atomic_table): Make const.
18139 2013-06-05 Richard Biener <rguenther@suse.de>
18141 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
18142 index associated with the tree we are supposed to replace.
18143 * tree-streamer-out.c (pack_ts_base_value_fields): Output
18144 TREE_ASM_WRITTEN as zero for everything but SSA names.
18146 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18148 * tree-ssa-structalias.c (call_stmt_vars): Make static.
18150 2013-06-04 Jan Hubicka <jh@suse.cz>
18152 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
18153 (input_node, input_varpool_node): Handle correctly external same
18155 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
18156 nodes at ltrans stage.
18158 2013-06-04 Jan Hubicka <jh@suse.cz>
18160 * ipa-inline.c (update_caller_keys): Fix availability test.
18161 (update_callee_keys): Likewise.
18162 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
18163 to follow ELF standard.
18165 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
18167 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
18168 (mips64r5900el-*-elf*): New configurations.
18169 * config/mips/mips-cpus.def (r5900): New processor.
18170 * config/mips/mips-tables.opt: Regenerate.
18171 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
18172 (mips_issue_rate): Handle PROCESSOR_R5900.
18173 (mips_reorg_process_insns): Force reorder mode for the R5900.
18174 * config/mips/mips.h (TARGET_MIPS5900): Define.
18175 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
18177 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
18179 * config/mips/mips.md (processor): Add r5900.
18180 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
18182 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18184 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
18185 into function to generate MOVI instruction.
18186 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
18187 (aarch64_preferred_simd_mode): Turn into wrapper.
18188 (aarch64_output_scalar_simd_mov_immediate): New function.
18189 * config/aarch64/aarch64-protos.h: Add prototype for above.
18191 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18193 * config/aarch64/aarch64.c (simd_immediate_info): Remove
18194 element_char member.
18195 (sizetochar): Return signed char.
18196 (aarch64_simd_valid_immediate): Remove elchar and other
18197 unnecessary variables.
18198 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
18199 Calculate element_char as required.
18200 * config/aarch64/aarch64-protos.h: Update and move prototype
18201 for aarch64_output_simd_mov_immediate.
18202 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
18205 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18207 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
18208 information completed by aarch64_simd_valid_immediate.
18209 (aarch64_legitimate_constant_p): Update arguments.
18210 (aarch64_simd_valid_immediate): Work with struct rather than many
18212 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
18213 (aarch64_simd_make_constant): Update arguments.
18214 (aarch64_output_simd_mov_immediate): Work with struct rather than
18215 many pointers. Output immediate directly rather than as operand.
18216 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18218 * config/aarch64/constraints.md (Dn): Update arguments.
18220 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18222 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
18224 (aarch64_simd_immediate_valid_for_move): Remove.
18225 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
18226 (aarch64_simd_make_constant): Update call.
18227 (aarch64_output_simd_mov_immediate): Update call.
18228 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18230 * config/aarch64/constraints.md (Dn): Update call.
18232 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18234 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
18235 return type to bool for prototype.
18236 (aarch64_legitimate_constant_p): Check for true instead of not -1.
18237 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
18238 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
18240 2013-06-04 Catherine Moore <clm@codesourcery.com>
18242 * config/mips/mips.opt (meva): New.
18243 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
18244 (ASM_SPEC): Handle -meva.
18245 * doc/invoke.texi (meva): Document.
18247 2013-06-04 Alan Modra <amodra@gmail.com>
18249 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
18252 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18254 * rtl.def: Add extra fourth optional field to define_cond_exec.
18255 * gensupport.c (process_one_cond_exec): Process attributes from
18257 * doc/md.texi: Document fourth field in define_cond_exec.
18259 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
18261 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
18262 out the processing order as in store_bit_field_1.
18264 2013-06-04 Jan Hubicka <jh@suse.cz>
18266 PR middle-end/57500
18267 * cgraphunit.c (cgraph_process_same_body_aliases): Create
18268 non-VAR_DECL node if it does not exist yet.
18270 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
18272 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
18273 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
18274 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
18275 target_cpu_default setting.
18277 2013-06-03 Teresa Johnson <tejohnson@google.com>
18279 * dumpfile.c (opt_info_switch_p): Change -fopt-info
18280 default to -fopt-info=optimized instead of all.
18281 * doc/invoke.texi: Ditto.
18282 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
18283 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
18284 (execute_vect_slp): Emit BB vectorization success under
18285 MSG_OPTIMIZED_LOCATIONS.
18286 * tree-vect-slp.c (vect_slp_transform_bb): Change
18287 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
18288 * tree-vect-loop.c (vect_transform_loop): Ditto.
18290 2013-06-03 Jason Merrill <jason@redhat.com>
18293 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
18294 Use TARGET_EXPR for C++.
18296 2013-06-03 Jakub Jelinek <jakub@redhat.com>
18298 PR rtl-optimization/57268
18299 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
18300 if DEBUG_INSN_P (insn).
18303 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18305 PR rtl-optimization/57268
18306 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18307 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18309 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18311 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
18312 (ix86_avoid_lea_for_addr): Likewise.
18313 (exact_dependency_1): Likewise.
18314 (ix86_adjust_cost): Likewise.
18315 (swap_top_of_ready_list): Fix formatting and !reload_completed check
18317 (do_reorder_for_imul): Fix typo, formatting and
18318 !reload_completed check removed.
18319 (ix86_sched_reorder): Fix typo and formatting.
18320 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
18323 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
18325 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
18327 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
18329 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
18330 <STRING_CST>: Likewise.
18331 <VECTOR_CST>: Likewise.
18333 2013-06-01 Janus Weil <janus@gcc.gnu.org>
18334 Mikael Morin <mikael@gcc.gnu.org>
18336 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
18337 * config.in: Regenerated.
18338 * configure: Regenerated.
18340 2013-06-01 Jan Hubicka <jh@suse.cz>
18342 PR middle-end/57366
18343 * cgraphunit.c (compile): When weakref is not supported,
18344 set up transparent aliases before final output pass.
18345 * varasm.c (assemble_alias): Do not try to do it here.
18347 2013-06-01 Jan Hubicka <jh@suse.cz>
18349 PR middle-end/57467
18350 * passes.c (for_per_function): Skip unanalyzed functions.
18352 2013-06-01 Jan Hubicka <jh@suse.cz>
18354 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
18355 (lto_symtab_merge_symbols_1): ... this one.
18356 (lto_symtab_merge_cgraph_nodes): Rename to ...
18357 (lto_symtab_merge_symbols): ... this one; simplify.
18358 * cgraph.c (same_body_aliases_done): Rename to ...
18359 (cpp_implicit_aliases_done): ... this one.
18360 (cgraph_create_function_alias): Update.
18361 (cgraph_same_body_alias): Update.
18362 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
18363 (verify_edge_corresponds_to_fndecl): Simplify.
18364 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
18365 (cgraph_node): Remove same_body_alias.
18366 (varpool_node): Remove alias_of and extra_name_alias.
18367 (same_body_aliases_done): Rename to ..
18368 (cpp_implicit_aliases_done): ... this one.
18369 (symtab_alias_ultimate_target): Add default parameter.
18370 (symtab_resolve_alias): New function.
18371 (fixup_same_cpp_alias_visibility): Declare.
18372 (cgraph_function_node): Add default parameter.
18373 (cgraph_node_asm_name): Likewise.
18374 (cgraph_function_or_thunk_node): Add default parameter; do
18375 not ICE when it is NULL.
18376 (varpool_variable_node): Likewise.
18377 * tree-emutls.c (create_emultls_var): Update.
18378 (ipa_lower_emutls): Update.
18379 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
18380 (cgraph_reset_node): Reset alias info.
18381 (cgraph_finalize_function): Update.
18382 (fixup_same_cpp_alias_visibility): Move to symtab.c.
18383 (analyze_function): Simplify.
18384 (cgraph_process_same_body_aliases): Simplify.
18385 (analyze_functions): Fixup same body aliases.
18386 (handle_alias_pairs): Simplify.
18387 (assemble_thunk): Update.
18388 (assemble_thunks_and_aliases): Update.
18389 (output_weakrefs): Rewrite.
18390 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
18391 (lto_output_varpool_node): Likewise.
18392 (compute_ltrans_boundary): Remve assert.
18393 (get_alias_symbol): New functoin.
18394 (input_node): Rewrite alias handling.
18395 (input_varpool_node): Likewise.
18396 * ipa-pure-const.c (propagate_pure_const): Fix formating.
18397 * ipa.c (process_references): Handle weakrefs correctly.
18398 (symtab_remove_unreachable_nodes): Likewise.
18399 * trans-mem.c (get_cg_data): Update.
18400 (ipa_tm_create_version_alias): Update.
18401 (ipa_tm_execute): Update.
18402 * symtab.c (dump_symtab_base): Dump aliases.
18403 (verify_symtab_base): Verify aliases.
18404 (symtab_node_availability): New function.
18405 (symtab_alias_ultimate_target): Simplify.
18406 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
18407 handle all the fixup cases.
18408 (symtab_resolve_alias): New function.
18409 * passes.c (ipa_write_summaries): Handle weakrefs.
18410 * varpool.c (varpool_analyze_node): Simplify.
18411 (assemble_aliases): Update.
18412 (varpool_create_variable_alias): Simplify.
18413 (varpool_extra_name_alias): Simplify.
18414 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
18415 (lto_symtab_merge_symbols): ... this one.
18417 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
18420 PR rtl-optimization/57268
18421 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18422 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18424 2013-06-01 Tobias Burnus <burnus@net-b.de>
18426 Partially reverted:
18427 2013-05-31 Tobias Burnus <burnus@net-b.de>
18429 PR middle-end/57073
18430 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
18433 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18435 PR rtl-optimization/57268
18436 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18437 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18439 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18441 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
18442 unordered comparison operators when -fno-trapping-math is in effect
18444 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
18445 and implement unordered comparison operators properly on the e500.
18447 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18449 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
18450 for constant scalar integers.
18451 (simplify_relational_operation_1): Likewise.
18453 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
18455 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
18456 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
18459 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18460 Igor Zamyatin <igor.zamyatin@intel.com>
18462 Silvermont (SLM) architecture performance tuning.
18463 * config/i386/i386.h (enum ix86_tune_indices): Add
18464 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
18465 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
18467 * config/i386/i386.c (initial_ix86_tune_features)
18468 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
18469 (ix86_lea_outperforms): Handle Silvermont tuning.
18470 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
18472 (ix86_use_lea_for_mov): Likewise.
18473 (ix86_avoid_lea_for_addr): Likewise.
18474 (ix86_lea_for_add_ok): Likewise.
18475 (exact_dependency_1): New function.
18476 (exact_store_load_dependency): Likewise.
18477 (ix86_adjust_cost): Handle Silvermont tuning.
18478 (do_reoder_for_imul): Likewise.
18479 (swap_top_of_ready_list): New function.
18480 (ix86_sched_reorder): Changed to handle Silvermont tuning.
18482 * config/i386/i386.md (peepholes that split memory operand in fp
18485 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18487 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
18488 Remove un-necessary braces.
18490 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18492 * config/aarch64/aarch64.c (aarch64_classify_symbol):
18493 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
18495 2013-05-31 Tobias Burnus <burnus@net-b.de>
18497 PR middle-end/57073
18498 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
18500 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18503 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
18504 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
18505 * config/arm/neon.md (iordi3_neon): Remove.
18506 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
18507 * config/arm/predicates.md (imm_for_neon_logic_operand):
18508 Move to earlier in the file.
18509 (neon_logic_op2): Likewise.
18510 (arm_iordi_operand_neon): New predicate.
18512 2013-05-31 Richard Biener <rguenther@suse.de>
18514 PR tree-optimization/57478
18515 PR tree-optimization/57453
18516 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
18519 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18521 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
18522 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
18524 2013-05-30 Tobias Burnus <burnus@net-b.de>
18525 Thomas Koenig <tkoenig@gcc.gnu.org>
18527 PR middle-end/57073
18528 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
18529 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
18531 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
18533 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
18535 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
18537 * target.def (register_usage_leveling_p): New hook.
18538 * targhooks.c (default_register_usage_leveling_p): New.
18539 * targhooks.h (default_register_usage_leveling_p): New prototype.
18540 * lra-assigns.c (register_usage_leveling_p): Use the hook.
18541 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
18542 * doc/tm.texi: Update.
18543 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
18545 2013-05-30 Ian Bolton <ian.bolton@arm.com>
18547 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
18548 (*insv_reg<mode>): New define_insn.
18550 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
18552 PR rtl-optimization/57439
18553 * postreload.c (move2add_valid_value_p): Check that we have
18554 a zero subreg_regno_offset when accessing the register in
18555 the requested mode.
18557 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18558 Igor Zamyatin <igor.zamyatin@intel.com>
18560 Silvermont (SLM) architecture pipeline model, tuning and
18562 * config.gcc: Add slm config options and target.
18564 * config/i386/slm.md: New.
18566 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
18568 * config/i386/i386-c.c (ix86_target_macros_internal): New case
18570 (ix86_target_macros_internal): Likewise.
18572 * config/i386/i386.c (slm_cost): New cost.
18573 (m_SLM): New macro flag.
18574 (initial_ix86_tune_features): Set m_SLM.
18575 (x86_accumulate_outgoing_args): Likewise.
18576 (x86_arch_always_fancy_math_387): Likewise.
18577 (processor_target_table): Add slm cost.
18578 (cpu_names): Add slm cpu name.
18579 (x86_option_override_internal): Set SLM ISA.
18580 (ix86_issue_rate): New case PROCESSOR_SLM.
18581 (ia32_multipass_dfa_lookahead): Likewise.
18582 (fold_builtin_cpu): Add slm.
18584 * config/i386/i386.h (TARGET_SLM): New target macro.
18585 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
18586 (processor_type): Add PROCESSOR_SLM.
18588 * config/i386/i386.md (cpu): Add new value "slm".
18589 (slm.md): Include slm.md.
18591 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
18592 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18594 * config/arm/arm-protos.h: Add and update function protos.
18595 * config/arm/arm.c (use_simple_return_p): New added.
18596 (thumb2_expand_return): Check simple_return flag.
18597 * config/arm/arm.md: Add simple_return and conditional simple_return.
18598 * config/arm/iterators.md: Add iterator for return and simple_return.
18600 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18602 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
18603 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
18604 (arm_emit_vfp_multi_reg_pop): Likewise.
18605 (thumb2_emit_ldrd_pop): Likewise.
18606 (arm_expand_epilogue): Add misc REG_CFA notes.
18607 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
18609 2013-05-29 Lawrence Crowl <crowl@google.com>
18611 * config/arm/t-arm: Update for below.
18613 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
18614 Change type to hash_table. Update dependent calls and types.
18616 * config/i386/t-cygming: Update for below.
18618 * config/i386/t-interix: Update for below.
18620 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
18621 Change type to hash_table. Update dependent calls and types.
18622 (i386_find_on_wrapper_list::wrappers): Likewise.
18624 * config/ia64/t-ia64: Update for below.
18626 * config/ia64/ia64.c (bundle_state_table):
18627 Change type to hash_table. Update dependent calls and types.
18629 * config/mips/mips.c (mips_reorg_process_insns::htab):
18630 Change type to hash_table. Update dependent calls and types.
18632 * config/sol2.c (solaris_comdat_htab):
18633 Change type to hash_table. Update dependent calls and types.
18635 * config/t-sol2: Update for above.
18637 2013-05-29 Teresa Johnson <tejohnson@google.com>
18639 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
18640 functions are not yet marked as defined.
18642 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
18643 Pat Haugen <pthaugen@us.ibm.com>
18644 Peter Bergner <bergner@vnet.ibm.com>
18646 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
18650 (vrotl<mode>3): Likewise.
18651 (vashl<mode>3): Likewise.
18652 (vlshr<mode>3): Likewise.
18653 (vashr<mode>3): Likewise.
18655 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18656 support for power8 V2DI builtins.
18658 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
18659 power8 V2DI builtins.
18660 (vupkhsw): Likewise.
18661 (vupklsw): Likewise.
18662 (vaddudm): Likewise.
18663 (vminsd): Likewise.
18664 (vmaxsd): Likewise.
18665 (vminud): Likewise.
18666 (vmaxud): Likewise.
18667 (vpkudum): Likewise.
18668 (vpksdss): Likewise.
18669 (vpkudus): Likewise.
18670 (vpksdus): Likewise.
18675 (vsubudm): Likewise.
18676 (vcmpequd): Likewise.
18677 (vcmpgtsd): Likewise.
18678 (vcmpgtud): Likewise.
18679 (vcmpequd_p): Likewise.
18680 (vcmpgtsd_p): Likewise.
18681 (vcmpgtud_p): Likewise.
18682 (vupkhsw): Likewise.
18683 (vupklsw): Likewise.
18684 (vaddudm): Likewise.
18685 (vmaxsd): Likewise.
18686 (vmaxud): Likewise.
18687 (vminsd): Likewise.
18688 (vminud): Likewise.
18689 (vpksdss): Likewise.
18690 (vpksdus): Likewise.
18691 (vpkudum): Likewise.
18692 (vpkudus): Likewise.
18697 (vsubudm): Likewise.
18699 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
18700 support for power8 V2DI instructions.
18702 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
18703 power8 V2DI instructions. Combine pack and unpack insns to use an
18704 iterator for each mode. Check whether a particular mode supports
18705 Altivec instructions instead of just checking TARGET_ALTIVEC.
18706 (UNSPEC_VPKUWUM): Likewise.
18707 (UNSPEC_VPKSHSS): Likewise.
18708 (UNSPEC_VPKSWSS): Likewise.
18709 (UNSPEC_VPKUHUS): Likewise.
18710 (UNSPEC_VPKSHUS): Likewise.
18711 (UNSPEC_VPKUWUS): Likewise.
18712 (UNSPEC_VPKSWUS): Likewise.
18713 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
18714 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
18715 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
18716 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
18717 (UNSPEC_VUPKHSB): Likewise.
18718 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
18719 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
18720 (UNSPEC_VUPKHSH): Likewise.
18721 (UNSPEC_VUPKLSB): Likewise.
18722 (UNSPEC_VUPKLSH): Likewise.
18724 (VI_char): Likewise.
18725 (VI_scalar): Likewise.
18726 (VI_unit): Likewise.
18728 (VP_small): Likewise.
18729 (VP_small_lc): Likewise.
18730 (VU_char): Likewise.
18731 (add<mode>3): Likewise.
18732 (altivec_vaddcuw): Likewise.
18733 (altivec_vaddu<VI_char>s): Likewise.
18734 (altivec_vadds<VI_char>s): Likewise.
18735 (sub<mode>3): Likewise.
18736 (altivec_vsubcuw): Likewise.
18737 (altivec_vsubu<VI_char>s): Likewise.
18738 (altivec_vsubs<VI_char>s): Likewise.
18739 (altivec_vavgs<VI_char>): Likewise.
18740 (altivec_vcmpbfp): Likewise.
18741 (altivec_eq<mode>): Likewise.
18742 (altivec_gt<mode>): Likewise.
18743 (altivec_gtu<mode>): Likewise.
18744 (umax<mode>3): Likewise.
18745 (smax<mode>3): Likewise.
18746 (umin<mode>3): Likewise.
18747 (smin<mode>3): Likewise.
18748 (altivec_vpkuhum): Likewise.
18749 (altivec_vpkuwum): Likewise.
18750 (altivec_vpkshss): Likewise.
18751 (altivec_vpkswss): Likewise.
18752 (altivec_vpkuhus): Likewise.
18753 (altivec_vpkshus): Likewise.
18754 (altivec_vpkuwus): Likewise.
18755 (altivec_vpkswus): Likewise.
18756 (altivec_vpks<VI_char>ss): Likewise.
18757 (altivec_vpks<VI_char>us): Likewise.
18758 (altivec_vpku<VI_char>us): Likewise.
18759 (altivec_vpku<VI_char>um): Likewise.
18760 (altivec_vrl<VI_char>): Likewise.
18761 (altivec_vsl<VI_char>): Likewise.
18762 (altivec_vsr<VI_char>): Likewise.
18763 (altivec_vsra<VI_char>): Likewise.
18764 (altivec_vsldoi_<mode>): Likewise.
18765 (altivec_vupkhsb): Likewise.
18766 (altivec_vupkhs<VU_char>): Likewise.
18767 (altivec_vupkls<VU_char>): Likewise.
18768 (altivec_vupkhsh): Likewise.
18769 (altivec_vupklsb): Likewise.
18770 (altivec_vupklsh): Likewise.
18771 (altivec_vcmpequ<VI_char>_p): Likewise.
18772 (altivec_vcmpgts<VI_char>_p): Likewise.
18773 (altivec_vcmpgtu<VI_char>_p): Likewise.
18774 (abs<mode>2): Likewise.
18775 (vec_unpacks_hi_v16qi): Likewise.
18776 (vec_unpacks_hi_v8hi): Likewise.
18777 (vec_unpacks_lo_v16qi): Likewise.
18778 (vec_unpacks_hi_<VP_small_lc>): Likewise.
18779 (vec_unpacks_lo_v8hi): Likewise.
18780 (vec_unpacks_lo_<VP_small_lc>): Likewise.
18781 (vec_pack_trunc_v8h): Likewise.
18782 (vec_pack_trunc_v4si): Likewise.
18783 (vec_pack_trunc_<mode>): Likewise.
18785 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
18787 (vec_vmaxsd): Likewise.
18788 (vec_vmaxud): Likewise.
18789 (vec_vminsd): Likewise.
18790 (vec_vminud): Likewise.
18791 (vec_vpksdss): Likewise.
18792 (vec_vpksdus): Likewise.
18793 (vec_vpkudum): Likewise.
18794 (vec_vpkudus): Likewise.
18795 (vec_vrld): Likewise.
18796 (vec_vsld): Likewise.
18797 (vec_vsrad): Likewise.
18798 (vec_vsrd): Likewise.
18799 (vec_vsubudm): Likewise.
18800 (vec_vupkhsw): Likewise.
18801 (vec_vupklsw): Likewise.
18803 2013-05-29 Jan Hubicka <jh@suse.cz>
18805 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
18806 flags; reorder rest of fields in more consistent way.
18807 (varpool_node): Remove analyzed, finalized and alias.
18808 (cgraph_ndoe): Likewise.
18809 (symtab_alias_ultimate_target): New function.
18810 (cgraph_function_node): Move offline.
18811 (cgraph_reset_node): Declare.
18812 (cgraph_comdat_can_be_unshared_p): Remove.
18813 (varpool_remove_initializer): Declare.
18814 (varpool_first_defined_variable, varpool_next_defined_variable
18815 cgraph_first_defined_function, cgraph_next_defined_function): Update.
18816 (cgraph_function_with_gimple_body_p): Update.
18817 (varpool_all_refs_explicit_p): Update.
18818 (symtab_alias_target): New function.
18819 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
18820 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
18821 (cgraph_function_or_thunk_node): Simplify using
18822 symtab_alias_ultimate_target.
18823 (varpool_variable_node): Likewise.
18824 * cgraph.c (cgraph_create_function_alias): Update.
18825 (cgraph_add_thunk): Update.
18826 (cgraph_remove_node): Update.
18827 (dump_cgraph_node): Do not dump removed flags.
18828 (cgraph_function_body_availability): Update.
18829 (cgraph_propagate_frequency): Update.
18830 (verify_cgraph_node): Check sanity of local flag.
18831 (cgraph_function_node): Move here from cgraph.h; revamp for
18832 cgraph_function_or_thunk_node.
18833 * lto-symtab.c (lto_varpool_replace_node): Update.
18834 (lto_symtab_resolve_can_prevail_p): Update.
18835 (lto_symtab_merge_cgraph_nodes): Update.
18836 * ipa-cp.c (determine_versionability, initialize_node_lattices,
18837 propagate_constants_accross_call, devirtualization_time_bonus,
18838 ipcp_propagate_stage): Update.
18839 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
18840 * ipa-inline-transform.c (clone_inlined_nodes,
18841 preserve_function_body_p): Update.
18842 * ipa-reference.c (propagate): Update.
18843 (write_node_summary_p): Update.
18844 * toplev.c (wrapup_global_declaration_2): Update.
18845 * cgraphunit.c (cgraph_analyze_function): Rename to ...
18846 (analyze_function) ... this one.
18847 (cgraph_process_new_functions): Update.
18848 (cgraph_reset_node): Export.
18849 (cgraph_finalize_function): Update.
18850 (cgraph_add_new_function): Update.
18851 (process_function_and_variable_attributes): Update.
18852 (varpool_finalize_decl): Update.
18853 (symbol_finalized): Remove.
18854 (symbol_finalized_and_needed): Rename to ...
18855 (symbol_defined_and_needed): ... update.
18856 (cgraph_analyze_functions): Update.
18857 (handle_alias_pairs): Update.
18858 (mark_functions_to_output): Update.
18859 (assemble_thunk): Update.
18860 (output_in_order): Update.
18861 (output_weakrefs): Update.
18862 (finalize_compilation_unit): Update.
18863 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
18864 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
18865 input_node, input_varpool_node): Update.
18866 * dbxout.c (dbxout_expand_expr): Update.
18867 * cgraphclones.c (cgraph_clone_node): Update.
18868 (cgraph_copy_node_for_versioning): Update.
18869 (cgraph_materialize_clone): Update.
18870 (cgraph_materialize_all_clones): Update.
18871 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
18872 propagate_pure_const, propagate_nothrow): Update.
18873 * lto-streamer-out.c (lto_output, write_symbol): Update.
18874 * ipa-utils.c (ipa_reverse_postorder): Update.
18875 * ipa-inline.c (can_inline_edge_p): Update.
18876 (update_caller_keys, ipa_inline): Update.
18877 * dwarf2out.c (reference_to_unused,
18878 premark_types_used_by_global_vars_helper): Update.
18879 * tree-eh.c (tree_could_trap_p): Update.
18880 * ipa-split.c (consider_split, execute_split_functions): Update.
18881 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
18882 has_addr_references_p): Update; move ahead in file for better
18884 (process_references): Simplify.
18885 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
18886 bodies are removed.
18887 (cgraph_comdat_can_be_unshared_p): Make static.
18888 (cgraph_externally_visible_p): Update.
18889 (varpool_externally_visible_p): Update.
18890 (function_and_variable_visibility): Update.
18891 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
18892 ipa_tm_mark_force_output_node): Update.
18893 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
18894 estimate_edge_devirt_benefit, inline_generate_summary,
18895 inline_write_summary): Update.
18896 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
18897 * ipa-prop.c (ipa_compute_jump_functions): Update.
18898 (ipa_print_node_params, ipa_prop_read_section,
18899 ipa_update_after_lto_read, read_replacements_section): Update.
18900 * varasm.c (mark_decl_referenced): Update.
18901 (assemble_alias, dump_tm_clone_pairs): Update.
18902 * tree-inline.c (copy_bb): Update.
18903 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
18905 * symtab.c (dump_symtab_base): Print new flags.
18906 (verify_symtab_base): Verify new flags.
18907 (symtab_alias_ultimate_target): New function.
18908 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
18909 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
18911 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
18913 * i386.c (ix86_get_function_versions_dispatcher,
18914 ix86_generate_version_dispatcher_body): Update.
18915 (fold_builtin_cpu): Use varpool_add_new_variable.
18916 * varpool.c (varpool_remove_initializer): Break out from ...
18917 (varpool_remove_node): ... this one.
18918 (dump_varpool_node, varpool_node_for_asm,
18919 cgraph_variable_initializer_availability, varpool_analyze_node,
18920 varpool_assemble_decl, varpool_remove_unreferenced_decls,
18921 varpool_finalize_named_section_flags, varpool_create_variable_alias):
18924 2013-05-29 Jan Hubicka <jh@suse.cz>
18926 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
18928 2013-05-29 Easwaran Raman <eraman@google.com>
18930 PR tree-optimization/57442
18931 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
18932 when control exits the main loop.
18934 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
18936 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
18938 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
18939 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
18940 * rx/t-rx: Add rx100 under multi library matches option for nofpu
18943 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18945 PR tree-optimization/57441
18946 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
18947 Don't limit size of incr_vec to number of candidates.
18949 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
18951 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
18952 and mips16 directories.
18953 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
18954 (MULTILIB_DIRNAMES): Ditto.
18955 (MULTILIB_EXCEPTIONS): Add new exceptions.
18956 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
18957 (MULTILIB_DIRNAMES): Ditto.
18958 (MULTILIB_EXCEPTIONS): Add new exceptions.
18960 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18961 Marcus Shawcroft <marcus.shawcroft@arm.com>
18963 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
18964 SYMBOL_TINY_ABSOLUTE.
18965 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
18966 SYMBOL_TINY_ABSOLUTE.
18967 (aarch64_expand_mov_immediate): Likewise.
18968 (aarch64_classify_symbol): Likewise.
18969 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
18970 Permit SYMBOL_TINY_ABSOLUTE.
18971 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
18973 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18974 Marcus Shawcroft <marcus.shawcroft@arm.com>
18976 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
18977 Refactor if/switch. Replace gcc_assert with if.
18979 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
18981 * config/i386/i386.c (initial_ix86_tune_features): Enable
18982 FP Reassociation for AMD bdver1 and bdver2.
18984 2013-05-29 Martin Jambor <mjambor@suse.cz>
18986 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
18987 and IMAGPART_EXPR do not occur within other handled_components.
18989 2013-05-29 Richard Biener <rguenther@suse.de>
18991 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
18992 access on whether the use is in the BB we currently try to
18994 (vect_bb_vectorization_profitable_p): Pass the BB we currently
18995 vectorize to vect_bb_slp_scalar_cost.
18997 2013-05-29 Richard Biener <rguenther@suse.de>
18999 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
19000 computing scalar cost offsetted by stmts that are kept live
19002 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
19003 for computation of scalar cost.
19005 2013-05-28 Steve Ellcey <sellcey@mips.com>
19007 * config/mips/mips-cpus.def (mips32r2): Change processor type.
19009 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
19011 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
19012 array notation built-in reduction functions.
19013 * doc/passes.texi (Passes): Added documentation about changes done
19015 * doc/invoke.texi (C Dialect Options): Added documentation about
19016 the -fcilkplus flag.
19017 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
19018 (BUILTINS_DEF): Depend on cilkplus.def.
19019 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
19020 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
19021 * cilkplus.def: New file.
19023 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
19025 PR rtl-optimization/57439
19026 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
19028 2013-05-28 Easwaran Raman <eraman@google.com>
19030 PR tree-optimization/57337
19031 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
19032 (find_insert_point): Correctly identify the insertion point
19033 when two statements with the same UID is compared.
19035 2013-05-28 Richard Biener <rguenther@suse.de>
19037 PR tree-optimization/56787
19038 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
19039 from the list of data references.
19040 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
19042 (vect_analyze_loop_operations): Likewise.
19043 (vect_transform_loop): Remove clobbers.
19045 2013-05-28 Martin Jambor <mjambor@suse.cz>
19047 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
19048 and REALPART_EXPRs have scalar type.
19050 2013-05-28 Richard Biener <rguenther@suse.de>
19052 PR tree-optimization/57411
19053 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
19055 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
19056 virtual operand propagation.
19058 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19060 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
19061 destination register for bmasksi_vis.
19062 (vector_init_bshuffle): Likewise.
19063 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
19065 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19067 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
19068 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
19069 mode if the instruction isn't available in the original mode.
19070 * config/sparc/sparc.opt (mfix-ut699): New option.
19071 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
19072 (divdf3): Turn into expander.
19073 (divdf3_nofix): New insn.
19074 (divdf3_fix): Likewise.
19075 (divsf3): Disable if -mfix-ut699.
19076 (sqrtdf2): Turn into expander.
19077 (sqrtdf2_nofix): New insn.
19078 (sqrtdf2_fix): Likewise.
19079 (sqrtsf2): Disable if -mfix-ut699.
19081 2013-05-27 Richard Biener <rguenther@suse.de>
19083 PR middle-end/57412
19084 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
19085 block for the new loop.
19087 2013-05-27 Richard Biener <rguenther@suse.de>
19089 PR tree-optimization/57343
19090 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
19091 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
19092 (number_of_iterations_cond): Do not build the folded tree.
19094 2013-05-27 Richard Biener <rguenther@suse.de>
19097 PR middle-end/57381
19098 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19099 OEP_CONSTANT_ADDRESS_OF retained.
19101 PR tree-optimization/57417
19102 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
19103 for unchanged base.
19104 (set_ssa_val_to): Compare addresses using
19105 get_addr_base_and_unit_offset.
19107 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
19109 PR rtl-optimization/56833
19110 * postreload.c (move2add_record_mode): New function.
19111 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
19112 (move2add_use_add2_insn): Use move2add_record_sym_value.
19113 (move2add_use_add3_insn): Likewise.
19114 (reload_cse_move2add): Use move2add_valid_value_p and
19115 move2add_record_mode. Invalidate call-clobbered and REG_INC
19116 affected regs by setting reg_mode to VOIDmode.
19117 (move2add_note_store): Don't pretend the inside of a SUBREG is
19118 the actual destination. Invalidate single/leading registers by
19119 setting reg_mode to VOIDmode.
19120 Use move2add_record_sym_value, move2add_valid_value_p and
19121 move2add_record_mode.
19123 2013-05-27 Richard Biener <rguenther@suse.de>
19125 PR tree-optimization/57396
19126 * tree-affine.c (double_int_constant_multiple_p): Properly
19127 return false for val == 0 and div != 0.
19129 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19131 * config/mips/mips.h: Use #elif in preprocessor conditions.
19133 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19136 * config/mips/constraints.md (kl): New constraint.
19137 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
19138 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
19139 constraint for operand 0. Split after CSE for MIPS16. Emit a move
19140 from LO for MIPS16.
19141 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
19143 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19146 * config/mips/mips.c (mips_can_inline_p): New function.
19147 (TARGET_CAN_INLINE_P): Define.
19149 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
19151 * sched-int.h (ds_t, dw_t): Make unsigned int.
19152 Fix documentation that describes how all the ds_t bits are used.
19153 Reserve the last bit for delayed-branch scheduling.
19154 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
19155 (BITS_PER_DEP_WEAK): Fix definition and documentation.
19156 (gen_dep_weak_1): Remove prototype.
19157 * sched-deps.c (get_dep_weak_1): Make static.
19158 * target.def (speculate_insn, needs_block_p, gen_spec_check,
19159 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
19160 * doc/tm.texi: Regenerate.
19161 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
19163 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
19166 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
19168 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
19169 Sandra Loosemore <sandra@codesourcery.com>
19171 * config.gcc (powerpc-*): Allow native for with-cpu.
19173 2013-05-24 Jeff Law <law@redhat.com>
19175 PR tree-optimization/57124
19176 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
19177 conversion feeding a condition if the range has an overflow
19178 if -fstrict-overflow. Add warnings for when we do make the
19181 2013-05-24 Dehao Chen <dehao@google.com>
19183 * tree-cfg.c (locus_discrim_map): Fix the typo.
19184 (locus_discrim_hasher): Likewise.
19185 (locus_discrim_hasher::hash): Likewise.
19186 (locus_discrim_hasher::equal): Likewise.
19188 2013-05-24 Martin Jambor <mjambor@suse.cz>
19190 PR tree-optimization/57294
19191 * cgraph.h (ipa_record_stmt_references): Declare.
19192 * cgraphbuild.c (ipa_record_stmt_references): New function.
19193 (build_cgraph_edges): Use ipa_record_stmt_references.
19194 (rebuild_cgraph_edges): Likewise.
19195 (cgraph_rebuild_references): Likewise.
19196 * ipa-prop.c (ipa_modify_call_arguments): Discard references
19197 associated with the old statement and build references from the
19198 newly built statements.
19199 * ipa-ref.c (ipa_remove_stmt_references): New function.
19200 * ipa-ref.h (ipa_remove_stmt_references): Declare.
19202 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
19204 * lra-constraints.c (emit_spill_move): Use smaller mode for
19206 (check_and_process_move): Consider mem-reg moves for secondary
19208 (curr_insn_transform): Don't lose insns emitted before for
19209 secondary memory moves.
19210 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
19211 reg set up in the current insn.
19213 2013-05-24 Dehao Chen <dehao@google.com>
19215 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
19217 (locus_descrim_hasher::equal): Likewise.
19218 (build_gimple_cfg): New discriminator assignment algorithm.
19219 (make_edges): Likewise.
19220 (next_discriminator_for_locus): Likewise.
19221 (same_line_p): Likewise.
19222 (assign_discriminators): Likewise.
19223 (make_cond_expr_edges): Likewise.
19224 (make_gimple_switch_edges): Likewise.
19225 (make_goto_expr_edges): Likewise.
19226 (make_gimple_asm_edges): Likewise.
19228 2013-05-24 Ian Bolton <ian.bolton@arm.com>
19230 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
19231 X format specifier to only display bottom 16 bits.
19232 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
19233 immediate to match for operand 2, since it will be masked.
19235 2013-05-24 Richard Biener <rguenther@suse.de>
19237 PR tree-optimization/57287
19238 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
19239 all SSA names that occur in abnormal PHIs.
19241 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
19244 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
19245 that index is not negative.
19247 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
19249 PR rtl-optimization/55177
19250 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
19251 (simplify_byte_swapping_operation): New.
19252 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
19253 (simplify_relational_operation_1): Deal with BSWAP.
19255 2013-05-23 Richard Henderson <rth@redhat.com>
19258 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
19259 (ix86_reorg): Call it.
19261 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
19264 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
19265 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
19266 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
19268 2013-05-23 Christian Bruel <christian.bruel@st.com>
19271 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
19273 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19274 Marcus Shawcroft <marcus.shawcroft@arm.com>
19276 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
19277 * config/aarch64/constraints.md (Usa): Remove.
19278 * doc/md.texi (AArch64 Usa): Remove.
19280 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19281 Marcus Shawcroft <marcus.shawcroft@arm.com>
19283 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
19284 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
19285 * config/aarch64/predicates.md (aarch64_const_address): Remove.
19286 (aarch64_mov_operand): Use aarch64_mov_operand_p.
19288 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
19290 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
19291 instruction (AdvSIMD).
19292 * config/aarch64/aarch64-builtins.c
19293 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
19294 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
19296 2013-05-23 Martin Jambor <mjambor@suse.cz>
19298 PR middle-end/57347
19299 * tree.h (contains_bitfld_component_ref_p): Declare.
19300 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
19301 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
19303 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
19304 not access a bit-field. Assert all final offsets are byte-aligned.
19306 2013-05-23 Richard Biener <rguenther@suse.de>
19308 PR tree-optimization/57380
19309 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
19310 least one invariant or re-used load.
19311 * passes.c (init_optimization_passes): Move pass_phiprop before
19314 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
19316 * config/aarch64/aarch64-simd.md
19317 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
19319 2013-05-23 Richard Biener <rguenther@suse.de>
19321 PR middle-end/57381
19322 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19323 OEP_CONSTANT_ADDRESS_OF retained.
19325 2013-05-23 Jakub Jelinek <jakub@redhat.com>
19327 PR middle-end/57344
19328 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
19329 don't lower unit. Handle unit not being always BITS_PER_WORD.
19331 2013-05-23 Richard Biener <rguenther@suse.de>
19333 PR rtl-optimization/57341
19334 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
19335 instead of true_dependence.
19337 2013-05-22 David Malcolm <dmalcolm@redhat.com>
19339 * bb-reorder.c (branch_threshold): Make const.
19340 (exec_threshold): Ditto.
19342 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19343 Pat Haugen <pthaugen@us.ibm.com>
19344 Peter Bergner <bergner@vnet.ibm.com>
19346 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
19347 documentation for the power8 crypto builtins.
19349 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
19351 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
19352 macros for defining power8 builtin functions.
19353 (BU_P8V_AV_2): Likewise.
19354 (BU_P8V_AV_P): Likewise.
19355 (BU_P8V_VSX_1): Likewise.
19356 (BU_P8V_OVERLOAD_1): Likewise.
19357 (BU_P8V_OVERLOAD_2): Likewise.
19358 (BU_CRYPTO_1): Likewise.
19359 (BU_CRYPTO_2): Likewise.
19360 (BU_CRYPTO_3): Likewise.
19361 (BU_CRYPTO_OVERLOAD_1): Likewise.
19362 (BU_CRYPTO_OVERLOAD_2): Likewise.
19363 (XSCVSPDP): Fix typo, point to the correct instruction.
19364 (VCIPHER): Add power8 crypto builtins.
19365 (VCIPHERLAST): Likewise.
19366 (VNCIPHER): Likewise.
19367 (VNCIPHERLAST): Likewise.
19368 (VPMSUMB): Likewise.
19369 (VPMSUMH): Likewise.
19370 (VPMSUMW): Likewise.
19371 (VPERMXOR_V2DI): Likewise.
19372 (VPERMXOR_V4SI: Likewise.
19373 (VPERMXOR_V8HI: Likewise.
19374 (VPERMXOR_V16QI: Likewise.
19375 (VSHASIGMAW): Likewise.
19376 (VSHASIGMAD): Likewise.
19377 (VPMSUM): Likewise.
19378 (VPERMXOR): Likewise.
19379 (VSHASIGMA): Likewise.
19381 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
19382 __CRYPTO__ if the crypto instructions are available.
19383 (altivec_overloaded_builtins): Add support for overloaded power8
19386 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
19387 support for power8 crypto builtins.
19388 (builtin_function_type): Likewise.
19389 (altivec_init_builtins): Add support for builtins that take vector
19390 long long (V2DI) arguments.
19392 * config/rs6000/crypto.md: New file, define power8 crypto
19395 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19396 Pat Haugen <pthaugen@us.ibm.com>
19397 Peter Bergner <bergner@vnet.ibm.com>
19399 * doc/invoke.texi (Option Summary): Add power8 options.
19400 (RS/6000 and PowerPC Options): Likewise.
19402 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
19403 constraints.md instead of rs6000.h. Reorder w* constraints. Add
19404 wm, wn, wr documentation.
19406 * config/rs6000/constraints.md (wm): New constraint for VSX
19407 registers if direct move instructions are enabled.
19408 (wn): New constraint for no registers.
19409 (wq): New constraint for quad word even GPR registers.
19410 (wr): New constraint if 64-bit instructions are enabled.
19411 (wv): New constraint if power8 vector instructions are enabled.
19412 (wQ): New constraint for quad word memory locations.
19414 * config/rs6000/predicates.md (const_0_to_15_operand): New
19415 constraint for 0..15 for crypto instructions.
19416 (gpc_reg_operand): If VSX allow registers in VSX registers as well
19417 as GPR and floating point registers.
19418 (int_reg_operand): New predicate to match only GPR registers.
19419 (base_reg_operand): New predicate to match base registers.
19420 (quad_int_reg_operand): New predicate to match even GPR registers
19421 for quad memory operations.
19422 (vsx_reg_or_cint_operand): New predicate to allow vector logical
19423 operations in both GPR and VSX registers.
19424 (quad_memory_operand): New predicate for quad memory operations.
19425 (reg_or_indexed_operand): New predicate for direct move support.
19427 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
19428 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
19429 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
19430 (POWERPC_MASKS): Add power8 options.
19431 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
19434 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
19435 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
19437 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
19438 (-mpower8-fusion): New power8 options.
19439 (-mpower8-fusion-sign): Likewise.
19440 (-mpower8-vector): Likewise.
19441 (-mcrypto): Likewise.
19442 (-mdirect-move): Likewise.
19443 (-mquad-memory): Likewise.
19445 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
19446 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
19448 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
19449 (rs6000_debug_vector_unit): Add p8_vector.
19450 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
19451 definitions. Also print fusion state.
19452 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
19453 (rs6000_builtin_mask_calculate): Add power8 builtin support.
19454 (rs6000_option_override_internal): Add support for power8.
19455 (rs6000_common_init_builtins): Add debugging for skipped builtins
19456 if -mdebug=builtin.
19457 (rs6000_adjust_cost): Add power8 support.
19458 (rs6000_issue_rate): Likewise.
19459 (insn_must_be_first_in_group): Likewise.
19460 (insn_must_be_last_in_group): Likewise.
19461 (force_new_group): Likewise.
19462 (rs6000_register_move_cost): Likewise.
19463 (rs6000_opt_masks): Likewise.
19465 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
19466 power8 capable assembler, default to power7 options.
19467 (TARGET_DIRECT_MOVE): Likewise.
19468 (TARGET_CRYPTO): Likewise.
19469 (TARGET_P8_VECTOR): Likewise.
19470 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
19471 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
19472 (VECTOR_MEM_P8_VECTOR_P): Likewise.
19473 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
19474 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
19475 (TARGET_XSCVDPSPN): Likewise.
19476 (TARGET_XSCVSPDPN): Likewsie.
19477 (TARGET_SYNC_HI_QI): Likewise.
19478 (TARGET_SYNC_TI): Likewise.
19479 (MASK_CRYPTO): Likewise.
19480 (MASK_DIRECT_MOVE): Likewise.
19481 (MASK_P8_FUSION): Likewise.
19482 (MASK_P8_VECTOR): Likewise.
19483 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
19484 temporary used by some of the direct move instructions to get two FP
19485 temporary registers does not force creation of a stack frame.
19486 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
19487 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
19488 that any VSX registers are tieable, even if they are also an
19489 Altivec vector mode.
19490 (r6000_reg_class_enum): Add wm, wr, wv constraints.
19491 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
19492 (RS6000_BTM_CRYPTO): Likewise.
19493 (RS6000_BTM_COMMON): Likewise.
19495 * config/rs6000/rs6000.md (cpu attribute): Add power8.
19496 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
19497 (enum rs6000_vector): Add power8 vector support.
19499 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19503 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
19504 (any_sibcall_could_use_r3): this and handle indirect calls.
19505 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
19507 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19509 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
19511 2013-05-22 Richard Biener <rguenther@suse.de>
19513 PR middle-end/57349
19514 * profile.c (branch_prob): Do not split blocks that are
19515 abnormally receiving from ECF_RETURNS_TWICE functions.
19517 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19519 * recog.c (offsettable_address_addr_space_p): Fix calculation of
19520 address mode. Move pointer mode initialization to the same place.
19522 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
19524 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
19525 while it has any effect.
19527 2013-05-21 Easwaran Raman <eraman@google.com>
19529 PR tree-optimization/57322
19530 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
19531 UID of the statement added to the BB to be 1.
19533 2013-05-21 Jakub Jelinek <jakub@redhat.com>
19535 PR tree-optimization/57331
19536 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
19537 of conversion from pointer type to integral type with integer.
19539 2013-05-21 Martin Jambor <mjambor@suse.cz>
19542 * ipa-prop.c (ipa_read_node_info): Process param_used and
19543 controlled_uses in the same order as when writing.
19545 2013-05-21 Magnus Granberg <baldrick@free.fr>
19548 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
19550 2013-05-21 Richard Biener <rguenther@suse.de>
19552 PR tree-optimization/57318
19553 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
19554 estimate stmts with side-effects as likely eliminated.
19556 2013-05-21 Richard Biener <rguenther@suse.de>
19558 PR tree-optimization/57330
19559 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
19560 preserve the call stmts fntype.
19562 2013-05-21 Richard Biener <rguenther@suse.de>
19564 PR tree-optimization/57303
19565 * tree-ssa-sink.c (statement_sink_location): Improve killing
19566 stmt detection and properly handle self-assignments.
19568 2013-05-21 Christian Bruel <christian.bruel@st.com>
19570 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
19571 spanning registers. LEAF_REG_REMAP is supported only for contiguous
19572 registers. Set register size out of the PARALLEL loop.
19574 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
19577 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
19578 (*fmasf4, *fmasf4_media): New insns.
19580 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19582 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
19583 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
19584 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
19585 (mips_idiv_insns): Update the comments to say that the returned
19586 instruction counts are in units of BASE_INSN_LENGTH.
19587 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
19588 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
19589 using 2 rather than 4 as the length of indirect MIPS16 and
19590 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
19591 length of a NOP. Don't divide MIPS16 lengths by 2.
19592 (mips16_split_long_branches): Assume a branch is long if the
19593 length is greater than 4 rather than 8.
19594 * config/mips/mips.md (length): Give MIPS16 lengths directly,
19595 rather than multiplying them by 2. Multiply instruction counts
19596 by BASE_INSN_LENGTH rather than 4.
19597 (*jump_mips16, tls_get_tp_mips16_<mode>)
19598 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
19600 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19602 * config/mips/mips.md (extended_mips16): Remove branch case.
19603 (length): Remove duplicated extended_mips16 test.
19605 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19607 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
19609 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
19611 * recog.h (Recog_data): Rename to...
19612 (recog_data_d): ...this.
19613 (recog_data): Update accordingly.
19614 * recog.c (recog_data): Likewise.
19615 * reload.c (save_recog_data): Likewise.
19616 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
19617 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
19619 2013-05-17 Julian Brown <julian@codesourcery.com>
19621 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
19622 found in a REG_EQUAL note, invalidate it.
19624 2013-05-17 Easwaran Raman <eraman@google.com>
19626 * tree-ssa-reassoc.c (find_insert_point): New function.
19627 (insert_stmt_after): Likewise.
19628 (get_def_stmt): Likewise.
19629 (ensure_ops_are_available): Likewise.
19630 (not_dominated_by): Likewise.
19631 (rewrite_expr_tree): Do not move statements beyond what is
19632 necessary. Remove call to swap_ops_for_binary_stmt...
19633 (reassociate_bb): ... and move it here.
19634 (build_and_add_sum): Assign UIDs for new statements.
19635 (linearize_expr): Likewise.
19636 (do_reassoc): Renumber gimple statement UIDs.
19638 2013-05-17 Jan Hubicka <jh@suse.cz>
19640 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
19642 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
19643 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
19645 (output_weakrefs): Update.
19647 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
19648 Martin Jambor <mjambor@suse.cz>
19650 PR middle-end/57276
19651 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
19652 value that corresponds to the given aggval is found in values vector.
19654 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
19656 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
19657 sse, sse2, sse3, ssse3 and sse4a flags to options.
19659 2013-05-17 David Malcolm <dmalcolm@redhat.com>
19661 * gengtype-state.c: (s_expr_writer): New class, to handle
19662 prettifying of output layout of s-expressions.
19663 (state_writer): New class, to write out gtype.state.
19664 (state_written_type_count): Move this variable into member data of
19666 (s_expr_writer::s_expr_writer): New code: constructor for new class
19667 (state_writer::state_writer(): ditto
19668 (s_expr_writer::write_new_line): New function
19669 (s_expr_writer::write_any_indent): ditto
19670 (s_expr_writer::begin_s_expr): ditto
19671 (s_expr_writer::end_s_expr): ditto
19672 (write_state_fileloc): convert to method of state_writer...
19673 (state_writer:: write_state_fileloc): ...and use methods of
19674 s_expr_writer to write indentation into the gtype.state output file
19675 to visually represent the hierarchical structure of the list
19677 (write_state_fields): ditto, renaming to...
19678 (state_writer::write_state_fields)
19679 (write_state_a_string): ditto, renaming to...
19680 (state_writer::write_state_a_string)
19681 (write_state_string_option): ditto, renaming to...
19682 (state_writer::write_state_string_option)
19683 (write_state_type_option): ditto, renaming to...
19684 (state_writer::write_state_type_option)
19685 (write_state_nested_option): ditto, renaming to...
19686 (state_writer::write_state_nested_option)
19687 (write_state_option): ditto, renaming to...
19688 (state_writer::write_state_option)
19689 (write_state_options): ditto, renaming to...
19690 (state_writer::write_state_options)
19691 (write_state_lang_bitmap): ditto, renaming to...
19692 (state_writer::write_state_lang_bitmap)
19693 (write_state_version): ditto, renaming to...
19694 (state_writer::write_state_version)
19695 (write_state_scalar_type): ditto, renaming to...
19696 (state_writer::write_state_scalar_type)
19697 (write_state_string_type): ditto, renaming to...
19698 (state_writer::write_state_string_type)
19699 (write_state_undefined_type): ditto, renaming to...
19700 (state_writer::write_state_undefined_type)
19701 (write_state_struct_union_type): ditto, renaming to...
19702 (state_writer::write_state_struct_union_type)
19703 (write_state_struct_type): ditto, renaming to...
19704 (state_writer::write_state_struct_type)
19705 (write_state_user_struct_type): ditto, renaming to...
19706 (state_writer::write_state_user_struct_type)
19707 (write_state_lang_struct_type): ditto, renaming to...
19708 (state_writer::write_state_lang_struct_type)
19709 (write_state_param_struct_type): ditto, renaming to...
19710 (state_writer::write_state_param_struct_type)
19711 (write_state_pointer_type): ditto, renaming to...
19712 (state_writer::write_state_pointer_type)
19713 (write_state_array_type): ditto, renaming to...
19714 (state_writer::write_state_array_type)
19715 (write_state_gc_used): ditto, renaming to...
19716 (state_writer::write_state_gc_used)
19717 (write_state_common_type_content): ditto, renaming to...
19718 (state_writer::write_state_common_type_content)
19719 (write_state_type): ditto, renaming to...
19720 (state_writer::write_state_type)
19721 (write_state_pair_list): ditto, renaming to...
19722 (state_writer::write_state_pair_list)
19723 (write_state_pair): ditto, renaming to...
19724 (state_writer::write_state_pair)
19725 (write_state_typedefs): ditto, renaming to...
19726 (state_writer::write_state_typedefs)
19727 (write_state_structures): ditto, renaming to...
19728 (state_writer::write_state_structures)
19729 (write_state_param_structs): ditto, renaming to...
19730 (state_writer::write_state_param_structs)
19731 (write_state_variables): ditto, renaming to...
19732 (state_writer::write_state_variables)
19733 (write_state_srcdir): ditto, renaming to...
19734 (state_writer::write_state_srcdir)
19735 (write_state_files_list): ditto, renaming to...
19736 (state_writer::write_state_files_list)
19737 (write_state_languages): ditto, renaming to...
19738 (state_writer::write_state_languages)
19739 (write_state): create a state_writer instance and use it when
19740 writing out the state file
19742 2013-05-17 Mike Stump <mikestump@comcast.net>
19744 PR rtl-optimization/57304
19745 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
19746 accessing DF_REF_REAL_LOC.
19748 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19750 PR rtl-optimization/57281
19751 PR rtl-optimization/57300
19752 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
19753 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
19754 what the other splitter did if the registers are dead.
19756 2013-05-17 Richard Biener <rguenther@suse.de>
19758 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
19761 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19763 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
19766 2013-05-17 Marek Polacek <polacek@redhat.com>
19768 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
19769 length when doing non-zero store of storing '\0' to '\0'.
19771 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19773 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
19774 vect_external_def oprnd1 with loop_vinfo, try to emit
19775 optional cast, negation and and stmts on the loop preheader
19776 edge instead of into the pattern def seq.
19778 PR tree-optimization/57051
19779 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
19780 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
19782 2013-05-16 Nick Clifton <nickc@redhat.com>
19784 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
19785 (rl78_is_naked_func): New function.
19786 (rl78_expand_prologue): Skip prologue generation for naked functions.
19787 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
19788 * doc/extend.texi (naked): Add RL78 to the list of processors
19789 that supports this attribute.
19791 2013-05-16 Jeff Law <law@redhat.com>
19793 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
19795 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19797 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
19798 cache parameters using detect_caches_amd also for CYRIX,
19799 NSC and TM2 signatures.
19801 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19802 Dzianis Kahanovich <mahatma@eu.by>
19806 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
19807 VIA/Centaur processors and determine their cache parameters
19808 using detect_caches_amd.
19810 2013-05-16 Teresa Johnson <tejohnson@google.com>
19812 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
19813 (rtl_verify_edges): New function.
19814 (rtl_verify_bb_insns): Ditto.
19815 (rtl_verify_bb_pointers): Ditto.
19816 (rtl_verify_bb_insn_chain): Ditto.
19817 (rtl_verify_fallthru): Ditto.
19818 (rtl_verify_bb_layout): Ditto.
19819 (rtl_verify_flow_info_1): Outline checks into new functions.
19820 (rtl_verify_flow_info): Ditto.
19822 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
19824 * cfghooks.c (copy_bbs): Add update_dominance argument.
19825 * cfghooks.h (copy_bbs): Update prototype.
19826 * tree-cfg.c (gimple_duplicate_sese_region):
19827 Add update_dominance argument.
19828 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
19829 * tree-ssa-loop-ch.c (copy_loop_headers): Update
19830 gimple_duplicate_sese_region call.
19831 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
19832 Update copy_bbs call.
19833 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
19834 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
19836 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19838 * tree-vectorizer.h (NUM_PATTERNS): Increment.
19839 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
19840 vect_recog_rotate_pattern.
19841 (vect_recog_rotate_pattern): New function.
19843 2013-05-16 Jason Merrill <jason@redhat.com>
19845 * Makefile.in (LLINKER): New variable.
19846 (mostlyclean): Remove link mutex.
19847 * configure.ac: Handle --enable-link-mutex.
19848 * lock-and-run.sh: New script.
19850 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19853 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
19856 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19858 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
19860 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
19862 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
19863 * config/arm/arm.c (next_consecutive_mem): New function.
19864 (gen_movmem_ldrd_strd): Likewise.
19865 * config/arm/arm.md (movmemqi): Update condition and code.
19866 (unaligned_loaddi, unaligned_storedi): New patterns.
19868 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19870 * config.gcc: Obsolete *-*-solaris2.9*.
19871 * doc/install.texi (Specific, *-*-solaris2*): Document it.
19873 2013-05-16 Richard Biener <rguenther@suse.de>
19875 * passes.c (init_optimization_passes): Move pass_parallelize_loops
19876 earlier, after GRAPHITE transforms and IV canonicalization.
19878 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19880 * omp-low.c (extract_omp_for_data): For collapsed loops,
19881 if at least one of the loops is known at compile time to
19882 iterate zero times, set count to 0.
19883 (expand_omp_regimplify_p): New function.
19884 (expand_omp_for_generic): For collapsed loops, if at least
19885 one of the loops isn't known to iterate at least once,
19886 add runtime check with setting count to 0.
19887 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
19888 For unsigned types if it isn't known at compile time that
19889 the loop will iterate at least once, add runtime check to bypass
19890 the whole loop if initial condition isn't true.
19892 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
19894 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
19896 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
19898 PR middle-end/57286
19899 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
19900 transformations to avoid an infinite loop.
19902 2013-05-16 Marek Polacek <polacek@redhat.com>
19904 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
19906 2013-05-15 Leif Ekblad <leif@rdos.net>
19908 * config/i386/i386.c (ix86_decompose_address): Use
19909 DEFAULT_TLS_SEG_REG to access TLS segment register.
19910 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
19911 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
19912 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
19914 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
19917 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
19918 sibling calls to functions that would normally be lazily bound,
19919 unless $gp is call-clobbered.
19921 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
19923 * config/i386/i386.c (ix86_option_override_internal): Update
19924 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
19925 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
19926 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
19927 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
19929 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
19931 2013-05-15 Andreas Schwab <schwab@suse.de>
19933 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
19934 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
19937 2013-05-15 Teresa Johnson <tejohnson@google.com>
19939 * loop-unroll.c (report_unroll_peel): Check decision before
19940 emitting unroll/peel message.
19942 2013-05-15 Teresa Johnson <tejohnson@google.com>
19944 * function.h (has_bb_partition): New rtl_data flag.
19945 (bb_reorder_complete): Ditto.
19946 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
19947 instead of flag_reorder_blocks_and_partition.
19948 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
19949 with some enhancements.
19950 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
19951 * bb-reorder.c (connect_traces): Check for has_bb_partition
19952 instead of flag_reorder_blocks_and_partition.
19953 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
19954 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
19955 verify_hot_cold_block_grouping.
19956 (partition_hot_cold_basic_blocks): Set has_bb_partition.
19958 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19961 * config/arm/predicates.md (call_insn_operand): New predicate.
19962 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
19963 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
19964 if insn is not a tail call.
19965 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
19967 * config/arm/arm.h (enum reg_class): New caller save register class.
19968 (REG_CLASS_NAMES): Likewise.
19969 (REG_CLASS_CONTENTS): Likewise.
19970 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
19973 2013-05-15 Richard Biener <rguenther@suse.de>
19975 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
19976 of MSG_OPTIMIZED_LOCATIONS.
19977 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
19978 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
19980 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
19981 of MSG_OPTIMIZED_LOCATIONS.
19982 (execute_vect_slp): Likewise.
19983 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
19984 (vect_create_cond_for_alias_checks): Likewise.
19985 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
19986 (vect_recog_widen_mult_pattern): Likewise.
19987 (vect_recog_widen_sum_pattern): Likewise.
19988 (vect_recog_over_widening_pattern): Likewise.
19989 (vect_recog_widen_shift_pattern): Likewise.
19990 (vect_recog_vector_vector_shift_pattern): Likewise.
19991 (vect_recog_divmod_pattern): Likewise.
19992 (vect_recog_mixed_size_cond_pattern): Likewise.
19993 (vect_recog_bool_pattern): Likewise.
19994 (vect_pattern_recog_1): Likewise.
19996 2013-05-15 Martin Jambor <mjambor@suse.cz>
19998 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
19999 non-functions to builtin_unreachable.
20000 * ipa-inline-transform.c (inline_call): Do not assert estimates were
20001 correct when new direct edges were discovered.
20003 2013-05-15 Martin Jambor <mjambor@suse.cz>
20005 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
20006 header, print symbol order instead of node uid, print more information
20007 about indirect edge targets.
20008 (ipa_make_edge_direct_to_target): Print symbol order instead of node
20010 (ipa_make_edge_direct_to_target): Likewise.
20011 (remove_described_reference): Likewise.
20012 (propagate_controlled_uses): Likewise.
20013 (ipa_print_node_params): Also print symbol order.
20014 (ipcp_transform_function): Print symbol order instead of node uids.
20015 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
20016 (cgraph_get_create_real_symbol_node): Likewise.
20017 * ipa-cp.c (print_lattice): Likewise.
20018 (print_all_lattices): Likewise.
20019 (determine_versionability): Likewise.
20020 (initialize_node_lattices): Likewise.
20021 (estimate_local_effects): Likewise.
20022 (update_profiling_info): Likewise.
20023 (create_specialized_node): Likewise.
20024 (perhaps_add_new_callers): Likewise.
20025 (decide_about_value): Likewise.
20026 (decide_whether_version_node): Likewise.
20027 (identify_dead_nodes): Likewise.
20028 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
20029 (dump_inline_summary): Likewise.
20030 (estimate_node_size_and_time): Likewise.
20031 (inline_analyze_function): Likewise.
20032 * ipa-inline.c (report_inline_failed_reason): Likewise.
20033 (want_early_inline_function_p): Likewise.
20034 (edge_badness): Likewise.
20035 (update_edge_key): Likewise.
20036 (inline_small_functions): Likewise. Add dumping of order to two other
20038 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
20039 instead of node uids.
20040 (propagate_pure_const): Likewise.
20041 (propagate_pure_const): Likewise.
20042 * ipa-utils.c (dump_cgraph_node_set): Likewise.
20043 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
20044 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
20046 * tree-pretty-print.c (dump_function_header): Likewise.
20047 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
20048 Print symbol order instead of node uids.
20050 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20052 * config/s390/s390.c (s390_register_move_cost): Don't impose the
20053 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
20055 2013-05-15 Richard Biener <rguenther@suse.de>
20057 PR tree-optimization/57275
20058 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
20059 return value for fail to do runtime alias checks for gather loads.
20061 2013-05-15 Jan Hubicka <jh@suse.cz>
20065 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
20066 weakrefs are not external.
20067 (lto_symtab_merge_decls): Fix thinko when dealing with
20068 non-lto_symtab decls.
20069 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
20070 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
20071 * varpool.c (dump_varpool_node): Dump more flags.
20073 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20075 * config/i386/i386.c (processor_alias_table): Add instruction
20076 FSGSBASE for AMD bdver3 architecture.
20078 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20080 * tree.c (warn_deprecated_use): Print file:line using locus color.
20081 * diagnostic.c (diagnostic_report_current_module): Print file:line
20082 and file:line:column using locus color.
20084 2013-05-14 Mike Stump <mikestump@comcast.net>
20086 * gdbinit.in: Add __null.
20088 2013-05-14 Mike Stump <mikestump@comcast.net>
20090 * recog.h: Rename struct recog_data to Recog_data.
20091 * recog.c: Likewise.
20092 * reload.c (can_reload_into): Likewise.
20093 * config/picochip/picochip.c: Likewise.
20095 2013-05-14 Mike Stump <mikestump@comcast.net>
20097 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
20099 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
20101 * resource.h (struct resources): Remove unch_memory member.
20102 (CLEAR_RESOURCE): Don't clear unch_memory.
20103 * resource.c (mark_referenced_resources): Don't set it.
20104 (mark_set_resources): Likewise.
20105 (mark_target_live_regs): Don't clear it.
20106 (init_resource_info): Likewise.
20107 * reorg.c (resource_conflicts_p): Don't compare it.
20108 (redundant_insn): Don't set it.
20110 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
20112 * emit-rtl.c (next_label): Remove unused function.
20113 (skip_consecutive_labels, link_cc0_insns): Move to ...
20114 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
20115 only place where these functions are used, and make them static.
20117 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20119 * fold-const.c (fold_negate_expr): Handle vectors.
20120 (fold_truth_not_expr): Make it static.
20121 (fold_invert_truthvalue): New static function.
20122 (invert_truthvalue_loc): Handle vectors. Do not call
20123 fold_truth_not_expr directly.
20124 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
20125 <TRUTH_NOT_EXPR>: Do not cast to boolean.
20126 (fold_comparison): Handle vector constants.
20127 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
20128 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
20129 * tree.h (fold_truth_not_expr): Remove declaration.
20131 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
20133 * config/aarch64/aarch64-simd.md
20134 (aarch64_vcond_internal<mode>): Rename to...
20135 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
20136 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
20137 float modes. Clarify all iterator modes.
20138 (vcond<mode><mode>): Use new name for vcond expanders.
20139 (vcond<v_cmp_result><mode>): Likewise.
20140 (vcondu<mode><mode>: Likewise.
20141 * config/aarch64/iterators.md (VDQF_COND): New.
20143 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20146 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
20147 variable for the shift amount. Check that we shift by non-negative
20150 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
20153 * config/arm/arm.h (EPILOGUE_USES): Only return true
20154 for LR_REGNUM after epilogue_completed.
20156 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
20158 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
20159 is error_mark_node.
20161 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20164 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
20165 and Solaris 11+/x86 with gld.
20166 * configure: Regenerate.
20168 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20170 * expmed.c (expand_shift_1): Canonicalize rotates by
20171 constant bitsize / 2 to bitsize - 1.
20172 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
20173 case ROTATERT>: Likewise.
20176 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20178 * config/i386/i386.md (rotateinv): New code attr.
20179 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20180 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20181 roll $31, %eax, etc.
20183 2013-05-14 Richard Biener <rguenther@suse.de>
20185 PR middle-end/57235
20186 * tree-eh.c (sink_clobbers): Give up for successors with
20187 multiple predecessors and no virtual uses.
20189 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
20191 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
20192 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
20194 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20196 PR middle-end/57251
20197 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
20198 the case when both op0 and op1 have VOIDmode.
20200 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20202 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
20203 in multiply-accumulate mode.
20205 2013-05-13 Guozhi Wei <carrot@google.com>
20207 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
20209 2013-05-13 Kai Tietz <ktietz@redhat.com>
20212 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
20213 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
20214 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
20215 * config/i386/i386.c (ix86_option_override_internal): Likewise.
20216 (ix86_expand_prologue): Likewise.
20217 (ix86_expand_split_stack_prologue): Likewise.
20218 (legitimate_pic_address_disp_p): Likewise.
20219 (legitimize_pic_address): Likewise.
20220 (legitimize_tls_address): Likewise.
20221 (legitimize_pe_coff_symbol): Likewise.
20222 (output_pic_addr_const): Likewise.
20223 (construct_plt_address): Likewise.
20224 (ix86_expand_call): Likewise.
20225 (x86_output_mi_thunk): Likewise.
20226 (x86_function_profiler): Likewise.
20228 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
20230 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
20231 similar switch cases.
20232 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
20233 (aarch64_simd_mov_to_<mode>low): Delete.
20234 (aarch64_simd_mov_to_<mode>high): Delete.
20235 (move_lo_quad_<mode>): Add w<-r alternative.
20236 (aarch64_simd_move_hi_quad_<mode>): Likewise.
20237 (aarch64_simd_mov_from_*): Update type attribute.
20238 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
20241 2013-05-13 Jan Hubicka <jh@suse.cz>
20243 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
20244 * config/i386/i386.c (ix86_compute_frame_layout,
20245 ix86_expand_epilogue, emit_i387_cw_initialization,
20246 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
20247 ix86_local_alignment): Fix use of size/speed predicates.
20249 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20251 PR tree-optimization/45216
20252 PR tree-optimization/57157
20253 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
20254 the (-Y) & (B - 1) variant if OP is |.
20255 * expmed.c (expand_shift_1): For rotations by const0_rtx just
20256 return shifted. Use (-op1) & (prec - 1) as other_amount
20257 instead of prec - op1.
20259 2013-05-13 Martin Jambor <mjambor@suse.cz>
20261 PR middle-end/42371
20262 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
20263 (ipa_constant_data): New type.
20264 (ipa_jump_func): Use ipa_constant_data to hold information about
20265 constant jump functions.
20266 (ipa_get_jf_constant): Adjust to jump function type changes.
20267 (ipa_get_jf_constant_rdesc): New function.
20268 (ipa_param_descriptor): New field controlled_uses.
20269 (ipa_get_controlled_uses): New function.
20270 (ipa_set_controlled_uses): Likewise.
20271 * ipa-ref.h (ipa_find_reference): Declare.
20272 * ipa-prop.c (ipa_cst_ref_desc): New type.
20273 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
20275 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
20276 New parameter cs. Adjust all callers.
20277 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
20278 (remove_described_reference): New function.
20279 (jfunc_rdesc_usable): Likewise.
20280 (try_make_edge_direct_simple_call): Decrement controlled use count,
20281 attempt to remove reference if it hits zero.
20282 (combine_controlled_uses_counters): New function.
20283 (propagate_controlled_uses): Likewise.
20284 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
20285 (ipa_edge_duplication_hook): Duplicate reference descriptions.
20286 (ipa_print_node_params): Print described use counter.
20287 (ipa_write_jump_function): Adjust to jump function type changes.
20288 (ipa_read_jump_function): New parameter CS, pass it to
20289 ipa_set_jf_constant. Adjust caller.
20290 (ipa_write_node_info): Stream controlled use count
20291 (ipa_read_node_info): Likewise.
20292 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
20294 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
20295 count. Remove cloning-added reference if it reaches zero.
20296 * ipa-ref.c (ipa_find_reference): New function.
20298 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20300 * config/i386/i386.c (processor_target_table): Modified default
20301 alignment values for AMD BD and BT architectures.
20303 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
20305 * tree-vect-generic.c (uniform_vector_p): Move ...
20306 * tree.c (uniform_vector_p): ... here.
20307 * tree.h (uniform_vector_p): Declare it.
20308 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
20311 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20313 PR tree-optimization/57230
20314 * tree-ssa-strlen.c (handle_char_store): Record length for
20315 array store from STRING_CST.
20317 PR tree-optimization/57230
20318 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
20321 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
20323 * config/epiphany/epiphany.c (epiphany_init): Check size of
20324 NUM_MODES_FOR_MODE_SWITCHING.
20325 (epiphany_expand_prologue):
20326 Remove CONFIG_REGNUM initial value handling code.
20327 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
20328 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
20329 (emit_set_fp_mode, epiphany_mode_after): Likewise.
20330 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
20331 Don't return 1 for FP_MODE_NONE.
20332 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
20333 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
20334 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
20335 * config/epiphany/epiphany.md (save_config): New pattern.
20337 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
20339 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
20341 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
20343 * config/i386/i386.md (memory): Handle sseishft1.
20344 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
20345 (*vec_extractv2di_1): Ditto.
20347 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
20349 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
20352 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
20354 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
20355 Add mthumb/march=armv7-a multilib.
20356 Add mthumb/march=armv7-r multilib.
20357 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
20359 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
20361 * config/v850/t-rtems: Add more multilibs.
20363 2013-05-10 Richard Biener <rguenther@suse.de>
20365 PR tree-optimization/57214
20366 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
20367 not propagate from SSA names that occur in abnormal PHI nodes.
20369 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
20371 * stor-layout.c (element_precision): New function.
20372 * machmode.h (element_precision): Declare it.
20373 * tree.c (build_minus_one_cst): New function.
20374 (element_precision): Likewise.
20375 * tree.h (build_minus_one_cst): Declare new function.
20376 (element_precision): Likewise.
20377 * fold-const.c (operand_equal_p): Use element_precision.
20378 (fold_binary_loc): Handle vector types.
20379 * convert.c (convert_to_integer): Use element_precision.
20380 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
20383 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
20385 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
20386 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
20387 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
20388 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20389 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
20390 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
20391 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
20392 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20393 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
20394 (Uuw8): New constraints.
20395 (Usb4): Move into alphabetical order.
20396 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
20397 (sd8_operand, ub8_operand, uw8_operand): New predicates.
20398 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
20399 previously unnamed patterns.
20400 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
20401 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
20402 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
20403 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
20404 of set_attr_alternative/if_then_else. Use extended_mips16 instead
20405 of specific lengths.
20407 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20409 * config/i386/i386.md (rotateinv): New code attr.
20410 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20411 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20412 roll $31, %eax, etc.
20414 PR tree-optimization/45216
20415 PR tree-optimization/57157
20416 * tree-ssa-forwprop.c (simplify_rotate): New function.
20417 (ssa_forward_propagate_and_combine): Call it.
20419 2013-05-10 Richard Biener <rguenther@suse.de>
20421 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
20422 disable peeling when we version for aliasing.
20423 (vector_alignment_reachable_p): Honor explicit user alignment.
20424 (vect_supportable_dr_alignment): Likewise.
20425 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
20426 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
20427 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
20428 then peeling to arrange for the cost-model check to come first.
20430 2013-05-10 Alan Modra <amodra@gmail.com>
20432 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
20433 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
20434 * configure: Regenerate.
20436 2013-05-10 Alan Modra <amodra@gmail.com>
20439 * varasm.c (default_elf_select_section): Move !DECL_P check..
20440 (get_named_section): ..to here before calling get_section_name.
20442 (default_section_type_flags): Add DECL_P check.
20443 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
20444 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
20446 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
20448 * config/epiphany/epiphany.c (epiphany_expand_prologue):
20449 When using gen_stack_adjust_str with a register offset, add a
20450 REG_FRAME_RELATED_EXPR note.
20452 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
20454 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
20455 (*vec_extractv4si_zext_mem): Ditto.
20456 (*vec_extractv2di): Add 0->x and x->x alternatives.
20457 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
20458 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
20460 2013-05-09 Jason Merrill <jason@redhat.com>
20462 N3639 C++1y VLA support
20463 * gimplify.c (gimplify_vla_decl): Don't touch an existing
20466 * tree.c (build_constructor_va): New.
20467 * tree.h: Declare it.
20469 2013-05-09 Martin Jambor <mjambor@suse.cz>
20472 * gimple-fold.c (canonicalize_constructor_val): Call
20473 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
20475 2013-05-09 Jan Hubicka <jh@suse.cz>
20476 Richard Biener <rguenther@suse.de>
20479 * symtab.c (symtab_make_decl_local): Do not add private names.
20481 2013-05-09 Jan Hubicka <jh@suse.cz>
20484 * symtab.c (insert_to_assembler_name_hash): Handle clones.
20485 (unlink_from_assembler_name_hash): Likewise.
20486 (symtab_prevail_in_asm_name_hash, symtab_register_node,
20487 symtab_unregister_node, symtab_initialize_asm_name_hash,
20488 change_decl_assembler_name): Update.
20490 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
20492 * config/aarch64/aarch64.md: New movtf split.
20493 (*movtf_aarch64): Update.
20494 (aarch64_movdi_tilow): Handle TF modes and rename to
20495 aarch64_movdi_<mode>low.
20496 (aarch64_movdi_tihigh): Handle TF modes and rename to
20497 aarch64_movdi_<mode>high
20498 (aarch64_movtihigh_di): Handle TF modes and rename to
20499 aarch64_mov<mode>high_di
20500 (aarch64_movtilow_di): Handle TF modes and rename to
20501 aarch64_mov<mode>low_di
20502 (aarch64_movtilow_tilow): Remove spurious whitespace.
20503 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
20505 (aarch64_print_operand): Update.
20507 2013-05-09 Alan Modra <amodra@gmail.com>
20509 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
20511 * configure: Regenerate.
20513 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20515 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
20516 splitter preparation statements.
20517 * config/i386/sse.md (*vec_extract* splitters): Ditto.
20518 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
20521 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20523 * gimple-ssa-strength-reduction.c (count_candidates): Change
20524 return value to int.
20525 (analyze_candidates_and_replace): Change type of length to int.
20527 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20529 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
20530 (*vec_extract<mode>): Use VI12_128 mode iterator.
20531 (*vec_extract<mode>_mem): Ditto.
20532 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
20535 2013-05-08 Diego Novillo <dnovillo@google.com>
20540 2012-08-17 Diego Novillo <dnovillo@google.com>
20543 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
20544 * config.in: Regenerate.
20545 * configure: Regenerate.
20546 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
20548 2013-05-08 Jan Hubicka <jh@suse.cz>
20551 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
20552 * cgraph.h (symtab_node_base): Add unique_name.
20553 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20554 input_overwrite_node, input_varpool_node): Stream unique_name.
20555 * cgraphclones.c (cgraph_create_virtual_clone,
20556 cgraph_function_versioning): Set unique_name.
20557 * ipa.c (function_and_variable_visibility): Set unique_name.
20559 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20561 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
20562 (alloc_cand_and_find_basis): Restrict conditional candidate
20563 processing to CAND_MULTs.
20565 2013-05-08 Jan Hubicka <jh@suse.cz>
20568 lto-symtab.c (lto_symtab_symbol_p): New function.
20569 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
20570 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
20571 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
20572 Skip static symbols.
20574 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
20576 PR tree-optimization/57200
20577 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
20578 Only call inform if the preceding warning_at returns true.
20580 2013-05-07 Han Shen <shenhan@google.com>
20582 * cfgexpand.c (record_or_union_type_has_array_p): New function.
20583 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
20584 * common.opt (fstack-protector-strong): New option.
20585 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
20586 * doc/invoke.texi (Optimization Options): Document
20587 "-fstack-protector-strong".
20588 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
20590 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20592 * config/mips/mips.c (mips_machine_reorg2): Return 0.
20594 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
20596 * ira.c (update_equiv_regs): Add insn having equiv memory even if
20597 it is not lhs of the insn.
20598 (setup_reg_equiv): Remove insn having equiv memory which it is not
20600 * lra-constraints.c (process_address): Try to improve generation
20601 code for address base + disp.
20602 (lra_constraints): Make correct the code for checking insn setting
20603 up backward equivalence. Remove insn only if it is in the init
20605 * lra-eliminations.c (update_reg_eliminate): Change return value.
20606 (lra_eliminate): Use the result.
20608 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
20610 * config/i386/sse.md (ssescalarnummask): New mode attribute.
20611 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
20612 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
20613 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
20614 register target operands.
20615 (*vec_extractv8hi_sse2): New pattern.
20616 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
20617 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
20618 (*vec_extract<mode>_mem): New insn and split pattern.
20620 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
20622 * config/arm/arm.c (arm_asan_shadow_offset): New function.
20623 (TARGET_ASAN_SHADOW_OFFSET): Define.
20624 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
20625 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
20627 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20629 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
20630 (incr_vec_index): Return -1 if increment not found.
20631 (create_add_on_incoming_edge): Assert if increment not found.
20632 (record_increment): Limit number of increments recorded.
20633 (all_phi_incrs_profitable): Return false if an increment not found.
20634 (replace_profitable_candidates): Don't process increments that were
20636 (analyze_candidates_and_replace): Limit size of incr_vec.
20638 2013-05-07 Richard Biener <rguenther@suse.de>
20640 * calls.c (special_function_p): setjmp-like functions are leaf.
20641 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
20642 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
20644 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
20646 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
20647 (aarch64_simd_mov<mode>): New expander.
20648 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
20649 (aarch64_simd_mov_to_<mode>high): Likewise.
20650 (aarch64_simd_mov_from_<mode>low): Likewise.
20651 (aarch64_simd_mov_from_<mode>high): Likewise.
20652 (aarch64_dup_lane<mode>): Update.
20653 (aarch64_dup_lanedi): New instruction pattern.
20654 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
20655 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
20657 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20659 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
20660 (replace_mult_candidate): Remove unnecessary argument; remove
20661 unnecessary parameter from call to introduce_cast_before_cand.
20662 (replace_unconditional_candidate): Remove unnecessary parameter
20663 from call to replace_mult_candidate.
20664 (replace_conditional_candidate): Likewise.
20665 (insert_initializers): Use make_temp_ssa_name.
20666 (introduce_cast_before_cand): Remove unnecessary argument; use
20667 make_temp_ssa_name.
20668 (replace_one_candidate): Remove unnecessary argument; remove
20669 unnecessary parameter from calls to introduce_cast_before_cand.
20670 (replace_profitable_candidates): Remove unnecessary parameters
20671 from calls to replace_one_candidate.
20673 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20675 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
20676 phi def as possibly hiding a basis for a CAND_ADD whose operands
20677 have been commuted in the analysis.
20678 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
20680 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20682 * config/aarch64/aarch64.md
20683 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
20684 shift value between 0-4.
20686 2013-05-07 Richard Biener <rguenther@suse.de>
20688 * double-int.h (rshift): New overload.
20689 * double-int.c (rshift): New function.
20690 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
20691 (create_reference_ops_from_ref): Remove.
20692 (vn_reference_insert): Use shared ops for constructing the
20693 reference and copy it.
20695 2013-05-07 Richard Biener <rguenther@suse.de>
20697 PR middle-end/57190
20698 * tree-eh.c (sink_clobbers): Properly propagate
20699 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
20701 2013-05-07 Jakub Jelinek <jakub@redhat.com>
20703 PR tree-optimization/57149
20704 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
20705 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
20706 collect_phi_def_edges, execute_late_warn_uninitialized): Use
20707 uninit_undefined_value_p instead of ssa_undefined_value_p.
20710 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
20711 for modifier == EXPAND_INITIALIZER.
20713 2013-05-07 Anton Blanchard <anton@samba.org>
20715 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
20716 for powerpc64 little endian.
20717 * configure: Regenerate.
20719 2013-05-06 Graham Stott <grahams@btinternet.com>
20721 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
20722 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
20723 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
20724 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
20726 2013-05-06 Graham Stott <grahams@btinternet.com>
20728 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
20729 codes which allow non-lvalues.
20731 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
20733 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
20734 components are all 1s.
20735 (integer_minus_onep): New function.
20736 * tree.h (integer_minus_onep): Declare it.
20737 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
20738 integer_minus_onep instead of integer_all_onesp.
20740 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20743 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
20744 variations of these patterns.
20746 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20748 * config/i386/i386.md (isa): Add x64_sse4 member.
20749 (enabled): Handle x64_sse4.
20750 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
20751 instruction for 64bit SSE4_1 targets. Update insn attributes.
20752 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
20753 instruction for SSE4_1 targets. Update insn attributes.
20754 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20755 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
20756 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
20758 (*vec_extractv4si): Rename from *sse4_1_pextrd.
20759 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
20760 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
20762 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20765 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
20767 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
20769 * final.c (do_assembler_dialects): Don't handle curly braces and
20770 vertical bar escaped by % as dialect delimiters.
20771 (output_asm_insn): Print curly braces and vertical bar if escaped
20772 by % and ASSEMBLER_DIALECT defined.
20773 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
20774 * doc/tm.texi: Regenerated.
20776 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20778 * config/mips/mips.c: Include tree-pass.h.
20779 (mips_reorg): Split in pre- and post-dbr_schedule parts.
20780 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
20781 (pass_mips_machine_reorg2): New machine specific pass.
20782 (insert_pass_mips_machine_reorg2): New pass plugin definition.
20783 (mips_option_override): Register the new pass.
20784 * rtl.h (cleanup_barriers): Remove prototype.
20785 (dbr_schedule): Likewise.
20786 * jump.c (cleanup_barriers): Make static.
20787 * reorg.c (dbr_schedule): Likewise.
20789 2013-05-06 Richard Biener <rguenther@suse.de>
20791 PR tree-optimization/57185
20792 * tree-parloops.c (add_field_for_reduction): Handle anonymous
20793 SSA names properly.
20795 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20798 * config/i386/i386.c (add_parameter_dependencies): Add dependence
20799 between "first_arg" and "insn", not "last" and "insn".
20801 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20803 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
20804 (find_candidates_in_block): Re-enable slsr_process_phi.
20805 (create_phi_basis): Fix double counting of candidate adjustment.
20807 2013-05-06 Richard Biener <rguenther@suse.de>
20809 PR middle-end/57147
20810 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
20811 the edge is also fallthru, preserve it and just clear the
20813 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
20814 also complex, preserve that and just clear the fallthru flag.
20815 * tree-inline.c (update_ssa_across_abnormal_edges): Also
20816 update virtual operands.
20818 2013-05-06 Alan Modra <amodra@gmail.com>
20820 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
20821 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
20822 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
20823 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
20824 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20825 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
20827 2013-05-06 Alan Modra <amodra@gmail.com>
20829 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
20830 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
20831 (DEFAULT_ASM_ENDIAN): Define.
20832 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20833 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
20834 Update -K PIC clause from sysv4.h.
20835 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
20836 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
20838 2013-05-06 Alan Modra <amodra@gmail.com>
20840 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
20841 twice for little-endian.
20842 (ashrdi3_no_power, ashrdi3): Support little-endian.
20844 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20847 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
20848 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
20850 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
20851 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
20852 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
20855 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
20856 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20858 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
20859 * config/pa/pa.opt: Make mbig-switch a no-op.
20860 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
20861 (CASE_VECTOR_MODE): Always return SImode.
20862 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
20863 for the !TARGET_BIG_SWITCH case.
20864 * config/pa/pa-linux.h: Likewise.
20865 * config/pa/pa-openbsd.h: Likewise.
20866 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
20867 * config/pa/pa.md (short_jump): Remove define_insn.
20868 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
20869 (casesi0): Remove define_insn.
20870 (type): Remove btable_branch.
20871 (pa_combine_type): Likewise.
20872 (in_nullified_branch_delay): Likewise.
20873 (in_call_delay): Likewise.
20874 (define_delay): Likewise.
20875 (define_insn_reservation "Z3"): Likewise.
20876 (define_insn_reservation "Z4"): Likewise.
20877 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
20878 (pa_adjust_insn_length): Remove adjustment for btable branches.
20879 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
20882 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
20884 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20885 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
20886 Add m->r,x alternatives.
20887 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
20888 splitters using SWI48x mode iterator.
20889 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
20890 TARGET_64BIT. Add m->x alternative.
20891 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
20892 Add o->x alternative. Enable for TARGET_SSE.
20893 (sse_storeq): Remove expander.
20894 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
20895 with memory input operand.
20896 (*vec_extractv2di_1 splitter): New.
20897 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
20898 * config/i386/i386.md (ssevecmodelower): New mode attribute.
20900 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20902 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
20903 (INT_LOWPART): Delete.
20904 (extract_MB): Adjust.
20905 (extract_ME): Adjust.
20906 (print_operand): Adjust.
20908 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20910 * config/rs6000/predicates.md (reg_or_add_cint_operand,
20911 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20912 (reg_or_logical_cint_operand, easy_fp_constant,
20913 logical_const_operand): Delete "CONST_DOUBLE" case.
20914 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
20915 "HOST_BITS_PER_WIDE_INT == 64" test.
20916 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
20917 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
20918 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
20919 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
20921 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
20922 CONST_DOUBLE DImode/VOIDmode case.
20923 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
20924 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20925 CONST_DOUBLE VOIDmode case.
20926 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20927 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
20928 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
20929 Delete CONST_DOUBLE case.
20930 (splitters for mov FMOVE64 const_double): Delete
20931 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20932 "HOST_BITS_PER_WIDE_INT >= 64" test.
20933 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
20935 (mov DI const_double): Delete.
20937 2013-05-04 Jakub Jelinek <jakub@redhat.com>
20939 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
20940 on op shows all bits zero in mode of a lowpart subreg, return zero.
20942 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
20945 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
20946 to save TFmode registers and DImode to save TImode registers for
20947 caller save operations.
20948 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
20949 mark being partially clobbered since they only use the first
20952 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
20953 and TDmode only use the upper 64-bits of each VSX register.
20955 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20957 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
20958 (find_candidates_in_block): Disable slsr_process_phi.
20960 2013-05-03 Guozhi Wei <carrot@google.com>
20962 * coverage.c (coverage_obj_init): Move the construction of gcov
20964 (build_init_ctor): ... here.
20966 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20968 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
20969 (slsr_cand_d): Redefine def_phi.
20970 (stride_status, phi_adjust_status, count_phis_status): New enums.
20971 (find_phi_def): New.
20972 (find_basis_for_base_expr): New.
20973 (find_basis_for_candidate): Handle hidden bases.
20974 (alloc_cand_and_find_basis): Handle phi candidates.
20975 (slsr_process_phi): New.
20976 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
20977 (create_mul_imm_cand): Likewise.
20978 (create_add_ssa_cand): Exclude phi base candidates.
20979 (create_add_imm_cand): Likewise.
20980 (slsr_process_cast): Likewise.
20981 (slsr_process_copy): Likewise.
20982 (find_candidates_in_block): Handle phi candidates.
20983 (dump_candidate): Likewise.
20984 (unconditional_cands): Delete.
20985 (unconditional_cands_with_known_stride_p): Delete.
20986 (phi_dependent_cand_p): New.
20987 (cand_increment): Handle phi-dependent candidates.
20988 (replace_dependent): Delete.
20989 (replace_mult_candidate): New.
20990 (replace_unconditional_candidate): New.
20991 (incr_vec_index): Move to avoid forward reference.
20992 (create_add_on_incoming_edge): New.
20993 (create_phi_basis): New.
20994 (replace_dependents): Delete.
20995 (replace_conditional_candidate): New.
20996 (phi_add_costs): New.
20997 (replace_uncond_cands_and_profitable_phis): New.
20998 (record_increment): Handle phi adjustments.
20999 (record_phi_increments): New.
21000 (record_increments): Handle phi adjustments.
21001 (phi_incr_cost): New.
21002 (lowest_cost_path): Handle phis.
21003 (total_savings): Likewise.
21004 (analyze_increments): Likewise.
21005 (ncd_with_phi): New.
21006 (ncd_of_cand_and_phis): New.
21007 (nearest_common_dominator_for_cands): Handle phi increments.
21008 (all_phi_incrs_profitable): New.
21009 (replace_profitable_candidates): Handle phi-dependent candidates.
21010 (analyze_candidates_and_replace): Likewise.
21012 2013-05-03 Teresa Johnson <tejohnson@google.com>
21015 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
21016 do not exceed REG_BR_PROB_BASE.
21018 2013-05-03 Jeff Law <law@redhat.com>
21020 PR tree-optimization/57144
21021 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
21022 operand of the condition will bit into the new type when eliminating
21023 a cast feeding a condition.
21025 2013-05-03 Jakub Jelinek <jakub@redhat.com>
21027 PR rtl-optimization/57130
21028 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
21029 of COMPARE as in_code to the recursive call if needed.
21031 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
21033 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
21034 (enabled): Handle new members.
21035 * config/i386/sse.md (*vec_concatv2si): Merge from
21036 *vec_concatv2si_sse2 and vec_concatv2si_sse.
21037 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
21039 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
21041 PR tree-optimization/57027
21042 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
21043 for fnms opportunity, check we got the prerequisite kind
21044 of tree / gimple before using accessor functions.
21046 2013-05-03 Richard Biener <rguenther@suse.de>
21048 * double-int.h (lshift): New overload without precision
21049 and arith argument.
21050 (operator *=, operator +=, operator -=): Move ...
21051 * double-int.c (operator *=, operator +=, operator -=): ... here
21052 and implement more efficiently.
21053 (mul_double_with_sign): Remove.
21054 (lshift_double): Adjust to take unsinged shift argument, push
21055 dispatching code to callers.
21056 (mul_double_wide_with_sign): Add early out for callers that
21057 are not interested in high parts or overflow.
21058 (lshift): New function.
21059 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
21060 dispatch code here.
21061 (lrotate, rrotate): Use logical shifts.
21062 * expr.c (get_inner_reference): Use lshift.
21063 * fixed-value.c (do_fixed_divide): Likewise.
21064 * tree-dfa.c (get_ref_base_and_extent): Likewise.
21065 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
21066 (indirect_refs_may_alias_p): Likewise.
21067 (stmt_kills_ref_p_1): Likewise.
21069 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21071 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
21073 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21075 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
21076 scalar form of FABD instruction.
21078 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
21080 * lra-constraints.c (process_alt_operands): Add checking alt
21081 number to choose the best alternative.
21083 2013-05-02 Richard Biener <rguenther@suse.de>
21085 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
21086 bitmap and its handling.
21087 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
21089 2013-05-02 Richard Biener <rguenther@suse.de>
21091 PR middle-end/57140
21092 * tree-inline.c (copy_loops): Properly handle removed loops.
21093 (copy_cfg_body): Mark destination loops for fixup if source
21094 loops needed fixup.
21096 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
21099 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
21100 generating simple_return for naked functions.
21102 2013-05-02 Martin Jambor <mjambor@suse.cz>
21104 PR middle-end/56988
21105 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
21106 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
21108 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
21109 ipa_agg_replacement_value structures.
21110 (known_aggs_to_agg_replacement_list): Likewise.
21111 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
21112 (read_agg_replacement_chain): Likewise.
21113 (ipcp_transform_function): Also check that by_ref flags match.
21115 2013-05-02 Richard Biener <rguenther@suse.de>
21117 * graphds.h (struct graph): Add obstack member.
21118 * graphds.c (new_graph): Initialize obstack and allocate
21120 (add_edge): Allocate edge from the obstack.
21121 (free_graph): Free the obstack instead of all edges and vertices.
21123 2013-05-02 Teresa Johnson <tejohnson@google.com>
21125 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
21127 * cfg.c (update_bb_profile_for_threading): Ditto.
21128 * tree-inline.c (copy_bb): Ditto.
21129 (copy_edges_for_bb): Ditto.
21130 (initialize_cfun): Ditto.
21131 (copy_cfg_body): Ditto.
21132 (expand_call_inline): Ditto.
21133 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
21134 (estimate_node_size_and_time): Ditto.
21135 (inline_merge_summary): Ditto.
21136 * cgraphclones.c (cgraph_clone_edge): Ditto.
21137 (cgraph_clone_node): Ditto.
21138 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21139 (compute_trg_info): Ditto.
21141 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21143 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
21144 S reg when fp attribute set.
21145 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
21147 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21149 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
21151 (*and_one_cmplsi3_compare0_uxtw): Likewise.
21152 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
21153 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
21155 2013-05-02 Richard Biener <rguenther@suse.de>
21157 * tree-scalar-evolution.c (scev_info_hasher): Remove.
21158 (struct instantiate_cache_entry): New type.
21159 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
21160 (struct instantiate_cache_type): New type.
21161 (set_instantiated_value, get_instantiated_value): Remove.
21162 (get_instantiated_value_entry): New function.
21163 (instantiate_scev_name): Use the new cache and adjust.
21164 (instantiate_scev_poly): Adjust.
21165 (instantiate_scev_binary): Likewise.
21166 (instantiate_array_ref): Likewise.
21167 (instantiate_scev_convert): Likewise.
21168 (instantiate_scev_not): Likewise.
21169 (instantiate_scev_3): Likewise.
21170 (instantiate_scev_2): Likewise.
21171 (instantiate_scev_r): Likewise.
21172 (instantiate_scev): Likewise.
21173 (resolve_mixers): Likewise.
21175 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
21178 * lra-constraints.c (best_small_class_operands_num): Remove.
21179 (process_alt_operands): Remove small_class_operands_num. Take
21180 small classes operands into losers and only if the operand is not
21181 matched. Modify debugging output.
21182 (curr_insn_transform): Remove best_small_class_operands_num.
21185 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21187 * config/aarch64/aarch64-builtins.c
21188 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
21189 * config/aarch64/aarch64-simd-builtins.def
21190 (reduc_splus_): Add new modes.
21191 (reduc_uplus_): New.
21192 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
21193 (reduc_uplus_v4sf): Likewise.
21194 (reduc_splus_v4sf): Likewise.
21195 (aarch64_addv<mode>): Likewise.
21196 (reduc_uplus_<mode>): Likewise.
21197 (reduc_splus_<mode>): Likewise.
21198 (aarch64_addvv2di): Likewise.
21199 (reduc_uplus_v2di): Likewise.
21200 (reduc_splus_v2di): Likewise.
21201 (aarch64_addvv2si): Likewise.
21202 (reduc_uplus_v2si): Likewise.
21203 (reduc_splus_v2si): Likewise.
21204 (reduc_<sur>plus_<mode>): New.
21205 (reduc_<sur>plus_v2di): Likewise.
21206 (reduc_<sur>plus_v2si): Likewise.
21207 (reduc_<sur>plus_v4sf): Likewise.
21208 (aarch64_addpv4sf): Likewise.
21209 * config/aarch64/arm_neon.h
21210 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
21211 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
21212 add UNSPEC_SADDV, UNSPEC_UADDV.
21214 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
21216 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21218 * config/aarch64/arm_neon.h
21219 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
21221 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21223 * config/aarch64/aarch64-builtins
21224 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
21226 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21228 * config/aarch64/aarch64-simd-builtins.def
21229 (reduc_smax_): New.
21230 (reduc_smin_): Likewise.
21231 (reduc_umax_): Likewise.
21232 (reduc_umin_): Likewise.
21233 (reduc_smax_nan_): Likewise.
21234 (reduc_smin_nan_): Likewise.
21237 (smax): Update for V2SF, V4SF and V2DF modes.
21240 (smin_nan): Likewise.
21241 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
21242 (<su><maxmin><mode>3): ...This, refactor.
21243 (s<maxmin><mode>3): New.
21244 (<maxmin_uns><mode>3): Likewise.
21245 (reduc_<maxmin_uns>_<mode>): Refactor.
21246 (reduc_<maxmin_uns>_v4sf): Likewise.
21247 (reduc_<maxmin_uns>_v2si): Likewise.
21248 (aarch64_<fmaxmin><mode>: Remove.
21249 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
21251 (vmin<q>_f<32,64>): Likewise.
21252 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21254 (su): Add mappings for smax, smin, umax, umin.
21256 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21257 (FMAXMIN): Rename as...
21258 (FMAXMIN_UNS): ...This.
21260 (fmaxminv): Likewise.
21261 (fmaxmin): Likewise.
21263 (maxmin_uns_op): Likewise.
21265 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21267 * config/aarch64/arm_neon.h
21268 (vac<ge, gt><sd>_f<32, 64>): Rename to...
21269 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
21270 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
21272 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21274 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
21275 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
21277 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21279 * config/aarch64/aarch64-simd.md
21280 (vcond<mode>_internal): Handle special cases for constant masks.
21281 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
21282 (vcondu<mode><mode>): Likewise.
21283 (vcond<v_cmp_result><mode>): New.
21285 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21287 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
21288 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
21289 * config/aarch64/aarch64-simd-builtins.def
21290 (cmeq): Update to BUILTIN_VALLDI.
21295 * config/aarch64/arm_neon.h
21296 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
21297 to builtins or C as appropriate.
21299 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21301 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
21303 (cmhi): Rename to...
21305 * config/aarch64/aarch64-simd.md
21306 (simd_mode): Add SF.
21307 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
21308 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
21309 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
21310 (cstore<mode>_neg): ...This.
21311 * config/aarch64/iterators.md
21313 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
21314 (COMPARISONS): New.
21315 (UCOMPARISONS): Likewise.
21316 (optab): Add missing comparisons.
21322 (VCMP_S): Likewise.
21323 (VCMP_U): Likewise.
21324 (V_cmp_result): Add DF, SF modes.
21325 (v_cmp_result): Likewise.
21327 (vmtype): Likewise.
21328 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
21330 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
21332 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
21333 define_insn to define_insn_and_split.
21334 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
21335 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
21336 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
21337 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
21338 (thumb2_negscc): Likewise.
21340 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21342 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
21344 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21346 * config/arm/thumb2.md: Remove trailing whitespaces.
21348 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21350 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
21351 Use gen_int_mode rather than GEN_INT.
21353 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
21355 * value-prof.c (stream_in_histogram_value): Remove the strayed
21358 2013-04-30 Richard Biener <rguenther@suse.de>
21360 PR middle-end/57122
21361 * cfghooks.c (split_edge): Properly check for the loop latch edge.
21363 2013-04-30 Richard Biener <rguenther@suse.de>
21365 PR middle-end/57107
21366 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
21368 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
21370 PR rtl-optimization/56957
21371 PR rtl-optimization/57105
21372 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
21373 variable. Use just INSN_UID for determining whether an insn
21374 should be only disconnected from the insn stream.
21375 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
21377 2013-04-30 Jakub Jelinek <jakub@redhat.com>
21379 PR tree-optimization/57104
21380 * tsan.c (instrument_expr): Don't instrument accesses to
21381 DECL_HARD_REGISTER VAR_DECLs.
21383 2013-04-30 Richard Biener <rguenther@suse.de>
21385 * function.h (loops_for_fn): New inline function.
21386 (set_loops_for_fn): Likewise.
21387 * cfgloop.h (place_new_loop): Add struct function parameter.
21388 (get_loop): Likewise.
21389 (get_loops): Likewise.
21390 (number_of_loops): Likewise.
21391 (fel_next): Adjust.
21392 (fel_init): Likewise.
21393 * cfg.c (get_loop_copy): Adjust.
21394 * cfgloop.c (flow_loops_dump): Likewise.
21395 (record_loop_exits): Likewise.
21396 (verify_loop_structure): Likewise.
21397 * cfgloopanal.c (mark_irreducible_loops): Likewise.
21398 (estimate_reg_pressure_cost): Likewise.
21399 (mark_loop_exit_edges): Likewise.
21400 * cfgloopmanip.c (place_new_loop): Likewise.
21401 (add_loop): Likewise.
21402 (duplicate_loop): Likewise.
21403 * graph.c (draw_cfg_nodes): Likewise.
21404 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
21405 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
21406 (extract_affine_chrec): Likewise.
21407 (build_scop_iteration_domain): Likewise.
21408 * graphite.c (graphite_initialize): Likewise.
21409 * ira-build.c (create_loop_tree_nodes): Likewise.
21410 (more_one_region_p): Likewise.
21411 (rebuild_regno_allocno_maps): Likewise.
21412 (mark_loops_for_removal): Likewise.
21413 (mark_all_loops_for_removal): Likewise.
21414 (remove_unnecessary_regions): Likewise.
21415 (ira_build): Likewise.
21416 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21417 * loop-init.c (fix_loop_structure): Likewise.
21418 (gate_rtl_move_loop_invariants): Likewise.
21419 (gate_rtl_unswitch): Likewise.
21420 (gate_rtl_unroll_and_peel_loops): Likewise.
21421 (rtl_doloop): Likewise.
21422 * lto-streamer-in.c (input_cfg): Likewise.
21423 * lto-streamer-out.c (output_cfg): Likewise.
21424 * modulo-sched.c (sms_schedule): Likewise.
21425 * predict.c (tree_estimate_probability): Likewise.
21426 (tree_estimate_probability_driver): Likewise.
21427 (estimate_loops): Likewise.
21428 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
21429 (move_sese_region_to_fn): Likewise.
21430 (debug_loop_num): Likewise.
21431 * tree-chrec.c (chrec_evaluate): Likewise.
21432 (hide_evolution_in_other_loops_than_loop): Likewise.
21433 (chrec_component_in_loop_num): Likewise.
21434 (reset_evolution_in_loop): Likewise.
21435 (evolution_function_is_invariant_rec_p): Likewise.
21436 * tree-if-conv.c (main_tree_if_conversion): Likewise.
21437 * tree-inline.c (copy_loops): Likewise.
21438 (copy_cfg_body): Likewise.
21439 (tree_function_versioning): Likewise.
21440 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
21441 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
21443 (add_to_evolution_1): Likewise.
21444 (scev_const_prop): Likewise.
21445 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
21446 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
21447 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
21448 (tree_ssa_lim_initialize): Likewise.
21449 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
21450 (verify_loop_closed_ssa): Likewise.
21451 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
21452 (tree_ssa_loop_im): Likewise.
21453 (tree_ssa_loop_unswitch): Likewise.
21454 (tree_vectorize): Likewise.
21455 (check_data_deps): Likewise.
21456 (tree_ssa_loop_ivcanon): Likewise.
21457 (tree_ssa_loop_bounds): Likewise.
21458 (tree_complete_unroll): Likewise.
21459 (tree_complete_unroll_inner): Likewise.
21460 (tree_parallelize_loops): Likewise.
21461 (tree_ssa_loop_prefetch): Likewise.
21462 (tree_ssa_loop_ivopts): Likewise.
21463 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
21464 * tree-vectorizer.c (vectorize_loops): Likewise.
21466 2013-04-29 Mike Frysinger <vapier@gentoo.org>
21468 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
21469 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
21470 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
21471 with EABI_LINK_SPEC.
21473 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21476 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
21479 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
21482 * lra-constraints.c (process_alt_operands): Discourage a bit more
21483 using memory for pseudos. Print cost dump for alternatives.
21484 Modify cost values for conflicts with early clobbers.
21485 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
21487 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21490 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
21492 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21494 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
21495 from/to S register.
21496 (movdi_aarch64): Support LDR/STR from/to D register.
21498 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21500 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
21501 or higher by default.
21503 2013-04-29 Richard Biener <rguenther@suse.de>
21505 PR middle-end/57075
21506 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
21507 even if not adding abnormal edges for calls that can make
21510 2013-04-29 Richard Biener <rguenther@suse.de>
21512 PR middle-end/57103
21513 * tree-cfg.c (move_stmt_op): Fix condition under which to update
21515 (move_stmt_r): Remove redundant checking.
21517 2013-04-29 Teresa Johnson <tejohnson@google.com>
21520 * basic-block.h (apply_scale): New function.
21521 (apply_probability): Use apply_scale.
21522 * gimple-streamer-in.c (input_bb): Ditto.
21523 * lto-streamer-in.c (input_cfg): Ditto.
21524 * lto-cgraph.c (merge_profile_summaries): Ditto.
21525 * tree-optimize.c (execute_fixup_cfg): Ditto.
21526 * tree-inline.c (copy_bb): Update comment to use apply_scale.
21527 (copy_edges_for_bb): Ditto.
21528 (copy_cfg_body): Ditto.
21530 2013-04-29 Tom de Vries <tom@codesourcery.com>
21532 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
21533 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
21534 (tail_merge_optimize): Handle current_loops == NULL.
21536 2013-04-26 Jeff Law <law@redhat.com>
21538 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
21539 (simplify_cond_using_ranges): Generalize code to simplify
21540 COND_EXPRs where one argument is a constant and the other
21541 is an SSA_NAME created by an integral type conversion.
21543 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21545 * config/arm/arm.md (store_minmaxsi): Use only when
21546 optimize_insn_for_size_p.
21548 2013-04-29 Christian Bruel <christian.bruel@st.com>
21551 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
21553 2013-04-29 Richard Biener <rguenther@suse.de>
21555 PR middle-end/57089
21556 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
21557 loop tree make sure to schedule a fixup for the child as well.
21558 (expand_omp_for_generic): Properly add loops.
21559 (expand_omp_for_static_nochunk): Likewise.
21560 (expand_omp_for_static_chunk): Likewise.
21561 (expand_omp_for): For the degenerate case fixup loops.
21562 (expand_omp_sections): Fix default bb placement in loops.
21563 (expand_omp_atomic_pipeline): Properly add loops.
21565 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21567 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
21569 2013-04-29 Tom de Vries <tom@codesourcery.com>
21571 * tree-ssa-tail-merge.c: Update header comment.
21573 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21575 * config/aarch64/arm_neon.h
21576 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
21577 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
21578 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
21579 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
21580 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
21581 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
21582 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
21583 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
21585 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21587 * config/aarch64/aarch64-simd.md
21588 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
21589 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
21590 fix_trunc, fixuns_trunc.
21591 (ftrunc<VDQF:mode>2): New.
21592 * config/aarch64/iterators.md (optab): Add fix, fixuns.
21593 (fix_trunc_optab): New.
21595 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21597 * config/aarch64/aarch64-builtins.c
21598 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
21599 iceilf, lround, iroundf.
21601 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21604 * config/i386/i386.h (enum ix86_tune_indices)
21605 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
21606 New, split from X86_TUNE_INTER_UNIT_MOVES.
21607 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
21608 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
21609 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
21610 (TARGET_INTER_UNIT_MOVES): Remove.
21611 * config/i386/i386.c (initial_ix86_tune_features): Update.
21612 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
21613 (ix86_expand_convert_uns_didf_sse): Use
21614 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21615 (ix86_expand_vector_init_one_nonzero): Ditto.
21616 (ix86_expand_vector_init_interleave): Ditto.
21617 (inline_secondary_memory_needed): Return true for moves from SSE class
21618 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
21619 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
21620 * config/i386/constraints.md (Yi, Ym): Depend on
21621 TARGET_INTER_UNIT_MOVES_TO_VEC.
21622 (Yj, Yn): New constraints.
21623 * config/i386/i386.md (*movdi_internal): Change constraints of
21624 operand 1 from Yi to Yj and from Ym to Yn.
21625 (*movsi_internal): Ditto.
21626 (*movdf_internal): Ditto.
21627 (*movsf_internal): Ditto.
21628 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
21629 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21630 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
21631 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
21632 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
21633 * config/i386/sse.md (movdi_to_sse): Ditto.
21634 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
21635 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
21636 TARGET_INTER_UNIT_MOVES.
21637 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
21638 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
21639 instead of TARGET_INTER_UNIT_MOVES.
21640 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
21641 operand 1 from Yi to Yj and from Ym to Yn.
21643 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21645 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
21646 (float_truncate_hi_): Likewise.
21647 (float_extend_lo_): Likewise.
21648 (float_truncate_lo_): Likewise.
21649 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
21650 (aarch64_float_extend_lo_v2df): Likewise.
21651 (vec_unpacks_hi_v4sf): Likewise.
21652 (aarch64_float_truncate_lo_v2sf): Likewise.
21653 (aarch64_float_truncate_hi_v4sf): Likewise.
21654 (vec_pack_trunc_v2df): Likewise.
21655 (vec_pack_trunc_df): Likewise.
21657 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21659 * config/aarch64/aarch64-builtins.c
21660 (aarch64_fold_builtin): Fold float conversions.
21661 * config/aarch64/aarch64-simd-builtins.def
21662 (floatv2si, floatv4si, floatv2di): New.
21663 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
21664 * config/aarch64/aarch64-simd.md
21665 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
21666 * config/aarch64/iterators.md (FLOATUORS): New.
21667 (optab): Add float, floatuns.
21668 (su_optab): Likewise.
21670 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21672 * config/aarch64/aarch64-builtins.c
21673 (aarch64_builtin_vectorized_function): Use new names for
21675 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
21676 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
21677 (fcvtzu): Split as...
21678 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
21679 (fcvtas): Split as...
21680 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
21681 (fcvtau): Split as...
21682 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
21683 (fcvtps): Split as...
21684 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
21685 (fcvtpu): Split as...
21686 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
21687 (fcvtms): Split as...
21688 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
21689 (fcvtmu): Split as...
21690 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
21691 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
21692 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
21693 (lfrintnusf, lfrintnudf): Likewise.
21694 * config/aarch64/aarch64-simd.md
21695 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
21697 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
21698 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
21699 (fcvt_pattern): Likewise.
21701 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21703 * config/aarch64/aarch64-simd.md
21704 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
21705 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
21707 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21709 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
21710 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
21711 (vrnd<a,m,n,p>_f32): Implement using builtins.
21712 (vrnd<i,x><q>_f<32, 64>): New.
21714 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21716 * config/aarch64/aarch64-builtins.c
21717 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
21718 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
21719 (frintz): Rename to...
21721 (frintp): Rename to...
21723 (frintm): Rename to...
21725 (frinti): Rename to...
21726 (nearbyint): ...this.
21727 (frintx): Rename to...
21729 (frinta): Rename to...
21731 * config/aarch64/aarch64-simd.md
21732 (aarch64_frint<frint_suffix><mode>): Delete.
21733 (<frint_pattern><mode>2): Convert to insn.
21734 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
21735 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
21736 (frint_pattern): Likewise.
21737 (frint_suffix): Likewise.
21739 2013-04-29 Richard Biener <rguenther@suse.de>
21741 PR tree-optimization/57081
21742 * loop-init.c: Include tree-flow.h.
21743 (loop_optimizer_finalize): Free number of iteration estimates.
21744 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
21746 2013-04-29 Jakub Jelinek <jakub@redhat.com>
21748 PR tree-optimization/57083
21749 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
21750 non-singleton shift count range, zero extend low_bound for uns case.
21752 * config/i386/predicates.md (general_vector_operand): New predicate.
21753 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
21754 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
21755 if they aren't nonimmediate operands. If their original values
21756 satisfy const_vector_equal_evenodd_p, don't shift them.
21757 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
21758 predicates. For the SSE4.1 case force operands[{1,2}] into registers
21759 if not nonimmediate_operand.
21760 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
21761 instead of register_operand.
21762 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
21764 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
21766 * stor-layout.c (finalize_size_functions): Allocate a structure and
21767 reset cfun before dumping the functions.
21769 2013-04-27 Jakub Jelinek <jakub@redhat.com>
21771 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
21774 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
21775 use xop_pmacsdqh if uns_p.
21776 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
21777 the immediate rotate count.
21779 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
21781 * rtl.h (struct rtx_def): Add comment for field jump.
21782 (LRA_SUBREG_P): New macro.
21783 * recog.c (register_operand): Check LRA_SUBREG_P.
21784 * lra.c (lra): Add note at the end of RTL code. Align non-empty
21786 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
21787 (lra_final_code_change): Skip subreg change for operators.
21788 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
21789 if there are no operand changes.
21790 * lra-constraints.c (curr_insn_set): New.
21791 (match_reload): Set LRA_SUBREG_P.
21792 (emit_spill_move): Ditto.
21793 (check_and_process_move): Use curr_insn_set. Process only single
21794 set insns. Don't initialize sec_mem_p and change_p.
21795 (simplify_operand_subreg): Use LRA_SUBREG_P.
21796 (reg_in_class_p): New function.
21797 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
21798 of #ifdef. Add code to remove cycling.
21799 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
21800 non-null disp. Reload inner instead of disp when base and index
21801 are null. Try to put lo_sum into register.
21802 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
21803 (check_and_process_move): Move code for move cost check to
21804 simple_move_p. Remove equiv_substitution.
21805 (simple_move_p): New function.
21806 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
21807 curr_insn_set. Call check_and_process_move only for single set
21808 insns. Use the new function. Move call of check_and_process_move
21809 after operand equiv substitution and address process.
21811 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21814 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
21815 with nonlocal goto receivers or returns twice calls, ignore
21816 unininitialized values from abnormal edges to nl goto receiver
21817 or returns twice call.
21819 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21821 PR tree-optimization/57051
21822 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
21823 and VEC_RSHIFT_EXPR if shift count is a multiple of element
21826 2013-04-26 Richard Biener <rguenther@suse.de>
21828 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
21829 (expand_omp_taskreg): Likewise. Mark loops for fixup.
21830 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
21831 (fixup_loop_arrays_after_move): New function.
21832 (move_sese_region_to_fn): Properly outline the loop tree parts
21833 of the SESE region.
21835 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
21837 * config/i386/i386.md (type, unit): Fix long lines.
21839 2013-04-26 Richard Biener <rguenther@suse.de>
21841 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
21842 (lto-streamer-out.o): Likewise.
21843 * cfgloop.c (init_loops_structure): Export, add struct function
21844 argument and adjust.
21845 (flow_loops_find): Adjust.
21846 * cfgloop.h (enum loop_estimation): Add EST_LAST.
21847 (init_loops_structure): Declare.
21848 * lto-streamer-in.c: Include cfgloop.h.
21849 (input_cfg): Input the loop tree.
21850 * lto-streamer-out.c: Include cfgloop.h.
21851 (output_cfg): Output the loop tree.
21852 (output_struct_function_base): Do not drop PROP_loops.
21854 2013-03-26 Richard Biener <rguenther@suse.de>
21856 * tree-cfg.c (execute_build_cfg): Build the loop tree.
21857 (pass_build_cfg): Provide PROP_loops.
21858 (move_sese_region_to_fn): Remove loops that are outlined into fn
21860 * tree-inline.c: Include cfgloop.h.
21861 (initialize_cfun): Do not drop PROP_loops.
21862 (copy_loops): New function.
21863 (copy_cfg_body): Copy loop structure.
21864 (tree_function_versioning): Initialize destination loop tree.
21865 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
21866 (pass_parallelize_loops): Do IL verification.
21867 * loop-init.c (loop_optimizer_init): Fixup loops if required.
21868 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
21869 the CFG make sure we fixup loops as well.
21870 * tree-ssa-tail-merge.c: Include cfgloop.h.
21871 (replace_block_by): When merging loop latches mark loops for fixup.
21872 * lto-streamer-out.c (output_struct_function_base): Drop
21873 PROP_loops for now.
21874 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
21875 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
21876 * ipa-split.c: Include cfgloop.h.
21877 (split_function): Add the new return block to the loop tree root.
21878 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
21879 whether we have removed the forwarder block.
21880 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
21881 * cfgloop.h (place_new_loop): Declare.
21882 * cfgloopmanip.c (place_new_loop): Export.
21883 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
21884 (tree-switch-conversion.o): Likewise.
21885 (tree-complex.o): Likewise.
21886 (tree-inline.o): Likewise.
21887 (tree-ssa-tailmerge.o): Likewise.
21888 (ipa-split.o): Likewise.
21889 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
21890 (tree-ssa-copy.o): Likewise.
21891 * tree-switch-conversion.c: Include cfgloop.h
21892 (process_switch): If we emit a bit-test cascade, schedule loops
21894 * tree-complex.c: Include cfgloop.h.
21895 (expand_complex_div_wide): Properly add new basic-blocks to loops.
21896 * asan.c: Include cfgloop.h.
21897 (create_cond_insert_point): Properly add new basic-blocks to
21898 loops, schedule loop fixup.
21899 * cfgloop.c (verify_loop_structure): Check that looks are not
21901 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
21903 (expand_omp_for_generic): Likewise.
21904 (expand_omp_sections): Likewise.
21905 (expand_omp_atomic_pipeline): Schedule loops for fixup.
21906 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
21907 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
21908 is initialized, not when loops are present.
21909 * tree-parloops.c (parallelize_loops): Remove checking here.
21910 * passes.c (init_optimization_passes): Schedule a copy-propagation
21911 pass before complete unrolling of inner loops.
21913 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21915 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
21916 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
21917 (colorize_init): Add argument to _WIN32 version.
21918 * toplev.c: Include diagnostic-color.h.
21919 (process_options): Default to -fdiagnostics-color=auto if
21920 GCC_COLORS env var is in the environment.
21921 * common.opt (fdiagnostics-color=): Add Var and Init.
21922 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
21923 env var is in the environment, the default is auto rather than never.
21925 * diagnostic.h (file_name_as_prefix): Add context argument.
21926 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
21927 the string as locus.
21928 * langhooks.c (lhd_print_error_function): Adjust caller.
21930 2013-04-25 Lawrence Crowl <crowl@google.com>
21932 * var-tracking.c (shared_hash_def::htab):
21933 Change type to hash_table. Update dependent calls and types.
21935 2013-04-25 Lawrence Crowl <crowl@google.com>
21937 * Makefile.in: Update as needed below.
21939 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
21940 Move declaration to after the type's method definitons.
21942 * attribs.c (htab_t scoped_attributes::attribute_hash):
21943 Change type to hash_table. Update dependent calls and types.
21945 * bitmap.c (htab_t bitmap_desc_hash):
21946 Change type to hash_table. Update dependent calls and types.
21948 * cselib.c (htab_t cselib_hash_table):
21949 Change type to hash_table. Update dependent calls and types.
21951 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
21952 (hash_string_slot_node): Move implementation into lto-streamer.h
21953 struct string_slot_hasher.
21954 (eq_string_slot_node): Likewise.
21956 * data-streamer-out.c: Update output_block::string_hash_table
21957 dependent calls and types.
21959 * dwarf2cfi.c (htab_t trace_index):
21960 Change type to hash_table. Update dependent calls and types.
21962 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
21963 Change type to hash_table. Update dependent calls and types.
21964 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
21965 (htab_t optimize_external_refs::map): Likewise.
21966 (htab_t output_comp_unit::extern_map): Likewise.
21967 (htab_t output_comdat_type_unit::extern_map): Likewise.
21968 (htab_t output_macinfo::macinfo_htab): Likewise.
21969 (htab_t optimize_location_lists::htab): Likewise.
21970 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
21972 * except.c (htab_t ehspec_hash_type):
21973 Change type to hash_table. Update dependent calls and types.
21974 (assign_filter_values::ttypes): Likewise.
21975 (assign_filter_values::ehspec): Likewise.
21976 (sjlj_assign_call_site_values::ar_hash): Likewise.
21977 (convert_to_eh_region_ranges::ar_hash): Likewise.
21979 * gcse.c (htab_t pre_ldst_table):
21980 Change type to hash_table. Update dependent calls and types.
21982 * ggc-common.c (htab_t saving_htab):
21983 Change type to hash_table. Update dependent calls and types.
21984 (htab_t loc_hash): Likewise.
21985 (htab_t ptr_hash): Likewise.
21986 (call_count): Rename ggc_call_count.
21987 (call_alloc): Rename ggc_call_alloc.
21988 (loc_descriptor): Rename make_loc_descriptor.
21989 (add_statistics): Rename ggc_add_statistics.
21991 * ggc-common.c (saving_htab):
21992 Change type to hash_table. Update dependent calls and types.
21994 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
21995 (push_gimplify_context): Likewise.
21996 (pop_gimplify_context): Likewise.
21997 (struct gimple_temp_hash_elt): Added.
21998 (struct gimplify_hasher): Likewise.
21999 (struct gimplify_ctx.temp_htab):
22000 Change type to hash_table. Update dependent calls and types.
22002 * gimple-fold.c: Include gimplify-ctx.h.
22004 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
22005 Change type to hash_table. Update dependent calls and types.
22006 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
22007 avoid potential global name collision.
22009 * gimplify.c: Include gimplify-ctx.h.
22010 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
22011 (htab_t gimplify_ctx::temp_htab):
22012 Update dependent calls and types for new type hash_table.
22013 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
22014 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
22016 * gimplify-ctx.h: New.
22017 (struct gimple_temp_hash_elt): Move from gimplify.c.
22018 (class gimplify_hasher): New.
22019 (struct gimplify_ctx): Move from gimple.h.
22020 (htab_t gimplify_ctx::temp_htab):
22021 Change type to hash_table. Update dependent calls and types.
22023 * graphite-clast-to-gimple.c: Include graphite-htab.h.
22024 (htab_t ivs_params::newivs_index):
22025 Change type to hash_table. Update dependent calls and types.
22026 (htab_t ivs_params::params_index): Likewise.
22027 (htab_t print_generated_program::params_index): Likewise.
22028 (htab_t gloog::newivs_index): Likewise.
22029 (htab_t gloog::params_index): Likewise.
22031 * graphite.c: Include graphite-htab.h.
22032 4htab_t graphite_transform_loops::bb_pbb_mapping):
22033 Change type to hash_table. Update dependent calls and types.
22035 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
22036 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
22037 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
22039 * graphite-dependences.c: Include graphite-htab.h.
22040 (loop_is_parallel_p): Change hash table type of parameter.
22042 * graphite-htab.h: New.
22043 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
22044 (extern find_pbb_via_hash): Move from graphite-poly.h.
22045 (extern loop_is_parallel_p): Move from graphite-poly.h.
22046 (extern get_loop_body_pbbs): Move from graphite-poly.h.
22048 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
22049 (extern loop_is_parallel_p): Move to graphite-htab.h.
22050 (extern get_loop_body_pbbs): Move to graphite-htab.h.
22052 * haifa-sched.c (htab_t delay_htab):
22053 Change type to hash_table. Update dependent calls and types.
22054 (htab_t delay_htab_i2): Likewise.
22056 * ira-color.c (htab_t allocno_hard_regs_htab):
22057 Change type to hash_table. Update dependent calls and types.
22059 * ira-costs.c (htab_t cost_classes_htab):
22060 Change type to hash_table. Update dependent calls and types.
22062 * loop-invariant.c (htab_t merge_identical_invariants::eq):
22063 Change type to hash_table. Update dependent calls and types.
22065 * loop-iv.c (htab_t bivs):
22066 Change type to hash_table. Update dependent calls and types.
22068 * loop-unroll.c (htab_t opt_info::insns_to_split):
22069 Change type to hash_table. Update dependent calls and types.
22070 (htab_t opt_info::insns_with_var_to_expand): Likewise.
22072 * lto-streamer.h (struct string_slot): Move from data-streamer.h
22073 (struct string_slot_hasher): New.
22074 (htab_t output_block::string_hash_table):
22075 Change type to hash_table. Update dependent calls and types.
22077 * lto-streamer-in.c (freeing_string_slot_hasher): New.
22078 (htab_t file_name_hash_table):
22079 Change type to hash_table. Update dependent calls and types.
22081 * lto-streamer-out.c: Update output_block::string_hash_table dependent
22084 * lto-streamer.c (htab_t tree_htab):
22085 Change type to hash_table. Update dependent calls and types.
22087 * omp-low.c: Include gimplify-ctx.h.
22089 * passes.c (htab_t name_to_pass_map):
22090 Change type to hash_table. Update dependent calls and types.
22091 (pass_traverse): Rename to passes_pass_traverse.
22093 * plugin.c (htab_t event_tab):
22094 Change type to hash_table. Update dependent calls and types.
22096 * postreload-gcse.c (htab_t expr_table):
22097 Change type to hash_table. Update dependent calls and types.
22098 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
22100 * sese.c (debug_rename_map_1): Make extern.
22101 (htab_t copy_bb_and_scalar_dependences::rename_map):
22102 Change type to hash_table. Update dependent calls and types.
22104 * sese.h (extern debug_rename_map): Move to .c file.
22106 * store-motion.c (htab_t store_motion_mems_table):
22107 Change type to hash_table. Update dependent calls and types.
22109 * trans-mem.c (htab_t tm_new_mem_hash):
22110 Change type to hash_table. Update dependent calls and types.
22112 * tree-browser.c (htab_t TB_up_ht):
22113 Change type to hash_table. Update dependent calls and types.
22115 * tree-cfg.c (htab_t discriminator_per_locus):
22116 Change type to hash_table. Update dependent calls and types.
22118 * tree-complex.c: Include tree-hasher.h
22119 (htab_t complex_variable_components):
22120 Change type to hash_table. Update dependent calls and types.
22122 * tree-eh.c (htab_t finally_tree):
22123 Change type to hash_table. Update dependent calls and types.
22125 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
22126 struct int_tree_hasher.
22127 (extern int_tree_map_eq): Likewise.
22128 (uid_decl_map_hash): Removed.
22129 (extern decl_tree_map_eq): Likewise.
22131 * tree-hasher.h: New.
22132 (struct int_tree_hasher): New.
22133 (typedef int_tree_htab_type): New.
22135 * tree-inline.c: Include gimplify-ctx.h.
22137 * tree-mudflap.c: Include gimplify-ctx.h.
22139 * tree-parloops.c: Include tree-hasher.h.
22140 (htab_t eliminate_local_variables_stmt::decl_address):
22141 Change type to hash_table. Update dependent calls and types.
22142 (htab_t separate_decls_in_region::decl_copies): Likewise.
22144 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
22145 Change type to hash_table. Update dependent calls and types.
22147 * tree-sra.c (candidates):
22148 Change type to hash_table. Update dependent calls and types.
22150 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
22152 (int_tree_map_hash): Likewise.
22154 * tree-ssa-dom.c (htab_t avail_exprs):
22155 Change type to hash_table. Update dependent calls and types.
22157 * tree-ssa-live.c (var_map_base_init::tree_to_index):
22158 Change type to hash_table. Update dependent calls and types.
22160 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
22161 Change type to hash_table. Update dependent calls and types.
22163 * tree-ssa-phiopt.c (seen_ssa_names):
22164 Change type to hash_table. Update dependent calls and types.
22166 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
22167 Change type to hash_table. Update dependent calls and types.
22169 * tree-ssa-uncprop.c (equiv):
22170 Change type to hash_table. Update dependent calls and types.
22172 2013-04-25 Jakub Jelinek <jakub@redhat.com>
22174 PR rtl-optimization/57003
22175 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
22176 call note_stores with kill_clobbered_value callback again after
22177 killing regs_invalidated_by_call.
22179 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22181 * config/aarch64/aarch64-simd.md
22182 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
22183 (aarch64_simd_bsl<mode>): Likewise.
22184 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
22186 2013-04-25 Marek Polacek <polacek@redhat.com>
22188 PR tree-optimization/57066
22189 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
22191 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22193 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
22195 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22197 * config/aarch64/aarch64-builtins.c
22198 (aarch64_fold_builtin): New.
22199 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
22200 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
22201 * config/aarch64/aarch64-simd-builtins.def (abs): New.
22202 * config/aarch64/arm_neon.h
22203 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
22205 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22206 Tejas Belagod <tejas.belagod@arm.com>
22208 * config/aarch64/aarch64-builtins.c
22209 (aarch64_gimple_fold_builtin): New.
22210 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
22211 * config/aarch64/aarch64-simd-builtins.def (addv): New.
22212 * config/aarch64/aarch64-simd.md (addpv4sf): New.
22213 (addvv4sf): Update.
22214 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
22216 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22218 * config/aarch64/aarch64.md
22219 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
22221 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22223 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
22224 (*ngcsi_uxtw): New pattern.
22226 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22227 Julian Brown <julian@codesourcery.com>
22229 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
22230 (TB_DREG): Add T_V4HF.
22231 (v4hf_UP): New macro.
22232 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22233 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22234 Handle initialisation of V4HF. Adjust initialisation of reinterpret
22236 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22237 (arm_vector_mode_supported_p): Handle V4HF.
22238 (arm_mangle_map): Handle V4HFmode.
22239 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
22240 * config/arm/arm_neon_builtins.def: Add entries for
22241 vcvtv4hfv4sf, vcvtv4sfv4hf.
22242 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
22243 (neon_vcvtv4hfv4sf): Likewise.
22244 * config/arm/neon-gen.ml: Handle half-precision floating point
22246 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
22247 * config/arm/arm_neon.h: Regenerate.
22248 * config/arm/neon.ml (type elts): Add F16.
22249 (type vectype): Add T_float16x4, T_floatHF.
22250 (type vecmode): Add V4HF.
22251 (type features): Add Requires_FP_bit feature.
22252 (elt_width): Handle F16.
22253 (elt_class): Likewise.
22254 (elt_of_class_width): Likewise.
22255 (mode_of_elt): Refactor.
22256 (type_for_elt): Handle F16, fix error messages.
22257 (vectype_size): Handle T_float16x4.
22258 (vcvt_sh): New function.
22259 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
22260 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
22261 (string_of_mode): Handle V4HF.
22262 * doc/arm-neon-intrinsics.texi: Regenerate.
22264 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22266 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
22267 format specifier in 'X' case.
22269 2013-04-25 Alan Modra <amodra@gmail.com>
22272 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
22273 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
22274 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
22275 Repeat for many other rotate/shift and mask patterns using subregs.
22276 Name lshiftrt insns.
22277 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
22278 on WORDS_BIG_ENDIAN.
22280 2013-04-25 Alan Modra <amodra@gmail.com>
22282 * config.gcc: Support little-endian powerpc-linux targets.
22283 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
22284 (LINK_OS_LINUX_SPEC): Define.
22285 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
22286 Preserve MASK_LITTLE_ENDIAN.
22287 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
22288 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
22289 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
22290 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
22291 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
22292 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
22293 Correct fp word order for little-endian. Don't shift toc entries
22294 smaller than a word for little-endian.
22295 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
22296 (bswapdi2 splits): Correct low-part subreg for little-endian.
22297 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
22298 low/high where such is correct only for be.
22299 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
22300 little-endian for -mcall-aixdesc.
22302 2013-04-25 Alan Modra <amodra@gmail.com>
22304 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
22305 replace_equiv_address_nv.
22307 2013-04-25 Alan Modra <amodra@gmail.com>
22309 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
22311 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22314 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22315 * rtl.h (struct rtx_def): ...
22317 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22319 PR rtl-optimizations/57046
22320 * lra-constraints (split_reg): Set up lra_risky_transformations_p
22321 for multi-reg splits.
22323 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
22325 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
22327 2013-04-24 Sterling Augustine <saugustine@google.com>
22329 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
22330 (comp_dir_string, debug_str_dwo_section): New.
22331 (DEBUG_STR_DWO_SECTION): Rename to ...
22332 (DEBUG_DWO_STR_SECTION): ... this.
22333 (DEBUG_NORM_STR_SECTION): Delete.
22334 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
22335 (DEBUG_STR_DWO_SECTION_FLAGS): New.
22336 (find_AT_string): Move most logic to ...
22337 (find_AT_string_in_table): ... here. New.
22338 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
22339 add_skeleton_AT_string. Delete logic.
22340 (output_skeleton_debug_sections): Remove call to
22341 add_top_level_skeleton_die_attrs.
22342 (add_comp_dir_attribute): Move logic to comp_dir_string.
22343 (dwarf2out_init): Initialize debug_str_dwo_section.
22344 (output_indirect_string): Call find_string_form.
22345 (output_indirect_strings): Rewrite.
22346 (prune_unused_types): Empty skeleton_debug_str_hash.
22347 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
22348 (dwarf2out_finish): Call output_indirect_strings.
22350 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22352 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
22354 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22356 * rtl.h (struct rtx_def): Add comment for field jump.
22357 (LRA_SUBREG_P): New macro.
22358 * recog.c (register_operand): Check LRA_SUBREG_P.
22359 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22361 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22362 (lra_final_code_change): Skip subreg change for operators.
22363 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22364 if there are no operand changes.
22365 * lra-constraints.c (curr_insn_set): New.
22366 (match_reload): Set LRA_SUBREG_P.
22367 (emit_spill_move): Ditto.
22368 (check_and_process_move): Use curr_insn_set. Process only single
22369 set insns. Don't initialize sec_mem_p and change_p.
22370 (simplify_operand_subreg): Use LRA_SUBREG_P.
22371 (reg_in_class_p): New function.
22372 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22373 of #ifdef. Add code to remove cycling.
22374 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22375 non-null disp. Reload inner instead of disp when base and index
22376 are null. Try to put lo_sum into register.
22377 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22378 (check_and_process_move): Move code for move cost check to
22379 simple_move_p. Remove equiv_substitution.
22380 (simple_move_p): New function.
22381 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22382 curr_insn_set. Call check_and_process_move only for single set
22383 insns. Use the new function. Move call of check_and_process_move
22384 after operand equiv substitution and address process.
22386 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
22388 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
22389 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
22390 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
22392 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22394 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
22396 2013-04-24 Marek Polacek <polacek@redhat.com>
22398 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
22399 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
22400 (select_loops_exit_conditions): Likewise.
22401 (number_of_iterations_for_all_loops): Likewise.
22402 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
22403 (scev_analysis): Likewise.
22405 2013-04-02 Catherine Moore <clm@codesourcery.com>
22406 Chao-ying Fu <fu@mips.com>
22408 * config/mips/micromips.md (jraddiusp): New pattern.
22409 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
22410 instruction if possible.
22412 2013-04-24 Alan Modra <amodra@gmail.com>
22414 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
22416 2013-04-24 Julian Brown <julian@codesourcery.com>
22417 Chung-Lin Tang <cltang@codesourcery.com>
22419 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
22420 dependency behavior in enumeration type DIE generation. Add TODO note
22421 to comments about future DW_FORM_sdata/udata re-work of related code.
22423 2013-04-23 Lawrence Crowl <crowl@google.com>
22425 * Makefile.in: Update as needed below.
22427 * hash-table.h (class hash_table):
22428 Correct many methods with parameter types compare_type to the correct
22429 value_type. (Correct code was unlikely to notice the change.)
22430 (hash_table::elements_with_deleted) New.
22431 (class hashtable::iterator): New.
22432 (hashtable::begin()): New.
22433 (hashtable::end()): New.
22434 (FOR_EACH_HASH_TABLE_ELEMENT): New.
22436 * statistics.c (statistics_hashes):
22437 Change type to hash_table. Update dependent calls and types.
22439 * tree-into-ssa.c (var_infos):
22440 Change type to hash_table. Update dependent calls and types.
22442 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
22443 Change type to hash_table. Update dependent calls and types.
22445 * tree-ssa-loop-im.c (struct mem_ref.refs):
22446 Change type to hash_table. Update dependent calls and types.
22448 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
22449 Change type to hash_table. Update dependent calls and types.
22451 * tree-ssa-sccvn.c (vn_tables_s::nary):
22452 Change type to hash_table. Update dependent calls and types.
22453 (vn_tables_s::phis): Likewise.
22454 (vn_tables_s::references): Likewise.
22456 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
22457 (vn_reference_eq): Update parameter and return types.
22459 * tree-ssa-structalias.c (pointer_equiv_class_table):
22460 Change type to hash_table. Update dependent calls and types.
22461 (location_equiv_class_table): Likewise.
22463 * tree-vect-data-refs.c: Consequential changes for making
22464 peeling a hash_table.
22466 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
22467 (destroy_loop_vec_info): Dependent hash_table update.
22469 * tree-vectorizer.h (peeling_htab):
22470 Change type to hash_table. Update dependent calls and types.
22472 2013-04-23 Shiva Chen <shiva0217@gmail.com>
22474 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
22475 to check the register content is equal or not.
22476 * lra-constraints.c (match_reload): Use lra_assign_reg_val
22477 to assign register content record.
22478 * lra-eliminations.c (update_reg_eliminate): Use
22479 lra_update_reg_val_offset to update register content offset.
22480 * lra-int.h (struct lra_reg): Add offset member.
22481 (lra_reg_val_equal_p): New static inline function.
22482 (lra_update_reg_val_offset): New static inline function.
22483 (lra_assign_reg_val): New static inline function.
22484 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
22485 to assign register content record.
22486 (initialize_lra_reg_info_element): Initial offset to zero.
22488 2013-04-23 Catherine Moore <clm@codesourcery.com>
22490 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
22491 operands. Record compression.
22493 2013-04-23 Xinliang David Li <davidxl@google.com>
22495 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
22497 2013-04-23 Richard Biener <rguenther@suse.de>
22499 PR middle-end/57036
22500 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
22501 parameter, only add abnormal goto edges from the copied body
22502 if the call could perform abnormal gotos.
22503 (copy_cfg_body): Adjust.
22505 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
22507 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
22509 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
22511 * coretypes.h (gimple_stmt_iterator): Add struct to make
22514 2013-04-23 Richard Biener <rguenther@suse.de>
22516 PR tree-optimization/57026
22517 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
22518 from SSA names occuring in abnormal PHI nodes.
22520 2013-04-22 Andi Kleen <ak@linux.intel.com>
22522 * lto/lto.c (print_lto_report_1): Fix LTO report names.
22524 2013-04-22 Andi Kleen <ak@linux.intel.com>
22526 * lto/lto.c (print_lto_report_1): Declare early.
22527 (read_cgraph_and_symbols): Call print_lto_report_1 early.
22529 2013-04-22 Andi Kleen <ak@linux.intel.com>
22531 * common.opt (-flto-report-wpa): Add.
22532 * doc/invoke.texi (-flto-report-wpa): Add.
22533 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
22536 2013-04-22 Xinliang David Li <davidxl@google.com>
22538 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
22539 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
22540 * Makefile.in: New dependency
22542 David Daney <ddaney.cavm@gmail.com>
22544 * configure.ac (gcc_cv_as_micromips_support): Use the
22545 --fatal-warnings option.
22546 * configure: Regenerate.
22548 2013-04-22 Marek Polacek <polacek@redhat.com>
22551 * tsan.c (instrument_expr): Don't instrument expression
22552 in case its size is zero.
22554 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
22558 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22560 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22562 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22564 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
22565 (gimple_stmt_iterator): New typedef.
22566 * gimple.h (gimple_stmt_iterator): Rename to...
22567 (gimple_stmt_iterator_d): ... This.
22568 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
22569 trees be valid for GIMPLE and GENERIC.
22570 (TARGET_GIMPLE_FOLD_BUILTIN): New.
22571 * gimple-fold.c (gimple_fold_call): Call target hook
22572 gimple_fold_builtin.
22573 * hooks.c (hook_bool_gsiptr_false): New.
22574 * hooks.h (hook_bool_gsiptr_false): New.
22575 * target.def (fold_stmt): New.
22576 * doc/tm.texi: Regenerate.
22578 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
22581 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
22582 a set sp if no stack realignment.
22584 2013-04-22 Nick Clifton <nickc@redhat.com>
22586 * config.gcc (tilegx-linux): Extend extra_objs rather than
22588 (tilepro-linux): Likewise.
22590 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22592 * config/aarch64/aarch64-builtins.c
22594 (CF0, CF1, CF2, CF3, CF4, CF10): New.
22595 (VAR<1-12>): Add MAP parameter.
22596 (BUILTIN_*): Likewise.
22597 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
22598 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
22599 (aarch64_ushl_n<mode>): Likewise.
22600 (aarch64_sshr_n<mode>): Likewise.
22601 (aarch64_ushr_n<mode>): Likewise.
22602 (aarch64_<maxmin><mode>): Likewise.
22603 (aarch64_sqrt<mode>): Likewise.
22604 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
22605 (vshr<q>_n_*): Likewise.
22607 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22609 * config/aarch64/aarch64-builtins.c
22610 (aarch64_simd_builtin_type_mode): Handle SF types.
22612 (BUILTIN_GPF): Define.
22613 (aarch64_init_simd_builtins): Handle SF types.
22614 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
22615 (frecps): Likewise.
22616 (frecpx): Likewise.
22617 * config/aarch64/aarch64-simd.md
22618 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
22619 (aarch64_frecpe<mode>): New.
22620 (aarch64_frecps<mode>): Likewise.
22621 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
22622 (v8type): Add frecp<esx>.
22623 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
22624 (aarch64_frecps<mode>): Likewise.
22625 * config/aarch64/iterators.md (FRECP): New.
22626 (frecp_suffix): Likewise.
22627 * config/aarch64/arm_neon.h
22628 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
22630 2013-04-22 Christian Bruel <christian.bruel@st.com>
22633 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
22634 (REG_CLASS_NAMES): Idem.
22635 (REG_CLASS_CONTENTS): Idem.
22636 (REGCLASS_HAS_FP_REG): Idem.
22637 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
22638 (sh_conditional_register_usage): Idem.
22640 2013-04-21 Jeff Law <law@redhat.com>
22642 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
22643 (ssa_forward_propagate_and_combine): Use it.
22645 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22647 * lra.c: Update the flow chart diagram.
22649 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22651 PR rtl-optimization/56847
22652 * lra-constraints.c (process_alt_operands): Discourage alternative
22653 with non-matche doffsettable memory constraint fro memory with
22656 2013-04-19 Richard Biener <rguenther@suse.de>
22658 PR tree-optimization/56982
22659 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
22661 * gimplify.c (gimplify_call_expr): Notice special calls.
22662 (gimplify_modify_expr): Likewise.
22663 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
22664 abnormal control flow receivers.
22665 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
22666 in the same way as cfun->has_nonlocal_labels.
22667 (gimple_purge_dead_abnormal_call_edges): Likewise.
22668 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
22669 receivers start a basic-block.
22671 2013-04-19 Richard Biener <rguenther@suse.de>
22673 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
22675 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
22676 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
22677 (SLP_TREE_LOAD_PERMUTATION): Add.
22678 (vect_transform_slp_perm_load): Adjust prototype.
22679 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
22680 (vect_free_slp_instance): Likewise.
22681 (vect_create_new_slp_node): Likewise.
22682 (vect_supported_slp_permutation_p): Remove.
22683 (vect_slp_rearrange_stmts): Adjust.
22684 (vect_supported_load_permutation_p): Likewise. Inline
22685 vect_supported_slp_permutation_p here.
22686 (vect_analyze_slp_instance): Compute load permutations per
22687 slp node instead of per instance.
22688 (vect_get_slp_defs): Adjust.
22689 (vect_transform_slp_perm_load): Likewise.
22690 (vect_schedule_slp_instance): Remove redundant code.
22691 (vect_schedule_slp): Remove hack for PR56270, add it ...
22692 * tree-vect-stmts.c (vectorizable_load): ... here, do not
22693 CSE loads for SLP. Adjust.
22695 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22697 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
22698 spelling in two comments.
22700 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22703 * config/arm/arm.c (load_multiple_sequence): Require SP
22704 as base register for loads if SP is in the register list.
22706 2013-04-19 Martin Jambor <mjambor@suse.cz>
22708 PR tree-optimization/56718
22709 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
22710 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
22711 and made public. Adjusted all callers.
22712 (ipa_intraprocedural_devirtualization): New function.
22713 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
22714 (ipa_intraprocedural_devirtualization): Likewise.
22715 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
22717 2013-04-19 Richard Biener <rguenther@suse.de>
22719 PR tree-optimization/57000
22720 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
22722 2013-04-19 Terry Guo <terry.guo@arm.com>
22724 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
22726 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
22727 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
22728 (cortex_m4_fmacs): Use new reservations.
22729 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
22731 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
22733 PR rtl-optimization/56999
22734 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
22736 (lra_coalesce): Remove split_origin_bitmap and related code.
22737 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
22738 ranges if necessary.
22740 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
22742 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
22744 (ix86_expand_call): Remove clobbered_registers array and use
22745 x86_64_ms_sysv_extra_clobbered_registers instead.
22746 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
22748 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
22750 * config/i386/i386.md (*call_rex64_ms_sysv): Use
22751 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
22752 (*call_value_rex64_ms_sysv): Ditto.
22754 2013-04-18 Cary Coutant <ccoutant@google.com>
22756 * dwarf2out.c (output_pubnames): Check die_perennial_p of
22757 parent instead of die_mark.
22759 2013-04-18 Diego Novillo <dnovillo@google.com>
22761 * gimple.c (create_gimple_tmp): New.
22762 (get_expr_type): New.
22763 (build_assign): New.
22764 (build_type_cast): New.
22765 * gimple.h (enum ssa_mode): Define.
22766 (gimple_seq_set_location): New.
22767 * asan.c (build_check_stmt): Change some gimple_build_* calls
22768 to use build_assign and build_type_cast.
22770 2013-04-18 Richard Biener <rguenther@suse.de>
22772 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
22773 handle negative step. Remove redundant checks.
22774 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
22775 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
22776 for negative step and grouped loads fail to vectorize.
22778 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22780 * emit-rtl.c (reset_insn_used_flags): New function.
22781 (reset_all_used_flags): Use it.
22782 (verify_insn_sharing): New function.
22783 (verify_rtl_sharing): Fix verification for SEQUENCEs.
22785 2013-04-18 Jakub Jelinek <jakub@redhat.com>
22787 PR tree-optimization/56984
22788 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
22789 and (x >> M) >= N don't register any assertion if N << M is the
22792 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22794 * lower-subreg.c (resolve_simple_move): If called self-recursive,
22795 do not delete_insn insns that have not yet been emitted, only
22796 unlink them with remove_insn.
22797 * df-scan.c (df_insn_delete): Revert r197492.
22799 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
22801 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
22802 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
22804 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22806 * config/arm/arm.md (movsicc_insn): Convert define_insn into
22807 define_insn_and_split.
22808 (and_scc,ior_scc,negscc): Likewise.
22809 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
22811 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22813 * config/arm/arm.c (use_return_insn): Return 0 for targets that
22814 can benefit from using a sequence of LDRD instructions in epilogue
22815 instead of a single LDM instruction.
22817 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
22820 * doc/extend.texi: Fix typo.
22822 2013-04-17 Richard Biener <rguenther@suse.de>
22824 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
22825 (vect_build_slp_tree): ... here.
22826 (vect_build_slp_tree_1): Compute which stmts of the SLP group
22827 match. Remove special-casing of mismatched complex loads.
22828 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
22829 re-try the match with swapped commutative operands.
22830 (vect_supported_load_permutation_p): Remove special-casing of
22831 mismatched complex loads.
22832 (vect_analyze_slp_instance): Adjust.
22834 2013-04-17 Richard Biener <rguenther@suse.de>
22836 PR rtl-optimization/56921
22837 * cfgloop.h (struct loop): Add simple_loop_desc member.
22838 (struct niter_desc): Mark with GTY(()).
22839 (simple_loop_desc): Do not use aux field but simple_loop_desc.
22840 * loop-iv.c (get_simple_loop_desc): Likewise.
22841 (free_simple_loop_desc): Likewise.
22844 2013-04-16 Richard Biener <rguenther@suse.de>
22846 PR rtl-optimization/56921
22847 * loop-init.c (pass_rtl_move_loop_invariants): Add
22848 TODO_do_not_ggc_collect to todo_flags_finish.
22849 (pass_rtl_unswitch): Same.
22850 (pass_rtl_unroll_and_peel_loops): Same.
22851 (pass_rtl_doloop): Same.
22853 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
22855 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
22856 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
22857 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
22859 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
22860 * tree-streamer.c (record_common_node): Adjust reference in comment.
22862 2013-04-17 Terry Guo <terry.guo@arm.com>
22864 * config/arm/cortex-m4.md: Add a new bypass.
22866 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22868 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
22870 (*subs_<optab><mode>_multp2): New pattern.
22871 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22872 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22874 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22876 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
22877 (*subs_mul_imm_<mode>): New pattern.
22879 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
22882 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
22883 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
22884 (vsx_movti_32bit): Change j->wa to O->wa.
22886 2013-04-16 Richard Biener <rguenther@suse.de>
22888 PR rtl-optimization/56921
22889 * loop-init.c (pass_rtl_move_loop_invariants): Add
22890 TODO_do_not_ggc_collect to todo_flags_finish.
22891 (pass_rtl_unswitch): Same.
22892 (pass_rtl_unroll_and_peel_loops): Same.
22893 (pass_rtl_doloop): Same.
22895 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
22897 * config/arm/arm.c (emit_multi_reg_push): New declaration
22898 for an existing function.
22899 (arm_emit_strd_push): New function.
22900 (arm_expand_prologue): Used here.
22901 (arm_emit_ldrd_pop): New function.
22902 (arm_expand_epilogue): Used here.
22903 (arm_get_frame_offsets): Update condition.
22904 (arm_emit_multi_reg_pop): Add a special case for load of a single
22905 register with writeback.
22907 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
22909 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
22912 2013-04-16 Richard Biener <rguenther@suse.de>
22914 PR tree-optimization/56756
22915 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
22916 (first_mem_ref_loc): New.
22917 (execute_sm): Place the load temporarily before a previous
22918 access instead of in the latch edge to ensure its SSA dependencies
22919 are defined at points dominating the load.
22921 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
22923 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
22924 correct fix by moving header and footer insn to the footer of
22925 the merged basic block. Clear BB_END of the merged-away block.
22927 PR middle-end/43631
22928 * emit-rtl.c (make_note_raw): New function.
22929 (link_insn_into_chain): New static inline function.
22930 (add_insn): Use it.
22931 (add_insn_before, add_insn_after): Factor insn chain linking code...
22932 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
22933 using link_insn_into_chain.
22934 (note_outside_basic_block_p): New helper function for emit_note_after
22935 and emit_note_before.
22936 (emit_note_after): Use nobb variant of add_insn_after if the note
22937 should not be contained in a basic block.
22938 (emit_note_before): Use nobb variant of add_insn_before if the note
22939 should not be contained in a basic block.
22940 (emit_note_copy): Use make_note_raw.
22941 (emit_note): Likewise.
22942 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
22943 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
22944 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
22945 the moved barrier the tail of the basic block it follows.
22946 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
22948 2013-04-15 Jakub Jelinek <jakub@redhat.com>
22950 PR tree-optimization/56962
22951 * gimple-ssa-strength-reduction.c (record_increment): Only set
22952 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
22953 either rhs1 or rhs2 is equal to c->base_expr.
22955 2013-04-15 Richard Biener <rguenther@suse.de>
22957 PR tree-optimization/56933
22958 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
22960 (GROUP_READ_WRITE_DEPENDENCE): Remove.
22961 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
22962 * tree-vect-data-refs.c (vect_analyze_group_access): Move
22963 dependence check ...
22964 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
22966 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
22967 GROUP_READ_WRITE_DEPENDENCE.
22969 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22971 * emit-rtl.c (reset_all_used_flags): New function.
22972 (verify_rtl_sharing): Call reset_all_used_flags before and after
22973 performing the checks.
22975 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22977 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
22978 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
22979 * config/arm/constraints.md (De): New constraint.
22980 * config/arm/neon.md (anddi3_neon): Delete.
22981 (neon_vand<mode>): Expand to standard anddi3 pattern.
22982 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
22983 Move earlier in the file.
22984 (neon_inv_logic_op2): Likewise.
22985 (arm_anddi_operand_neon): New predicate.
22987 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22989 * configure.ac (gcc_cv_ld_as_needed): Set
22990 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
22991 Use -z ignore, -z record on *-*-solaris2*.
22992 (HAVE_LD_AS_NEEDED): Update comment.
22993 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
22994 * configure: Regenerate.
22995 * config.in: Regenerate.
22996 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
22997 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
22998 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
22999 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
23000 equivalents. Fix markup.
23001 * doc/tm.texi: Regenerate.
23003 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
23005 * config/i386/i386.opt: New option mstack-protector-guard=.
23006 * config/i386/i386-opts.h: Add enum stack_protector_guard.
23007 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
23008 TARGET_SSP_TLS_GUARD.
23009 * config/i386/i386.c (ix86_option_override_internal): Set
23010 ix86_stack_protector_guard.
23011 * config/i386/i386.md (stack_protect_set): Enable for
23012 TARGET_SSP_TLS_GUARD only.
23013 (stack_protect_set_<mode>): Ditto.
23014 (stack_protect_test): Ditto.
23015 (stack_protect_test_<mode>): Ditto.
23016 * doc/invoke.texi (i386 Option): Document.
23018 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
23021 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
23022 (S_MODES): Set H_MODE bit.
23023 (SF_MODES): Set only S_MODE and SF_MODE bits.
23024 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
23025 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
23026 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
23027 <MODE_FLOAT>: Likewise.
23029 2013-04-15 Joey Ye <joey.ye@arm.com>
23031 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
23033 2013-04-15 Joey Ye <joey.ye@arm.com>
23035 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
23037 (thumb_far_jump_used_p): Count instruction size and set
23040 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
23042 * reorg.c (fill_simple_delay_slots): Reindent block of code.
23043 * resource.c (mark_target_live_regs): Reformat conditional block.
23045 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23047 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
23048 notes, they are emitted only just before final.
23049 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
23051 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23053 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
23054 * cfgrtl.c (delete_insn): Call it here instead.
23055 * lra-spills.c (lra_final_code_change): Use delete_insn.
23056 * haifa-sched.c (sched_remove_insn): Likewise.
23057 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
23058 returning to the nop pool.
23059 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
23060 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
23062 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23064 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
23065 * doc/tm.texi: Regenerated.
23067 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
23069 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
23072 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23074 * df-core.c (df_find_def): Compare register numbers.
23075 (df_find_use): Likewise.
23077 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
23080 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
23081 lra_in_progress for return.
23083 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23085 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
23086 define_insn into define_insn_and_split and emit movsicc patterns.
23088 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23090 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
23092 2013-04-12 Richard Biener <rguenther@suse.de>
23094 * tree-pass.h (TODO_do_not_ggc_collect): New.
23095 * passes.c (execute_one_ipa_transform_pass): Honor
23096 TODO_do_not_ggc_collect.
23097 (execute_one_pass): Likewise.
23100 2013-04-10 Richard Biener <rguenther@suse.de>
23102 * passes.c (init_optimization_passes): Remove reload pass.
23103 * ira.c (do_reload): Merge into ...
23105 (rest_of_handle_reload): Remove.
23106 (pass_reload): Likewise.
23107 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23108 of reload for vzeroupper pass placement.
23110 2013-04-12 Jakub Jelinek <jakub@redhat.com>
23112 PR tree-optimization/56918
23113 PR tree-optimization/56920
23114 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
23115 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
23116 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
23117 use wide_mul_with_sign method.
23119 2013-04-12 Richard Biener <rguenther@suse.de>
23121 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
23122 not be considered a gimple constant.
23124 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
23126 * fold-const.c (const_binop): Handle vector shifts by a scalar.
23127 (fold_binary_loc): Call const_binop also for mixed vector-scalar
23130 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
23131 Jakub Jelinek <jakub@redhat.com>
23133 * opts.c: Include diagnostic-color.h.
23134 (common_handle_option): Handle OPT_fdiagnostics_color_.
23135 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
23136 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
23137 (diagnostic-color.o): New.
23138 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
23139 (diagnostic_color_rule): New enum.
23140 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
23141 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
23142 the location string.
23143 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
23144 either NULL, or color kind.
23145 * diagnostic-color.c: New file.
23146 * diagnostic-color.h: New file.
23147 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
23149 * doc/invoke.texi (-fdiagnostics-color): Document.
23150 * pretty-print.h (pp_show_color): Define.
23151 (struct pretty_print_info): Add show_color field.
23152 * diagnostic.c: Include diagnostic-color.h.
23153 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
23154 macros. Colorize error:, warning: etc. strings and also the location
23156 (diagnostic_show_locus): Colorize the caret line.
23157 * pretty-print.c: Include diagnostic-color.h.
23158 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
23159 inside of %< %> quotes or quoted through q format modifier.
23161 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23163 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
23165 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23167 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
23168 code in CC_NZ mode.
23169 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
23172 2013-04-11 Marek Polacek <polacek@redhat.com>
23174 PR tree-optimization/48184
23175 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
23177 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
23179 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
23180 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
23181 (skip_simple_arithmetic): Tidy up.
23182 * tree.h (skip_simple_constant_arithmetic): Declare.
23184 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23186 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
23188 2013-04-11 Richard Biener <rguenther@suse.de>
23190 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23191 generate vector constants.
23193 2013-04-11 Richard Biener <rguenther@suse.de>
23195 PR tree-optimization/56878
23196 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
23197 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
23199 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
23200 Prefer to align the DR with the most invariant base address.
23202 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23204 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
23207 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
23209 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
23210 floating-point vector comparisons against 0.
23212 2013-04-11 Jakub Jelinek <jakub@redhat.com>
23214 PR tree-optimization/56899
23215 * fold-const.c (extract_muldiv_1): Apply distributive law
23216 only if TYPE_OVERFLOW_WRAPS (ctype).
23218 2013-04-11 Bin Cheng <bin.cheng@arm.com>
23221 * ira-costs.c (scan_one_insn): Check whether the source rtx of
23222 loading has side effect.
23224 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
23226 * config/sparc/sparc.c: Include tree-pass.h.
23227 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
23228 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
23229 head of file. Change return type. Split off gate function.
23230 (sparc_gate_work_around_errata): New function.
23231 (pass_work_around_errata): New pass definition.
23232 (insert_pass_work_around_errata) New pass insert definition to
23233 insert pass_work_around_errata just after delayed-branch scheduling.
23234 (sparc_option_override): Insert the pass.
23235 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
23237 2013-04-10 David S. Miller <davem@davemloft.net>
23239 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
23240 or -mcpu=hypersparc.
23242 * target.def (cstore_mode): New hook.
23243 * target.h: Include insn-codes.h
23244 * targhooks.c: Likewise.
23245 (default_cstore_mode): New function.
23246 * targhooks.h: Declare it.
23247 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
23248 * doc/tm.texi: Rebuild.
23249 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
23250 target hook, rather than inspecting the insn_data.
23251 * config/sparc/sparc.c (sparc_cstore_mode): New function.
23252 (TARGET_CSTORE_MODE): Redefine.
23253 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
23255 * config/sparc/predicates.md (cstore_result_operand): New special
23257 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
23258 Use it for operand 0.
23259 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
23260 (*snesi_special): Likewise.
23261 (*snesi_zero): Likewise.
23262 (*seqsi_zero): Likewise.
23263 (*sltu_insn): Likewise.
23264 (*sgeu_insn): Likewise.
23265 (*seqdi_special): Make operand 0 and comparison operation be of
23267 (*snedi_special): Likewise.
23268 (*snedi_special_vis3): Likewise.
23269 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
23270 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
23271 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
23272 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
23273 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
23274 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
23275 (*sltu_extend_sp64): Likewise.
23276 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
23277 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
23278 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
23279 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
23280 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
23282 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
23284 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
23285 (aarch64_start_file): Use the new function.
23287 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23288 Jason Merrill <jason@redhat.com>
23290 * common.opt: Add -gdwarf.
23291 * opts.c (common_handle_option): Handle it.
23292 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
23294 2013-04-10 Richard Biener <rguenther@suse.de>
23296 * passes.c (execute_todo): Do not call ggc_collect conditional here.
23297 (execute_one_ipa_transform_pass): But unconditionally here.
23298 (execute_one_pass): And here.
23299 (init_optimization_passes): Remove reload pass.
23300 * tree-pass.h (TODO_ggc_collect): Remove.
23301 (pass_reload): Likewise.
23302 * ira.c (do_reload): Merge into ...
23304 (rest_of_handle_reload): Remove.
23305 (pass_reload): Likewise.
23306 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23307 of reload for vzeroupper pass placement.
23308 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
23309 and todo_flags_finish of all passes.
23311 2013-04-10 Richard Biener <rguenther@suse.de>
23313 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
23314 first_const_oprnd field, rename first_def_type to first_op_type.
23315 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
23316 (vect_get_and_check_slp_defs): Always use the type of the
23317 operand. Allow mixed vect_external_def, vect_constant_def types.
23318 (vect_get_constant_vectors): Handle mixed vect_external_def,
23319 vect_constant_def types.
23321 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
23323 PR tree-optimization/55524
23324 * tree-ssa-math-opts.c
23325 (convert_mult_to_fma): Don't use an fms construct
23326 when we don't have an fms operation, but fnma, and it looks
23327 likely that we'll be able to use the latter.
23329 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
23331 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
23333 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
23334 inline fail caused by overwritable functions.
23336 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
23338 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
23339 unnecessary bits in the constant power of two case.
23341 2013-04-10 Richard Biener <rguenther@suse.de>
23343 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
23344 broken code swapping operands.
23345 (vect_build_slp_tree): Do not compute load permutations here.
23346 (vect_analyze_slp_instance): Compute load permutations here,
23347 after building the SLP tree.
23349 2013-04-09 Christian Bruel <christian.bruel@st.com>
23351 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
23352 of next/prev_real_insn.
23354 2013-04-09 Jan Hubicka <jh@suse.cz>
23356 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
23357 Drop aliased parameter.
23358 (function_and_variable_visibility): Do not handle alias pairs.
23359 * cgraph.c (varpool_externally_visible_p): Update prototype.
23360 * varpool.c (varpool_add_new_variable): Update.
23362 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23364 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
23366 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
23368 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
23370 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
23372 2013-04-09 Marek Polacek <polacek@redhat.com>
23374 PR tree-optimization/48762
23375 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
23377 2013-04-09 Richard Biener <rguenther@suse.de>
23379 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
23381 (vect_build_slp_tree): Likewise.
23382 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
23383 calculating the cost of a SLP instance.
23384 (vect_analyze_slp_instance): Use it from here, after building
23387 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23389 PR middle-end/56883
23390 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
23391 expand_omp_for_static_chunk): Use simple_p = true in
23392 force_gimple_operand_gsi calls when assigning to addressable decls.
23394 2013-04-09 Jeff Law <law@redhat.com>
23396 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
23397 when the boolean was created by converting a wider object which
23398 had a boolean range.
23400 2013-04-09 Richard Biener <rguenther@suse.de>
23402 * tree-vectorizer.h (slp_void_p): Remove.
23403 (slp_tree): Typedef before _slp_tree declaration.
23404 (struct _slp_tree): Use a vector of slp_tree as children.
23405 (vect_get_place_in_interleaving_chain): Remove.
23406 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
23408 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
23410 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
23411 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
23412 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
23413 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
23414 Use slp_node instead of slp_void_p and adjust.
23416 2013-04-09 Richard Biener <rguenther@suse.de>
23418 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
23419 work that is not necessary.
23421 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23423 PR tree-optimization/56854
23424 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
23425 forward into clobber stmts if it would change MEM_REF lhs into
23428 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
23430 * tree.c (type_hash_lookup, type_hash_add): Make static.
23431 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
23433 2013-04-09 Richard Biener <rguenther@suse.de>
23435 * tree.h (unsave_expr_now): Remove.
23436 * tree-inline.c (mark_local_for_remap_r): Remove.
23437 (unsave_expr_1): Likewise.
23438 (unsave_r): Likewise.
23439 (unsave_expr_now): Likewise.
23440 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
23441 (propagate_tree_value): Likewise.
23443 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
23445 * doc/rtl.texi (sequence): Rewrite documentation to match the
23446 current use of SEQUENCE rtl objects.
23447 * rtl.def (SEQUENCE): Likewise.
23449 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
23450 Update documentation.
23451 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
23452 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
23454 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
23456 2013-04-08 Teresa Johnson <tejohnson@google.com>
23458 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
23459 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
23461 (estimate_edge_size_and_time): Add comment to suggest using rounding
23463 (estimate_node_size_and_time): Ditto.
23464 (remap_edge_change_prob): Use helper rounding divide methods.
23465 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
23466 (gimple_mod_pow2_value_transform): Ditto.
23467 (gimple_mod_subtract_transform): Ditto.
23468 (gimple_ic_transform): Ditto.
23469 (gimple_stringops_transform): Ditto.
23470 * stmt.c (conditional_probability): Ditto.
23471 (emit_case_dispatch_table): Ditto.
23472 * lto-cgraph.c (merge_profile_summaries): Ditto.
23473 * tree-optimize.c (execute_fixup_cfg): Ditto.
23474 * cfgcleanup.c (try_forward_edges): Ditto.
23475 * cfgloopmanip.c (scale_loop_profile): Ditto.
23477 (duplicate_loop_to_header_edge): Ditto.
23478 (lv_adjust_loop_entry_edge): Ditto.
23479 * tree-vect-loop.c (vect_transform_loop): Ditto.
23480 * profile.c (compute_branch_probabilities): Ditto.
23481 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
23482 * lto-streamer-in.c (input_cfg): Ditto.
23483 * gimple-streamer-in.c (input_bb): Ditto.
23484 * ipa-cp.c (update_profiling_info): Ditto.
23485 (update_specialized_profile): Ditto.
23486 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
23487 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
23489 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23490 (compute_trg_info): Ditto.
23491 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
23492 (purge_dead_edges): Ditto.
23493 * loop-unswitch.c (unswitch_loop): Ditto.
23494 * cgraphclones.c (cgraph_clone_edge): Ditto.
23495 (cgraph_clone_node): Ditto.
23496 * tree-inline.c (copy_bb): Ditto.
23497 (copy_edges_for_bb): Ditto.
23498 (initialize_cfun): Ditto.
23499 (copy_cfg_body): Ditto.
23500 (expand_call_inline): Ditto.
23502 2013-04-08 Kai Tietz <ktietz@redhat.com>
23504 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
23505 TARGET_CYGWIN64 by TARGET_64BIT.
23507 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23509 * config/epiphany/epiphany.md (GPR_1): New constant.
23510 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
23511 * config/epiphany/epiphany.c (gen_compare_reg):
23512 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
23513 is already in place.
23514 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
23515 Don't require being called during rtl expansion; If y operlaps r0,
23517 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
23518 (epiphany_expand_epilogue): Likewise.
23520 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
23521 Don't use CC_FPmode for ORDERED / UNORDERED.
23522 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
23524 * config/epiphany/constraints.md (CnL): New constraint.
23525 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
23526 * config/epiphany/predicates.md (add_operand): Allow 1024.
23528 * config/epiphany/epiphany.md (logical_op): New code iterator.
23529 (op_mnc): New code attribute.
23530 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
23531 (mov_f+1, mov_f+2): New peephole2 patterns.
23533 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
23534 (cstoresi4): Also allow re-use of zero result when doing a NE
23535 comparison to a non-zero operand.
23536 Use (clobber (scratch)) for first insn if the gpr output is not needed.
23538 * config/epiphany/epiphany.md (<insn_opname>v2si3):
23539 Use gen_addsi3_i / gen_subsi3_i.
23541 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23545 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
23546 contain anything but clobbers, at most one __builtin_stack_restore,
23547 optionally debug stmts and final resx, and if it has at least one
23548 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
23549 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
23550 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
23551 which isn't defaut definition, remove them.
23552 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
23553 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
23554 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
23555 with MEM_REF LHS with SSA_NAME address.
23557 2013-04-08 Jeff Law <law@redhat.com>
23559 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
23561 2013-04-08 Richard Biener <rguenther@suse.de>
23563 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
23565 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
23566 determined vector type.
23567 (vect_analyze_data_refs): Likewise.
23568 (vect_get_new_vect_var): Adjust.
23569 (vect_create_destination_var): Preserve SSA name versions.
23570 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
23571 not dump anything here.
23573 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23575 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
23576 Add member lr_slot_known.
23577 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
23579 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
23580 Remove code that sets lr_slot_offset according to what a previous
23581 version of epiphany_emit_save_restore used to do.
23582 (epiphany_emit_save_restore): When doing an lr save or restore,
23583 set/verify lr_slot_known and lr_slot_offset.
23585 2013-04-08 Xinyu Qi <xyqi@marvell.com>
23588 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
23591 2013-04-08 Richard Biener <rguenther@suse.de>
23593 * alias.c (find_base_term): Fix thinko in previous change.
23595 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23597 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
23598 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
23599 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
23600 if possible to compute val.
23601 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
23602 For QImode integers don't require anything about precision. Use
23603 const_with_all_bytes_same to find out if the constant doesn't have
23604 repeated bytes in it.
23606 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23608 * config/s390/s390.c (s390_expand_insv): Only accept insertions
23611 2013-04-08 Marek Polacek <polacek@redhat.com>
23613 PR rtl-optimization/48182
23614 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
23617 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23620 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
23621 nuses, make sure we have a label.
23623 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23626 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
23627 (rs6000_emit_swdiv_low_precision): Remove.
23628 (rs6000_emit_swdiv): Rewrite to handle between one and four
23629 iterations of Newton-Raphson generally; modify required number of
23630 iterations for some cases.
23631 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
23633 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
23635 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
23636 set-but-unused variable.
23638 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
23639 basic blocks of released function bodies garbage-collectable.
23641 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
23642 (struct rtl_opt_pass): Add TODO_df_finish.
23644 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
23646 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23648 * config/arm/constraints.md (q): New constraint.
23649 * config/arm/ldrdstrd.md: New file.
23650 * config/arm/arm.md (ldrdstrd.md) New include.
23651 (arm_movdi): Use "q" instead of "r" constraint
23652 for double-word memory access.
23653 (movdf_soft_insn): Likewise.
23654 * config/arm/vfp.md (movdi_vfp): Likewise.
23655 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
23656 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
23657 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
23658 (mem_ok_for_ldrd_strd): Likewise.
23659 (output_move_double): Update assertion.
23661 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23663 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
23665 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23667 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
23668 define_insn_and_split.
23669 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
23671 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23673 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
23674 define_insn_and_split.
23675 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
23676 (shiftsi3_compare): New pattern.
23677 (rrx): New pattern.
23678 * config/arm/unspecs.md (UNSPEC_RRX): New.
23680 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23682 * config/arm/arm.md (negdi_extendsidi): New pattern.
23683 (negdi_zero_extendsidi): Likewise.
23685 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23687 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
23688 define_insn_and_split.
23689 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
23690 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
23692 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23694 * config/arm/arm.md (arm_subdi3): Convert define_insn into
23695 define_insn_and_split.
23696 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
23697 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
23699 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23701 * config/arm/arm.md (subsi3_carryin): New pattern.
23702 (subsi3_carryin_const): Likewise.
23703 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
23704 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
23706 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23708 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
23710 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23712 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
23713 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
23715 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23717 * config/arm/arm.c (arm_expand_builtin): Change fcode
23718 type to unsigned int.
23720 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23722 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
23724 2013-04-04 Ian Lance Taylor <iant@google.com>
23726 * doc/standards.texi (Standards): The Go frontend supports the Go 1
23729 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
23731 PR middle-end/56729
23732 * df-scan.c (df_insn_delete): Disable failing assert.
23734 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23736 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
23737 New function prototype.
23738 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
23739 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
23740 (arm_builtin_vectorized_function): New function.
23742 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23744 * config/arm/arm_neon_builtins.def: New file.
23745 * config/arm/arm.c (neon_builtin_data): Move contents to
23746 arm_neon_builtins.def.
23747 (enum arm_builtins): Include neon builtin definitions.
23748 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
23749 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
23751 2013-04-04 Marek Polacek <polacek@redhat.com>
23753 PR tree-optimization/48186
23754 * predict.c (maybe_hot_frequency_p): Return false if
23755 HOT_BB_FREQUENCY_FRACTION is 0.
23756 (cgraph_maybe_hot_edge_p): Likewise.
23758 2013-04-04 Richard Biener <rguenther@suse.de>
23760 PR tree-optimization/56826
23761 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
23764 2013-04-04 Richard Biener <rguenther@suse.de>
23766 PR tree-optimization/56213
23767 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
23768 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
23770 2013-04-04 Richard Biener <rguenther@suse.de>
23772 PR tree-optimization/56837
23773 * tree-loop-distribution.c (classify_partition): For non-zero
23774 values require that the value has the same precision as its
23775 mode to be useful as memset value.
23777 2013-04-03 Nick Clifton <nickc@redhat.com>
23779 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
23780 (fmssf4): Use fmsf.s on E3V5 architectures.
23781 (fnmasf4): Use fnmaf.s on E3V5 architectures.
23782 (fnmssf4): Use fnmsf.s on E3V5 architectures.
23784 2013-04-03 Jeff Law <law@redhat.com>
23786 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
23787 (lra-eliminations.o): Likewise.
23789 2013-04-03 Teresa Johnson <tejohnson@google.com>
23791 * gcov-io.c (compute_working_sets): Moved most of body of old
23792 compute_working_sets here from profile.c.
23793 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
23794 (gcov_working_set_t): Moved typedef here from basic-block.h
23795 (compute_working_set): Declare.
23796 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
23797 (get_working_sets): Renamed from compute_working_set,
23798 replace most of body with call to new compute_working_sets.
23799 (get_exec_counts): Replace call to compute_working_sets
23800 to get_working_sets.
23801 * profile.h (get_working_sets): Renamed from compute_working_set.
23802 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
23803 to get_working_sets.
23804 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
23805 * gcov-dump.c (dump_working_sets): New function.
23807 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
23809 * hwint.c (sext_hwi, zext_hwi): New functions.
23810 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
23811 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
23812 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
23813 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
23814 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
23815 (sext_hwi, zext_hwi): New functions.
23817 2013-04-03 Jeff Law <law@redhat.com>
23819 PR tree-optimization/56799
23820 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
23821 back test for widening conversion erroneously dropped in prior change.
23823 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23826 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
23827 instead of next_real_insn.
23829 2013-04-03 Marek Polacek <polacek@redhat.com>
23832 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
23834 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23837 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
23839 (thumb1_output_casesi): Likewise.
23840 (thumb2_output_casesi): Likewise.
23842 2013-04-03 Richard Biener <rguenther@suse.de>
23844 PR tree-optimization/56817
23845 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
23847 (tree_unroll_loops_completely_1): ... new function to manually
23848 walk the loop tree, properly defering outer loops of unrolled
23849 loops to later iterations.
23851 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23853 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
23854 (vectorizable_load): Likewise.
23855 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
23856 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
23858 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23860 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
23863 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
23865 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
23867 2013-04-03 Bin Cheng <bin.cheng@arm.com>
23869 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
23871 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23873 PR tree-optimization/56790
23874 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
23877 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23879 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
23881 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
23882 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
23885 2013-04-03 Jakub Jelinek <jakub@redhat.com>
23888 * tree.h (force_folding_builtin_constant_p): New decl.
23889 * builtins.c (force_folding_builtin_constant_p): New variable.
23890 (fold_builtin_constant_p): Fold immediately also if
23891 force_folding_builtin_constant_p.
23893 2013-04-03 Richard Biener <rguenther@suse.de>
23895 PR tree-optimization/56812
23896 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
23897 DRs of the same interleaving chain are independent.
23899 2013-04-02 Jason Merrill <jason@redhat.com>
23901 * gdbinit.in (pbb): Use debug fn.
23903 2013-04-02 Lawrence Crowl <crowl@google.com>
23905 * sese.h (struct ivtype_map_elt_s): Remove unused.
23906 (extern debug_ivtype_map): Remove unused.
23907 (extern eq_ivtype_map_elts): Remove unused.
23908 * sese.c (debug_ivtype_map): Removed unused.
23909 (debug_ivtype_map_1): Removed unused.
23910 (debug_ivtype_elt): Remove unused.
23911 (eq_ivtype_map_elts): Remove unused.
23914 2013-04-02 Kai Tietz <ktietz@redhat.com>
23917 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
23918 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
23919 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
23921 (legitimize_pe_coff_symbol): Likewise.
23922 (is_imported_p): New helper-function.
23923 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
23924 for Windows x64 targets.
23925 (ix86_expand_prologue): Optimize for pe-coff targets.
23926 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
23927 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
23928 medium/large code-model.
23929 (legitimize_pic_address): Likewise.
23930 (legitimize_tls_address): Likewise.
23931 (ix86_expand_call): Likewise.
23932 (x86_output_mi_thunk): Likewise.
23933 (get_dllimport_decl): Add new beimport argument.
23934 (construct_plt_address): Don't assert for x64 pe-coff targets.
23935 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
23937 (SYMBOL_FLAG_STUBVAR): New macro.
23938 (SYMBOL_REF_STUBVAR_P): Likewise.
23939 * config/i386/winnt.c (stub_list): New structure.
23940 (stub_head): New local variable.
23941 (i386_pe_record_stub): New function.
23942 (i386_pe_file_end): Emit refptr-stubs.
23944 2013-04-02 Jakub Jelinek <jakub@redhat.com>
23946 PR rtl-optimization/56745
23947 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
23948 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
23951 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
23952 and both of them are MEM_REFs, just compare first argument for
23953 equality and attempt to deal even with differing offsets.
23956 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
23957 of gimple_clobber_p to be MEM_REF.
23958 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
23959 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
23960 after gimplification.
23961 * asan.c (get_mem_ref_of_assignment): Don't instrument
23962 gimple_clobber_p stmts.
23963 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
23964 gimple_clobber_p stmt if they have MEM_REF lhs and
23965 are dead because of another gimple_clobber_p stmt.
23966 * tree-ssa-live.c (clear_unused_block_pointer): Treat
23967 gimple_clobber_p stmts like debug stmts.
23968 (remove_unused_locals): Remove clobbers with MEM_REF lhs
23969 that refer to unused VAR_DECLs or uninitialized values.
23970 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
23971 gimple_clobber_p stmts if they refer to removed parameters.
23972 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
23975 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
23977 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
23978 using SWI48 mode attribute.
23980 2013-04-02 Wei Mi <wmi@google.com>
23982 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
23983 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
23984 *<rotate_insn><mode>3_mask in i386.md.
23986 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
23988 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
23990 2013-04-02 Richard Biener <rguenther@suse.de>
23992 PR tree-optimization/56778
23993 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
23994 Runtime alias tests are not supported for gather loads.
23995 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
23996 stmts referenced from SSA operands before updating SSA form.
23998 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
23999 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24001 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
24002 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
24003 * config/arm/cortex-a53.md: New file.
24004 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
24005 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
24006 * config/arm/arm.c (arm_issue_rate): Likewise.
24007 * config/arm/arm-tune.md: Regenerate
24008 * config/arm/arm-tables.opt: Regenerate.
24009 * config/arm/arm-cores.def: Add cortex-a53.
24011 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
24013 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
24016 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24018 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
24019 scalar load/store operations using B/H registers.
24020 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
24022 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24024 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
24026 * config/aarch64/aarch64.c
24027 (aarch64_simd_scalar_immediate_valid_for_move): New.
24028 * config/aarch64/aarch64-protos.h
24029 (aarch64_simd_scalar_immediate_valid_for_move): New.
24030 * config/aarch64/constraints.md (Dh, Dq): New.
24031 * config/aarch64/iterators.md (hq): New.
24033 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
24035 * reorg.c (get_branch_condition): Deal with conditional returns.
24036 (fill_simple_delay_slots): Remove dead code dealing with jumps.
24038 2013-04-01 Wei Mi <wmi@google.com>
24040 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
24041 Truncate operand 2 using %b asm operand modifier.
24042 (*<shift_insn><mode>3_mask): Ditto.
24043 (*<rotate_insn><mode>3_mask): Ditto.
24045 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
24047 PR middle-end/56798
24048 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
24050 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
24052 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
24054 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
24056 2013-03-30 Lawrence Crowl <crowl@google.com>
24058 * dse.c (clear_alias_sets): Remove never set.
24059 (disqualified_clear_alias_sets): Remove never set.
24060 (clear_alias_mode_pool): Remove never set.
24061 (dse_step0): Remove condition that is never true.
24062 (canon_address): Remove condition that is never true.
24063 (dse_step7): Remove condition that is never true.
24064 (rest_of_handle_dse): Remove condition that is never true.
24065 (rest_of_handle_dse::did_global): Remove never read from above.
24066 (dse_step2_spill): Remove never called from above.
24067 (dse_step5_spill): Remove never called from above.
24069 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
24071 * doc/md.texi (Standard Names) <casesi>: Update documentation for
24072 JUMP_TABLE_DATA changes.
24073 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
24074 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
24075 (Insns) <jump_table_data>: New entry.
24076 * doc/tm.texi: Regenerate.
24078 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
24080 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
24081 for table jump at the end of a basic block using tablejump_p.
24082 * targhooks.c (default_invalid_within_doloop): Likewise.
24083 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
24084 target hook implementation that is identical to the default hook.
24085 (rs6000_invalid_within_doloop): Remove.
24087 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
24088 unused variable from tablejump_p call.
24090 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
24091 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
24092 (INSN_DELETED_P): Likewise.
24093 (emit_jump_table_data): New prototype.
24094 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
24095 after 4th as unused.
24096 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
24097 * sched-vis.c (print_insn): Likewise.
24098 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
24099 insn for compatibility with back ends that use next_active_insn to
24100 identify jump table data.
24101 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
24102 (remove_insn): Likewise.
24103 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
24105 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
24106 (emit_jump_table_data): New function.
24108 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
24109 basic block, a JUMP_TABLE_DATA never is.
24110 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
24111 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
24112 off from code handling real insns.
24113 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
24114 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
24115 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
24116 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
24117 is not a NONDEBUG_INSN_P.
24118 * ira-costs.c (scan_one_insn): Likewise.
24119 * jump.c (mark_all_labels): Likewise.
24120 (mark_jump_label_1): Likewise.
24121 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24122 * lra.c (get_insn_freq): Expect all insns reaching here to be in
24124 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
24125 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
24126 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
24127 JUMP_TABLE_DATA_P insns.
24128 (calculate_elim_costs_all_insns): Likewise.
24129 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
24130 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
24131 (delete_output_reload): Code style fixups.
24132 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
24133 insn flags on this non-insn.
24134 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
24135 as scheduling barriers, for pre-change compatibility.
24136 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
24137 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
24139 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
24140 redundant JUMP_TABLE_DATA_P test.
24141 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
24142 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24143 (frv_for_each_packet): Likewise.
24144 * config/i386/i386.c (min_insn_size): Likewise.
24145 (ix86_avoid_jump_mispredicts): Likewise.
24146 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24147 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
24148 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24149 (mips16_insn_length): Robustify.
24150 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
24151 (mips16_split_long_branches): Likewise.
24152 * config/pa/pa.c (pa_combine_instructions): Likewise.
24153 * config/rs6000/rs6000.c (get_next_active_insn): Treat
24154 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
24155 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
24156 as contributing to pool range lengths.
24157 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
24158 Remove redundant JUMP_TABLE_DATA_P test.
24159 (sh_loop_align): Likewise.
24160 (split_branches): Likewise.
24161 (sh_insn_length_adjustment): Likewise.
24162 * config/spu/spu.c (get_branch_target): Likewise.
24164 2013-03-29 Jan Hubicka <jh@suse.cz>
24166 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
24167 gcov streaming; stream hot bb threshold to ltrans.
24168 * predict.c (get_hot_bb_threshold): Break out from ....
24169 (maybe_hot_count_p): ... here.
24170 (set_hot_bb_threshold): New function.
24171 * lto-section-in.c (lto_section_name): Add profile.
24172 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
24173 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
24174 and data-streamer.h
24175 (histogram_entry): New structure.
24176 (histogram, histogram_pool): New global vars.
24177 (histogram_hash): New structure.
24178 (histogram_hash::hash): New method.
24179 (histogram_hash::equal): Likewise.
24180 (account_time_size): New function.
24181 (cmp_counts): New function.
24182 (dump_histogram): New function.
24183 (ipa_profile_generate_summary): New function.
24184 (ipa_profile_write_summary): New function.
24185 (ipa_profile_read_summary): New function.
24186 (ipa_profile): Decide on threshold.
24187 (pass_ipa_profile): Add ipa_profile_write_summary and
24188 ipa_profile_read_summary.
24189 * Makefile.in (ipa.o): Update dependencies.
24190 * lto-streamer.h (LTO_section_ipa_profile): New section.
24192 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
24194 * tree.h (VAR_P): New.
24196 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
24199 * doc/invoke.texi ([-fwhole-program]): Fix typo.
24201 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
24203 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
24204 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24205 (control_flow_insn_p): Likewise.
24206 * cfgrtl.c (duplicate_insn_chain): Likewise.
24207 * final.c (get_attr_length_1): Likewise.
24208 (shorten_branches): Likewise.
24209 (final_scan_insn): Likewise.
24210 * function.c (instantiate_virtual_regs): Likewise.
24211 * gcse.c (insert_insn_end_basic_block): Likewise.
24212 * ira-costs.c (scan_one_insn): Likewise.
24213 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24214 * lra.c (check_rtl): Likewise.
24215 * reload1.c (elimination_costs_in_insn): Likewise.
24216 * reorg.c (follow_jumps): Likewise.
24218 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
24219 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24220 (thumb_far_jump_used_p): Likewise.
24221 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
24222 (workaround_speculation): Likewise.
24223 (add_sched_insns_for_speculation): Likewise.
24224 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
24225 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24226 (frv_for_each_packet): Likewise.
24227 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
24228 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
24229 (final_emit_insn_group_barriers): Likewise.
24230 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24231 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24232 (mips16_insn_length): Likewise.
24233 * config/pa/pa.c (pa_reorg): Likewise.
24234 (pa_combine_instructions): Likewise.
24235 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
24236 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
24237 (sh_reorg): Likewise.
24238 (split_branches): Likewise.
24239 * config/spu/spu.c (get_branch_target): Likewise.
24241 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
24244 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
24246 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
24247 Fix declaration name.
24249 2013-03-28 Lawrence Crowl <crowl@google.com>
24251 * graphds.h (struct graph.indicies): Remove unused.
24252 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
24253 (SCOP_ORIGINAL_PDDRS): Remove unused.
24254 * sese.h (extern insert_loop_close_phis): Removed unused.
24255 (extern insert_guard_phis): Removed unused.
24256 (extern ivtype_map_elt_info): Removed unused.
24257 (new_ivtype_map_elt): Removed unused.
24258 * sese.c (ivtype_map_elt_info): Removed unused.
24260 2013-03-28 Lawrence Crowl <crowl@google.com>
24262 * Makefile.in: Add several missing include dependences.
24264 (test-dump.o): New. This object is not added to any executable,
24265 but is present for ad-hoc testing.
24267 (debug (const bitmap_head_def &)): New.
24268 (debug (const bitmap_head_def *)): New.
24270 (extern debug (const bitmap_head_def &)): New.
24271 (extern debug (const bitmap_head_def *)): New.
24273 (debug (edge_def &)): New.
24274 (debug (edge_def *)): New.
24276 (debug (basic_block_def &)): New.
24277 (debug (basic_block_def *)): New.
24279 (dump_node (const_tree, int, FILE *)): Correct source file.
24281 (debug (die_struct &)): New.
24282 (debug (die_struct *)): New.
24284 (extern debug (die_struct &)): New.
24285 (extern debug (die_struct *)): New.
24286 * gimple-pretty-print.c
24287 (debug (gimple_statement_d &)): New.
24288 (debug (gimple_statement_d *)): New.
24289 * gimple-pretty-print.h
24290 (extern debug (gimple_statement_d &)): New.
24291 (extern debug (gimple_statement_d *)): New.
24293 (debug (ira_allocno_copy &)): New.
24294 (debug (ira_allocno_copy *)): New.
24295 (debug (ira_allocno &)): New.
24296 (debug (ira_allocno *)): New.
24298 (extern debug (ira_allocno_copy &)): New.
24299 (extern debug (ira_allocno_copy *)): New.
24300 (extern debug (ira_allocno &)): New.
24301 (extern debug (ira_allocno *)): New.
24303 (debug (live_range &)): New.
24304 (debug (live_range *)): New.
24306 (debug (lra_live_range &)): New.
24307 (debug (lra_live_range *)): New.
24309 (debug (lra_live_range &)): New.
24310 (debug (lra_live_range *)): New.
24312 (debug (omega_pb_d &)): New.
24313 (debug (omega_pb_d *)): New.
24315 (extern debug (omega_pb_d &)): New.
24316 (extern debug (omega_pb_d *)): New.
24318 (debug (const rtx_def &)): New.
24319 (debug (const rtx_def *)): New.
24321 (debug_tree (tree): Move within file.
24322 (debug_raw (const tree_node &)): New.
24323 (debug_raw (const tree_node *)): New.
24324 (dump_tree_via_hooks (const tree_node *, int)): New.
24325 (debug (const tree_node &)): New.
24326 (debug (const tree_node *)): New.
24327 (debug_verbose (const tree_node &)): New.
24328 (debug_verbose (const tree_node *)): New.
24329 (debug_head (const tree_node &)): New.
24330 (debug_head (const tree_node *)): New.
24331 (debug_body (const tree_node &)): New.
24332 (debug_body (const tree_node *)): New.
24333 (debug_vec_tree (tree): Move and reimplement in terms of dump.
24334 (debug (vec<tree, va_gc> &)): New.
24335 (debug (vec<tree, va_gc> *)): New.
24337 (extern debug (const rtx_def &)): New.
24338 (extern debug (const rtx_def *)): New.
24340 (debug_raw (simple_bitmap_def &)): New.
24341 (debug_raw (simple_bitmap_def *)): New.
24342 (debug (simple_bitmap_def &)): New.
24343 (debug (simple_bitmap_def *)): New.
24345 (extern debug (simple_bitmap_def &)): New.
24346 (extern debug (simple_bitmap_def *)): New.
24347 (extern debug_raw (simple_bitmap_def &)): New.
24348 (extern debug_raw (simple_bitmap_def *)): New.
24350 (debug (vinsn_def &)): New.
24351 (debug (vinsn_def *)): New.
24352 (debug_verbose (vinsn_def &)): New.
24353 (debug_verbose (vinsn_def *)): New.
24354 (debug (expr_def &)): New.
24355 (debug (expr_def *)): New.
24356 (debug_verbose (expr_def &)): New.
24357 (debug_verbose (expr_def *)): New.
24358 (debug (vec<rtx> &)): New.
24359 (debug (vec<rtx> *)): New.
24361 (extern debug (vinsn_def &)): New.
24362 (extern debug (vinsn_def *)): New.
24363 (extern debug_verbose (vinsn_def &)): New.
24364 (extern debug_verbose (vinsn_def *)): New.
24365 (extern debug (expr_def &)): New.
24366 (extern debug (expr_def *)): New.
24367 (extern debug_verbose (expr_def &)): New.
24368 (extern debug_verbose (expr_def *)): New.
24369 (extern debug (vec<rtx> &)): New.
24370 (extern debug (vec<rtx> *)): New.
24372 (_list_iter_cond_expr): Make inline instead of static.
24374 (debug (sreal &)): New.
24375 (debug (sreal *)): New.
24377 (extern debug (sreal &)): New.
24378 (extern debug (sreal *)): New.
24380 (extern debug_raw (const tree_node &)): New.
24381 (extern debug_raw (const tree_node *)): New.
24382 (extern debug (const tree_node &)): New.
24383 (extern debug (const tree_node *)): New.
24384 (extern debug_verbose (const tree_node &)): New.
24385 (extern debug_verbose (const tree_node *)): New.
24386 (extern debug_head (const tree_node &)): New.
24387 (extern debug_head (const tree_node *)): New.
24388 (extern debug_body (const tree_node &)): New.
24389 (extern debug_body (const tree_node *)): New.
24390 (extern debug (vec<tree, va_gc> &)): New.
24391 (extern debug (vec<tree, va_gc> *)): New.
24393 (debug (struct loop &)): New.
24394 (debug (struct loop *)): New.
24395 (debug_verbose (struct loop &)): New.
24396 (debug_verbose (struct loop *)): New.
24397 * tree-dump.c: Add header dependence.
24399 (extern debug (struct loop &)): New.
24400 (extern debug (struct loop *)): New.
24401 (extern debug_verbose (struct loop &)): New.
24402 (extern debug_verbose (struct loop *)): New.
24404 (debug (data_reference &)): New.
24405 (debug (data_reference *)): New.
24406 (debug (vec<data_reference_p> &)): New.
24407 (debug (vec<data_reference_p> *)): New.
24408 (debug (vec<ddr_p> &)): New.
24409 (debug (vec<ddr_p> *)): New.
24411 (extern debug (data_reference &)): New.
24412 (extern debug (data_reference *)): New.
24413 (extern debug (vec<data_reference_p> &)): New.
24414 (extern debug (vec<data_reference_p> *)): New.
24415 (extern debug (vec<ddr_p> &)): New.
24416 (extern debug (vec<ddr_p> *)): New.
24418 (debug (pt_solution &)): New.
24419 (debug (pt_solution *)): New.
24421 (extern debug (pt_solution &)): New.
24422 (extern debug (pt_solution *)): New.
24424 (debug (_var_map &)): New.
24425 (debug (_var_map *)): New.
24426 (debug (tree_live_info_d &)): New.
24427 (debug (tree_live_info_d *)): New.
24429 (extern debug (_var_map &)): New.
24430 (extern debug (_var_map *)): New.
24431 (extern debug (tree_live_info_d &)): New.
24432 (extern debug (tree_live_info_d *)): New.
24434 2013-03-28 Jan Hubicka <jh@suse.cz>
24436 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
24438 2013-03-28 Ian Bolton <ian.bolton@arm.com>
24440 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
24441 record only when desired or required.
24443 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
24445 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
24446 *vec_extractv2di_1_rex64. Use x64 isa attribute.
24448 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24450 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
24451 (*andsi3_compare0_uxtw): New pattern.
24452 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
24453 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
24455 2013-03-28 Jan Hubicka <jh@suse.cz>
24457 * data-streamer-in.c (streamer_read_gcov_count): New function.
24458 * gimple-streamer-out.c: Include value-prof.h.
24459 (output_gimple_stmt): Output histogram.
24460 (output_bb): Use streamer_write_gcov_count.
24461 * value-prof.c: Include data-streamer.h
24462 (dump_histogram_value): Add HIST_TYPE_MAX.
24463 (stream_out_histogram_value): New function.
24464 (stream_in_histogram_value): New function.
24465 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
24466 (stream_out_histogram_value, stream_in_histogram_value): Declare.
24467 * data-streamer-out.c (streamer_write_gcov_count): New function.
24468 (streamer_write_gcov_count_stream): New function.
24469 * lto-cgraph.c (lto_output_edge): Update counter streaming.
24470 (lto_output_node): Likewise.
24471 (input_node, input_edge): Likewise.
24472 * lto-streamer-out.c (output_cfg): Update streaming.
24473 * lto-streamer-in.c (input_cfg): Likewise.
24474 * data-streamer.h (streamer_write_gcov_count,
24475 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
24476 * gimple-streamer-in.c: Include value-prof.h
24477 (input_gimple_stmt): Input histograms.
24478 (input_bb): Update profile streaming.
24480 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24482 * genmodes.c (emit_max_int): New function.
24483 (emit_insn_modes_h): Added call to emit_max_function.
24484 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
24486 * machmode.def: Fixed comment.
24488 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24490 * combine.c (try_combine): Removed useless assert.
24491 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
24493 2013-03-28 Marek Polacek <polacek@redhat.com>
24494 Richard Biener <rguenther@suse.de>
24496 PR tree-optimization/56695
24497 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
24498 build signed result of a vector comparison.
24499 * tree-cfg.c (verify_gimple_comparison): Check that a result
24500 of a vector comparison has signed type.
24502 2013-03-28 Richard Biener <rguenther@suse.de>
24504 PR tree-optimization/37021
24505 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
24506 do not restrict gaps between groups.
24507 * tree-vect-stmts.c (vectorizable_load): Properly account for
24508 a gap between groups.
24510 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
24512 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
24513 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
24516 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
24518 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
24519 * doc/extend.texi (Named Address Spaces): Ditto.
24520 (Variable Attributes): Ditto.
24522 2013-03-27 Kai Tietz <ktietz@redhat.com>
24524 * config.build: Add support for cygwin x64 target.
24525 * config.gcc: Likewise.
24526 * config.host: Likewise.
24527 * configure.ac: Likewise
24528 * configure: Regenerated.
24530 2013-03-27 Kai Tietz <ktietz@redhat.com>
24532 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
24533 * config/i386/t-cygwin-w64: New file.
24534 * config/i386/cygwin-w64.h: New file.
24535 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
24536 and add support for x64-cygwin target.
24537 (CPP_SPEC): Likewise.
24538 (CXX_WRAP_SPEC_LIST): Undefine before define.
24539 (LIBGCJ_SONAME): Use 15 as version.
24541 2013-03-27 Richard Biener <rguenther@suse.de>
24543 PR tree-optimization/56716
24544 * tree-ssa-structalias.c (perform_var_substitution): Adjust
24545 dumping for ref nodes.
24547 2013-03-27 Martin Jambor <mjambor@suse.cz>
24549 PR tree-optimization/55334
24550 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
24551 restricted pointers to arrays.
24553 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
24555 * Makefile.in (.SUFFIXES): Add .cc.
24556 (.c.o): Apply same recipe for implicit rule .cc.o.
24558 2013-03-27 Richard Biener <rguenther@suse.de>
24560 PR tree-optimization/37021
24561 * tree-vect-data-refs.c (vect_check_strided_load): Allow
24562 REALPART/IMAGPART_EXPRs around the supported refs.
24563 * tree-ssa-structalias.c (find_func_aliases): Assume that
24564 floating-point values are not used to transfer pointers.
24566 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
24568 * target.def (TARGET_HAS_IFUNC_P): New target hook.
24569 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
24570 * doc/tm.texi: Regenerate.
24571 * targhooks.h (default_has_ifunc_p): New.
24572 * targhooks.c (default_has_ifunc_p): Ditto.
24573 * config/linux-protos.h: New file.
24574 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
24575 hook for linux which disables support of indirect functions in android.
24576 * config/linux-android.c: New file.
24577 * config/t-linux-android.c: Ditto.
24578 * config.gcc: Added new object file linux-android.o.
24579 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
24580 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
24581 * varasm.c (do_assemble_alias): Likewise.
24582 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
24583 doesn't support indirect functions.
24584 * configure: Regenerate.
24586 2013-03-27 Bin Cheng <bin.cheng@arm.com>
24589 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
24590 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
24593 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24595 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
24597 2013-03-27 Terry Guo <terry.guo@arm.com>
24599 * config/arm/arm-cores.def: Added core cortex-r7.
24600 * config/arm/arm-tune.md: Regenerated.
24601 * config/arm/arm-tables.opt: Regenerated.
24602 * doc/invoke.texi: Added entry for core cortex-r7.
24604 2013-03-27 Walter Lee <walt@tilera.com>
24606 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
24607 double-decrement of next_scratch_regno.
24609 2013-03-27 Walter Lee <walt@tilera.com>
24611 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
24613 (insn_v1mulus): Ditto.
24614 (insn_v2muls): Ditto.
24616 2013-03-27 Walter Lee <walt@tilera.com>
24618 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
24619 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
24621 2013-03-27 Walter Lee <walt@tilera.com>
24623 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
24624 (*sibcall_value): Ditto.
24626 2013-03-27 Walter Lee <walt@tilera.com>
24628 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
24629 (insn_mnz_v8qi): ... this ...
24630 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
24632 (insn_v<n>mnz): Replaced by ...
24633 (insn_v1mnz): ... this ...
24634 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
24636 (insn_mz_<mode>): Replaced by ...
24637 (insn_mz_v8qi): ... this ...
24638 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
24640 (insn_v<n>mz): Replaced by ...
24641 (insn_v1mz): ... this ...
24642 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
24645 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
24647 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
24649 2013-03-26 Roland McGrath <mcgrathr@google.com>
24651 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
24652 than fprintf with a non-constant, non-format string.
24654 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24656 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
24657 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
24658 operand 0 predicate.
24659 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
24660 attribute. Use general_x64nomem_operand as operand 1 predicate.
24661 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
24662 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24663 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
24664 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24665 (mov<mode>_insv_1): Remove expander. Merge insn with
24666 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
24667 Use general_x64nomem_operand as operand 1 predicate.
24668 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
24669 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
24670 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24671 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
24672 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
24673 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
24674 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24675 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
24676 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
24677 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
24678 (general_x64nomem_operand): Ditto.
24680 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
24682 * config/rtems.opt: Add -pthread option.
24684 2013-03-26 Richard Biener <rguenther@suse.de>
24686 * alias.c (find_base_term): Avoid redundant and not used recursion.
24687 (base_alias_check): Get the initial base term from the caller.
24688 (true_dependence_1): Compute and pass base terms to base_alias_check.
24689 (write_dependence_p): Likewise.
24690 (may_alias_p): Likewise.
24692 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
24694 * config/aarch64/aarch64.c (aarch64_classify_address): Support
24695 PC-relative load in SI modes and above only.
24697 2013-03-26 Xinyu Qi <xyqi@marvell.com>
24699 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
24700 * config/arm/iwmmxt.md (WCGR0): Update.
24701 (WCGR1, WCGR2, WCGR3): Likewise.
24703 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24705 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
24706 Use x64 and nox64 isa attributes.
24708 2013-03-26 Richard Biener <rguenther@suse.de>
24710 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
24711 alignment computations and rely on get_object_alignment_1
24712 for the !TYPE_P case.
24713 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
24715 2013-03-26 Walter Lee <walt@tilera.com>
24717 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
24718 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
24720 2013-03-25 Jeff Law <law@redhat.com>
24722 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
24723 check for INTEGRAL_TYPE_P that was missing due to checking in
24724 wrong version of prior patch.
24726 2013-03-25 Walter Lee <walt@tilera.com>
24728 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
24729 TILEGX_INSN_SHUFFLEBYTES1.
24730 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
24732 (tilegx_builtins): Ditto.
24733 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
24735 2013-03-25 Walter Lee <walt@tilera.com>
24737 * config/tilegx/tilegx.md (floatsisf2): New pattern.
24738 (floatunssisf2): New pattern.
24739 (floatsidf2): New pattern.
24740 (floatunssidf2): New pattern.
24742 2013-03-25 Walter Lee <walt@tilera.com>
24744 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
24745 tests for constraint J, K, N, P.
24747 2013-03-25 Walter Lee <walt@tilera.com>
24749 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
24750 Use indirect/pcrel encoding.
24751 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
24754 2013-03-25 Steve Ellcey <sellcey@mips.com>
24756 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
24757 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
24758 * config/mips/mips.c (mips_option_override): Set IMADD default.
24759 * config/mips/mips.h (PTF_AVOID_IMADD): New.
24760 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
24761 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
24762 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
24763 * doc/invoke.texi (-mimadd/-mno-imadd): New.
24765 2013-03-25 Jeff Law <law@redhat.com>
24767 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
24768 slightly to avoid creating and folding useless trees. Simplify
24769 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
24771 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
24773 * config/i386/i386.md (*zero_extendsidi2): Merge with
24774 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
24775 * config/i386/predicates.md (x86_64_zext_operand): Rename from
24776 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
24777 targets. Clarify comment.
24779 2013-03-25 Martin Jambor <mjambor@suse.cz>
24781 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
24782 pass-through jump functions differently.
24783 (ipa_read_jump_function): Likewise. Also use setter functions to set
24786 2013-03-25 Martin Jambor <mjambor@suse.cz>
24788 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
24789 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
24791 (ipa_get_indirect_edge_target): New function.
24792 (devirtualization_time_bonus): New parameter known_aggs, pass it to
24793 ipa_get_indirect_edge_target. Update all callers.
24794 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
24795 ipa_get_indirect_edge_target_1 instead of calling
24796 ipa_get_indirect_edge_target.
24797 (create_specialized_node): Pass aggvlas to
24798 ipcp_discover_new_direct_edges.
24800 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24802 * config/arm/arm.md (f_sels, f_seld): New types.
24803 (*cmov<mode>): New pattern.
24804 * config/arm/predicates.md (arm_vsel_comparison_operator): New
24807 2013-03-25 Kai Tietz <ktietz@redhat.com>
24809 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
24810 POSIX-printf for mingw-hosted builds.
24812 2013-03-25 Richard Biener <rguenther@suse.de>
24814 PR middle-end/56694
24815 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
24816 must-not-throw stmt location.
24818 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24820 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
24821 Emit load-acquire versions when acq is true.
24822 (arm_emit_store_exclusive): Add rel parameter.
24823 Emit store-release versions when rel is true.
24824 (arm_split_compare_and_swap): Use acquire-release instructions
24826 of barriers when appropriate.
24827 (arm_split_atomic_op): Likewise.
24828 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
24829 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
24830 (VUNSPEC_SLX): Likewise.
24831 (VUNSPEC_LDA): Likewise.
24832 (VUNSPEC_STL): Likewise.
24833 * config/arm/sync.md (atomic_load<mode>): New pattern.
24834 (atomic_store<mode>): Likewise.
24835 (arm_load_acquire_exclusive<mode>): Likewise.
24836 (arm_load_acquire_exclusivesi): Likewise.
24837 (arm_load_acquire_exclusivedi): Likewise.
24838 (arm_store_release_exclusive<mode>): Likewise.
24840 2013-03-25 Catherine Moore <clm@codesourcery.com>
24842 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
24843 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
24844 * config/mip/predicates.md (lwsp_swsp_operand,
24845 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
24846 sb16_operand, db4_operand, db7_operand, ib3_operand,
24847 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
24848 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
24849 andi16_operand): New predicates.
24850 * config/mips/mips.md (compression): New attribute.
24851 (enabled): New attribute.
24852 (length): Consider compression in computing length.
24853 (shift_compression): New code attribute.
24854 (*add<mode>3): New operands. Record compression.
24855 (sub<mode>3): Likewise.
24856 (one_cmpl<mode>2): Likewise.
24857 (*and<mode>3): Likewise.
24858 (*ior<mode>3): Likewise.
24859 (unnamed pattern for xor): Likewise.
24860 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
24861 (*<optab><mode>3): Likewise.
24862 (*mov<mode>_internal: Likewise.
24863 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
24864 (mips_unsigned_immediate_p): New.
24865 (umips_lwsp_swsp_address_p): New.
24866 (m16_based_address_p): New.
24867 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
24868 (mips_unsigned_immediate_p): New prototype.
24869 (lwsp_swsp_address_p): New prototype.
24870 (m16_based_address_p): New prototype.
24871 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
24872 (mips_signed_immediate_p): New function.
24873 (m16_based_address_p): New function.
24874 (lwsp_swsp_address_p): New function.
24875 (mips_print_operand_punctuation): Recognize short delay slot insns
24876 for microMIPS.add<mode>3"
24878 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24881 * config/arm/iterators.md (v_cmp_result): New mode attribute.
24882 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
24884 2013-03-25 Richard Biener <rguenther@suse.de>
24886 PR tree-optimization/56689
24887 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
24890 2013-03-25 Richard Biener <rguenther@suse.de>
24892 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
24894 (memory_references): Likewise.
24895 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
24896 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
24897 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
24898 (gather_mem_refs_in_loops): Fold into ...
24899 (analyze_memory_references): ... this. Move initialization
24900 to tree_ssa_lim_initialize.
24901 (fill_always_executed_in): Rename to ...
24902 (fill_always_executed_in_1): ... this.
24903 (fill_always_executed_in): Move contains_call computation to
24904 this new function from ...
24905 (tree_ssa_lim_initialize): ... here.
24906 (tree_ssa_lim): Call fill_always_executed_in.
24908 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24910 * postreload.c (reload_combine): Fix code detecting returns.
24912 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24914 * function.c (emit_use_return_register_into_block): On cc0 targets,
24915 do not emit the sequence between cc0 setter and user.
24917 2013-03-25 Kai Tietz <ktietz@redhat.com>
24919 * config/i386/predicates.md (local_symbolic_operand): Interpret
24920 dll-imported symbols as none-local.
24922 2013-03-25 Richard Biener <rguenther@suse.de>
24924 * tree-ssa-loop-im.c (struct depend): Remove.
24925 (struct lim_aux_data): Make depends a vec of gimples.
24926 (free_lim_aux_data): Adjust.
24927 (add_dependency): Likewise.
24928 (set_level): Likewise.
24930 2013-03-25 Richard Biener <rguenther@suse.de>
24932 PR middle-end/56434
24933 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
24934 the pointer returned by calls with ECF_MALLOC set.
24936 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24938 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
24940 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24942 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
24943 using MMXMODE mode iterator.
24944 (*move<mode>_internal): Merge with *movv2sf_internal and
24945 *movv2sf_internal_rex64 using MMXMODE mode iterator.
24947 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24949 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
24950 (record_last_mem_set_info): Likewise.
24952 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
24953 of XNEWVEC followed by memset.
24954 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
24956 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24958 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
24959 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
24960 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
24961 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
24962 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
24963 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
24964 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
24965 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
24966 BARRIER_P instead of GET_CODE.
24968 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24970 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
24971 inaccuracy in the probing code.
24973 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
24974 (ctrapdi4): Likewise.
24976 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24978 * calls.c (expand_call): Add missing guard to code handling return
24979 of non-BLKmode structures in MSB.
24980 * function.c (expand_function_end): Likewise.
24982 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24984 * combine.c (try_combine): Adjust comment. Do not add the set of
24985 insn #0 if the destination indirectly is set or dies in insn #2.
24986 Tidy up code to distribute a new note.
24988 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24990 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
24991 also for alternatives 16 and 17.
24993 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24995 * config/i386/sse.md (*mov<mode>_internal): Merge with
24996 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
24997 Emit insn template depending on type attribute. Use
24998 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24999 movd instead of movq mnemonic for interunit moves. Rewrite mode
25000 attribute calculation. Remove unit attribute calculation.
25001 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
25002 Set prefix_data16 attribute for DImode ssemov types.
25003 Use Ym instead of y for SSE-MMX conversion alternatives.
25004 Reorder operand constraints.
25006 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
25008 * df.h (df_insn_delete): Adjust prototype.
25009 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
25010 and let it decide whether mark the basic block dirty.
25011 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
25012 * df-scan.c (df_insn_info_delete): New helper function, split
25013 off from df_insn_delete.
25014 (df_scan_free_bb_info): Use it.
25015 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
25017 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
25018 that the insn is actually an insn and it has a non-NULL basic block.
25019 Do not mark basic block dirty if only deleting a DEBUG_INSN.
25021 2013-03-22 Richard Biener <rguenther@suse.de>
25023 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
25025 (mem_ref_alloc): Do not allocate them.
25026 (refs_independent_p): Do not query or maintain a cache.
25028 2013-03-22 Richard Biener <rguenther@suse.de>
25030 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
25031 (gather_mem_refs_in_loops): Do not compute it.
25032 (analyze_memory_references): Do not allocate it.
25033 (tree_ssa_lim_finalize): Do not free it.
25034 (for_all_locs_in_loop): Do not query all_refs_in_loop.
25036 2013-03-22 Richard Biener <rguenther@suse.de>
25038 * is-a.h (as_a): Use gcc_checking_assert.
25040 2013-03-22 Ian Bolton <ian.bolton@arm.com>
25042 * config/aarch64/aarch64.c (aarch64_print_operand): New
25043 format specifier for printing a constant in hex.
25044 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
25045 format specifier for printing second operand.
25047 2013-03-22 Richard Biener <rguenther@suse.de>
25049 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
25051 (gather_mem_refs_in_loops): Perform store accumulation here.
25052 (create_vop_ref_mapping_loop): Remove.
25053 (create_vop_ref_mapping): Likewise.
25054 (analyze_memory_references): Initialize refs_stored_in_loop.
25055 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
25056 (record_indep_loop): Remove.
25057 (record_dep_loop): New function.
25058 (ref_indep_loop_p_1): Adjust to only walk over references
25059 in the loop, not its subloops.
25060 (ref_indep_loop_p): Rename to ...
25061 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
25062 maintaining a more fine-grained cache.
25063 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
25064 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
25066 2013-03-22 Richard Biener <rguenther@suse.de>
25068 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
25069 (struct mem_ref): Make accesses_in_loop a vec of a vec of
25070 aggregate mem_ref_loc.
25071 (free_mem_ref_locs): Inline into ...
25072 (memref_free): ... this and adjust.
25073 (mem_ref_alloc): Adjust.
25074 (mem_ref_locs_alloc): Remove.
25075 (record_mem_ref_loc): Adjust.
25076 (get_all_locs_in_loop): Rewrite into ...
25077 (for_all_locs_in_loop): ... this iterator.
25078 (rewrite_mem_ref_loc): New functor.
25079 (rewrite_mem_refs): Use for_all_locs_in_loop.
25080 (sm_set_flag_if_changed): New functor.
25081 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
25082 (ref_always_accessed): New functor.
25083 (ref_always_accessed_p): Use for_all_locs_in_loop.
25085 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25087 * tree-pass.h (PROP_gimple_lvec): New.
25088 * passes.c (dump_properties): Handle PROP_gimple_lvec.
25089 (init_optimization_passes): Move pass_lower_vector.
25090 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
25092 (pass_lower_vector): Provide PROP_gimple_lvec.
25093 (pass_lower_vector_ssa): Likewise.
25094 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
25096 2013-03-21 Mark Wielaard <mjw@redhat.com>
25098 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
25100 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25102 * config/i386/i386.md (*movdi_internal): Disparage slightly
25103 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
25104 conversion alternatives.
25106 2013-03-21 Jakub Jelinek <jakub@redhat.com>
25108 PR middle-end/48087
25109 * diagnostic.def (DK_WERROR): New kind.
25110 * diagnostic.h (werrorcount): Define.
25111 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
25112 promoted to DK_ERROR, increment DK_WERROR counter instead of
25114 * toplev.c (toplev_main): Call print_ignored_options even if
25115 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
25116 even if just werrorcount is non-zero.
25119 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
25121 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
25122 (string_cst_pool_decl): New function.
25123 (optimize_one_addr_into_implicit_ptr): New function.
25124 (resolve_addr_in_expr): Optimize DWARF location expression
25125 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
25126 which doesn't live in memory, but has DW_AT_location or
25127 DW_AT_const_value, or refers to a string literal, into
25128 DW_OP_GNU_implicit_pointer.
25129 (optimize_location_into_implicit_ptr): New function.
25130 (resolve_addr): If removing DW_AT_location of a variable because
25131 it was DW_OP_addr of address of the variable, but the variable doesn't
25132 live in memory, try to emit const value attribute for the initializer.
25134 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25136 * tree.h (VECTOR_TYPE_P): New macro.
25137 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
25138 TYPE_MODE): Use it.
25139 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
25140 VEC_COND_EXPR cannot be lvalues.
25141 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
25143 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25145 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
25146 Restrict the transformation to equal modes.
25148 2013-03-21 Richard Biener <rguenther@suse.de>
25150 PR tree-optimization/39326
25151 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
25152 (MEM_ANALYZABLE): Adjust.
25153 (record_mem_ref_loc): Move bitmap ops ...
25154 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
25155 unanalyzable refs, do not record locations for it.
25156 (analyze_memory_references): Allocate ref zero as shared
25158 (refs_independent_p): Do not test for unanalyzed mems here.
25159 (ref_indep_loop_p_1): Special-case disambiguation against
25160 the unanalyzed ref.
25161 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
25163 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
25165 * config/arm/arm-protos.h (tune_params): Add
25166 prefer_neon_for_64bits field.
25167 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
25168 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
25169 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
25170 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
25171 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
25172 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
25173 (arm_option_override): Handle -mneon-for-64bits new option.
25174 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
25175 (prefer_neon_for_64bits): Declare new variable.
25176 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
25177 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
25178 (arch_enabled): Handle new arch types. Remove support for onlya8
25180 (one_cmpldi2): Use new arch names.
25181 (zero_extend<mode>di2, extend<mode>di2): Ditto.
25182 * config/arm/arm.opt (mneon-for-64bits): Add option.
25183 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
25184 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
25185 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
25187 * doc/invoke.texi (-mneon-for-64bits): Document.
25189 2013-03-21 Richard Biener <rguenther@suse.de>
25191 PR tree-optimization/39326
25192 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
25193 (sort_bbs_in_loop_postorder_cmp): New function.
25194 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
25196 2013-03-21 Richard Biener <rguenther@suse.de>
25198 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
25199 (vect_insert_into_interleaving_chain): Likewise.
25200 (vect_drs_dependent_in_basic_block): Inline ...
25201 (vect_slp_analyze_data_ref_dependence): ... here. New function,
25203 (vect_analyze_data_ref_dependence): ... here. Simplify.
25204 (vect_check_interleaving): Simplify.
25205 (vect_analyze_data_ref_dependences): Likewise. Split out ...
25206 (vect_slp_analyze_data_ref_dependences): ... this new function.
25207 (dr_group_sort_cmp): New function.
25208 (vect_analyze_data_ref_accesses): Compute data-reference groups
25209 here instead of in vect_analyze_data_ref_dependence. Use
25210 a more efficient algorithm.
25211 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
25212 vect_slp_analyze_data_ref_dependences. Call
25213 vect_analyze_data_ref_accesses earlier.
25214 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
25215 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
25216 (vect_slp_analyze_data_ref_dependences): New prototype.
25218 2013-03-21 Richard Biener <rguenther@suse.de>
25220 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
25221 ref is stored in the loop.
25222 (find_refs_for_sm): Walk only over all stores.
25223 (store_motion_loop): Allocate from lim_bitmap_obstack.
25224 (store_motion): Likewise.
25226 2013-03-21 Richard Biener <rguenther@suse.de>
25228 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
25229 Update virtual SSA form.
25231 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25233 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
25234 * configure: Regenerate.
25235 * config.in: Regenerate.
25236 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
25237 if !HAVE_LD_EH_FRAME_CIEV3.
25239 2013-03-21 Richard Biener <rguenther@suse.de>
25241 * tree-cfg.c (verify_expr_no_block): New function.
25242 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
25243 nor DECL_VALUE_EXPR have locations with associated blocks.
25244 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
25245 (clear_unused_block_pointer): Remove code dealing with
25246 blocks in DECL_DEBUG_EXPR locations.
25248 2013-03-21 Richard Biener <rguenther@suse.de>
25250 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
25251 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
25252 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
25253 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
25254 instead of DECL_DEBUG_EXPR_IS_FROM.
25255 * gimplify.c (gimplify_modify_expr): Likewise.
25256 * tree-cfg.c (verify_expr_location_1): Likewise.
25257 * tree-complex.c (create_one_component_var): Likewise.
25258 * tree-sra.c (create_access_replacement): Likewise.
25259 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
25260 (clear_unused_block_pointer): Likewise.
25261 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
25262 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
25263 * var-tracking.c (var_debug_decl): Likewise.
25264 (track_expr_p): Likewise.
25265 * tree-inline.c (add_local_variables): Likewise. Set
25266 DECL_HAS_DEBUG_EXPR_P after copying it.
25267 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
25268 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
25270 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25273 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
25274 * configure: Regenerate.
25275 * config.in: Regenerate.
25276 * config/i386/i386.md (*movdf_internal): Use
25277 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25278 movd instead of movq mnemonic for interunit moves.
25279 (*movdi_internal): Ditto.
25281 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25283 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
25284 (abd<mode>_3): New pattern.
25285 (aba<mode>_3): New pattern.
25286 (fabd<mode>_3): New pattern.
25288 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25290 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
25291 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
25292 occurrence of REGISTER_PREFIX as its empty string.
25294 2013-03-20 Jeff Law <law@redhat.com>
25296 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
25297 addititional equivalences for equality comparisons between an SSA_NAME
25298 and a constant where the SSA_NAME was set from a widening conversion.
25300 2013-03-20 Walter Lee <walt@tilera.com>
25302 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
25304 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25306 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
25307 depending on type attribute.
25308 (*movti_internal): Ditto.
25309 (*movtf_internal): Ditto.
25310 (*movxf_internal): Ditto.
25311 (*movdf_internal): Ditto.
25312 (*movsf_internal): Ditto.
25314 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25316 * config/i386/i386.md (*movti_internal): Set prefix attribute to
25317 maybe_vex for sselog1 and ssemov types.
25318 (*movdi_internal): Reorder operand constraints.
25319 (*movsi_internal): Ditto. Set prefix attribute to
25320 maybe_vex for sselog1 and ssemov types.
25321 (*movtf_internal): Set prefix attribute to maybe_vex
25322 for sselog1 and ssemov types.
25323 (*movdf_internal): Ditto. Set prefix_data16 attribute for
25324 DImode ssemov types. Reorder operand constraints.
25325 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
25326 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
25327 attribute for SImode ssemov types. Reorder operand constraints.
25329 2013-03-20 Martin Jambor <mjambor@suse.cz>
25331 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
25332 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
25334 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
25336 * config/rs6000/predicates.md (indexed_address, update_address_mem
25337 update_indexed_address_mem): New predicates.
25338 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
25339 attribute for load/store instructions.
25340 * config/rs6000/dfp.md (movsd_store): Likewise.
25341 (movsd_load): Likewise.
25342 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
25343 (unnamed HI->DI extend define_insn): Likewise.
25344 (unnamed SI->DI extend define_insn): Likewise.
25345 (unnamed QI->SI extend define_insn): Likewise.
25346 (unnamed QI->HI extend define_insn): Likewise.
25347 (unnamed HI->SI extend define_insn): Likewise.
25348 (unnamed HI->SI extend define_insn): Likewise.
25349 (extendsfdf2_fpr): Likewise.
25350 (movsi_internal1): Likewise.
25351 (movsi_internal1_single): Likewise.
25352 (movhi_internal): Likewise.
25353 (movqi_internal): Likewise.
25354 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
25355 attribute for load/store instructions.
25356 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
25358 (mov<mode>_softfloat): Likewise.
25359 (mov<mode>_hardfloat32): Likewise.
25360 (mov<mode>_hardfloat64): Likewise.
25361 (mov<mode>_softfloat64): Likewise.
25362 (movdi_internal32): Likewise.
25363 (movdi_internal64): Likewise.
25364 (probe_stack_<mode>): Likewise.
25366 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
25368 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
25369 floating point, and decimal floating point to reload iterator.
25371 * config/rs6000/constraints.md (wl constraint): New constraints to
25372 return FLOAT_REGS if certain options are used to reduce the number
25373 of separate patterns that exist in the file.
25374 (wx constraint): Likewise.
25375 (wz constraint): Likewise.
25377 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25378 -mdebug=reg, print wg, wl, wx, and wz constraints.
25379 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25380 Initialize the reload functions for 64-bit binary/decimal floating
25382 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25383 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25384 create the buffer on the stack to overcome not having a 32-bit
25386 (rs6000_emit_move): Likewise.
25387 (rs6000_secondary_memory_needed_rtx): Likewise.
25388 (rs6000_alloc_sdmode_stack_slot): Likewise.
25389 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25390 via xxlxor, just like DFmode 0.0.
25392 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
25393 define as 1 if we are running on a power7 or newer.
25394 (enum r6000_reg_class_enum): Add new constraints.
25396 * config/rs6000/dfp.md (movsd): Delete, combine with binary
25397 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25398 with other moves by using conditional constraits (wg). Use LFIWZX
25399 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
25400 (movsd splitter): Likewise.
25401 (movsd_hardfloat): Likewise.
25402 (movsd_softfloat): Likewise.
25404 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
25405 binary and decimal floating point moves.
25406 (fmove_ok): New attributes to combine binary and decimal floating
25407 point moves, and to combine power6x (mfpgpr) moves along normal
25409 (real_value_to_target): Likewise.
25410 (f32_lr): Likewise.
25411 (f32_lm): Likewise.
25412 (f32_li): Likewise.
25413 (f32_sr): Likewise.
25414 (f32_sm): Likewise.
25415 (f32_si): Likewise.
25416 (movsf): Combine binary and decimal floating point moves. Combine
25417 power6x (mfpgpr) moves with other moves by using conditional
25418 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
25419 (mov<mode> for SFmode/SDmode); Likewise.
25420 (SFmode/SDmode splitters): Likewise.
25421 (movsf_hardfloat): Likewise.
25422 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
25423 (movsf_softfloat): Likewise.
25424 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
25426 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
25427 wx and wz constraints.
25429 * config/rs6000/constraints.md (wg constraint): New constraint to
25430 return FLOAT_REGS if -mmfpgpr (power6x) was used.
25432 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
25435 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25436 -mdebug=reg, print wg, wl, wx, and wz constraints.
25437 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25438 Initialize the reload functions for 64-bit binary/decimal floating
25440 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25441 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25442 create the buffer on the stack to overcome not having a 32-bit
25444 (rs6000_emit_move): Likewise.
25445 (rs6000_secondary_memory_needed_rtx): Likewise.
25446 (rs6000_alloc_sdmode_stack_slot): Likewise.
25447 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25448 via xxlxor, just like DFmode 0.0.
25450 * config/rs6000/dfp.md (movdd): Delete, combine with binary
25451 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25452 with other moves by using conditional constraits (wg). Use LFIWZX
25453 and STFIWX for loading SDmode on power7.
25454 (movdd splitters): Likewise.
25455 (movdd_hardfloat32): Likewise.
25456 (movdd_softfloat32): Likewise.
25457 (movdd_hardfloat64_mfpgpr): Likewise.
25458 (movdd_hardfloat64): Likewise.
25459 (movdd_softfloat64): Likewise.
25461 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
25462 64-bit binary and decimal floating point moves.
25463 (FMOVE64X): Likewise.
25464 (movdf): Combine 64-bit binary and decimal floating point moves.
25465 Combine power6x (mfpgpr) moves with other moves by using
25466 conditional constraits (wg).
25467 (mov<mode> for DFmode/DDmode): Likewise.
25468 (DFmode/DDmode splitters): Likewise.
25469 (movdf_hardfloat32): Likewise.
25470 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
25471 (movdf_softfloat32): Likewise.
25472 (movdf_hardfloat64_mfpgpr): Likewise.
25473 (movdf_hardfloat64): Likewise.
25474 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
25475 (movdf_softfloat64): Likewise.
25476 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
25477 (reload_<mode>_load): Move to later in the file so they aren't in
25478 the middle of the floating point move insns.
25479 (reload_<mode>_store): Likewise.
25481 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
25484 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
25485 constraint if -mdebug=reg.
25486 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
25487 Enable using dd reload support if needed.
25489 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
25490 binary and decimal floating point moves in rs6000.md.
25491 (movtd_internal): Likewise.
25493 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
25494 decimal floating point moves.
25496 (movtf_internal): Likewise.
25497 (mov<mode>_internal, TDmode/TFmode): Likewise.
25498 (movtf_softfloat): Likewise.
25499 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
25501 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
25502 movdi_internal64, using wg constraint for move direct operations.
25503 (movdi_internal64): Likewise.
25505 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
25506 MODES_TIEABLE_P for selected modes. Print the numerical value of
25507 the various virtual registers. Use GPR/FPR first/last values,
25508 instead of hard coding the register numbers. Print which modes
25509 have reload functions registered.
25510 (rs6000_option_override_internal): If -mdebug=reg, trace the options
25511 settings before/after setting cpu, target and subtarget settings.
25512 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
25513 and for secondary reload failures in rs6000_secondary_reload_inner.
25514 (rs6000_secondary_reload_fail): Likewise.
25515 (rs6000_secondary_reload_inner): Likewise.
25517 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
25518 macros for first/last GPR and FPR registers.
25519 (LAST_GPR_REGNO): Likewise.
25520 (FIRST_FPR_REGNO): Likewise.
25521 (LAST_FPR_REGNO): Likewise.
25523 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
25524 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
25525 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
25526 (vcond<mode><mode>): Likewise.
25527 (vcondu<mode><mode>): Likewise.
25528 (vector_gtu<mode>): Likewise.
25529 (vector_gte<mode>): Likewise.
25530 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
25531 to prevent the compiler from converting DImode operations to TImode.
25532 (ior<mode>3): Likewise.
25533 (and<mode>3): Likewise.
25534 (one_cmpl<mode>2): Likewise.
25535 (nor<mode>3): Likewise.
25536 (andc<mode>3): Likewise.
25538 * config/rs6000/constraints.md (wt constraint): New constraint
25539 that returns VSX_REGS if TImode is allowed in VSX registers.
25541 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
25542 constant under VSX.
25544 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
25545 similar to TImode, but it is restricted to being in the GPRs.
25547 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
25548 TImode to occupy a single VSX register.
25550 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
25551 -mvsx-timode for power7/power8.
25552 (power7 cpu): Likewise.
25553 (power8 cpu): Likewise.
25555 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
25556 sure that TFmode/TDmode take up two registers if they are ever
25557 allowed in the upper VSX registers.
25558 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
25560 (rs6000_init_hard_regno_mode_ok): Likewise.
25561 (rs6000_debug_reg_global): Add debugging for PTImode and wt
25562 constraint. Print if LRA is turned on.
25563 (rs6000_option_override_internal): Give an error if -mvsx-timode
25564 and VSX is not enabled.
25565 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
25566 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
25567 to reg+offset addressing. Use PTImode when checking offset
25568 addresses for validity.
25569 (reg_offset_addressing_ok_p): Likewise.
25570 (rs6000_legitimate_offset_address_p): Likewise.
25571 (rs6000_legitimize_address): Likewise.
25572 (rs6000_legitimize_reload_address): Likewise.
25573 (rs6000_legitimate_address_p): Likewise.
25574 (rs6000_eliminate_indexed_memrefs): Likewise.
25575 (rs6000_emit_move): Likewise.
25576 (rs6000_secondary_reload): Likewise.
25577 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
25578 reloads to fpr registers to continue to use reg+offset addressing,
25579 but 64-bit reloads to altivec registers need reg+reg addressing.
25580 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
25581 it. Treat LO_SUM like a PLUS operation.
25582 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
25583 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
25584 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
25585 registers to share a register with a smaller sized type, since VSX
25586 puts scalars in the upper 64-bits.
25587 (print_operand): Add support for PTImode.
25588 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
25589 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
25590 registers, but don't have arithmetic support.
25591 (rs6000_memory_move_cost): Add test for VSX.
25592 (rs6000_opt_masks): Add -mvsx-timode.
25594 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
25597 (VSr): Use wt constraint for TImode.
25598 (VSv): Drop TImode support.
25599 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
25600 (vsx_movti_64bit): Likewise.
25601 (vsx_movti_32bit): Likewise.
25602 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
25603 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
25604 one '?' on the appropriate output constraint. Do not allow TImode
25605 logical operations on 32-bit systems.
25606 (vsx_ior<mode>3): Likewise.
25607 (vsx_xor<mode>3): Likewise.
25608 (vsx_one_cmpl<mode>2): Likewise.
25609 (vsx_nor<mode>3): Likewise.
25610 (vsx_andc<mode>3): Likewise.
25611 (vsx_concat_<mode>): Likewise.
25612 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
25614 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
25615 OPTION_MASK_VSX_TIMODE.
25616 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
25617 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
25619 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
25620 (TI2 iterator): New iterator for TImode, PTImode.
25621 (wd mode attribute): Add values for vector types.
25622 (movti_string): Replace TI move operations with operations for TImode
25623 and PTImode. Add support for TImode being allowed in VSX registers.
25624 (mov<mode>_string, TImode/PTImode): Likewise.
25625 (movti_ppc64): Likewise.
25626 (mov<mode>_ppc64, TImode/PTImode): Likewise.
25627 (TI mode splitters): Likewise.
25629 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
25632 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
25634 PR tree-optimization/56355
25635 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
25636 Also handle integers with undefined overflow.
25638 2013-03-20 Catherine Moore <clm@codesourcery.com>
25639 Maciej W. Rozycki <macro@codesourcery.com>
25640 Tom de Vries <tom@codesourcery.com>
25641 Nathan Sidwell <nathan@codesourcery.com>
25642 Iain Sandoe <iain@codesourcery.com>
25643 Nathan Froyd <froydnj@codesourcery.com>
25644 Chao-ying Fu <fu@mips.com>
25646 * doc/extend.texi: (micromips, nomicromips, nocompression):
25647 Document new function attributes.
25648 * doc/invoke.texi (minterlink-compressed, mmicromips,
25649 m14k, m14ke, m14kec): Document new options.
25650 (minterlink-mips16): Update documentation.
25651 * doc/md.texi (ZC, ZD): Document new constraints.
25652 * configure.ac (gcc_cv_as_micromips): Check if linker
25653 supports the .set micromips directive.
25654 * configure: Regenerate.
25655 * config.in: Regenerate.
25656 * config/mips/mips-tables.opt: Regenerate.
25657 * config/mips/micromips.md: New file.
25658 * constraints.md (ZC, ZD): New constraints.
25659 * config/mips/predicates.md (movep_src_register): New predicate.
25660 (movep_src_operand): New predicate.
25661 (non_volatile_mem_operand): New predicate.
25662 * config/mips/mips.md (multimem): New type.
25663 (length): Differentiate between 17-bit and 18-bit branch offsets.
25664 (MOVEP1, MOVEP2): New mode iterator.
25665 (mov_<load>l): Use ZC constraint.
25666 (mov_<load>r): Likewise.
25667 (mov_<store>l): Likewise.
25668 (mov_<store>r): Likewise.
25669 (*branch_equality<mode>_inverted): Add microMIPS support.
25670 (*branch_equality<mode>): Likewise.
25671 (*jump_absolute): Likewise.
25672 (indirect_jump_<mode>): Likewise.
25673 (tablejump_<mode>): Likewise.
25674 (<optab>_internal): Likewise.
25675 (sibcall_internal): Likewise.
25676 (sibcall_value_internal): Likewise.
25677 (prefetch): Use constraint ZD.
25678 * config/mips/mips.opt (minterlink-compressed): New option.
25679 (minterlink-mips16): Now an alias for minterlink-compressed.
25680 (mmicromips): New option.
25681 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
25682 (compare_and_swap_12): Likewise.
25683 (sync_add<mode>): Likewise.
25684 (sync_<optab>_12): Likewise.
25685 (sync_old_<optab>_12): Likewise.
25686 (sync_new_<optab>_12): Likewise.
25687 (sync_nand_12): Likewise.
25688 (sync_old_nand_12): Likewise.
25689 (sync_new_nand_12): Likewise.
25690 (sync_sub<mode>): Likewise.
25691 (sync_old_add<mode>): Likewise.
25692 (sync_old_sub<mode>): Likewise.
25693 (sync_new_add<mode>): Likewise.
25694 (sync_new_sub<mode>): Likewise.
25695 (sync_<optab><mode>): Likewise.
25696 (sync_old_<optab><mode>): Likewise.
25697 (sync_new_<optab><mode>): Likewise.
25698 (sync_nand<mode>): Likewise.
25699 (sync_old_nand<mode>): Likewise.
25700 (sync_new_nand<mode>): Likewise.
25701 (sync_lock_test_and_set<mode>): Likewise.
25702 (test_and_set_12): Likewise.
25703 (atomic_compare_and_swap<mode>): Likewise.
25704 (atomic_exchange<mode>_llsc): Likewise.
25705 (atomic_fetch_add<mode>_llsc): Likewise.
25706 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
25707 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
25708 (umips_save_restore_pattern_p): Likewise.
25709 (umips_load_store_pair_p): Likewise.
25710 (umips_output_load_store_pair): Likewise.
25711 (umips_movep_target_p): Likewise.
25712 (umips_12bit_offset_address_p): Likewise.
25713 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
25714 (mips_base_mips16): Rename this...
25715 (mips_base_compression_flags): ...to this. Update all uses.
25716 (mips_attribute_table): Add micromips, nomicromips and nocompression.
25717 (mips_mips16_decl_p): Delete.
25718 (mips_nomips16_decl_p): Delete.
25719 (mips_get_compress_on_flags): New function.
25720 (mips_get_compress_off_flags): New function.
25721 (mips_get_compress_mode): New function.
25722 (mips_get_compress_on_name): New function.
25723 (mips_get_compress_off_name): New function.
25724 (mips_insert_attributes): Support multiple compression types.
25725 (mips_merge_decl_attributes): Likewise.
25726 (umips_12bit_offset_address_p): New function.
25727 (mips_start_function_definition): Emit .set micromips directive.
25728 (mips_call_may_need_jalx_p): New function.
25729 (mips_function_ok_for_sibcall): Add microMIPS support.
25730 (mips_print_operand_punctuation): Support short delay slots and
25732 (umips_swm_mask, umips_swm_encoding): New.
25733 (umips_build_save_restore): New function.
25734 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
25735 (was_mips16_p): Remove.
25736 (old_compression_mode): New.
25737 (mips_set_compression_mode): New function.
25738 (mips_set_current_function): Add microMIPS support.
25739 (mips_option_override): Likewise.
25740 (umips_save_restore_pattern_p): New function.
25741 (umips_output_save_restore): New function.
25742 (umips_load_store_pair_p_1): New function.
25743 (umips_load_store_pair_p): New function.
25744 (umips_output_load_store_pair_1): New function.
25745 (umips_output_load_store_pair): New function.
25746 (umips_movep_target_p) New function.
25747 (mips_prepare_pch_save): Add microMIPS support.
25748 * config/mips/mips.h (TARGET_COMPRESSION): New.
25749 (TARGET_CPU_CPP_BUILTINS): Update macro
25750 to use new compression flags and to support microMIPS.
25751 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
25752 (MIPS_ARCH_FLOAT_SPEC): Likewise.
25753 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
25754 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
25755 (ASM_SPEC): Support mmicromips and mno-micromips.
25756 (M16STORE_REG_P): New macro.
25757 (MIPS_CALL): Support TARGET_MICROMIPS.
25758 (MICROMIPS_J): New macro.
25759 (mips_base_mips16): Rename this...
25760 (mips_base_compression_flags): ...to this.
25761 (UMIPS_12BIT_OFFSET_P): New macro.
25762 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
25763 (MULTILIB_DIRNAMES): Likewise.
25764 2013-03-20 Richard Biener <rguenther@suse.de>
25766 PR tree-optimization/56661
25767 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
25768 the result does not have to be distinct.
25770 2013-03-20 Richard Biener <rguenther@suse.de>
25772 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
25775 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
25776 Steven Bosscher <steven@gcc.gnu.org>
25778 PR rtl-optimization/56605
25779 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
25781 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25784 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
25785 that require movd instead of movq.
25787 2013-03-20 Richard Biener <rguenther@suse.de>
25789 * tree-ssa-structalias.c (struct variable_info): Add pointer
25790 to the first field of an aggregate with sub-vars. Make
25791 this and the pointer to the next subfield its ID.
25792 (vi_next): New function.
25793 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
25794 storedanything_id, integer_id): Increment by one.
25795 (new_var_info, get_call_vi, lookup_call_clobber_vi,
25796 get_call_clobber_vi): Adjust.
25797 (solution_set_expand): Simplify and speedup.
25798 (solution_set_add): Inline into ...
25799 (set_union_with_increment): ... this. Adjust accordingly.
25800 (do_sd_constraint): Likewise.
25801 (do_ds_constraint): Likewise.
25802 (do_complex_constraint): Simplify.
25803 (build_pred_graph): Adjust.
25804 (solve_graph): Likewise. Simplify and speedup.
25805 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
25806 get_constraint_for_component_ref, get_constraint_for_1,
25807 first_vi_for_offset, first_or_preceding_vi_for_offset,
25808 create_function_info_for, create_variable_info_for_1,
25809 create_variable_info_for, intra_create_variable_infos): Adjust.
25810 (init_base_vars): Push NULL for ID zero.
25811 (compute_points_to_sets): Adjust.
25813 2013-03-20 Richard Biener <rguenther@suse.de>
25815 * cfgloop.c (verify_loop_structure): Streamline and avoid
25816 ICEing on corrupt loop tree.
25817 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
25820 2013-03-20 Richard Biener <rguenther@suse.de>
25822 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
25823 check whether an SSA update is needed.
25825 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
25827 * config/mips/constraints.md (T): Rename to...
25831 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
25832 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
25834 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25836 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
25837 (*subsi3_carryin_uxtw): Likewise.
25839 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25841 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
25842 (*rorsi3_insn_uxtw): Likewise.
25844 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25846 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
25847 (*extrsi5_insn_uxtw): Likewise.
25849 2013-03-19 Richard Biener <rguenther@suse.de>
25851 PR tree-optimization/56273
25852 * passes.c (init_optimization_passes): Move second VRP after DOM.
25854 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
25856 * config/i386/i386.md (*movti_internal): Merge from
25857 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
25858 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
25859 nox64 isa attributes.
25861 2013-03-18 Richard Biener <rguenther@suse.de>
25863 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
25865 (merge_node_constraints): Likewise.
25866 (build_succ_graph): Likewise.
25867 (valid_graph_edge): Inline into single caller.
25868 (unify_nodes): Likewise. Use bitmap_set_bit return value
25870 (scc_visit): Fix formatting and variable use.
25871 (do_sd_constraint): Use gcc_checking_assert.
25872 (do_ds_constraint): Likewise.
25873 (do_complex_constraint): Likewise.
25874 (condense_visit): Likewise. Cleanup.
25875 (dump_pred_graph): New function.
25876 (perform_var_substitution): Dump the pred-graph before
25877 variable substitution.
25878 (find_equivalent_node): Use gcc_checking_assert.
25879 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
25881 2013-03-18 Richard Biener <rguenther@suse.de>
25883 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
25884 Remove cond_expr_stmt_list argument and do not gimplify the
25886 (vect_loop_versioning): Adjust.
25887 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
25888 Cleanup to use less temporaries.
25889 (vect_create_data_ref_ptr): Cleanup.
25891 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25893 PR tree-optimization/56635
25894 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
25895 require types_compatible_p types.
25897 2013-03-18 Nick Clifton <nickc@redhat.com>
25899 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
25900 spurious backslash.
25902 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
25903 Add missing line to comment describing function.
25905 2013-03-18 Richard Biener <rguenther@suse.de>
25907 PR tree-optimization/56210
25908 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
25909 Handle string / character search functions.
25910 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
25912 2013-03-18 Richard Biener <rguenther@suse.de>
25914 PR middle-end/56483
25915 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
25916 and implement properly.
25917 * gimple.h (gimple_cond_single_var_p): Remove.
25919 2013-03-18 Richard Biener <rguenther@suse.de>
25921 * tree-data-ref.h (find_data_references_in_loop): Declare.
25922 * tree-data-ref.c (get_references_in_stmt): Use a stack
25923 vector pre-allocated in the callers.
25924 (find_data_references_in_stmt): Adjust.
25925 (graphite_find_data_references_in_stmt): Likewise.
25926 (create_rdg_vertices): Likewise.
25927 (find_data_references_in_loop): Export.
25928 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
25929 Compute dependences here...
25930 (vect_analyze_data_refs): ...not here. When we encounter
25931 a non-vectorizable data reference in basic-block vectorization
25932 truncate the data reference vector. Do not bother to
25933 fixup data-dependence information for gather loads.
25934 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
25935 of data references, as reported.
25937 2013-03-18 Richard Biener <rguenther@suse.de>
25939 PR tree-optimization/3713
25940 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
25941 has_constants and expr.
25942 (stmt_has_constants): Properly valueize SSA names when deciding
25943 whether the stmt has constants.
25945 2013-03-18 Richard Biener <rguenther@suse.de>
25947 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
25948 whole function when there is nothing to do.
25949 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
25950 * tree-vectorizer.c (vectorize_loops): Update virtual and
25951 loop-closed SSA once.
25952 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
25954 2013-03-18 Richard Biener <rguenther@suse.de>
25956 PR middle-end/56113
25957 * domwalk.c (bb_postorder): New global static.
25958 (cmp_bb_postorder): New function.
25959 (walk_dominator_tree): Replace scheme imposing an order for
25960 visiting dominator sons by one sorting them at the time they
25961 are pushed on the stack.
25963 2013-03-18 Richard Biener <rguenther@suse.de>
25965 PR tree-optimization/39326
25966 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
25967 (struct mem_ref): Replace mem member with ao_ref typed member.
25968 (MEM_ANALYZABLE): Adjust.
25969 (memref_eq): Likewise.
25970 (mem_ref_alloc): Likewise.
25971 (gather_mem_refs_stmt): Likewise.
25972 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
25973 (execute_sm_if_changed_flag_set): Adjust.
25974 (execute_sm): Likewise.
25975 (ref_always_accessed_p): Likewise.
25976 (refs_independent_p): Likewise.
25977 (can_sm_ref_p): Likewise.
25979 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25982 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
25983 return 1 even for !unsignedp.
25985 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25987 * config/i386/i386.md (isa): Add x64 and nox64.
25988 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
25989 (*pushtf): Enable *roF alternative for x64 isa only.
25990 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
25991 mode attribute of integer alternatives to DImode for TARGET_64BIT.
25992 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
25993 (*movtf_internal): Merge from *movtf_internal_rex64 and
25994 *movtf_internal_sse. Use x64 and nox64 isa attributes.
25995 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
25996 nox64 isa attributes.
25997 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
25998 nox64 isa attributes.
25999 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
26001 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26003 * config/alpha/alpha.c (TARGET_LRA_P): New define.
26005 2013-03-17 Jakub Jelinek <jakub@redhat.com>
26008 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
26009 class names. Remove trailing comma after "ALL_REGS".
26011 2013-03-16 Jan Hubicka <jh@suse.cz>
26013 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
26014 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
26015 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
26016 of cgraph_get_create_node.
26017 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
26019 2013-03-16 Jason Merrill <jason@redhat.com>
26022 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
26023 with DW_AT_default_value.
26025 2013-03-16 Jakub Jelinek <jakub@redhat.com>
26027 * BASE-VER: Set to 4.9.0.
26029 2013-03-14 Andi Kleen <ak@linux.intel.com>
26032 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
26033 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
26034 Document _x* TSX intrinsics.
26036 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
26037 David Holsgrove <david.holsgrove@xilinx.com>
26039 * configure.ac: Add MicroBlaze TLS support detection.
26040 * configure: Regenerate.
26041 * config/microblaze/microblaze-protos.h
26042 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
26043 symbol_mentioned_p, label_mentioned_p): Add prototypes.
26044 * config/microblaze/microblaze.c (microblaze_address_type): Add
26045 ADDRESS_TLS and tls_reloc address types.
26046 (microblaze_address_info): Add tls_reloc.
26047 (TARGET_HAVE_TLS): Define.
26048 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
26049 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
26050 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
26051 load_tls_operand, microblaze_call_tls_get_addr,
26052 microblaze_legitimize_tls_address): New functions.
26053 (microblaze_classify_unspec): Handle UNSPEC_TLS.
26054 (get_base_reg): Use microblaze_tls_symbol_p.
26055 (microblaze_classify_address): Handle TLS.
26056 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
26057 label_mentioned_p and microblaze_tls_referenced_p.
26058 (microblaze_legitimize_address): Handle TLS.
26059 (microblaze_address_insns): Handle ADDRESS_TLS.
26060 (pic_address_needs_scratch): Handle TLS.
26061 (print_operand_address): Handle TLS.
26062 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
26063 (microblaze_expand_move): Handle TLS.
26064 (microblaze_legitimate_constant_p): Check
26065 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
26066 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
26067 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
26068 (PIC_OFFSET_TABLE_REGNUM): Set.
26069 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
26070 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
26071 (addsi3, movsi_internal2, movdf_internal): Update constraints
26072 * config/microblaze/predicates.md (arith_plus_operand): Define
26073 (move_operand): Redefine as move_src_operand,
26074 check microblaze_tls_referenced_p.
26076 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26078 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
26079 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
26081 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26083 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
26086 2013-03-14 Jakub Jelinek <jakub@redhat.com>
26088 PR tree-optimization/53265
26089 * common.opt (Waggressive-loop-optimizations): New option.
26090 * tree-ssa-loop-niter.c: Include tree-pass.h.
26091 (do_warn_aggressive_loop_optimizations): New function.
26092 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
26093 if number_of_latch_executions returned constant.
26094 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
26095 early. If number_of_latch_executions returned constant, set
26096 nb_iterations_upper_bound back to it.
26097 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
26099 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
26100 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
26102 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
26103 (MULTILIB_OSDIRNAMES): Set.
26104 * genmultilib: If defaultosdirname doesn't start with :: , set
26105 defaultosdirname2 instead, clear it and emit two . multilib_raw
26106 entries instead of just one.
26108 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
26110 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26111 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
26112 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26113 (SUBTARGET_OVERRIDE_OPTIONS): New.
26115 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
26118 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
26119 (musermode): Convert to Var(TARGET_USERMODE).
26120 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
26121 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
26122 * config/sh/sh.c (sh_option_override): Use
26123 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
26124 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
26126 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
26128 (udivsi3_i4_single, divsi3_i4_single): Use
26129 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
26131 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
26133 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
26136 2013-03-13 Richard Biener <rguenther@suse.de>
26138 PR tree-optimization/56608
26139 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
26140 calls when vectorizing basic-blocks.
26142 2013-03-13 Jakub Jelinek <jakub@redhat.com>
26145 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
26148 2013-03-12 Jakub Jelinek <jakub@redhat.com>
26150 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
26152 2013-03-11 Jan Hubicka <jh@suse.cz>
26155 * lto-streamer-out.c (output_symbol_p): Skip references from
26156 constructors of external variables.
26158 2013-03-11 Jan Hubicka <jh@suse.cz>
26160 PR middle-end/56571
26161 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
26163 * emit-rtl.c (verify_rtx_sharing): Likewise.
26164 (copy_insn_1): Likewise.
26165 * rtl.c (copy_rtx): Likewise.
26167 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
26170 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
26171 output_operand_lossage message.
26173 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
26176 * arm.c (shift_op): Validate RTL pattern on the fly.
26177 (arm_print_operand, case 'S'): Don't use shift_operator to validate
26180 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26183 * config/pa/pa.md (call_value): Check for calls to powf and direct to
26184 new call patterns that clobber %fr12.
26185 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
26186 split and postreload patterns.
26187 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
26188 registers %fr12 and %fr12R as call used.
26190 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
26192 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
26193 (canon_address, record_store, replace_read, check_mem_read_rtx,
26194 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
26195 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
26196 rest_of_handle_dse): Likewise.
26198 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
26200 PR middle-end/56524
26201 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
26203 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
26204 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
26205 (save_optabs_if_changed): Replace with...
26206 (init_tree_optimization_optabs): ...this.
26207 * optabs.c (save_optabs_if_changed): Rename to...
26208 (init_tree_optimization_optabs): ...this. Take the optimization node
26209 as argument. Do nothing if the base optabs are already correct.
26210 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
26211 to recompute optabs.
26212 * function.h (function): Remove optabs field.
26213 * function.c (invoke_set_current_function_hook): Call
26214 init_tree_optimization_optabs. Use the result to initialize
26217 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
26219 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
26220 if GTMA_HAS_NO_INSTRUMENTATION.
26221 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
26222 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
26223 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
26224 * gimple-pretty-print.c (dump_gimple_transaction): Handle
26225 GTMA_HAS_NO_INSTRUMENTATION.
26227 2013-03-08 Jakub Jelinek <jakub@redhat.com>
26229 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
26232 2013-03-08 Marek Polacek <polacek@redhat.com>
26233 Jakub Jelinek <jakub@redhat.com>
26235 PR tree-optimization/56478
26236 * predict.c (is_comparison_with_loop_invariant_p): Change the
26237 type of loop_step to tree.
26238 (predict_loops): Adjust.
26239 (predict_iv_comparison): Perform the computations on double_ints.
26241 2013-03-08 Richard Biener <rguenther@suse.de>
26243 PR tree-optimization/56570
26244 * tree-cfg.c (verify_expr_location_1): Verify locations for
26246 * tree-sra.c (create_access_replacement): Strip locations
26247 from DECL_DEBUG_EXPRs.
26249 2013-03-08 Richard Biener <rguenther@suse.de>
26251 * tree-inline.c (expand_call_inline): Do not associate
26252 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
26253 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
26255 2013-03-08 Richard Biener <rguenther@suse.de>
26257 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
26258 or block changes with -Og. Fix for location / block encoding
26259 changes and PHI arguments with locations.
26261 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
26263 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
26265 (struct output_info): Likewise.
26266 (register_overhead): Remove bad gcc_assert.
26267 (bitmap_find_bit): If there is only a single bitmap element, do not
26268 count a miss as a search.
26269 (print_statistics): Update for counter type changes.
26270 (dump_bitmap_statistics): Likewise. Print headers such that they
26271 are properly lined up with the printed counters.
26273 2013-03-07 Jakub Jelinek <jakub@redhat.com>
26275 PR tree-optimization/56559
26276 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
26277 check that it has only a single use.
26279 2013-03-07 Richard Biener <rguenther@suse.de>
26281 * doc/invoke.texi (fwhole-program): Discourage use in combination
26284 2013-03-06 Jakub Jelinek <jakub@redhat.com>
26286 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
26288 PR tree-optimization/56539
26289 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
26290 instead of GSI_CONTINUE_LINKING as last argument to
26291 force_gimple_operand_gsi. Adjust function comment.
26293 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
26296 PR middle-end/56548
26297 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
26298 promoted mode, convert the result back to the original mode.
26300 2013-03-06 Richard Biener <rguenther@suse.de>
26302 PR middle-end/56294
26303 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
26304 (insert_updated_phi_nodes_compare_uids): New function.
26305 (update_ssa): Sort symbols_to_rename after UID before
26306 traversing it to insert PHI nodes.
26308 2013-03-06 Richard Biener <rguenther@suse.de>
26310 PR middle-end/50494
26311 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
26312 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
26315 2013-02-13 Richard Biener <rguenther@suse.de>
26318 * varasm.c (output_constant_def_1): Get the decl representing
26319 the constant as argument.
26320 (output_constant_def): Wrap output_constant_def_1.
26321 (make_decl_rtl): Use output_constant_def_1 with the decl
26322 representing the constant.
26323 (build_constant_desc): Optionally re-use a decl already
26324 representing the constant.
26325 (tree_output_constant_def): Adjust.
26327 2013-03-06 Joey Ye <joey.ye@arm.com>
26330 * gcc.c (convert_white_space): New function.
26331 (main): Handles white space in function name.
26333 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
26336 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
26337 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
26338 to SH_DIV_CALL_TABLE for TARGET_SH2.
26339 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
26341 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
26342 call-table options.
26344 2013-03-05 Sterling Augustine <saugustine@google.com>
26345 Cary Coutant <ccoutant@google.com>
26348 * dwarf2out.c (resolve_addr): Don't call
26349 remove_loc_list_addr_table_entries a second time for the same
26352 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26355 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
26356 (avoid_complex_debug_insns): New function.
26357 (expand_debug_locations): Call it.
26359 PR rtl-optimization/56484
26360 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
26361 lifetimes of hard registers on small register class machines.
26363 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26365 * config/microblaze/microblaze-protos.h: Rename
26366 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
26367 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
26369 (microblaze_fast_interrupt_function_p): New function.
26370 (microblaze_is_interrupt_handler): Rename to
26371 microblaze_is_interrupt_variant and add fast_interrupt check.
26372 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
26373 (save_restore_insns): Likewise.
26374 (compute_frame_size): Likewise.
26375 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
26376 (microblaze_globalize_label): Likewise.
26377 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
26378 * config/microblaze/microblaze.md: Use wrapper
26379 microblaze_is_interrupt_variant.
26381 2013-03-05 Kai Tietz <ktietz@redhat.com>
26383 * sdbout.c (sdbout_one_type): Switch to current function's section
26384 supporting cold/hot.
26386 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26388 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
26391 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26393 PR middle-end/56461
26394 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
26395 if VALGRIND_GET_VBITS is defined, temporarily make object
26396 memory all defined, and restore previous valgrind addressability
26397 and definability afterwards. Free this_object at the end.
26399 PR middle-end/56461
26400 * lra.c (lra): Call lra_clear_live_ranges if live_p,
26401 right before calling lra_create_live_ranges, also call it
26402 when clearing live_p. Only call lra_clear_live_ranges
26403 at the end if live_p.
26405 PR middle-end/56461
26406 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
26408 2013-03-05 Richard Biener <rguenther@suse.de>
26410 PR tree-optimization/56521
26411 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
26414 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
26417 * except.h (remove_unreachable_eh_regions): New prototype.
26418 * except.c (remove_eh_handler_splicer): New function, split out
26419 of remove_eh_handler.
26420 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
26421 warning about running it on many EH regions one at a time.
26422 (remove_unreachable_eh_regions_worker): New function, walk the
26423 EH tree in depth-first order and remove non-marked regions.
26424 (remove_unreachable_eh_regions): New function.
26425 * tree-eh.c (mark_reachable_handlers): New function, split out
26426 from remove_unreachable_handlers.
26427 (remove_unreachable_handlers): Use mark_reachable_handlers and
26428 remove_unreachable_eh_regions.
26429 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
26430 and remove_unreachable_eh_regions.
26432 2013-03-05 Richard Biener <rguenther@suse.de>
26434 PR middle-end/56525
26435 * loop-init.c (fix_loop_structure): Remove loops in two stages,
26436 not freeing them until the end.
26438 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26440 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
26442 2013-03-05 Richard Biener <rguenther@suse.de>
26444 PR tree-optimization/56270
26445 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
26446 of loads after scheduling an SLP instance.
26448 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26450 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
26452 (check_gcc_parallelize): Run guality.exp as a separate job from
26453 vect.exp with unsorted.exp and $(dg_target_exps) separately from
26454 struct-layout-1.exp with stackalign.exp.
26456 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
26458 PR middle-end/56461
26459 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
26460 load_index sbitmap even if some bit in it isn't set.
26462 PR middle-end/56461
26463 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
26464 (discover_iteration_bound_by_body_walk): Change queues to
26465 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
26466 spelling in comment. Call safe_push on queues[bound_index] directly.
26467 Release queues[queue_index] in every iteration unconditionally.
26468 Release bounds vector.
26470 PR middle-end/56461
26471 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
26472 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
26473 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
26476 2013-03-05 Richard Biener <rguenther@suse.de>
26479 * tree-inline.c (remap_blocks_to_null): New function.
26480 (expand_call_inline): When expanding a call stmt without
26481 an associated block inline remap all callee blocks to NULL.
26483 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26485 PR rtl-optimization/56494
26486 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
26487 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
26488 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
26490 PR middle-end/56461
26491 * sel-sched-ir.c (free_sched_pools): Release
26492 succs_info_pool.stack[succs_info_pool.max_top] vectors too
26493 if succs_info_pool.max_top isn't -1.
26496 * opts.c (opts_obstack, opts_concat): Moved to...
26497 * opts-common.c (opts_obstack, opts_concat): ... here.
26499 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26501 PR middle-end/56461
26502 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
26504 2013-03-04 Martin Jambor <mjambor@suse.cz>
26506 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
26507 all appropriate places.
26509 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
26511 PR tree-optimization/56424
26512 * ipa-split.c (split_function): Do not set the RSO flag if result is
26513 not by reference and its type is a register type.
26515 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
26517 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
26518 (microblaze_legitimate_pic_operand): Likewise
26519 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
26520 new function microblaze_legitimate_pic_operand
26521 * config/microblaze/microblaze-protos.h
26522 (microblaze_legitimate_pic_operand): Declare.
26524 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26526 * config/microblaze/predicates.md (call_insn_simple_operand):
26527 New predicate for supported rtx code types.
26528 * config/microblaze/microblaze.md (call_internal1): Use
26529 call_insn_simple_operand predicate.
26531 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26533 PR middle-end/56461
26534 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
26535 partitions.ordered_remove.
26537 PR middle-end/56461
26538 * tree-vect-stmts.c (vectorizable_conversion): Don't call
26539 vec_oprnds0.create (1) for modifier == NONE.
26541 PR middle-end/56461
26542 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
26543 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
26544 vec_oprnds1 right before pushing anything to it for
26547 PR middle-end/56461
26548 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
26549 set nbbs to 0 instead of having separate code path.
26550 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
26551 instead of false as last argument if returning NULL.
26553 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
26555 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
26556 the attribute is now called "target" instead of "option".
26557 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
26558 * doc/tm.texi.in (Target Attributes): Likewise document the correct
26559 attribute/pragma name for TARGET_OPTION_VALID_P and
26560 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
26561 * doc/tm.texi: Regenerated.
26563 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
26565 * config/microblaze/microblaze.c:
26566 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
26567 * config/microblaze/microblaze.h: Add -mxl-reorder to
26569 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
26570 instructions emitted if TARGET_REORDER.
26571 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
26572 or 0 for -m/-mno case, but initialises as 2 to detect default use case
26575 2013-03-01 Xinliang David Li <davidxl@google.com>
26577 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
26580 2013-03-01 Jakub Jelinek <jakub@redhat.com>
26582 PR middle-end/56461
26583 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
26584 vector even when returning true. Fix up function comment formatting.
26586 PR middle-end/56461
26587 * ira-build.c (ira_loop_nodes_count): New variable.
26588 (create_loop_tree_nodes): Initialize it.
26589 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
26591 PR middle-end/56461
26592 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
26593 method on dr_chain and result_chain.
26594 * tree-vect-stmts.c (vectorizable_store): Only call
26595 result_chain.create if j == 0.
26597 PR middle-end/56461
26598 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
26599 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
26600 before overwriting it.
26602 2013-03-01 Tobias Burnus <burnus@net-b.de>
26604 * doc/extended.texi (C Extensions): Change order in @menu
26606 (Other MIPS Built-in Functions): Move last MIPS entry before
26607 "picoChip Built-in Functions".
26608 (SH Built-in Functions): Move after RX Built-in Functions.
26609 * doc/gcc.texi (Introduction): Change order in @menu
26611 * doc/md.texi (Constraints): Ditto.
26612 * gty.texi (Type Information): Ditto.
26613 (User-provided marking routines for template types): Make
26615 * doc/invoke.texi (AArch64 Options): Move before
26616 "Adapteva Epiphany Options".
26618 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
26619 Jakub Jelinek <jakub@redhat.com>
26622 * asan.c (gate_asan): Lookup no_sanitize_address instead of
26623 no_address_safety_analysis attribute.
26624 * doc/extend.texi (no_address_safety_attribute): Rename to
26625 no_sanitize_address attribute, mention no_address_safety_analysis
26626 attribute as deprecated alias.
26628 2013-02-28 Jakub Jelinek <jakub@redhat.com>
26630 PR middle-end/56461
26631 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
26632 type to vec<vec<tree> > *.
26633 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
26634 to be vec<tree> instead of vec<tree> *, set vec_defs
26635 to vNULL and call vec_defs.create (number_of_vects), adjust other
26637 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
26638 vectorizable_condition): Adjust vect_get_slp_defs callers.
26640 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26642 * config/aarch64/aarch64.c
26643 (aarch64_float_const_representable): Remove unused variable.
26645 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26647 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
26649 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26651 * config/aarch64/aarch64-builtins.c
26652 (aarch64_init_simd_builtins): Make static.
26654 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26656 * config/aarch64/aarch64.c
26657 (aarch64_simd_make_constant): Make static.
26659 2013-02-28 Martin Jambor <mjambor@suse.cz>
26661 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
26662 with no initialization to the RHS of debug statements.
26664 2013-02-28 Martin Jambor <mjambor@suse.cz>
26666 PR tree-optimization/56294
26667 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
26669 (get_access_replacement): Do not call create_access_replacement.
26670 Assert a replacement exists.
26671 (get_repl_default_def_ssa_name): Create the replacement declaration
26674 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26676 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
26677 final_end_function.
26679 2013-02-28 Marek Polacek <polacek@redhat.com>
26681 PR rtl-optimization/56466
26682 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
26683 if we're changing a loop.
26684 (peel_loops_completely): Likewise.
26686 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
26689 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
26691 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26694 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
26695 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
26696 INTX_FTYPE_FX, FX_FTYPE_INTX.
26697 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
26699 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26701 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
26702 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
26703 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
26704 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
26705 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
26706 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
26707 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
26708 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
26709 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
26710 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
26711 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
26712 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
26713 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
26714 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
26715 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
26716 (avrxmega6): Increase max flash segments from 5 to 6.
26717 * config/avr/t-multilib: Regenerate.
26718 * config/avr/avr-tables.opt: Regenerate.
26719 * doc/avr-mmcu.texi: Regenerate.
26721 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26723 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
26724 (avr_device_to_arch): Rename to avr_device_to_ld.
26725 (avr_device_to_as): New prototype.
26726 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
26727 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
26728 * config/avr/driver-avr.c (avr_device_to_as): New.
26729 (avr_device_to_arch): Rename to avr_device_to_ld.
26731 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26733 PR middle-end/56461
26734 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
26735 method on dr_chain and result_chain.
26737 PR middle-end/56461
26738 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
26739 pointer_set_destroy on not_executed_last_iteration.
26741 PR middle-end/56461
26742 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
26744 PR middle-end/56461
26745 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
26746 FOR_EACH_DEFINED_FUNCTION when freeing state.
26748 PR middle-end/56461
26749 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
26751 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
26754 PR middle-end/56461
26755 * ipa-cp.c (decide_whether_version_node): Call vec_free on
26756 known_aggs[i].items and release known_aggs vector.
26758 PR middle-end/56461
26759 * ipa-reference.c (propagate): Free node_info even for alias nodes.
26761 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26763 * config/microblaze/microblaze.c (microblaze_emit_compare):
26764 Use xor for EQ/NE comparisions.
26765 * config/microblaze/microblaze.md (cstoresf4): Add constraints
26766 (cbranchsf4): Adjust operator to comparison_operator.
26768 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26770 PR middle-end/56461
26771 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
26773 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
26774 vec_safe_push, always update *slot.
26775 (redirect_edge_var_map_clear): Use vec_free.
26776 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
26777 (free_var_map_entry): Use vec_free.
26778 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
26779 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
26781 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
26783 PR middle-end/45472
26784 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
26785 when the may_trap_p bit of the exprs being merged differs.
26786 Reorder tests for speculativeness in the logical and operator.
26788 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26790 * incpath.c (add_standard_paths): Use reconcat instead of concat
26791 where appropriate and avoid leaking memory.
26793 * opts.h: Include obstack.h.
26794 (opts_concat): New prototype.
26795 (opts_obstack): New declaration.
26796 * opts.c (opts_concat): New function.
26797 (opts_obstack): New variable.
26798 (init_options_struct): Call gcc_init_obstack on opts_obstack.
26799 (finish_options): Use opts_concat instead of concat
26800 and XOBNEWVEC instead of XNEWVEC.
26801 * opts-common.c (generate_canonical_option, decode_cmdline_option,
26802 generate_option): Likewise.
26803 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
26804 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
26807 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
26808 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
26810 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26812 PR middle-end/56461
26813 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
26815 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26817 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
26818 (arm_block_move_unaligned_straight): Likewise.
26819 (arm_adjust_block_mem): Likewise.
26821 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26824 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
26825 temp, cond and label.
26826 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
26829 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
26830 * config/c6x/c6x.h (dbx_register_map): Update declaration.
26833 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
26834 of prologue/epilogue functions.
26837 * config/tilegx/tilegx.c (tilegx_expand_prologue):
26838 Remove unused variable cfa_offset.
26839 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
26842 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
26843 type promotion to unsigned.
26846 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
26847 for HOST_WIDE_INT of 32 bit / same size as int.
26848 (arm_block_move_unaligned_straight): Likewise.
26849 (arm_adjust_block_mem): Likewise.
26852 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
26855 2013-02-26 Marek Polacek <polacek@redhat.com>
26857 PR tree-optimization/56426
26858 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
26860 2013-02-26 Richard Biener <rguenther@suse.de>
26863 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
26864 unused variable loops.
26866 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26868 PR tree-optimization/56448
26869 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
26870 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
26871 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
26872 later operands of the references, or even first operand for
26873 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
26875 PR tree-optimization/56443
26876 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
26877 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
26878 to type_for_mode langhook.
26880 2013-02-25 Matt Turner <mattst88@gmail.com>
26882 * doc/invoke.texi: Document r4700.
26884 2013-02-25 Richard Biener <rguenther@suse.de>
26886 PR tree-optimization/56175
26887 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
26889 (simplify_bitwise_binary): ... here. Also guard the conversion
26890 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
26892 2013-02-25 Catherine Moore <clm@codesourcery.com>
26895 2013-02-24 Catherine Moore <clm@codesourcery.com>
26896 Maciej W. Rozycki <macro@codesourcery.com>
26897 Tom de Vries <tom@codesourcery.com>
26898 Nathan Sidwell <nathan@codesourcery.com>
26899 Iain Sandoe <iain@codesourcery.com>
26900 Nathan Froyd <froydnj@codesourcery.com>
26901 Chao-ying Fu <fu@mips.com>
26903 * doc/extend.texi: (micromips, nomicromips, nocompression):
26904 Document new function attributes.
26905 * doc/invoke.texi (minterlink-compressed, mmicromips,
26906 m14k, m14ke, m14kec): Document new options.
26907 (minterlink-mips16): Update documentation.
26908 * doc/md.texi (ZC, ZD): Document new constraints.
26909 * configure.ac (gcc_cv_as_micromips): Check if linker
26910 supports the .set micromips directive.
26911 * configure: Regenerate.
26912 * config.in: Regenerate.
26913 * config/mips/mips-tables.opt: Regenerate.
26914 * config/mips/micromips.md: New file.
26915 * constraints.md (ZC, AD): New constraints.
26916 * config/mips/predicates.md (movep_src_register): New predicate.
26917 (movep_src_operand): New predicate.
26918 (non_volatile_mem_operand): New predicate.
26919 * config/mips/mips.md (multimem): New type.
26920 (length): Differentiate between 17-bit and 18-bit branch offsets.
26921 (MOVEP1, MOVEP2): New mode iterator.
26922 (mov_<load>l): Use ZC constraint.
26923 (mov_<load>r): Likewise.
26924 (mov_<store>l): Likewise.
26925 (mov_<store>r): Likewise.
26926 (*branch_equality<mode>_inverted): Add microMIPS support.
26927 (*branch_equality<mode>): Likewise.
26928 (*jump_absolute): Likewise.
26929 (indirect_jump_<mode>): Likewise.
26930 (tablejump_<mode>): Likewise.
26931 (<optab>_internal): Likewise.
26932 (sibcall_internal): Likewise.
26933 (sibcall_value_internal): Likewise.
26934 (prefetch): Use constraint ZD.
26935 * config/mips/mips.opt (minterlink-compressed): New option.
26936 (minterlink-mips16): Now an alias for minterlink-compressed.
26937 (mmicromips): New option.
26938 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26939 (compare_and_swap_12): Likewise.
26940 (sync_add<mode>): Likewise.
26941 (sync_<optab>_12): Likewise.
26942 (sync_old_<optab>_12): Likewise.
26943 (sync_new_<optab>_12): Likewise.
26944 (sync_nand_12): Likewise.
26945 (sync_old_nand_12): Likewise.
26946 (sync_new_nand_12): Likewise.
26947 (sync_sub<mode>): Likewise.
26948 (sync_old_add<mode>): Likewise.
26949 (sync_old_sub<mode>): Likewise.
26950 (sync_new_add<mode>): Likewise.
26951 (sync_new_sub<mode>): Likewise.
26952 (sync_<optab><mode>): Likewise.
26953 (sync_old_<optab><mode>): Likewise.
26954 (sync_new_<optab><mode>): Likewise.
26955 (sync_nand<mode>): Likewise.
26956 (sync_old_nand<mode>): Likewise.
26957 (sync_new_nand<mode>): Likewise.
26958 (sync_lock_test_and_set<mode>): Likewise.
26959 (test_and_set_12): Likewise.
26960 (atomic_compare_and_swap<mode>): Likewise.
26961 (atomic_exchange<mode>_llsc): Likewise.
26962 (atomic_fetch_add<mode>_llsc): Likewise.
26963 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26964 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26965 (umips_save_restore_pattern_p): Likewise.
26966 (umips_load_store_pair_p): Likewise.
26967 (umips_output_load_store_pair): Likewise.
26968 (umips_movep_target_p): Likewise.
26969 (umips_12bit_offset_address_p): Likewise.
26970 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26971 (mips_base_mips16): Rename this...
26972 (mips_base_compression_flags): ...to this. Update all uses.
26973 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26974 (mips_mips16_decl_p): Delete.
26975 (mips_nomips16_decl_p): Delete.
26976 (mips_get_compress_on_flags): New function.
26977 (mips_get_compress_off_flags): New function.
26978 (mips_get_compress_mode): New function.
26979 (mips_get_compress_on_name): New function.
26980 (mips_get_compress_off_name): New function.
26981 (mips_insert_attributes): Support multiple compression types.
26982 (mips_merge_decl_attributes): Likewise.
26983 (umips_12bit_offset_address_p): New function.
26984 (mips_start_function_definition): Emit .set micromips directive.
26985 (mips_call_may_need_jalx_p): New function.
26986 (mips_function_ok_for_sibcall): Add microMIPS support.
26987 (mips_print_operand_punctuation): Support short delay slots and
26989 (umips_swm_mask, umips_swm_encoding): New.
26990 (umips_build_save_restore): New function.
26991 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26992 (was_mips16_p): Remove.
26993 (old_compression_mode): New.
26994 (mips_set_compression_mode): New function.
26995 (mips_set_current_function): Add microMIPS support.
26996 (mips_option_override): Likewise.
26997 (umips_save_restore_pattern_p): New function.
26998 (umips_output_save_restore): New function.
26999 (umips_load_store_pair_p_1): New function.
27000 (umips_load_store_pair_p): New function.
27001 (umips_output_load_store_pair_1): New function.
27002 (umips_output_load_store_pair): New function.
27003 (umips_movep_target_p) New function.
27004 (mips_prepare_pch_save): Add microMIPS support.
27005 * config/mips/mips.h (TARGET_COMPRESSION): New.
27006 (TARGET_CPU_CPP_BUILTINS): Update macro
27007 to use new compression flags and to support microMIPS.
27008 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27009 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27010 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27011 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27012 (ASM_SPEC): Support mmicromips and mno-micromips.
27013 (M16STORE_REG_P): New macro.
27014 (MIPS_CALL): Support TARGET_MICROMIPS.
27015 (MICROMIPS_J): New macro.
27016 (mips_base_mips16): Rename this...
27017 (mips_base_compression_flags): ...to this.
27018 (UMIPS_12BIT_OFFSET_P): New macro.
27019 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27020 (MULTILIB_DIRNAMES): Likewise.
27022 2013-02-25 Tom de Vries <tom@codesourcery.com>
27024 PR rtl-optimization/56131
27025 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
27026 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
27027 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
27029 2013-02-25 Tobias Burnus <burnus@net-b.de>
27031 * doc/invoke.texi (-fsanitize=): Move from optimization
27032 to debugging options.
27034 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27036 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
27038 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27039 Alexander Monakov <amonakov@ispras.ru>
27041 PR middle-end/56077
27042 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
27043 flush pending lists also on non-jumps. Adjust comment.
27045 2013-02-24 Catherine Moore <clm@codesourcery.com>
27046 Maciej W. Rozycki <macro@codesourcery.com>
27047 Tom de Vries <tom@codesourcery.com>
27048 Nathan Sidwell <nathan@codesourcery.com>
27049 Iain Sandoe <iain@codesourcery.com>
27050 Nathan Froyd <froydnj@codesourcery.com>
27051 Chao-ying Fu <fu@mips.com>
27053 * doc/extend.texi: (micromips, nomicromips, nocompression):
27054 Document new function attributes.
27055 * doc/invoke.texi (minterlink-compressed, mmicromips,
27056 m14k, m14ke, m14kec): Document new options.
27057 (minterlink-mips16): Update documentation.
27058 * doc/md.texi (ZC, ZD): Document new constraints.
27059 * configure.ac (gcc_cv_as_micromips): Check if linker
27060 supports the .set micromips directive.
27061 * configure: Regenerate.
27062 * config.in: Regenerate.
27063 * config/mips/mips-tables.opt: Regenerate.
27064 * config/mips/micromips.md: New file.
27065 * constraints.md (ZC, AD): New constraints.
27066 * config/mips/predicates.md (movep_src_register): New predicate.
27067 (movep_src_operand): New predicate.
27068 (non_volatile_mem_operand): New predicate.
27069 * config/mips/mips.md (multimem): New type.
27070 (length): Differentiate between 17-bit and 18-bit branch offsets.
27071 (MOVEP1, MOVEP2): New mode iterator.
27072 (mov_<load>l): Use ZC constraint.
27073 (mov_<load>r): Likewise.
27074 (mov_<store>l): Likewise.
27075 (mov_<store>r): Likewise.
27076 (*branch_equality<mode>_inverted): Add microMIPS support.
27077 (*branch_equality<mode>): Likewise.
27078 (*jump_absolute): Likewise.
27079 (indirect_jump_<mode>): Likewise.
27080 (tablejump_<mode>): Likewise.
27081 (<optab>_internal): Likewise.
27082 (sibcall_internal): Likewise.
27083 (sibcall_value_internal): Likewise.
27084 (prefetch): Use constraint ZD.
27085 * config/mips/mips.opt (minterlink-compressed): New option.
27086 (minterlink-mips16): Now an alias for minterlink-compressed.
27087 (mmicromips): New option.
27088 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27089 (compare_and_swap_12): Likewise.
27090 (sync_add<mode>): Likewise.
27091 (sync_<optab>_12): Likewise.
27092 (sync_old_<optab>_12): Likewise.
27093 (sync_new_<optab>_12): Likewise.
27094 (sync_nand_12): Likewise.
27095 (sync_old_nand_12): Likewise.
27096 (sync_new_nand_12): Likewise.
27097 (sync_sub<mode>): Likewise.
27098 (sync_old_add<mode>): Likewise.
27099 (sync_old_sub<mode>): Likewise.
27100 (sync_new_add<mode>): Likewise.
27101 (sync_new_sub<mode>): Likewise.
27102 (sync_<optab><mode>): Likewise.
27103 (sync_old_<optab><mode>): Likewise.
27104 (sync_new_<optab><mode>): Likewise.
27105 (sync_nand<mode>): Likewise.
27106 (sync_old_nand<mode>): Likewise.
27107 (sync_new_nand<mode>): Likewise.
27108 (sync_lock_test_and_set<mode>): Likewise.
27109 (test_and_set_12): Likewise.
27110 (atomic_compare_and_swap<mode>): Likewise.
27111 (atomic_exchange<mode>_llsc): Likewise.
27112 (atomic_fetch_add<mode>_llsc): Likewise.
27113 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27114 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27115 (umips_save_restore_pattern_p): Likewise.
27116 (umips_load_store_pair_p): Likewise.
27117 (umips_output_load_store_pair): Likewise.
27118 (umips_movep_target_p): Likewise.
27119 (umips_12bit_offset_address_p): Likewise.
27120 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27121 (mips_base_mips16): Rename this...
27122 (mips_base_compression_flags): ...to this. Update all uses.
27123 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27124 (mips_mips16_decl_p): Delete.
27125 (mips_nomips16_decl_p): Delete.
27126 (mips_get_compress_on_flags): New function.
27127 (mips_get_compress_off_flags): New function.
27128 (mips_get_compress_mode): New function.
27129 (mips_get_compress_on_name): New function.
27130 (mips_get_compress_off_name): New function.
27131 (mips_insert_attributes): Support multiple compression types.
27132 (mips_merge_decl_attributes): Likewise.
27133 (umips_12bit_offset_address_p): New function.
27134 (mips_start_function_definition): Emit .set micromips directive.
27135 (mips_call_may_need_jalx_p): New function.
27136 (mips_function_ok_for_sibcall): Add microMIPS support.
27137 (mips_print_operand_punctuation): Support short delay slots and
27139 (umips_swm_mask, umips_swm_encoding): New.
27140 (umips_build_save_restore): New function.
27141 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27142 (was_mips16_p): Remove.
27143 (old_compression_mode): New.
27144 (mips_set_compression_mode): New function.
27145 (mips_set_current_function): Add microMIPS support.
27146 (mips_option_override): Likewise.
27147 (umips_save_restore_pattern_p): New function.
27148 (umips_output_save_restore): New function.
27149 (umips_load_store_pair_p_1): New function.
27150 (umips_load_store_pair_p): New function.
27151 (umips_output_load_store_pair_1): New function.
27152 (umips_output_load_store_pair): New function.
27153 (umips_movep_target_p) New function.
27154 (mips_prepare_pch_save): Add microMIPS support.
27155 * config/mips/mips.h (TARGET_COMPRESSION): New.
27156 (TARGET_CPU_CPP_BUILTINS): Update macro
27157 to use new compression flags and to support microMIPS.
27158 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27159 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27160 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27161 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27162 (ASM_SPEC): Support mmicromips and mno-micromips.
27163 (M16STORE_REG_P): New macro.
27164 (MIPS_CALL): Support TARGET_MICROMIPS.
27165 (MICROMIPS_J): New macro.
27166 (mips_base_mips16): Rename this...
27167 (mips_base_compression_flags): ...to this.
27168 (UMIPS_12BIT_OFFSET_P): New macro.
27169 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27170 (MULTILIB_DIRNAMES): Likewise.
27172 2013-02-24 Jakub Jelinek <jakub@redhat.com>
27175 * target-globals.c (save_target_globals): For init_reg_sets and
27176 target_reinit remporarily set this_fn_optabs to this_target_optabs.
27178 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
27180 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
27181 * config/aarch64/t-aarch64
27182 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
27184 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
27186 PR inline-asm/56148
27187 * lra-constraints.c (process_alt_operands): Reload operand
27188 conflicting with earlier clobber only if no more other conflicting
27191 2013-02-22 Jakub Jelinek <jakub@redhat.com>
27194 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
27195 if not linking a shared library.
27197 2013-02-22 Seth LaForge <sethml@google.com>
27199 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
27201 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
27203 * config/arm/arm.md (split for extendsidi): Update condition.
27204 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
27205 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
27206 (qhs_zextenddi_cstr): Likewise.
27208 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27210 PR middle-end/56420
27211 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
27212 avoid signed wrapping.
27213 (expand_mult): Handle properly multiplication by
27214 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
27215 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
27216 in the compiler if coeff is HOST_WIDE_INT_MIN.
27217 (expand_divmod): Don't make ext_op1 static, change it's type to
27218 uhwi. Avoid undefined behavior in -INTVAL (op1).
27220 PR rtl-optimization/50339
27221 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
27223 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
27224 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
27225 into splitting_ashiftrt field.
27226 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
27228 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
27231 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
27233 PR middle-end/56108
27234 * trans-mem.c (execute_tm_mark): Do not expand transactions that
27235 are sure to go irrevocable.
27237 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
27239 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
27240 scalars are valid operands.
27242 2013-02-21 Martin Jambor <mjambor@suse.cz>
27244 PR tree-optimization/56310
27245 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
27246 only matching indices and non-negative final offsets.
27247 (intersect_aggregates_with_edge): Pass src_idx to
27248 agg_replacements_to_vector. Pass src_idx insstead of index to
27249 intersect_with_agg_replacements.
27251 2013-02-21 Martin Jambor <mjambor@suse.cz>
27253 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
27254 instead of hard-wired defaults.
27256 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
27258 * doc/invoke.texi (MIPS Options): Update documentation of the
27259 floating-point multiply-accumulate instruction restrictions.
27261 2013-02-21 Kostya Serebryany <kcc@google.com>
27263 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
27264 asan_shadow_offset on x86_64 linux.
27266 2013-02-21 Richard Biener <rguenther@suse.de>
27268 PR tree-optimization/56415
27270 2013-02-11 Richard Biener <rguenther@suse.de>
27272 PR tree-optimization/56273
27273 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27276 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27279 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
27282 PR inline-asm/56405
27283 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
27284 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
27286 2013-02-20 Jan Hubicka <jh@suse.cz>
27288 PR tree-optimization/56265
27289 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
27290 when target is referenced for first time.
27292 2013-02-20 Richard Biener <rguenther@suse.de>
27294 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
27295 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
27296 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
27297 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
27298 not return anything.
27299 (rename_ssa_copies): Do not remove unused locals.
27300 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
27301 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
27302 * passes.c (execute_function_todo): Do not schedule unused locals
27303 removal if cleanup_tree_cfg did something.
27304 * tree-ssa-live.c (remove_unused_locals): Dump statistics
27305 about the number of removed locals.
27307 2013-02-20 Richard Biener <rguenther@suse.de>
27309 PR tree-optimization/56398
27310 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
27312 2013-02-20 Martin Jambor <mjambor@suse.cz>
27314 PR tree-optimization/55334
27315 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
27316 restricted pointers to arrays.
27318 2013-02-20 Richard Biener <rguenther@suse.de>
27319 Jakub Jelinek <jakub@redhat.com>
27321 PR tree-optimization/56396
27322 * tree-ssa-ccp.c (n_const_val): New static variable.
27323 (get_value): Return NULL for SSA names we don't have a lattice
27325 (ccp_initialize): Initialize n_const_val.
27326 * tree-ssa-copy.c (n_copy_of): New static variable.
27327 (init_copy_prop): Initialize n_copy_of.
27328 (get_value): Return NULL_TREE for SSA names we don't have a
27331 2013-02-20 Martin Jambor <mjambor@suse.cz>
27333 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
27335 2013-02-20 Richard Biener <rguenther@suse.de>
27337 * genpreds.c (write_lookup_constraint): Do not compare first
27338 letter of the constraint again.
27340 2013-02-20 Richard Biener <rguenther@suse.de>
27342 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
27344 (get_use_iv_cost): Terminate hashtable walk when coming across
27347 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
27349 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
27350 reassociation for avx2 targets.
27352 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27354 * config/microblaze/microblaze.c: microblaze_has_clz = 0
27355 Add version check for v8.10.a to enable microblaze_has_clz
27356 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
27357 version and TARGET_PATTERN_COMPARE check
27358 * config/microblaze/microblaze.md: New clzsi2 instruction
27360 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27362 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
27363 function before branching.
27365 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27367 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
27369 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
27371 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27373 PR middle-end/55889
27374 * sel-sched.c: Include ira.h.
27375 (implicit_clobber_conflict_p): New function.
27376 (moveup_expr): Use it.
27377 * Makefile.in (sel-sched.o): Depend on ira.h.
27379 2013-02-19 Richard Biener <rguenther@suse.de>
27381 PR tree-optimization/56384
27382 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
27383 (vn_hash_type): Split out from ...
27384 (vn_hash_constant_with_type): ... here.
27385 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
27386 (vn_phi_eq): Compare types from vn_phi_s structure.
27387 (vn_phi_lookup): Populate vn_phi_s type.
27388 (vn_phi_insert): Likewise.
27390 2013-02-19 Jakub Jelinek <jakub@redhat.com>
27392 PR tree-optimization/56350
27393 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
27394 if haven't found reduction or nested cycle operand, rather than
27395 asserting we must find it.
27397 PR tree-optimization/56381
27398 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
27401 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
27402 Jakub Jelinek <jakub@redhat.com>
27405 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
27406 (swap_optab_enable): Same.
27407 (init_all_optabs): Use argument instead of global.
27408 * tree.h (struct tree_optimization_option): New field target_optabs.
27409 * expr.h (init_all_optabs): Add argument to prototype.
27410 (TREE_OPTIMIZATION_OPTABS): New.
27411 (save_optabs_if_changed): Protoize.
27412 * optabs.h: Declare this_fn_optabs.
27413 * optabs.c (save_optabs_if_changed): New.
27414 Declare this_fn_optabs.
27415 (init_optabs): Add argument to init_all_optabs() call.
27416 * function.c (invoke_set_current_function_hook): Handle per
27418 * function.h (struct function): New field optabs.
27419 * config/mips/mips.c (mips_set_mips16_mode): Handle when
27420 optimization_current_node has changed.
27421 * target-globals.h (save_target_globals_default_opts): Protoize.
27422 * target-globals.c (save_target_globals_default_opts): New.
27424 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27427 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
27428 registers %fr12 and %fr12R as call used.
27431 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
27432 and HImode, require all displacements to be an integer multiple of
27434 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
27435 only allow QImode and HImode when reload is in progress and strict is
27436 true. Likewise for symbolic addresses. Use base14_operand to check
27437 displacements in REG+BASE addresses.
27439 2013-02-18 Richard Biener <rguenther@suse.de>
27441 PR tree-optimization/56366
27442 * tree-vect-loop.c (get_initial_def_for_induction): Properly
27443 handle sign-conversion of outer-loop initial induction value.
27445 2013-02-18 Richard Biener <rguenther@suse.de>
27447 PR middle-end/56349
27448 * cfghooks.c (merge_blocks): If we merge a latch into another
27449 block adjust references to it.
27450 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
27451 (verify_loop_structure): Verify that a recorded latch is in fact
27454 2013-02-18 Richard Biener <rguenther@suse.de>
27456 PR tree-optimization/56321
27457 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
27458 order SSA name release and virtual operand unlinking.
27460 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27462 * config/microblaze/microblaze.md (save_stack_block): Define.
27463 (restore_stack_block): Likewise.
27465 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27467 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
27468 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
27469 * config/microblaze/microblaze.c (microblaze_option_override):
27470 Bail out early for PIC modes when target does not support PIC.
27472 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27474 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
27475 Replace with a microblaze version.
27476 (microblaze_trampoline_init): Adapt for microblaze.
27477 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
27480 2013-02-16 Jakub Jelinek <jakub@redhat.com>
27481 Dodji Seketeli <dodji@redhat.com>
27484 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
27485 (instrument_mem_region_access): Do not forget to always put
27486 instrumentation of the of 'base' and 'base + len' in a "if (len !=
27487 0) statement, even for cases where either 'base' or 'base + len'
27488 are not instrumented -- because they have been previously
27489 instrumented. Simplify the logic by putting all the statements
27490 instrument 'base + len' inside a sequence, and then insert that
27491 sequence right before the current insertion point. Then, to
27492 instrument 'base + len', just get an iterator on that statement.
27493 And do not forget to update the pointer to iterator the function
27494 received as argument.
27496 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
27498 PR rtl-optimization/56348
27499 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
27501 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
27503 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
27504 (clean_graph_dump_file): Pass base to start_graph_dump.
27506 2013-02-14 Richard Henderson <rth@redhat.com>
27509 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
27511 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
27513 * collect2-aix.h: Define F_LOADONLY.
27515 2013-02-14 Richard Biener <rguenther@suse.de>
27518 * varasm.c (output_constant_def_1): Get the decl representing
27519 the constant as argument.
27520 (output_constant_def): Wrap output_constant_def_1.
27521 (make_decl_rtl): Use output_constant_def_1 with the decl
27522 representing the constant.
27523 (build_constant_desc): Optionally re-use a decl already
27524 representing the constant.
27525 (tree_output_constant_def): Adjust.
27527 2013-02-14 Dodji Seketeli <dodji@redhat.com>
27530 * asan.c (instrument_builtin_call): Really put the length of the
27531 second source argument into src1_len.
27533 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27535 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
27536 argument. If it is false, don't create edge from then_bb to
27538 (insert_if_then_before_iter): Pass true to it.
27539 (build_check_stmt): Pass false to it.
27540 (transform_statements): Flush hash table only on extended basic
27541 block boundaries, rather than at the beginning of every bb.
27542 Don't flush hash table on nonfreeing_call_p calls.
27543 * tree-flow.h (nonfreeing_call_p): New prototype.
27544 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
27546 2013-02-13 David S. Miller <davem@davemloft.net>
27548 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
27550 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
27553 * ira.c (max_regno_before_ira): Move from ...
27555 (fix_reg_equiv_init): Use max_regno_before_ira instead of
27558 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27560 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
27562 2013-02-13 Richard Biener <rguenther@suse.de>
27565 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
27566 globals in MEM_REFs.
27568 2013-02-13 Richard Biener <rguenther@suse.de>
27570 * loop-init.c (loop_optimizer_init): Clear loop state when
27571 re-initializing preserved loops.
27572 * loop-unswitch.c (unswitch_single_loop): Return whether
27573 we unswitched the loop. Do not verify loop state here.
27574 (unswitch_loops): When we unswitched a loop discover new loops.
27576 2013-02-13 Kostya Serebryany <kcc@google.com>
27578 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
27580 * sanitizer.def: Rename __asan_init to __asan_init_v1.
27582 2013-02-12 Dodji Seketeli <dodji@redhat.com>
27584 Avoid instrumenting duplicated memory access in the same basic block
27585 * Makefile.in (asan.o): Add new dependency on hash-table.h
27586 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
27587 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
27588 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
27589 (free_mem_ref_resources, has_mem_ref_been_instrumented)
27590 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
27591 (get_mem_ref_of_assignment): New functions.
27592 (get_mem_refs_of_builtin_call): Extract from
27593 instrument_builtin_call and tweak a little bit to make it fit with
27595 (instrument_builtin_call): Use the new
27596 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
27597 of is_gimple_builtin_call.
27598 (instrument_derefs, instrument_mem_region_access): Insert the
27599 instrumented memory reference into the hash table.
27600 (maybe_instrument_assignment): Renamed instrument_assignment into
27601 this, and change it to advance the iterator when instrumentation
27602 actually happened and return true in that case. This makes it
27603 homogeneous with maybe_instrument_assignment, and thus give a
27604 chance to callers to be more 'regular'.
27605 (transform_statements): Clear the memory reference hash table
27606 whenever we enter a new BB, when we cross a function call, or when
27607 we are done transforming statements. Use
27608 maybe_instrument_assignment instead of instrumentation. No more
27609 need to special case maybe_instrument_assignment and advance the
27610 iterator after calling it; it's now handled just like
27611 maybe_instrument_call. Update comment.
27613 2013-02-13 Richard Biener <rguenther@suse.de>
27615 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
27616 Fix loop discovery code.
27618 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
27620 PR inline-asm/56148
27621 * lra-constraints.c (process_alt_operands): Match early clobber
27622 operand with itself. Check conflicts with earlyclobber only if
27623 the operand is not reloaded. Prefer to reload conflicting operand
27624 if earlyclobber and matching operands are the same.
27626 2013-02-12 Richard Biener <rguenther@suse.de>
27629 * lto-streamer-out.c (write_symbol): Do not output symbols
27630 for hard register variables.
27632 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
27635 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
27636 (umulsidi3_insn, mulsidi3_insn): New insns.
27638 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
27640 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
27641 (struct tune_params): Add vec_costs field.
27642 * config/arm/arm.c (arm_builtin_vectorization_cost)
27643 (arm_add_stmt_cost): New functions.
27644 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
27645 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
27646 (arm_default_vec_cost): New struct of type cpu_vec_costs.
27647 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
27648 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
27649 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
27650 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
27652 2013-02-12 Richard Biener <rguenther@suse.de>
27655 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
27656 decls again if possible.
27658 2013-02-12 Richard Biener <rguenther@suse.de>
27660 PR middle-end/56288
27661 * tree-ssa.c (verify_ssa_name): Fix check, move
27662 SSA_NAME_IN_FREE_LIST check up.
27664 2013-02-12 Jakub Jelinek <jakub@redhat.com>
27665 Steven Bosscher <steven@gcc.gnu.org>
27667 PR rtl-optimization/56151
27668 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
27669 equal to op0 or op1, and last_insn pattern is CODE operation
27670 with MEM dest and one of the operands matches that MEM.
27672 2013-02-11 Sriraman Tallam <tmsriram@google.com>
27674 * doc/extend.texi: Document Function Multiversioning and "default"
27675 parameter string to target attribute.
27676 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
27677 target attribute parameter is "default".
27678 (ix86_compare_version_priority): Remove checks for target attribute.
27679 (ix86_mangle_function_version_assembler_name): Change error to sorry.
27680 Remove check for target attribute equal to NULL. Add assert.
27681 (ix86_generate_version_dispatcher_body): Change error to sorry.
27683 2013-02-11 Iain Sandoe <iain@codesourcery.com>
27684 Jack Howarth <howarth@bromo.med.uc.edu>
27685 Patrick Marlier <patrick.marlier@gmail.com>
27688 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
27689 define ENDFILE_SPEC as TM_DESTRUCTOR.
27690 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
27692 2013-02-11 Alexander Potapenko <glider@google.com>
27693 Jack Howarth <howarth@bromo.med.uc.edu>
27694 Jakub Jelinek <jakub@redhat.com>
27697 * config/darwin.c (cdtor_record): Rename ctor_record.
27698 (sort_cdtor_records): Rename sort_ctor_records.
27699 (finalize_dtors): New routine to sort destructors by
27700 priority before use in assemble_integer.
27701 (machopic_asm_out_destructor): Use finalize_dtors if needed.
27703 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
27705 PR rtl-optimization/56275
27706 * simplify-rtx.c (avoid_constant_pool_reference): Check that
27707 offset is non-negative and less than cmode size before
27708 calling simplify_subreg.
27710 2013-02-11 Richard Biener <rguenther@suse.de>
27712 PR tree-optimization/56264
27713 * cfgloop.h (fix_loop_structure): Adjust prototype.
27714 * loop-init.c (fix_loop_structure): Return the number of
27715 newly discovered loops.
27716 * tree-cfgcleanup.c (repair_loop_structures): When new loops
27717 are discovered, do a full loop-closed SSA rewrite.
27719 2013-02-11 Richard Biener <rguenther@suse.de>
27721 PR tree-optimization/56273
27722 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27724 (check_array_ref): Fix missing newline in dumps.
27725 (search_for_addr_array): Likewise.
27727 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
27729 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
27731 2013-02-09 Jakub Jelinek <jakub@redhat.com>
27734 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
27736 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
27738 PR rtl-optimization/56246
27739 * lra-constraints.c (simplify_operand_subreg): Try to reuse
27741 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
27742 constraints are satisfied.
27744 2013-02-08 Jeff Law <law@redhat.com>
27747 * emit-rtl.c (reg_is_parm_p): New function.
27748 * regs.h (reg_is_parm_p): New prototype.
27749 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
27750 callee-clobbered registers.
27752 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
27755 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
27756 If there is no implicit builtin declaration, just return NULL.
27758 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27760 * config/i386/sse.md (FMAMODEM): New mode iterator.
27761 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
27762 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
27764 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27766 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
27767 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
27768 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
27770 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27772 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
27773 (microblaze*-*-elf): Likewise.
27774 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
27776 * config/microblaze/microblaze-c.c: Add builtin defines for
27777 _LITTLE_ENDIAN and _BIG_ENDIAN.
27778 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
27779 add to TARGET_DEFAULT flags.
27780 Expand ASM_SPEC and LINK_SPEC.
27781 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
27782 * config/microblaze/microblaze.md: Update extendsidi2 and
27783 movdi_internal instructions to use low-order / high-order reg
27785 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
27786 options and inversemask / mask of LITTLE_ENDIAN.
27787 * config/microblaze/t-microblaze: Expand multilib options to
27788 include mlittle-endian (le) and update exceptions patterns.
27790 2013-02-08 Jakub Jelinek <jakub@redhat.com>
27792 PR rtl-optimization/56195
27793 * lra-constraints.c (get_reload_reg): Don't reuse regs
27794 if they have smaller mode than requested, if they have
27795 wider mode than requested, try to return a SUBREG.
27797 PR tree-optimization/56250
27798 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
27799 if type is unsigned and code isn't MULT_EXPR.
27801 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27803 PR tree-optimization/56064
27804 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
27805 bits according to mode.
27806 * fixed-value.h (fixed_from_double_int)
27807 (const_fixed_from_double_int): Adjust comments.
27809 2013-02-08 Richard Biener <rguenther@suse.de>
27812 * lto-streamer.h (struct data_in): Remove current_file, current_line
27813 and current_col members.
27814 * lto-streamer-out.c (lto_output_location): Stream changed bits
27815 en-block for efficiency.
27816 * lto-streamer-in.c (clear_line_info): Remove.
27817 (lto_input_location): Cache current file, line and column
27818 globally via local statics. Read changed bits en-block.
27819 (input_function): Do not call clear_line_info.
27820 (lto_read_body): Likewise.
27821 (lto_input_toplevel_asms): Likewise.
27823 2013-02-08 Michael Matz <matz@suse.de>
27825 PR tree-optimization/52448
27826 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
27827 (nt_call_phase): New static.
27828 (add_or_mark_expr): Only mark accesses with newer phase than any
27830 (nonfreeing_call_p): New.
27831 (nt_init_block): Update nt_call_phase, mark blocks as visited.
27832 (nt_fini_block): Keep blocks marked as visited.
27833 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
27835 2013-02-08 Richard Biener <rguenther@suse.de>
27837 * ira.c (ira): Free broken dominator information.
27839 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27841 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
27843 2013-02-08 Marek Polacek <polacek@redhat.com>
27845 * cfgloop.c (verify_loop_structure): Add more checking of headers.
27847 2013-02-08 Richard Biener <rguenther@suse.de>
27849 PR middle-end/56181
27850 * cfgloop.h (flow_loops_find): Adjust.
27851 (bb_loop_header_p): Declare.
27852 * cfgloop.c (bb_loop_header_p): New function split out from ...
27853 (flow_loops_find): ... here. Adjust function signature,
27854 support incremental loop structure update.
27855 (verify_loop_structure): Cleanup. Verify a loop is a loop.
27856 * cfgloopmanip.c (fix_loop_structure): Move ...
27857 * loop-init.c (fix_loop_structure): ... here.
27858 (apply_loop_flags): Split out from ...
27859 (loop_optimizer_init): ... here.
27860 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
27861 in incremental mode, only remove dead loops here.
27863 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27866 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
27867 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
27868 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
27869 (*round<mode>3.libgcc): New insns for fixed-modes.
27870 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
27871 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
27872 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
27873 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
27874 implementations. Define to __builtin_avr_absFX,
27875 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
27876 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
27877 __builtin_avr_countlsFX, respectively.
27878 * config/avr/avr-c.c (target.h): Include it.
27879 (enum avr_builtin_id): New enum.
27880 (avr_resolve_overloaded_builtin): New static function.
27881 (avr_register_target_pragmas): Use it to set
27882 targetm.resolve_overloaded_builtin.
27883 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
27884 tree nodes used by DEF_BUILTIN.
27885 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
27886 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
27887 <AVR_BUILTIN_xxBITS>: Same.
27889 2013-02-08 Richard Biener <rguenther@suse.de>
27891 * cfgloop.c (verify_loop_structure): Properly handle
27892 a loop exiting to another loop header.
27893 * ira-int.h (ira_loops): Remove.
27894 * ira.c (ira_loops): Remove.
27895 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
27896 (do_reload): Use loop_optimizer_finalize.
27897 * ira-build.c (create_loop_tree_nodes): Use get_loops and
27898 number_of_loops to access the loop tree.
27899 (more_one_region_p): Likewise.
27900 (finish_loop_tree_nodes): Likewise.
27901 (rebuild_regno_allocno_maps): Likewise.
27902 (mark_loops_for_removal): Likewise.
27903 (mark_all_loops_for_removal): Likewise.
27904 (remove_unnecessary_regions): Likewise.
27905 (ira_build): Likewise.
27906 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
27908 2013-02-08 Richard Biener <rguenther@suse.de>
27910 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
27911 * ipa-pure-const.c (analyze_function): Avoid calling
27912 mark_irreducible_loops twice.
27913 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
27915 2013-02-07 David S. Miller <davem@davemloft.net>
27917 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
27919 * var-tracking.c (vt_add_function_parameter): Test the presence of
27920 HAVE_window_save properly and do not remap argument registers when
27921 we have a leaf function.
27923 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
27926 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
27928 * config/i386/i386.c (ix86_print_operand): Ditto.
27930 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27932 * lra-constraints.c (process_alt_operands): Fix recently added comment.
27934 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27936 PR rtl-optimization/56225
27937 * lra-constraints.c (process_alt_operands): Check that reload hard
27938 reg can hold value for strict_low_part.
27940 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27943 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
27944 dwarf2out_end_function.
27945 (in_first_function_p, maybe_at_text_label_p,
27946 first_loclabel_num_not_at_text_label): New variables.
27947 (dwarf2out_var_location): In the first function find out
27948 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
27949 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
27952 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
27954 PR rtl-optimization/56178
27955 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
27956 SUBREG of a register. Tidy up related block of code.
27957 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
27958 note if the source is a register or a SUBREG of a register.
27960 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27963 * config/rs6000/rs6000.md (ptrm): New mode attr.
27964 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
27965 call_value_indirect_aix<pttrsize>,
27966 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
27969 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
27971 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
27972 if -bnortl. Convert to strcmp and strncmp.
27974 2013-02-07 Alan Modra <amodra@gmail.com>
27977 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
27978 addresses won't wrap when offsetting.
27979 (rs6000_secondary_reload): Provide secondary reloads needed for
27980 wrapping LO_SUM addresses.
27982 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
27984 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
27985 MACH, just __MACH__.
27987 2013-02-06 Richard Biener <rguenther@suse.de>
27989 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
27990 instead of calling fix_loop_structure.
27992 2013-02-06 Jakub Jelinek <jakub@redhat.com>
27994 PR middle-end/56217
27995 * omp-low.c (use_pointer_for_field): Return false if
27996 lower_send_shared_vars doesn't generate any copy-out code.
27998 2013-02-06 Tom de Vries <tom@codesourcery.com>
28000 PR rtl-optimization/56131
28001 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
28002 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
28003 of the label is NULL. Add comment.
28005 2013-02-05 Jakub Jelinek <jakub@redhat.com>
28007 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
28010 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
28011 (STATIC_LIBTSAN_LIBS): Likewise.
28012 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
28013 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
28014 is defined, don't add anything else beyond that.
28015 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
28016 (LINK_COMMAND_SPEC): Use them.
28018 PR tree-optimization/56205
28019 * tree-stdarg.c (check_all_va_list_escapes): Return true if
28020 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
28021 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
28023 2013-02-05 Richard Biener <rguenther@suse.de>
28025 PR tree-optimization/53342
28026 PR tree-optimization/53185
28027 * tree-vectorizer.h (vect_check_strided_load): Remove.
28028 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
28029 not disallow peeling for vectorized strided loads.
28030 (vect_check_strided_load): Make static and simplify.
28031 (vect_analyze_data_refs): Adjust.
28032 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
28033 correctly when vectorizing strided loads.
28035 2013-02-05 Richard Biener <rguenther@suse.de>
28037 * doc/install.texi: Refer to ISL, not PPL.
28039 2013-02-05 Jan Hubicka <jh@suse.cz>
28041 PR tree-optimization/55789
28042 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
28044 2013-02-05 Jan Hubicka <jh@suse.cz>
28046 PR tree-optimization/55789
28047 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
28048 the dead call anyway.
28050 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
28053 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
28055 2013-02-04 Alexander Potapenko <glider@google.com>
28056 Jack Howarth <howarth@bromo.med.uc.edu>
28057 Jakub Jelinek <jakub@redhat.com>
28060 * config/darwin.c (sort_ctor_records): Stabilized qsort
28061 on constructor priority by using original position.
28062 (finalize_ctors): New routine to sort constructors by
28063 priority before use in assemble_integer.
28064 (machopic_asm_out_constructor): Use finalize_ctors if needed.
28066 2013-02-04 Jakub Jelinek <jakub@redhat.com>
28069 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
28070 about visibility on artificial decls.
28071 * config/sol2.c (solaris_assemble_visibility): Likewise.
28073 2013-02-04 Kai Tietz <ktietz@redhat.com>
28076 * config/i386/i386.c (function_value_ms_64): Add additional valtype
28077 argument and improve checking of return-argument types for 16-byte
28079 (ix86_function_value_1): Add additional valtype argument on call
28080 of function_value_64.
28081 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
28082 handling infunction_value_64 function.
28084 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
28086 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
28088 2013-02-04 Richard Biener <rguenther@suse.de>
28090 PR tree-optimization/56188
28091 * tree-ssa-structalias.c (label_visit): Consider case with
28092 initially non-empty points-to set.
28093 (perform_var_substitution): Dump node mapping and clean up.
28095 2013-02-04 Richard Guenther <rguenther@suse.de>
28098 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
28099 node prevail as last resort.
28100 (lto_symtab_merge_decls): Remove guard on LTRANS here.
28101 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
28103 2013-02-04 Richard Biener <rguenther@suse.de>
28105 PR tree-optimization/56113
28106 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
28108 (equiv_class_lookup_or_add): ... this.
28109 (label_visit): Adjust and fix error in previous patch.
28110 (perform_var_substitution): Adjust.
28112 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
28114 * config/sh/divtab.c: Fix formatting and comments throughout the file.
28115 * config/sh/sh4-300.md: Likewise.
28116 * config/sh/sh4a.md: Likewise.
28117 * config/sh/constraints.md: Likewise.
28118 * config/sh/sh.md: Likewise.
28119 * config/sh/netbsd-elf.h: Likewise.
28120 * config/sh/predicates.md: Likewise.
28121 * config/sh/sh-protos.h: Likewise.
28122 * config/sh/ushmedia.h: Likewise.
28123 * config/sh/linux.h: Likewise.
28124 * config/sh/sh.c: Likewise.
28125 * config/sh/superh.h: Likewise.
28126 * config/sh/elf.h: Likewise.
28127 * config/sh/sh4.md: Likewise.
28128 * config/sh/sh.h: Likewise.
28130 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28132 * config/pa/constraints.md: Adjust unused letters. Change "T"
28133 constraint to match_test floating_point_store_memory_operand().
28134 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
28135 (base14_operand): New.
28136 (floating_point_store_memory_operand): New.
28137 (integer_store_memory_operand): Revise to use base14_operand and
28138 reg_plus_base_memory_operand.
28139 (move_dest_operand): Allow symbolic_memory_operands.
28140 (symbolic_memory_operand): Check for LO_SOM.
28141 (symbolic_operand): Change default case to break.
28142 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
28143 CONST_DOUBLE values to be reloaded by putting them into memory when
28144 the destination is a floating point register.
28145 (movdf): Remove code to handle CONST_DOUBLE.
28147 (reload_indf_r1): New.
28148 (reload_insf_r1): New.
28149 Consistently use "Q" and "T" constraints with integer and floating
28150 point move instructions, respectively.
28151 (movdi): Remove FAIL.
28152 Change predicate for source operand unamed DImode move from
28153 general_operand to move_src_operand.
28154 (umulsidi3): Change predicate for destination operand to
28156 Likewise for similar unamed patterns.
28157 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
28158 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
28159 (hppa_legitimize_address): Simplify mask calculation.
28160 (pa_emit_move_sequence): Revised handling of secondary reloads from
28161 REG+D addresses for floating point loads and stores. Directly handle
28162 loading CONST0_RTX (mode) to a floating point register.
28163 (pa_secondary_reload): Handle reloading DF and SFmode constant values
28164 to floating point registers. Don't restrict secondary reloads to
28165 floating point registers to integer modes. Revise some comments and
28167 (TARGET_LEGITIMATE_ADDRESS_P): Define.
28168 (pa_legitimate_address_p): New.
28169 (pa_legitimize_reload_address): New.
28170 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
28171 (STRICT_REG_OK_FOR_BASE_P): New.
28172 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
28173 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
28175 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
28176 Andrew Dixie <andrewd@gentrack.com>
28178 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
28181 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
28183 * expmed.c (extract_bit_field_1): Pass the full width of the
28184 structure to get_best_reg_extraction_insn.
28186 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28189 * configure.ac (use_cxa_atexit): Add AIX.
28190 * configure: Regenerate.
28192 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
28194 2013-02-01 Jakub Jelinek <jakub@redhat.com>
28197 * final.c (need_profile_function): New variable.
28198 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
28199 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
28200 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
28201 notes, targetm.asm_out.function_prologue doesn't emit anything,
28202 HAVE_prologue and profiler should be emitted before prologue,
28203 set need_profile_function instead of emitting it.
28204 (final_scan_insn): If need_profile_function, emit
28205 profile_function on the first NOTE_INSN_BASIC_BLOCK or
28206 NOTE_INSN_FUNCTION_BEG note.
28208 2013-02-01 Richard Henderson <rth@redhat.com>
28210 * config/rs6000/rs6000.md (smulditi3): New.
28213 * config/alpha/alpha.md (umulditi3): New.
28215 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28217 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
28218 (ASM_OUTPUT_ALIGNED_LOCAL): New.
28220 2013-02-01 Richard Biener <rguenther@suse.de>
28222 PR tree-optimization/56113
28223 * tree-ssa-structalias.c (label_visit): Reduce work for
28224 single-predecessor nodes.
28226 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
28228 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
28229 range isn't testing for zero.
28231 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
28233 PR middle-end/56113
28234 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
28236 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
28237 Nick Clifton <nickc@redhat.com>
28239 * config/v850/constraints.md (Q): Define as a memory constraint.
28240 * config/v850/predicates.md (label_ref_operand): New predicate.
28241 (e3v5_shift_operand): New predicate.
28242 (ior_operator): New predicate.
28243 * config/v850/t-v850: Add e3v5 multilib.
28244 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
28245 (v850_gen_movdi): Prototype.
28246 * config/v850/v850.c: Add support for e3v5 architecture.
28247 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
28249 (construct_save_jarl): Add e3v5 long JARL support.
28250 (v850_adjust_insn_length): New function. Adjust length of call
28251 insns when using e3v5 instructions.
28252 (v850_gen_movdi): New function: Generate instructions to move a
28254 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
28255 (CPP_SPEC): Define __v850e3v5__ as appropriate.
28256 (TARGET_USE_FPU): Enable for e3v5.
28257 (CONST_OK_FOR_W): New macro.
28258 (ADJUST_INSN_LENGTH): Define.
28259 * config/v850/v850.md (UNSPEC_LOOP): Define.
28260 (attr cpu): Add v850e3v5.
28261 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
28262 (movdi): New pattern.
28263 (movdi_internal): New pattern.
28264 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
28265 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
28266 (cstoresf4): Likewise.
28267 (cstoredf4): Likewise.
28268 (insv): New pattern.
28269 (rotlso3_a): New pattern.
28270 (rotlsi3_b): New pattern
28271 (rotlsi3_v850e3v5): New pattern.
28272 (doloop_begin): New pattern.
28273 (fix_loop_counter): New pattern.
28274 (doloop_end): New pattern.
28275 (branch_normal): Add e3v5 long branch support.
28276 (branch_invert): Likewise.
28277 (branch_z_normal): Likewise.
28278 (branch_z_invert): Likewise.
28279 (branch_nz_normal): Likewise.
28280 (branch_nz_invert): Likewise.
28281 (call_internal_short): Add e3v5 register-indirect JARL support.
28282 (call_internal_long): Likewise.
28283 (call_value_internal_short): Likewise.
28284 (call_value_internal_long): Likewise.
28285 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
28286 (mloop): New option.
28287 * config.gcc: Add support for configuring v840e3v5 target.
28288 * doc/invoke.texi: Document new v850 specific command line options.
28290 2013-01-31 Paul Koning <ni1d@arrl.net>
28294 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
28295 children if parent is a class.
28296 (prune_unused_types_prune): Don't add DW_AT_declaration.
28298 2013-01-31 Richard Biener <rguenther@suse.de>
28300 PR tree-optimization/56157
28301 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
28302 match up operand with SLP child.
28304 2013-01-31 Jason Merrill <jason@redhat.com>
28307 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
28308 parameters the first time.
28309 (gen_scheduled_generic_parms_dies): Check completeness here.
28311 2013-01-31 Richard Biener <rguenther@suse.de>
28313 PR middle-end/53073
28314 * common.opt (faggressive-loop-optimizations): New flag,
28315 enabled by default.
28316 * doc/invoke.texi (faggressive-loop-optimizations): Document.
28317 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
28318 infer_loop_bounds_from_undefined by it.
28320 2013-01-31 Richard Biener <rguenther@suse.de>
28322 PR tree-optimization/56150
28323 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
28324 visit virtual operands.
28325 (find_uses_to_rename_bb): Likewise.
28327 2013-01-31 Richard Biener <rguenther@suse.de>
28329 PR tree-optimization/56150
28330 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
28331 mixed store non-store stmts.
28333 2013-01-30 Jakub Jelinek <jakub@redhat.com>
28336 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
28337 LIBASAN_EARLY_SPEC is defined.
28338 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
28339 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
28341 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
28344 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
28345 invalid args instead of ICEing on it.
28346 (ix86_valid_target_attribute_tree): Return error_mark_node if
28347 ix86_valid_target_attribute_inner_p failed.
28348 (ix86_valid_target_attribute_p): Return false only if
28349 ix86_valid_target_attribute_tree returned error_mark_node. Allow
28350 target("default") attribute.
28351 (sorted_attr_string): Change argument from const char * to tree,
28352 merge in all target attribute arguments rather than just one.
28353 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
28354 instead of free. Avoid using strcat.
28355 (ix86_mangle_function_version_assembler_name): Mangle
28356 target("default") as if no target attribute is present. Adjust
28357 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
28358 instead of xmalloc and XDELETEVEC instead of free.
28359 (ix86_function_versions): Don't return true if one of the decls
28360 doesn't have target attribute. If they don't and one of the decls
28361 is DECL_FUNCTION_VERSIONED, report an error. Adjust
28362 sorted_attr_string caller. Use XDELETEVEC instead of free.
28363 (ix86_supports_function_versions): Remove.
28364 (make_name): Fix up formatting.
28365 (make_dispatcher_decl): Remove resolver_name and its initialization.
28366 Avoid leaking memory.
28367 (is_function_default_version): Return true if there is
28368 target("default") attribute rather than no target attribute at all.
28369 (make_resolver_func): Avoid leaking memory.
28370 (ix86_generate_version_dispatcher_body): Likewise.
28371 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
28372 * target.def (supports_function_versions): Remove.
28373 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
28374 * doc/tm.texi: Regenerated.
28376 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
28378 PR rtl-optimization/56144
28379 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
28380 for values with side effects.
28382 2013-01-30 Richard Biener <rguenther@suse.de>
28384 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
28385 (sparseset_pop): Likewise.
28386 * cfganal.c (compute_idf): Likewise. Increase work-stack size
28387 to be able to use quick_push in the worker loop.
28389 2013-01-30 Marek Polacek <polacek@redhat.com>
28391 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
28393 2013-01-30 Richard Biener <rguenther@suse.de>
28396 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
28398 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
28400 PR tree-optimization/56064
28401 * fixed-value.c (fixed_from_double_int): New function.
28402 * fixed-value.h (fixed_from_double_int): New prototype.
28403 (const_fixed_from_double_int): New static inline function.
28404 * fold-const.c (native_interpret_fixed): New static function.
28405 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
28406 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
28407 (native_encode_fixed): New static function.
28408 (native_encode_expr) <FIXED_CST>: Use it.
28409 (native_interpret_int): Move double_int worker code to...
28410 * double-int.c (double_int::from_buffer): ...this new static method.
28411 * double-int.h (double_int::from_buffer): Prototype it.
28413 2013-01-30 Richard Biener <rguenther@suse.de>
28415 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
28416 New pointer-map and obstack.
28417 (init_alias_vars): Allocate pointer-map and obstack.
28418 (delete_points_to_sets): Free them.
28419 (find_what_var_points_to): Cache result.
28420 (find_what_p_points_to): Adjust for changed interface of
28421 find_what_var_points_to.
28422 (compute_points_to_sets): Likewise.
28423 (ipa_pta_execute): Likewise.
28425 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
28427 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
28428 * configure: Regenerate.
28429 * config.in: Regenerate.
28430 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
28431 #nobits/#progbits if supported.
28433 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
28436 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
28437 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
28438 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
28440 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28442 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
28443 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
28445 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28447 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
28448 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
28450 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28452 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
28454 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
28455 * config/arm/cortex-a7.md: New bypasses using
28456 arm_mac_accumulator_is_result.
28458 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28460 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
28461 (cortex_a7_neon_mla): Likewise.
28462 (cortex_a7_fpfmad): New reservation.
28463 (cortex_a7_fpmacs): Use ffmas and update required units.
28464 (cortex_a7_fpmuld): Update required units and latency.
28465 (cortex_a7_fpmacd): Likewise.
28466 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
28467 (cortex_a7_neon). Likewise.
28468 (bypass) Update participating units.
28470 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28472 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
28473 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
28475 * config/arm/vfp11.md (vfp_farith): Use ffmas.
28476 (vfp_fmul): Use ffmad.
28477 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
28478 (cortex_r4_fmacd): Use ffmad.
28479 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
28480 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
28481 (cortex_a9_fmacd): Use ffmad.
28482 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
28483 (cortex_a8_vfp_macd): Use ffmad.
28484 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
28485 (cortex_a5_fpmacd): Use ffmad.
28486 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
28487 (cortex_a15_vfp_macd): Use ffmad.
28488 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
28490 2013-01-29 Jason Merrill <jason@redhat.com>
28493 * varasm.c (default_assemble_visibility): Don't warn about
28494 visibility on artificial decls.
28496 2013-01-29 Richard Biener <rguenther@suse.de>
28498 PR tree-optimization/56113
28499 * tree-ssa-structalias.c (equiv_class_lookup): Also return
28501 (label_visit): Free duplicate bitmaps and record the leader instead.
28502 (perform_var_substitution): Adjust.
28504 2013-01-29 Richard Biener <rguenther@suse.de>
28506 PR tree-optimization/55270
28507 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
28508 the CFG, schedule loops for fixup.
28510 2013-01-29 Nick Clifton <nickc@redhat.com>
28512 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
28515 2013-01-28 Leif Ekblad <leif@rdos.net>
28517 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
28518 * config/i386/i386.h (TARGET_RDOS): New macro.
28519 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
28520 * config/i386/i386.c (ix86_option_override_internal): For 64bit
28521 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
28522 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
28523 DEFAULT_LARGE_SECTION_THRESHOLD.
28524 * config/i386/i386.md (R14_REG, R15_REG): New constants.
28525 * config/i386/rdos.h: New file.
28526 * config/i386/rdos64.h: New file.
28528 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
28531 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
28534 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28536 PR rtl-optimization/56117
28537 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
28538 call cselib_lookup_from_insn on the MEM before calling
28539 add_insn_mem_dependence.
28541 2013-01-28 Richard Biener <rguenther@suse.de>
28543 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
28544 to a stmt that didn't have one.
28545 (copy_phis_for_bb): Likewise for PHI arguments.
28546 (copy_debug_stmt): Likewise for debug stmts.
28548 2013-01-28 Richard Biener <rguenther@suse.de>
28550 PR tree-optimization/56034
28551 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
28552 (partition_builtin_p): Adjust.
28553 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
28554 it is the last partition.
28555 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
28556 up the vertex for the definition.
28557 (classify_partition): Classify whether a partition is a
28558 PKIND_REDUCTION, thus has uses outside of the loop.
28559 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
28560 Merge all PKIND_REDUCTION partitions into the last partition.
28561 (tree_loop_distribution): Seed partitions from reductions as well.
28563 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28565 PR tree-optimization/56125
28566 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
28567 pow(x,c) into sqrt(x) * powi(x, n/2) or
28568 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
28569 optimizing for size.
28570 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
28571 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
28574 PR tree-optimization/56094
28575 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
28576 to UNKNOWN_LOCATION while gimplifying expr.
28578 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
28581 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
28582 operand 0 in movabs insn template for -masm=intel asm alternative.
28583 (*movabs<mode>_2): Ditto for operand 1.
28585 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
28588 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
28591 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28593 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
28596 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
28598 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
28599 Undef to avoid warning.
28601 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28603 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
28604 * configure: Regenerate.
28606 2013-01-25 Jakub Jelinek <jakub@redhat.com>
28608 PR tree-optimization/56098
28609 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
28610 for stmts with volatile ops.
28611 (cond_store_replacement): Don't optimize if assign has volatile ops.
28612 (cond_if_else_store_replacement_1): Don't optimize if either
28613 then_assign or else_assign have volatile ops.
28614 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
28617 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28619 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
28621 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28623 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
28624 missing ':' in asm example.
28626 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
28628 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
28629 entries into lane and laneq entries.
28630 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
28631 Remove AdvSIMD scalar modes.
28632 (aarch64_sq<r>dmulh_laneq<mode>): New.
28633 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
28635 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
28636 builtin implementations to relfect changes in RTL in aarch64-simd.md.
28637 * config/aarch64/iterators.md (VCOND): New.
28640 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28643 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
28644 Add NULL LIBNAME argument to existing definitions.
28645 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
28646 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
28647 * config/avr/avr.c (DEF_BUILTIN): Same.
28648 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
28649 (avr_expand_builtin): Expand to a vanilla call if a libgcc
28650 implementation is available (DECL_ASSEMBLER_NAME is set).
28651 (avr_fold_absfx): New static function.
28652 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
28653 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
28654 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
28655 AVR_BUILTIN_ABSLLK.
28656 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
28657 (abshk, absk, abslk, absllk): Provide as static inline functions.
28659 2013-01-25 Marek Polacek <polacek@redhat.com>
28661 PR tree-optimization/56035
28662 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
28664 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
28666 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
28667 (*movtf_internal_rex64): Add (!o,C) alternative
28668 (*movxf_internal_rex64): Ditto.
28669 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
28671 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
28673 * doc/invoke.texi: fix typo.
28674 * doc/objc.texi: fix typo.
28676 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
28678 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
28679 for the first two alternatives.
28681 2013-01-24 Diego Novillo <dnovillo@google.com>
28683 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
28684 (ggc-zone.o): Remove.
28685 * configure.ac: Remove option --with-gc.
28686 * configure: Re-generate.
28687 * doc/install.texi: Remove documentation for --with-gc.
28688 * gengtype.c (write_enum_defn): Remove. Update all users.
28689 (write_Types_process_field): Remove generation of gt_e_* argument.
28690 (output_type_enum): Remove. Update all users.
28691 (write_enum_defn): Remove. Update all users.
28692 (enum alloc_zone): Remove. Update all users.
28693 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
28694 * ggc-common.c (ggc_splay_alloc): Remove first argument.
28695 Update all callers.
28696 (struct ptr_data): Remove field TYPE. Update all users.
28697 (gt_pch_note_object): Remove argument TYPE. Update all users.
28698 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
28700 * ggc-none.c (ggc_alloc_typed_stat): Remove.
28701 (struct alloc_zone): Remove.
28702 (ggc_internal_alloc_zone_stat): Remove.
28703 (ggc_internal_cleared_alloc_zone_stat): Remove.
28704 * ggc-page.c (ggc_alloc_typed_stat): Remove.
28705 (ggc_pch_count_object): Remove last argument. Update all users.
28706 (ggc_pch_alloc_object): Remove last argument. Update all users.
28707 (struct alloc_zone): Remove.
28708 * ggc-zone.c: Remove.
28709 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
28710 (struct alloc_zone): Remove.
28711 (ggc_alloc_typed_stat): Remove.
28712 (ggc_alloc_typed): Remove.
28713 (ggc_splay_alloc): Remove first argument.
28714 (rtl_zone): Remove. Update all users.
28715 (tree_zone): Remove. Update all users.
28716 (tree_id_zone): Remove. Update all users.
28717 (ggc_internal_zone_alloc_stat): Remove. Update all users.
28718 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
28719 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
28720 * tree-ssanames.c: Remove references to zone allocator in comments.
28722 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
28724 * config/avr/avr.c (avr_out_fract): Make register numbers that
28725 might be outside of source operand signed.
28727 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
28729 * config/i386/constraints.md (Yf): New constraint.
28730 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
28731 of f constraint to conditionaly disable x87 register preferences.
28732 (*movdf_internal): Ditto.
28733 (*movsf_internal): Ditto.
28735 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
28737 PR inline-asm/55934
28738 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
28739 that have operands with impossible constraints.
28740 Add a FIXME for a speed-up opportunity.
28741 * lra-constraints.c (process_alt_operands): Verify that a class
28742 selected from constraints on asms is valid for the operand mode.
28743 (curr_insn_transform): Remove incorrect comment.
28745 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
28747 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
28748 TOC operand is a valid symbol ref in the constant pool.
28750 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28752 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
28754 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
28757 * config/avr/stdfix.h: New file.
28758 * t-avr (stdfix-gcc.h): New rule to build it.
28759 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
28761 2013-01-23 Kostya Serebryany <kcc@google.com>
28763 * config/darwin.h: remove dependency on
28764 CoreFoundation (asan on Mac OS).
28766 2013-01-23 Jakub Jelinek <jakub@redhat.com>
28769 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
28770 instead of cmpdi_operand for first comparison operand.
28771 Don't assert that comparison operands aren't both constants.
28773 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28775 * doc/install.texi (Downloading the Source): Update references to
28776 downloading separate components.
28778 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28780 * doc/extend.texi (__int128): Improve grammar.
28782 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
28785 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
28786 alternative to (o,r).
28787 (*movdi_internal_rex64): Remove (!o,n) alternative.
28788 (DImode immediate->memory splitter): Remove.
28789 (DImode immediate->memory peephole2): Remove.
28790 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
28791 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
28792 alternative to (!o,*r).
28793 (*movtf_internal_sse): New pattern.
28794 (*movxf_internal_rex64): New pattern.
28795 (*movxf_internal): Disable for TARGET_64BIT.
28796 (*movdf_internal_rex64): Remove (!o,F) alternative.
28798 2013-01-22 Jakub Jelinek <jakub@redhat.com>
28800 PR middle-end/56074
28801 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
28802 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
28803 * tree-vect-loop-manip.c (find_loop_location): Also ignore
28804 stmt locations where LOCATION_LOCUS of the stmt location is
28805 UNKNOWN_LOCATION or BUILTINS_LOCATION.
28808 * config/i386/i386.md (UNSPEC_STOS): New.
28809 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
28810 *strsetqi_1): Add UNSPEC_STOS.
28812 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
28815 * doc/invoke.texi: Remove left over -Wsynth example.
28817 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28819 PR tree-optimization/56051
28820 * fold-const.c (fold_binary_loc): Don't fold
28821 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
28822 a narrowing conversion, or widening conversion from signed
28825 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28827 PR rtl-optimization/56023
28828 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
28829 dependent on debug instruction.
28831 2013-01-21 Martin Jambor <mjambor@suse.cz>
28833 PR middle-end/56022
28834 * function.c (allocate_struct_function): Call
28835 invoke_set_current_function_hook earlier.
28837 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28839 * reload1.c (init_reload): Only initialize reload_obstack
28840 during the first call.
28842 2013-01-21 Marek Polacek <polacek@redhat.com>
28844 * cfgloop.c (verify_loop_structure): Fix up grammar.
28846 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
28848 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
28849 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
28851 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28854 * config/arm/marvell-pj4.md: Update copyright year.
28855 Fix up use of alu to alu_reg and simple_alu_imm.
28857 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28859 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
28861 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
28864 * lra-constraints.c (curr_insn_transform): Don't reuse original
28865 insn for secondary memory move when memory mode should be different.
28867 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28869 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
28870 atomic_storedi_1): New patterns.
28872 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
28874 btver2 pipeline descriptions.
28875 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
28877 * config/i386/i386.md (btver2_decode): New type attributes.
28878 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
28880 * config/i386/btver2.md: New file describing btver2 pipelines.
28882 2013-01-19 Andrew Pinski <apinski@cavium.com>
28884 PR tree-optimization/52631
28885 * tree-ssa-sccvn (visit_use): Before looking up the original
28886 statement, try looking up the simplified expression.
28888 2013-01-19 Anthony Green <green@moxielogic.com>
28890 * config/moxie/moxie.c (moxie_expand_prologue): Set
28891 current_function_static_stack_size.
28893 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28895 PR tree-optimization/56029
28896 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
28897 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
28899 2013-01-18 Sharad Singhai <singhai@google.com>
28901 PR tree-optimization/55995
28902 * dumpfile.c (dump_loc): Print location only if available.
28903 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
28905 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
28908 * lra-constraints.c (curr_insn_transform): Reuse original insn for
28909 secondary memory move.
28910 (inherit_reload_reg): Use rclass instead of cl for
28911 check_secondary_memory_needed_p.
28913 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28915 PR middle-end/56015
28916 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
28917 the case where writing real complex part of target modifies op1.
28919 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
28921 * config/aarch64/aarch64-simd.md
28922 (aarch64_vcond_internal<mode>): Handle unordered cases.
28923 * config/aarch64/iterators.md (v_cmp_result): New.
28925 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
28926 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28928 * config/arm/marvell-pj4.md: New file.
28929 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
28930 * config/arm/arm.md (generic_sched): Add marvell_pj4.
28931 (generic_vfp): Likewise.
28932 * config/arm/arm-cores.def: Add marvell-pj4.
28933 * config/arm/arm-tune.md: Regenerate.
28934 * config/arm/arm-tables.opt: Regenerate.
28935 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
28936 * doc/invoke.texi: Document marvell-pj4.
28938 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
28940 * config/aarch64/arm_neon.h: Map scalar types to standard types.
28942 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28947 * var-tracking.c (negative_power_of_two_p): New.
28948 (global_get_addr_cache, local_get_addr_cache): New.
28949 (get_addr_from_global_cache, get_addr_from_local_cache): New.
28950 (vt_canonicalize_addr): Rewrite using the above. Adjust the
28952 (vt_stack_offset_p): Remove.
28953 (vt_canon_true_dep): Always canonicalize loc's address.
28954 (clobber_overlapping_mems): Make sure we have a MEM.
28955 (local_get_addr_clear_given_value): New.
28956 (val_reset): Clear local cached entries.
28957 (compute_bb_dataflow): Create and release the local cache.
28958 Disable duplicate MEMs clobbering.
28959 (emit_notes_in_bb): Clobber MEMs likewise.
28960 (vt_emit_notes): Create and release the local cache.
28961 (vt_initialize, vt_finalize): Create and release the global
28962 cache, respectively.
28963 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
28965 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28967 PR libmudflap/53359
28968 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
28969 not found in the symtab.
28971 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28974 PR rtl-optimization/55547
28975 PR rtl-optimization/53827
28978 * alias.c (offset_overlap_p): New, factored out of...
28979 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
28980 the conservative special case for symbolic constants. Don't
28981 adjust zero sizes on alignment.
28983 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
28985 PR rtl-optimization/52573
28986 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
28987 REG_UNUSED for the same register.
28989 2013-01-17 Richard Biener <rguenther@suse.de>
28990 Marek Polacek <polacek@redhat.com>
28992 PR rtl-optimization/55833
28993 * loop-unswitch.c (unswitch_loops): Move loop verification...
28994 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
28995 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
28996 Set it to true when we're removing a loop from hierarchy tree in
28997 an irreducible region.
28998 (fix_bb_placements): Adjust caller.
28999 (fix_loop_placements): Likewise.
29001 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
29003 * config/avr/builtins.def (DEF_BUILTIN): Factor out
29004 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
29005 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
29006 Remove ID. Adjust comments.
29007 * config/avr/avr-c.c (avr_builtin_name): Remove.
29008 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
29009 * config/avr/avr.c (avr_tolower): New static function.
29010 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
29011 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
29012 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
29015 2013-01-17 Jan Hubicka <jh@suse.cz>
29017 PR tree-optimization/55273
29018 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
29020 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
29023 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
29024 store through atomic_store<mode>_1.
29025 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
29027 2013-01-17 Martin Jambor <mjambor@suse.cz>
29029 PR tree-optimizations/55264
29030 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
29031 for virtual methods.
29032 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
29033 virtual methods before inlining is over.
29034 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
29036 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
29039 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
29041 PR rtl-optimization/56005
29042 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
29043 pending reads for prefetch.
29045 2013-01-16 Ian Bolton <ian.bolton@arm.com>
29047 * config/aarch64/aarch64.md
29048 (*cstoresi_neg_uxtw): New pattern.
29049 (*cmovsi_insn_uxtw): New pattern.
29050 (*<optab>si3_uxtw): New pattern.
29051 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
29052 (*<optab>si3_insn_uxtw): New pattern.
29053 (*bswapsi2_uxtw): New pattern.
29055 2013-01-16 Richard Biener <rguenther@suse.de>
29057 * tree-inline.c (tree_function_versioning): Remove set but
29058 never used variable.
29060 2013-01-16 Richard Biener <rguenther@suse.de>
29062 PR tree-optimization/55964
29063 * tree-flow.h (rename_variables_in_loop): Remove.
29064 (rename_variables_in_bb): Likewise.
29065 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
29066 (copy_loop_before): Adjust and delete update-ssa status.
29067 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
29068 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
29069 (rename_variables_in_loop): Remove.
29070 (slpeel_update_phis_for_duplicate_loop): Likewise.
29071 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
29072 use available cfg machinery instead of duplicating it.
29073 Update PHI nodes and perform poor-mans SSA update here.
29074 (slpeel_tree_peel_loop_to_edge): Adjust.
29076 2013-01-16 Richard Biener <rguenther@suse.de>
29078 PR tree-optimization/54767
29079 PR tree-optimization/53465
29080 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
29081 (vrp_visit_phi_node): For PHI arguments coming via backedges
29082 drop all symbolical range information.
29083 (execute_vrp): Compute backedges.
29085 2013-01-16 Richard Biener <rguenther@suse.de>
29087 * doc/install.texi: Update CLooG and ISL requirements to
29090 2013-01-16 Christian Bruel <christian.bruel@st.com>
29093 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
29094 (broken_move): Handle UNSPECV_SP_SWITCH_B.
29095 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
29097 2013-01-16 DJ Delorie <dj@redhat.com>
29099 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
29100 (UNSPECV_SP_SWITCH_E): New.
29101 (sp_switch_1): Change to an unspec.
29102 (sp_switch_2): Change to an unspec. Don't use post-inc when we
29105 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
29107 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
29108 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
29109 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
29110 (expand_mem_thread_fence): Ditto.
29111 (expand_mem_signal_fence): Ditto.
29112 (expand_atomic_load): Ditto.
29113 (expand_atomic_store): Ditto.
29115 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
29117 PR rtl-optimization/55547
29118 PR rtl-optimization/53827
29121 * alias.c (memrefs_conflict_p): Set sizes to negative after
29124 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29127 * function.c (thread_prologue_and_epilogue_insns): Always
29128 add crtl->drap_reg to set_up_by_prologue.set, even if
29129 stack_realign_drap is false.
29131 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
29133 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
29134 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
29135 *call): Fix indention.
29137 2013-01-15 Tom de Vries <tom@codesourcery.com>
29140 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
29143 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
29145 PR rtl-optimization/55153
29146 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
29148 2013-01-15 Martin Jambor <mjambor@suse.cz>
29150 PR tree-optimization/55920
29151 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
29152 accesses as grp_to_be_debug_replaced.
29154 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29156 PR tree-optimization/55920
29157 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
29158 there is non-useless type conversion needed from debug rhs to lhs,
29159 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
29161 2013-01-15 Joseph Myers <joseph@codesourcery.com>
29162 Mikael Pettersson <mikpe@it.uu.se>
29165 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
29167 (ASM_OUTPUT_CASE_LABEL): Remove.
29168 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
29169 * final.c (shorten_branches): Update alignment of labels before
29170 jump tables if CASE_VECTOR_SHORTEN_MODE.
29172 2013-01-15 Richard Biener <rguenther@suse.de>
29175 * system.h: Do not include gmp.h for building host tools.
29177 2013-01-15 Richard Biener <rguenther@suse.de>
29179 PR middle-end/55882
29180 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
29181 account for bitpos when computing alignment.
29183 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
29185 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
29186 (ix86_target_macros_internal): Likewise.
29188 * config/i386/i386.c (m_CORE2I7): Removed.
29189 (m_CORE_HASWELL): New macro.
29190 (m_CORE_ALL): Likewise.
29191 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
29192 (initial_ix86_arch_features): Likewise.
29193 (processor_target_table): Initializations for Core avx2.
29194 (cpu_names): New names "core-avx2".
29195 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
29196 PROCESSOR_CORE_HASWELL.
29197 (ix86_issue_rate): New case.
29198 (ia32_multipass_dfa_lookahead): Likewise.
29199 (ix86_sched_init_global): Likewise.
29201 * config/i386/i386.h (TARGET_HASWELL): New macro.
29202 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
29203 (processor_type): New PROCESSOR_HASWELL.
29205 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29207 PR tree-optimization/55955
29208 * tree-vect-loop.c (vectorizable_reduction): Give up early on
29209 *SHIFT_EXPR and *ROTATE_EXPR codes.
29211 PR tree-optimization/48766
29212 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
29213 -ftrapv disable -fwrapv.
29215 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29218 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
29219 etc. to 1 and not to __flash.
29220 Use LL suffix for __INT24_MAX__ with -mint8.
29221 Use ULL suffix for __UINT24_MAX__ with -mint8.
29223 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29225 * config/avr/avr-arch.h
29226 (struct base_arch_s): Use typedef avr_arch_t instead.
29227 (struct arch_info_s): Use typedef avr_arch_info_t instead.
29228 (struct mcu_type_s): Use typedef avr_mcu_t instead.
29229 * config/avr/avr.c: Same.
29230 * config/avr/avr-devices.c: Same.
29231 * config/avr/driver-avr.c: Same.
29232 * config/avr/gen-avr-mmcu-texi.c: Same.
29233 * config/avr/avr-mcus.def: Adjust comment.
29235 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
29237 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
29238 * config/aarch64/iterators.md (VALLDI): New.
29240 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
29241 Andi Kleen <ak@linux.intel.com>
29244 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
29245 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
29248 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29250 * config/avr/avr-stdint.h: Remove trailing blanks.
29251 * config/avr/avr-log.h: Same.
29252 * config/avr/avr-arch.h: Same.
29253 * config/avr/avr-devices.c: Same.
29254 * config/avr/avr-dimode.md: Same.
29255 * config/avr/predicates.md: Same.
29256 * config/avr/avr-c.c: Same. And fix typo.
29258 * config/avr/avr-protos.h: Same. And:
29259 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
29260 (init_cumulative_args): Rename to avr_init_cumulative_args.
29261 (expand_prologue): Rename to avr_expand_prologue.
29262 (expand_epilogue): Rename to avr_expand_epilogue.
29263 (adjust_insn_length): Rename to avr_adjust_insn_length.
29264 (notice_update_cc): Rename to avr_notice_update_cc.
29265 (final_prescan_insn): Rename to avr_final_prescan_insn.
29266 * config/avr/avr.c: Same.
29267 * config/avr/avr.h: Same.
29268 * config/avr/avr.md: Remove trailing blanks.
29269 (prologue): Use avr_expand_prologue.
29270 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
29272 2013-01-14 Richard Biener <rguenther@suse.de>
29274 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
29275 verify_location, collect_subblocks): New functions.
29276 (verify_gimple_in_cfg): Verify that locations only reference
29277 BLOCKs in the functions BLOCK tree.
29279 2013-01-14 Richard Biener <rguenther@suse.de>
29281 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
29283 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
29285 (insert_out_of_ssa_copy_on_edge): Likewise.
29286 (rewrite_close_phi_out_of_ssa): Likewise.
29287 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
29289 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
29290 propagated constants.
29291 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
29294 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29296 * config/avr/avr-modes.def: Add GPL copyright notice.
29298 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
29300 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
29301 MEMMODEL_MASK to determine memory model.
29302 (atomic_store<mode>): Ditto from operands[2].
29303 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
29305 2013-01-13 Jakub Jelinek <jakub@redhat.com>
29308 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
29309 (fold_gimple_assign): Don't call unshare_expr here.
29310 (fold_ctor_reference): Call unshare_expr.
29312 2013-01-13 Terry Guo <terry.guo@arm.com>
29314 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
29315 * doc/fragments.texi: Document MULTILIB_REUSE.
29316 * gcc.c (multilib_reuse): New internal spec.
29317 (set_multilib_dir): Also search multilib from multilib_reuse.
29318 * genmultilib (tmpmultilib3): Refactor code.
29319 (tmpmultilib4): Ditto.
29320 (multilib_reuse): New multilib argument.
29322 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
29324 * Makefile.in: Update copyright.
29326 2013-01-12 Tom de Vries <tom@codesourcery.com>
29328 PR middle-end/55890
29329 * calls.c (expand_call): Check if arg_nr is valid.
29331 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
29333 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
29334 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
29335 documentation. Add missing '__' in front of
29336 __builtin_ia32_packssdw256.
29338 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29341 * config/s390/s390.c (s390_preferred_reload_class): Do not return
29342 NO_REGS for larl operands.
29343 (s390_reload_larl_operand): Use s390_load_address instead of
29346 2013-01-11 Richard Biener <rguenther@suse.de>
29348 * tree-cfg.c (verify_node_sharing_1): Split out from ...
29349 (verify_node_sharing): ... here.
29350 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
29352 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
29354 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
29355 Substitute TREECHECKING.
29356 * configure: Regenerate.
29357 * Makefile.in (TREECHECKING): New.
29359 2013-01-11 Richard Guenther <rguenther@suse.de>
29361 PR tree-optimization/44061
29362 * tree-vrp.c (extract_range_basic): Compute zero as
29363 value-range for __builtin_constant_p of function parameters.
29365 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
29367 Update copyright years.
29369 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
29371 PR rtl-optimization/55672
29372 * lra-eliminations.c (mark_not_eliminable): Permit addition with
29373 const to be eliminable.
29375 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
29377 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
29378 * configure: Regenerate.
29380 2013-01-10 Richard Biener <rguenther@suse.de>
29382 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
29384 2013-01-10 Richard Biener <rguenther@suse.de>
29387 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
29388 locations for virtual PHI arguments.
29389 (rewrite_update_phi_arguments): Likewise.
29391 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
29393 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
29396 2013-01-10 Jakub Jelinek <jakub@redhat.com>
29398 PR tree-optimization/55921
29399 * tree-complex.c (expand_complex_asm): New function.
29400 (expand_complex_operations_1): Call it for GIMPLE_ASM.
29402 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29405 * config/s390/s390.c (s390_symref_operand_p)
29406 (s390_loadrelative_operand_p): Merge the two functions.
29407 (s390_check_qrst_address, print_operand_address): Add parameters
29408 to s390_loadrelative_operand_p invokation.
29409 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
29410 (s390_reload_larl_operand, s390_secondary_reload): Use
29411 s390_loadrelative_operand_p instead of s390_symref_operand_p.
29412 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
29414 2013-01-09 Mike Stump <mikestump@comcast.net>
29416 * dse.c (record_store): Remove unnecessary assert.
29418 2013-01-09 Jan Hubicka <jh@suse.cz>
29420 PR tree-optimization/55569
29421 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
29422 * cfgloop.h (scale_loop_profile): Likewise.
29424 2013-01-09 Jan Hubicka <jh@suse.cz>
29427 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
29429 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
29431 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
29433 PR middle-end/55114
29434 * expr.h (maybe_emit_group_store): Declare.
29435 * expr.c (maybe_emit_group_store): New function.
29436 * builtins.c (expand_builtin_int_roundingfn): Call it.
29437 (expand_builtin_int_roundingfn_2): Likewise.
29439 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
29441 PR rtl-optimization/55829
29442 * lra-constraints.c (match_reload): Add code for absent output.
29443 (curr_insn_transform): Add code for reloads of matched inputs
29446 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
29448 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
29449 attribute of movddup insn to DF.
29450 (*vec_interleave_lowv2df): Ditto.
29451 (vec_dupv2df): Ditto.
29453 2013-01-09 Jan Hubicka <jh@suse.cz>
29455 PR tree-optimiation/55875
29456 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
29457 EVERY_ITERATION parameter.
29458 (number_of_iterations_exit): Check if exit is executed every iteration.
29459 (idx_infer_loop_bounds): Similarly here.
29460 (n_of_executions_at_most): Simplify
29461 to only test for cases where statement is dominated by the
29462 particular bound; handle correctly the "postdominance" test.
29463 (scev_probably_wraps_p): Use max loop iterations info
29464 as a global bound first.
29466 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
29467 Nick Clifton <nickc@redhat.com>
29469 * config/v850/v850.md (cbranchsf4): New pattern.
29470 (cstoresf4): New pattern.
29471 (cbranchdf4): New pattern.
29472 (cstoredf4): New pattern.
29473 (movsicc): Disallow floating point comparisons.
29474 (cmpsf_le_insn): Fix order of operators.
29475 (cmpsf_lt_insn): Likewise.
29476 (cmpsf_eq_insn): Likewise.
29477 (cmpdf_le_insn): Likewise.
29478 (cmpdf_lt_insn): Likewise.
29479 (cmpdf_eq_insn): Likewise.
29480 (cmpsf_ge_insn): Use LE comparison.
29481 (cmpdf_ge_insn): Likewise.
29482 (cmpsf_gt_insn): Use LT comparison.
29483 (cmpdf_gt_insn): Likewise.
29484 (cmpsf_ne_insn): Delete pattern.
29485 (cmpdf_ne_insn): Delete pattern.
29486 * config/v850/v850.c (v850_gen_float_compare): Use
29487 gen_cmpdf_eq_insn for NE comparison.
29488 (v850_float_z_comparison_operator)
29489 (v850_float_nz_comparison_operator): Move from here ...
29490 * config/v850/predicates.md: ... to here. Move GT and GE
29491 comparisons into v850_float_z_comparison_operator.
29492 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
29494 (v850_float_nz_comparison_operator): Likewise.
29496 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29498 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
29499 with calls to gen_insvsi/gen_insvdi.
29501 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29503 * config/i386/i386.c (initial_ix86_tune_features): Set up
29504 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
29506 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
29507 Jakub Jelinek <jakub@redhat.com>
29509 PR tree-optimization/48189
29510 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
29511 If nitercst is 0, don't predict the exit edge.
29513 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29515 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
29516 in asm_fprintf with reg_names.
29517 (aarch64_print_operand_address): Likewise.
29518 (aarch64_return_addr): Likewise.
29519 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
29521 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29523 * config/pa/pa.h (VAL_U6_BITS_P): Define.
29524 (INT_U6_BITS): Likewise.
29525 * config/pa/predicates.md (uint6_operand): New predicate.
29526 (shift5_operand, shift6_operand): Likewise.
29527 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
29529 (lshrdi3): Use shift6_operand.
29530 (shrpsi4, shrpdi4): New insn patterns.
29531 (extzv): Delete expander.
29532 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
29533 predicates in unamed zero extract patterns. Tighten common constraint.
29534 (extv): Delete expander.
29535 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
29536 predicates in unamed sign extract patterns. Tighten common constraint.
29537 (insv): Delete expander.
29538 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
29539 predicates in unamed insert patterns. Tighten common constraint.
29540 Change uint32_operand predicate to uint6_operand predicate in unamed
29541 DImode pattern to insert constant values of type 1...1xxxx.
29543 2013-01-04 Jan Hubicka <jh@suse.cz>
29545 PR tree-optimization/55823
29546 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
29549 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29550 Uros Bizjak <ubizjak@gmail.com>
29552 PR rtl-optimization/55845
29553 * df-problems.c (can_move_insns_across): Stop scanning at
29554 volatile_insn_p source instruction or give up if
29555 across_from .. across_to range contains any volatile_insn_p
29558 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29560 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
29561 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
29563 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
29564 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
29566 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29569 * asan.c (asan_clear_shadow): New function.
29570 (asan_emit_stack_protection): Use it.
29572 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29574 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
29575 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
29576 with tab instead of space.
29578 2013-01-08 Nick Clifton <nickc@redhat.com>
29580 * config/rl78/rl78.c (rl78_expand_prologue): Always select
29581 register bank 0 at the start of an interrupt handler.
29582 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
29585 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29587 * config/aarch64/aarch64-simd.md
29588 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
29589 (aarch64_simd_bsl): Likewise.
29590 (aarch64_vcond_internal<mode>): Likewise.
29591 (vcond<mode><mode>): Likewise.
29592 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
29593 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
29595 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29597 * config/aarch64/aarch64-builtins.c
29598 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
29600 2013-01-08 Martin Jambor <mjambor@suse.cz>
29603 * tree-sra.c (analyze_access_subtree): Return true also after
29604 potentially creating a debug-only replacement.
29606 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29608 PR middle-end/55890
29609 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
29611 PR tree-optimization/54120
29612 * tree-vrp.c (range_fits_type_p): Don't allow
29613 src_precision < precision from signed vr to unsigned_p
29614 if vr->min or vr->max is negative.
29615 (simplify_float_conversion_using_ranges): Test can_float_p
29616 against CODE_FOR_nothing.
29618 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29619 Richard Biener <rguenther@suse.de>
29621 PR middle-end/55851
29622 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
29623 types instead of just INTEGER_TYPE types.
29625 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
29627 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
29630 2013-01-07 Steve Ellcey <sellcey@mips.com>
29633 * config/mips/mips.opt: Change mad to mmad to match documentation.
29635 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29638 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
29639 .progmemx.data now.
29641 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29644 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
29645 (avr_addrspace_t): Add .section_name field.
29646 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
29648 (avr_addrspace): Same. Initialize .section_name. Remove last
29649 NULL entry. Put __memx into .progmemx.data.
29650 (progmem_section_prefix): Remove.
29651 (avr_asm_init_sections): No need to initialize progmem_section.
29652 (avr_asm_named_section): Use avr_addrspace[].section_name to get
29653 section name prefix.
29654 (avr_asm_select_section): Ditto. And use get_unnamed_section to
29655 retrieve the progmem section.
29656 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
29657 boundary to run over avr_addrspace[].
29658 (avr_register_target_pragmas): Ditto.
29660 2013-01-06 Jakub Jelinek <jakub@redhat.com>
29662 * varasm.c (output_constant_def_contents): For asan_protect_global
29663 protected strings, adjust DECL_ALIGN if needed, before testing for
29665 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
29666 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
29668 (output_object_block): For asan protected decls, emit asan padding
29669 after their contents.
29670 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
29671 (asan_finish_file): Test it here instead.
29673 2013-01-07 Nick Clifton <nickc@redhat.com>
29674 Matthias Klose <doko@debian.org>
29675 Doug Kwan <dougkwan@google.com>
29676 H.J. Lu <hongjiu.lu@intel.com>
29679 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
29681 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
29683 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
29685 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
29687 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
29689 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29692 * doc/install.texi (Cross-Compiler-Specific Options): Document
29695 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
29697 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
29698 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
29699 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
29700 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
29701 vqmovun_high_s64): Fix source operand number and update copyright.
29703 2013-01-07 Richard Biener <rguenther@suse.de>
29705 PR middle-end/55890
29706 * gimple.h (gimple_call_builtin_p): New overload.
29707 * gimple.c (validate_call): New function.
29708 (gimple_call_builtin_p): Likewise.
29709 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
29710 Use gimple_call_builtin_p.
29711 (find_func_clobbers): Likewise.
29712 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
29713 (strlen_optimize_stmt): Likewise.
29715 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29717 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
29718 (vld1q_dup_*): Likewise.
29719 (vld1_*): Likewise.
29720 (vld1q_*): Likewise.
29721 (vld1_lane_*): Likewise.
29722 (vld1q_lane_*): Likewise.
29724 2013-01-07 Richard Biener <rguenther@suse.de>
29726 * lto-streamer.h (LTO_minor_version): Bump to 2.
29728 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29730 * config/aarch64/aarch64-protos.h
29731 (aarch64_const_double_zero_rtx_p): Rename to...
29732 (aarch64_float_const_zero_rtx_p): ...this.
29733 (aarch64_float_const_representable_p): New.
29734 (aarch64_output_simd_mov_immediate): Likewise.
29735 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
29736 move immediate case.
29737 * config/aarch64/aarch64.c
29738 (aarch64_const_double_zero_rtx_p): Rename to...
29739 (aarch64_float_const_zero_rtx_p): ...this.
29740 (aarch64_print_operand): Allow printing of new constants.
29741 (aarch64_valid_floating_const): New.
29742 (aarch64_legitimate_constant_p): Check for valid floating-point
29744 (aarch64_simd_valid_immediate): Likewise.
29745 (aarch64_vect_float_const_representable_p): New.
29746 (aarch64_float_const_representable_p): Likewise.
29747 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
29748 (aarch64_output_simd_mov_immediate): New.
29749 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
29750 (*movdf_aarch64): Likewise.
29751 * config/aarch64/constraints.md (Ufc): New.
29752 (Y): call aarch64_float_const_zero_rtx.
29753 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
29755 2013-01-07 Richard Biener <rguenther@suse.de>
29757 PR tree-optimization/55888
29758 PR tree-optimization/55862
29759 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
29760 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
29761 not if it is contained therein.
29763 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29765 * config/avr/t-avr: Typo.
29767 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29770 * config/avr/t-avr: Don't automatically rebuild
29771 $(srcdir)/config/avr/t-multilib
29772 $(srcdir)/config/avr/avr-tables.opt
29773 $(srcdir)/doc/avr-mmcu.texi
29774 (avr-mcus): New phony target to build them on request.
29775 (s-avr-mlib, s-avr-mmcu-texi): Remove.
29776 * avr/avr-mcus.def: Adjust comments.
29778 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
29780 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
29782 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29784 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
29786 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29788 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
29790 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
29792 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
29793 to generate profiling.
29794 * config/rs6000/aix64.h (LIB_SPEC): Same.
29796 2013-01-04 Andrew Pinski <apinski@cavium.com>
29798 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
29800 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
29802 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
29804 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
29806 (ix86_expand_move): Ditto.
29807 (ix86_zero_extend_to_Pmode): Ditto.
29808 (ix86_expand_call): Ditto.
29809 (ix86_expand_special_args_builtin): Ditto.
29810 (ix86_expand_builtin): Ditto.
29812 2013-01-04 Richard Biener <rguenther@suse.de>
29814 PR tree-optimization/55862
29815 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
29816 translating them through PHI nodes.
29818 2013-01-04 Martin Jambor <mjambor@suse.cz>
29820 PR tree-optimization/55755
29821 * tree-sra.c (sra_modify_assign): Do not check that an access has no
29822 children when trying to avoid producing a VIEW_CONVERT_EXPR.
29824 2013-01-04 Marek Polacek <polacek@redhat.com>
29826 PR middle-end/55859
29827 * opts.c (default_options_optimization): Clarify error message.
29829 2013-01-04 Richard Biener <rguenther@suse.de>
29831 PR middle-end/55863
29832 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
29835 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29838 * config/pa/pa.md (movsi): Revert previous change.
29839 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
29842 2013-01-03 Richard Henderson <rth@redhat.com>
29844 * config/i386/i386.c (ix86_expand_move): Always assign to op1
29845 after eliminating TLS symbols.
29847 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
29850 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
29851 * graphite-poly.c (debug_gmp_value): Likewise.
29853 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
29856 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
29857 selected code model, define __code_mode_small__, __code_model_medium__,
29858 __code_model_large__, __code_model_32__ or __code_model_kernel__.
29859 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
29860 xchg temporary register with %k. Declare temporary register as
29862 [__x86_64__]: For medium and large code models, preserve %rbx register.
29864 2013-01-03 Richard Biener <rguenther@suse.de>
29866 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
29867 (dump_subscript): Adjust.
29868 (finalize_ddr_dependent): Do not dump redundant info.
29869 (analyze_siv_subscript): Adjust.
29870 (subscript_dependence_tester): Likewise.
29871 (compute_affine_dependence): Likewise.
29873 2013-01-03 Richard Biener <rguenther@suse.de>
29876 2013-01-03 Richard Biener <rguenther@suse.de>
29878 PR tree-optimization/55857
29879 * tree-vect-stmts.c (vectorizable_load): Do not setup
29880 re-alignment for invariant loads.
29882 2013-01-02 Richard Biener <rguenther@suse.de>
29884 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29885 invariant load do not generate a vector load from the scalar location.
29887 2013-01-03 Richard Biener <rguenther@suse.de>
29889 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
29890 for not vectorizing.
29891 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
29892 not build INDIRECT_REFs, call get_name once only.
29893 (vect_create_data_ref_ptr): Likewise. Dump base object kind
29894 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
29896 2013-01-03 Richard Biener <rguenther@suse.de>
29898 PR tree-optimization/55857
29899 * tree-vect-stmts.c (vectorizable_load): Do not setup
29900 re-alignment for invariant loads.
29902 2013-01-03 Richard Biener <rguenther@suse.de>
29905 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
29906 prefer a built-in decl.
29908 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29910 * gcc.c (process_command): Update copyright notice dates.
29911 * gcov.c (print_version): Likewise.
29912 * gcov-dump.c (print_version): Likewise.
29914 PR rtl-optimization/55838
29915 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
29916 iv0.step, iv1.step and step.
29918 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29919 Marc Glisse <marc.glisse@inria.fr>
29921 PR tree-optimization/55832
29922 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
29923 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
29924 integer_{one,zero}_node.
29926 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29929 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
29930 * var-tracking.c (reverse_op): Don't add reverse ops to
29931 VALUEs that have already
29932 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
29934 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
29936 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
29938 2013-01-02 Teresa Johnson <tejohnson@google.com>
29940 * dumpfile.c (dump_loc): Print filename with location.
29941 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
29942 new location_t parameter to emit complete unroll message with
29943 new dump framework.
29944 (canonicalize_loop_induction_variables): Compute loops location
29945 and pass to try_unroll_loop_completely.
29946 * loop-unroll.c (report_unroll_peel): New function.
29947 (peel_loops_completely): Use new dump format with location
29948 for main dumpfile message, and invoke report_unroll_peel on success.
29949 (decide_unrolling_and_peeling): Ditto.
29950 (decide_peel_once_rolling): Remove old dumpfile message subsumed
29951 by report_unroll_peel.
29952 (decide_peel_completely): Ditto.
29953 (decide_unroll_constant_iterations): Ditto.
29954 (decide_unroll_runtime_iterations): Ditto.
29955 (decide_peel_simple): Ditto.
29956 (decide_unroll_stupid): Ditto.
29957 * cfgloop.c (get_loop_location): New function.
29958 * cfgloop.h (get_loop_location): Declare.
29960 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29962 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
29965 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29967 PR middle-end/55198
29968 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
29969 BLKmode objects when EXPAND_MEMORY is specified.
29971 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29973 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
29975 (fold_builtin_cpu): Do not share cpu model decls across statements.
29977 2013-01-02 Jason Merrill <jason@redhat.com>
29980 * tree.c (build_array_type_1): Revert earlier change.
29982 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
29984 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
29986 * config/aarch64/aarch64-tune.md: Re-generate.
29988 2013-01-02 Richard Biener <rguenther@suse.de>
29990 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29991 invariant load do not generate a vector load from the scalar location.
29993 2013-01-02 Richard Biener <rguenther@suse.de>
29996 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
29997 * configure: Regenerate.
29999 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30001 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
30002 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
30003 (expand_builtin_int_roundingfn_2): Keep the original target around
30004 for the fallback case.
30006 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30008 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
30009 to be clear for sign changes.
30011 2013-01-01 Jan Hubicka <jh@suse.cz>
30013 * ipa-inline-analysis.c: Fix formatting.
30015 2013-01-01 Jakub Jelinek <jakub@redhat.com>
30017 PR tree-optimization/55831
30018 * tree-vect-loop.c (get_initial_def_for_induction): Use
30019 gsi_after_labels instead of gsi_start_bb.
30021 Copyright (C) 2013 Free Software Foundation, Inc.
30023 Copying and distribution of this file, with or without modification,
30024 are permitted in any medium without royalty provided the copyright
30025 notice and this notice are preserved.