1 2013-11-06 Richard Biener <rguenther@suse.de>
3 PR tree-optimization/58653
4 * tree-predcom.c (ref_at_iteration): Rewrite to generate
6 (prepare_initializers_chain): Adjust.
8 2013-11-06 Andrew MacLeod <amacleod@redhat.com>
10 * gimple.h (block_in_transaction): Move to basic-block.h and rename.
11 (gimple_in_transaction): Use bb_in_transaction.
12 * basic-block.h (bb_in_transaction): Relocate here and rename.
13 * tree-ssa-loop-im.c (execute_sm): Use bb_in_transaction.
15 2013-11-06 Richard Biener <rguenther@suse.de>
17 * tree.c (drop_tree_overflow): New function.
18 * tree.h (drop_tree_overflow): Declare.
19 * gimplify.c (gimplify_expr): Drop TREE_OVERFLOW.
20 * tree-vrp.c (range_int_cst_singleton_p): Use
21 is_overflow_infinity instead of testing TREE_OVERFLOW.
22 (extract_range_from_assert): Likewise.
23 (zero_nonzero_bits_from_vr): Likewise.
24 (extract_range_basic): Likewise.
25 (register_new_assert_for): Use drop_tree_overflow.
26 (vrp_visit_phi_node): Likewise.
28 2013-11-06 Dodji Seketeli <dodji@redhat.com>
31 * input.h (location_get_source_line): Take an additional line_size
33 * input.c (get_line): New static function definition.
34 (read_line): Take an additional line_length output parameter to be
35 set to the size of the line. Use the new get_line function do the
37 (location_get_source_line): Take an additional output line_len
38 parameter. Update the use of read_line to pass it the line_len
40 * diagnostic.c (adjust_line): Take an additional input parameter
41 for the length of the line, rather than calculating it with
43 (diagnostic_show_locus): Adjust the use of
44 location_get_source_line and adjust_line with respect to their new
45 signature. While displaying a line now, do not stop at the first
46 null byte. Rather, display the zero byte as a space and keep
47 going until we reach the size of the line.
49 2013-11-06 Eric Botcazou <ebotcazou@adacore.com>
51 * config/i386/i386.c (ix86_expand_prologue): Optimize stack checking for
52 leaf functions without dynamic stack allocation.
53 * config/ia64/ia64.c (ia64_emit_probe_stack_range): Adjust.
54 (ia64_expand_prologue): Likewise.
55 * config/mips/mips.c (mips_expand_prologue): Likewise.
56 * config/rs6000/rs6000.c (rs6000_emit_prologue): Likewise.
57 * config/sparc/sparc.c (sparc_expand_prologue): Likewise.
58 (sparc_flat_expand_prologue): Likewise.
60 2013-11-06 James Greenhalgh <james.greenhalgh@arm.com>
62 * config/aarch64/arm_neon.h (__ST2_LANE_FUNC): Better model data size.
63 (__ST3_LANE_FUNC): Likewise.
64 (__ST4_LANE_FUNC): Likewise.
66 2013-11-06 Nick Clifton <nickc@redhat.com>
68 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define the
69 name returned by msp430_mcu_name.
70 (LIB_SPEC): If a -T option has not been specified then set a
71 default, mcu-specific, linker script.
72 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more mcu names.
73 * config/msp430/msp430.c (msp430x_names): Likewise.
74 Alpha sort the names for ease of comparison.
75 (msp430_mcu_name): New function: Returns a string suitable for
76 use as a C preprocessor symbol based upon the name of the MCU
78 (msp430_option_override): Accept msp430x and msp430xv2 as generic
80 * config/msp430/msp430-protos.h (msp430_mcu_name): Prototype.
82 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
83 unless it is the last part of a spec substring.
84 * doc/invoke.texi (Spec Files): Document space insertion
87 2013-11-06 Christian Bruel <christian.bruel@st.com>
89 * config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
90 Factorize probabilities, Use adjust_address instead of
91 adjust_automodify_address when possible. Enable for optimize.
92 (sh_expand_strlen): New function.
93 * config/sh/sh-protos.h (sh_expand_strlen): Declare.
94 * config/sh/sh.md (strlensi): New pattern.
95 (UNSPEC_BUILTIN_STRLEN): Define.
97 2013-11-06 Jakub Jelinek <jakub@redhat.com>
100 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
101 (expand_assignment): If *bitpos is negative, set *offset
102 and adjust *bitpos, so that it is not negative.
104 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
106 * config/i386/bdver3.md : Added two additional decoder units
107 to support issue rate of 4 and remodeled vector unit.
108 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
109 architectures is set to 4.
110 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
111 lookahead is set to 4 for BD architectures.
113 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
115 * config/rs6000/rs6000.c (rs6000_option_override_internal):
116 Remove restriction against use of VSX instructions when generating
117 code for little endian mode.
119 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
121 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
122 for both big and little endian.
123 (mulv8hi3): Swap input operands for merge high and merge low
124 instructions for little endian.
126 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
128 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
129 define_insn to define_expand that uses even patterns for big
130 endian and odd patterns for little endian.
131 (vec_widen_smult_even_v16qi): Likewise.
132 (vec_widen_umult_even_v8hi): Likewise.
133 (vec_widen_smult_even_v8hi): Likewise.
134 (vec_widen_umult_odd_v16qi): Likewise.
135 (vec_widen_smult_odd_v16qi): Likewise.
136 (vec_widen_umult_odd_v8hi): Likewise.
137 (vec_widen_smult_odd_v8hi): Likewise.
138 (altivec_vmuleub): New define_insn.
139 (altivec_vmuloub): Likewise.
140 (altivec_vmulesb): Likewise.
141 (altivec_vmulosb): Likewise.
142 (altivec_vmuleuh): Likewise.
143 (altivec_vmulouh): Likewise.
144 (altivec_vmulesh): Likewise.
145 (altivec_vmulosh): Likewise.
147 2013-11-05 Mike Stump <mikestump@comcast.net>
149 * Makefile.in (mostlyclean): Remove c-family objects.
151 2013-11-05 Ian Lance Taylor <iant@google.com>
153 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
154 If possible, add .cfi directives to record change to bx.
155 * config/i386/i386.c (ix86_emit_cfi): New function.
156 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
158 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
160 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
161 immediately after a label for a tablejump pattern.
163 * config/arm/arm.c (is_jump_table): Remove.
164 (create_fix_barrier): Use tablejump_p instead.
165 (arm_reorg): Likewise.
166 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
167 (thumb2_output_casesi): Likewise.
168 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
169 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
170 casesi_shift_media, casesi_load_media): Likewise.
171 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
172 * config/microblaze/microblaze.md: Likewise.
174 2013-11-05 Tobias Burnus <burnus@net-b.de>
176 * doc/invoke.texi (-Wdate-time): Document.
178 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
180 * double-int.c (lshift_double, rshift_double): Remove
181 SHIFT_COUNT_TRUNCATED handling.
183 2013-11-05 Jeff Law <law@redhat.com>
185 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
186 * common.opt (-fisolate-erroneous-paths): Add option and
188 * gimple-ssa-isolate-paths.c: New file.
189 * gimple.c (check_loadstore): New function.
190 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
191 Verify OP is in the argument list and the argument corresponding
192 to OP is a pointer type. Use operand_equal_p rather than
193 pointer equality when testing if OP is on the nonnull list.
194 Use check_loadstore rather than count_ptr_derefs. Handle
195 GIMPLE_RETURN statements.
196 * tree-vrp.c (infer_nonnull_range): Remove.
197 * gimple.h (infer_nonnull_range): Declare.
198 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
199 * passes.def: Add pass_isolate_erroneous_paths.
200 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
201 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
202 * tree-ssa.c (struct count_ptr_d): Remove.
203 (count_ptr_derefs, count_uses_and_derefs): Remove.
204 * tree-ssa.h (count_uses_and_derefs): Remove.
206 2013-11-05 Jakub Jelinek <jakub@redhat.com>
208 PR rtl-optimization/58997
209 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
210 get_iv_value to be in iv->mode rather than iv->extend_mode.
211 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
212 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
213 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
216 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
218 * gimple.h: Move some prototypes to gimple-expr.h and add to include
220 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
221 Move to gimple-expr.h.
222 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
223 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
224 is_gimple_id, virtual_operand_p, is_gimple_addressable,
225 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
227 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
228 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
229 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
230 is_gimple_address, is_gimple_invariant_address,
231 is_gimple_ip_invariant_address, is_gimple_min_invariant,
232 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
233 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
234 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
235 gimple_decl_printable_name, useless_type_conversion_p,
236 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
238 * gimple-expr.c: New File.
239 (useless_type_conversion_p, gimple_set_body, gimple_body,
240 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
241 gimple_can_coalesce_p, extract_ops_from_tree_1,
242 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
243 is_gimple_address, is_gimple_invariant_address,
244 is_gimple_ip_invariant_address, is_gimple_min_invariant,
245 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
246 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
247 is_gimple_mem_ref_addr): Relocate here.
248 * Makefile.in (OBJS): Add gimple-expr.o.
250 2013-11-05 David Malcolm <dmalcolm@redhat.com>
252 * gengtype-parse.c (struct_field_seq): Support empty structs.
254 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
256 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
258 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
260 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
262 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
264 (LONG_DOUBLE_TYPE_SIZE): New define.
265 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
267 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
270 * doc/extend.texi [visibility ("visibility_type")]: Add example
271 about visibility attribute on namespace declaration.
273 2013-11-05 Richard Biener <rguenther@suse.de>
276 * passes.def (all_passes): Start with pass_fixup_cfg again.
278 2013-11-05 Richard Biener <rguenther@suse.de>
280 PR tree-optimization/58955
281 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
284 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
286 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
288 (vec_pack_ufix_trunc_v2df): Likewise.
290 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
293 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
294 pattern, instead of word_mode.
296 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
298 (set_storage_via_setmem): Likewise.
300 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
302 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
303 gimple.h and the rest of the condition in eliminate_build.
304 (eliminate_build): Call new routine.
305 * gimple.h (phi_ssa_name_p): Delete.
307 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
309 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
310 case of no prefix and reserving zero slots, because when that's the
311 case we'll never get here.
312 * vec.h (va_heap::reserve): Don't try and handle
313 vec_prefix::calculate_allocation returning zero because that should
316 2013-11-05 Richard Biener <rguenther@suse.de>
319 * tree-dfa.c (get_ref_base_and_extent): Merge common code
320 in MEM_REF and TARGET_MEM_REF handling. Make sure to
321 process trailing array detection before diving into the
322 view-converted object (and possibly apply some extra offset).
324 2013-11-05 Joseph Myers <joseph@codesourcery.com>
326 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
328 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
330 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
332 PR tree-optimization/58958
333 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
334 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
336 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
338 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
339 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
341 2013-11-05 Jakub Jelinek <jakub@redhat.com>
343 PR tree-optimization/58984
344 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
345 set *SIZE_P if non-NULL on success.
346 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
348 (ipcp_transform_function): Likewise. Punt if size of access
349 is different from TYPE_SIZE on v->value's type.
351 2013-11-05 Tobias Burnus <burnus@net-b.de>
353 * doc/invoke.texi (-fopenmp-simd): Document new option.
354 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
355 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
356 * tree.c (attribute_value_equal): Ditto.
358 2013-11-04 Wei Mi <wmi@google.com>
360 * sched-rgn.c (add_branch_dependences): Keep insns in
361 a SCHED_GROUP at the end of BB to remain their location.
363 2013-11-04 Wei Mi <wmi@google.com>
365 * config/i386/i386.c (memory_address_length): Extract a part
366 of code to rip_relative_addr_p.
367 (rip_relative_addr_p): New Function.
368 (ix86_macro_fusion_p): Ditto.
369 (ix86_macro_fusion_pair_p): Ditto.
370 * config/i386/i386.h: Add new tune features about macro-fusion.
371 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
372 * doc/tm.texi: Generated.
373 * doc/tm.texi.in: Ditto.
374 * haifa-sched.c (try_group_insn): New Function.
375 (group_insns_for_macro_fusion): Ditto.
376 (sched_init): Call group_insns_for_macro_fusion.
377 * target.def: Add two hooks: macro_fusion_p and
380 2013-11-04 Kostya Serebryany <kcc@google.com>
382 Update to match the changed asan API.
383 * asan.c (asan_function_start): New function.
384 (asan_emit_stack_protection): Update the string stored in the
385 stack red zone to match new API. Store the PC of the current
386 function in the red zone.
387 (asan_global_struct): Update the __asan_global definition to match
389 (asan_add_global): Ditto.
390 * asan.h (asan_function_start): New prototype.
391 * final.c (final_start_function): Call asan_function_start.
392 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
395 2013-11-04 Wei Mi <wmi@google.com>
397 * config/i386/i386-c.c (ix86_target_macros_internal): Separate
398 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
399 * config/i386/i386.c (ix86_option_override_internal): Ditto.
400 (ix86_issue_rate): Ditto.
401 (ix86_adjust_cost): Ditto.
402 (ia32_multipass_dfa_lookahead): Ditto.
403 (ix86_sched_init_global): Ditto.
404 (get_builtin_code_for_version): Ditto.
405 * config/i386/i386.h (enum target_cpu_default): Ditto.
406 (enum processor_type): Ditto.
407 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
409 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
411 PR rtl-optimization/58967
412 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
413 !lra_in_progress for mode sizes bigger word.
415 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
417 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
418 arguments to merge instruction for little endian.
419 (vec_widen_umult_lo_v16qi): Likewise.
420 (vec_widen_smult_hi_v16qi): Likewise.
421 (vec_widen_smult_lo_v16qi): Likewise.
422 (vec_widen_umult_hi_v8hi): Likewise.
423 (vec_widen_umult_lo_v8hi): Likewise.
424 (vec_widen_smult_hi_v8hi): Likewise.
425 (vec_widen_smult_lo_v8hi): Likewise.
427 2013-11-04 Ian Lance Taylor <iant@google.com>
429 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
430 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
431 builtins that take pointers.
432 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
433 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
434 (run_gcc): Pass -fnon-call-exceptions.
436 2013-11-04 Jakub Jelinek <jakub@redhat.com>
438 * optabs.c (expand_vec_perm): Revert one incorrect line from
441 PR tree-optimization/58978
442 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
443 use_stmt by single_imm_use directly. Only call single_imm_use
446 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
448 PR rtl-optimization/58968
449 * lra-spills.c (return_regno_p): New function.
450 (lra_final_code_change): Use it.
452 2013-11-04 Joseph Myers <joseph@codesourcery.com>
454 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
455 * target.def (float_exceptions_rounding_supported_p): New hook.
456 * targhooks.c (default_float_exceptions_rounding_supported_p): New
458 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
459 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
461 * doc/tm.texi: Regenerate.
462 * config.gcc (powerpc*-*-linux*): Set extra_objs.
463 * config/rs6000/rs6000-linux.c: New file.
464 * config/rs6000/rs6000-protos.h
465 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
466 * config/rs6000/linux.h
467 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
468 * config/rs6000/linux64.h
469 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
470 * config/rs6000/t-linux (rs6000-linux.o): New rule.
471 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
473 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
475 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
476 Replace the define_insn_and_split with a define_insn and two
477 define_splits, with the split after reload re-permuting the source
478 register to its original value.
479 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
480 (*vsx_le_perm_store_v8hi): Likewise.
481 (*vsx_le_perm_store_v16qi): Likewise.
483 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
485 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
488 2013-11-04 Jakub Jelinek <jakub@redhat.com>
490 PR tree-optimization/58946
491 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
492 bbs with bbinfo[idx].op != NULL before all blocks with
493 bbinfo[idx].op == NULL.
495 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
497 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
498 (avr_dump_double_int_hex): Likewise.
499 (avr_log_vadump): Remove %D and %X handling.
500 * config/avr/avr.c (avr_double_int_push_digit): Delete.
501 (avr_map_op_t): Change map from double_int to unsigned int.
502 (avr_map_op): Update accordingly.
503 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
504 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
505 unsigned ints rather than double_ints.
507 2013-11-03 Marek Polacek <polacek@redhat.com>
509 Implement -fsanitize=vla-bound.
510 * opts.c (common_handle_option): Handle vla-bound.
511 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
512 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
513 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
515 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
517 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
520 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
522 * config/i386/constraints.md (Ts, Tv): New address constrains.
523 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
524 constraint for address_no_seg_operand.
525 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
526 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
527 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
528 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
529 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
530 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
531 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
532 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
533 constraint for vsib_address_operand.
535 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
537 * gcse.c (pre_delete): Remove references to regmove from comments.
538 * recog.c: (validate_replace_rtx_1): Likewise.
539 * config/rl78/rl78.c: Likewise.
540 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
541 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
542 * common/config/mmix/mmix-common.c: Likewise.
544 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
546 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
547 * gimplify.c (gimplify_compound_lval): Likewise.
548 * graphite-clast-to-gimple.c (gloog): Likewise.
549 * graphite-dependences.c (loop_is_parallel_p): Likewise.
550 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
551 (limit_scop); Likewise.
552 (build_scops): Likewise.
553 (dot_scop): Likewise.
554 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
555 (build_scop_drs): Likewise.
556 (insert_stmts): Likewise.
557 (insert_out_of_ssa_copy): Likewise.
558 (remove_phi): Likewise.
559 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
560 * hw-doloop.c (discover_loop): Likewise.
561 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
562 * tree-dfa.c (dump_enumerated_decls): Likewise.
563 * tree-if-conv.c (if_convertable_loop_p): Likewise.
564 * tree-inline.c (tree_function_versioning): Likewise.
565 * tree-loop-distribution.c (build_rdg): Likewise.
566 (rdg_flag_vertex_and_dependent): Likewise.
567 (distribute_loop): Likewise.
568 * tree-parloops.c (loop_parallel_p): Likewise.
569 (eliminate_local_variables): Likewise.
570 (separate_decls_in_region): Likewise.
571 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
572 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
573 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
574 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
575 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
576 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
577 (vectorizable_condition): Likewise.
579 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
581 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
582 * configure: Regenerate.
583 * config/i386/i386.md (*movdi_internal): Change
584 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
585 (*movdf_internal): Ditto.
586 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
587 * config/i386/sse.md (vec_concatv2di): Output interunit movq
588 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
590 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
592 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
595 2013-10-31 David Malcolm <dmalcolm@redhat.com>
597 Automated part of renaming of symtab_node_base to symtab_node.
599 Patch autogenerated by rename_symtab.py from
600 https://github.com/davidmalcolm/gcc-refactoring-scripts
601 revision 58bb219cc090b2f4516a9297d868c245495ee622
602 with ChangeLog entry fixed up by hand.
604 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
606 (cgraph_node_for_asm): Likewise.
607 * cgraph.h (symtab_node_base): Likewise.
608 (cgraph_node): Likewise.
609 (varpool_node): Likewise.
610 (is_a_helper <cgraph_node>::test): Likewise.
611 (is_a_helper <varpool_node>::test): Likewise.
612 (symtab_nodes): Likewise.
613 (symtab_register_node): Likewise.
614 (symtab_unregister_node): Likewise.
615 (symtab_remove_node): Likewise.
616 (symtab_get_node): Likewise.
617 (symtab_node_for_asm): Likewise.
618 (symtab_node_asm_name): Likewise.
619 (symtab_node_name): Likewise.
620 (symtab_insert_node_to_hashtable): Likewise.
621 (symtab_add_to_same_comdat_group): Likewise.
622 (symtab_dissolve_same_comdat_group_list): Likewise.
623 (dump_symtab_node): Likewise.
624 (debug_symtab_node): Likewise.
625 (dump_symtab_base): Likewise.
626 (verify_symtab_node): Likewise.
627 (verify_symtab_base): Likewise.
628 (symtab_used_from_object_file_p): Likewise.
629 (symtab_alias_ultimate_target): Likewise.
630 (symtab_resolve_alias): Likewise.
631 (fixup_same_cpp_alias_visibility): Likewise.
632 (symtab_for_node_and_aliases): Likewise.
633 (symtab_nonoverwritable_alias): Likewise.
634 (availability symtab_node_availability): Likewise.
635 (symtab_semantically_equivalent_p): Likewise.
636 (fixup_same_cpp_alias_visibility): Likewise.
637 (symtab_prevail_in_asm_name_hash): Likewise.
640 (varpool_first_variable): Likewise.
641 (varpool_next_variable): Likewise.
642 (varpool_first_static_initializer): Likewise.
643 (varpool_next_static_initializer): Likewise.
644 (varpool_first_defined_variable): Likewise.
645 (varpool_next_defined_variable): Likewise.
646 (cgraph_first_defined_function): Likewise.
647 (cgraph_next_defined_function): Likewise.
648 (cgraph_first_function): Likewise.
649 (cgraph_next_function): Likewise.
650 (cgraph_first_function_with_gimple_body): Likewise.
651 (cgraph_next_function_with_gimple_body): Likewise.
652 (symtab_alias_target): Likewise.
653 (symtab_real_symbol_p): Likewise.
654 (symtab_can_be_discarded): Likewise.
655 * cgraphbuild.c (mark_address): Likewise.
656 (mark_load): Likewise.
657 (mark_store): Likewise.
658 * cgraphunit.c (decide_is_symbol_needed): Likewise.
660 (enqueue_node): Likewise.
661 (referred_to_p): Likewise.
662 (cgraph_process_same_body_aliases): Likewise.
663 (analyze_functions): Likewise.
664 (handle_alias_pairs): Likewise.
665 (output_weakrefs): Likewise.
667 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
668 * ipa-inline-analysis.c (inline_write_summary): Likewise.
669 * ipa-prop.c (remove_described_reference): Likewise.
670 (try_decrement_rdesc_refcount): Likewise.
671 (ipa_edge_duplication_hook): Likewise.
672 * ipa-ref.c (ipa_record_reference): Likewise.
673 (ipa_maybe_record_reference): Likewise.
674 (ipa_clone_ref): Likewise.
675 (ipa_clone_references): Likewise.
676 (ipa_clone_referring): Likewise.
677 (ipa_find_reference): Likewise.
678 (ipa_remove_stmt_references): Likewise.
679 (ipa_clear_stmts_in_references): Likewise.
680 * ipa-ref.h (symtab_node_base): Likewise.
682 (ipa_record_reference): Likewise.
683 (ipa_maybe_record_reference): Likewise.
684 (ipa_clone_references): Likewise.
685 (ipa_clone_referring): Likewise.
686 (ipa_clone_ref): Likewise.
687 (ipa_find_reference): Likewise.
688 (ipa_remove_stmt_references): Likewise.
689 (ipa_clear_stmts_in_references): Likewise.
690 * ipa-reference.c (ipa_reference_write_optimization_summary):
692 * ipa.c (enqueue_node): Likewise.
693 (process_references): Likewise.
694 (walk_polymorphic_call_targets): Likewise.
695 (symtab_remove_unreachable_nodes): Likewise.
696 (address_taken_from_non_vtable_p): Likewise.
697 (comdat_can_be_unshared_p_1): Likewise.
698 (comdat_can_be_unshared_p): Likewise.
699 (can_replace_by_local_alias): Likewise.
700 (function_and_variable_visibility): Likewise.
701 * is-a.h: Likewise (within example in comment).
702 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
703 (lto_symtab_encoder_encode): Likewise.
704 (lto_symtab_encoder_delete_node): Likewise.
705 (lto_symtab_encoder_in_partition_p): Likewise.
706 (lto_set_symtab_encoder_in_partition): Likewise.
707 (output_refs): Likewise.
708 (compute_ltrans_boundary): Likewise.
709 (output_symtab): Likewise.
710 (input_node): Likewise.
711 (input_ref): Likewise.
712 (input_edge): Likewise.
713 (input_cgraph_1): Likewise.
714 (input_refs): Likewise.
715 (output_cgraph_opt_summary): Likewise.
716 (input_node_opt_summary): Likewise.
717 (input_cgraph_opt_section): Likewise.
718 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
720 * lto-streamer-out.c (lto_output): Likewise.
721 (output_symbol_p): Likewise.
722 (produce_symtab): Likewise.
723 * lto-streamer.h (lto_encoder_entry): Likewise.
724 (lto_free_function_in_decl_state_for_node): Likewise.
725 (lto_symtab_encoder_encode): Likewise.
726 (lto_symtab_encoder_delete_node): Likewise.
727 (lto_symtab_encoder_in_partition_p): Likewise.
728 (lto_set_symtab_encoder_in_partition): Likewise.
729 (lto_symtab_encoder_lookup): Likewise.
730 (lsei_node): Likewise.
731 (lto_symtab_encoder_deref): Likewise.
732 * symtab.c (symtab_hash): Likewise.
733 (assembler_name_hash): Likewise.
734 (symtab_nodes): Likewise.
735 (hash_node): Likewise.
737 (hash_node_by_assembler_name): Likewise.
738 (eq_assembler_name): Likewise.
739 (insert_to_assembler_name_hash): Likewise.
740 (unlink_from_assembler_name_hash): Likewise.
741 (symtab_prevail_in_asm_name_hash): Likewise.
742 (symtab_register_node): Likewise.
743 (symtab_insert_node_to_hashtable): Likewise.
744 (symtab_unregister_node): Likewise.
745 (symtab_get_node): Likewise.
746 (symtab_remove_node): Likewise.
747 (symtab_initialize_asm_name_hash): Likewise.
748 (symtab_node_for_asm): Likewise.
749 (symtab_add_to_same_comdat_group): Likewise.
750 (symtab_dissolve_same_comdat_group_list): Likewise.
751 (symtab_node_asm_name): Likewise.
752 (symtab_node_name): Likewise.
753 (dump_symtab_base): Likewise.
754 (dump_symtab_node): Likewise.
755 (dump_symtab): Likewise.
756 (debug_symtab_node): Likewise.
757 (verify_symtab_base): Likewise.
758 (verify_symtab_node): Likewise.
759 (verify_symtab): Likewise.
760 (symtab_used_from_object_file_p): Likewise.
761 (symtab_node_availability): Likewise.
762 (symtab_alias_ultimate_target): Likewise.
763 (fixup_same_cpp_alias_visibility): Likewise.
764 (symtab_resolve_alias): Likewise.
765 (symtab_for_node_and_aliases): Likewise.
766 (symtab_for_node_and_aliases): Likewise.
767 (symtab_nonoverwritable_alias_1): Likewise.
768 (symtab_nonoverwritable_alias): Likewise.
769 (symtab_semantically_equivalent_p): Likewise.
770 * value-prof.c (init_node_map): Likewise.
771 * varasm.c (find_decl): Likewise.
772 * varpool.c (varpool_node_for_asm): Likewise.
773 (varpool_remove_unreferenced_decls): Likewise.
775 2013-10-31 David Malcolm <dmalcolm@redhat.com>
777 Manual part of renaming of symtab_node_base to symtab_node.
779 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
780 clashes with the preferred name for the base class.
781 (const_symtab_node): Remove redundant typedef.
783 2013-10-31 Jakub Jelinek <jakub@redhat.com>
785 * optabs.c (expand_vec_perm): Avoid vector mode punning
787 * expmed.c (store_bit_field_1): Likewise.
788 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
789 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
790 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
791 * config/i386/i386.c (ix86_expand_vector_move_misalign,
792 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
793 ix86_expand_sse_unpack, ix86_expand_args_builtin,
794 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
795 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
796 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
797 expand_vec_perm_vpshufb2_vpermq,
798 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
799 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
800 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
801 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
804 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
806 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
807 Use gcc_unreachable for unhandled modes. Do not check results of
808 expand_simple_binop. If not expanded to target, move the result.
810 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
811 Shiva Chen <shiva0217@gmail.com>
813 * config.gcc (nds32*-*-*): Add nds32 target.
814 * config/nds32/nds32.c: New file.
815 * config/nds32/nds32.h: New file.
816 * config/nds32/nds32.md: New file.
817 * config/nds32/constants.md: New file.
818 * config/nds32/constraints.md: New file.
819 * config/nds32/iterators.md: New file.
820 * config/nds32/nds32-doubleword.md: New file.
821 * config/nds32/nds32-intrinsic.md: New file.
822 * config/nds32/nds32_intrinsic.h: New file.
823 * config/nds32/nds32-modes.def: New file.
824 * config/nds32/nds32-multiple.md: New file.
825 * config/nds32/nds32.opt: New file.
826 * config/nds32/nds32-opts.h: New file.
827 * config/nds32/nds32-protos.h: New file.
828 * config/nds32/nds32-peephole2.md: New file.
829 * config/nds32/pipelines.md: New file.
830 * config/nds32/predicates.md: New file.
831 * config/nds32/t-mlibs: New file.
832 * common/config/nds32: New directory and files.
834 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
835 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
836 * doc/install.texi (Cross-Compiler-Specific Options): Document
837 --with-nds32-lib for nds32 target.
838 * doc/extend.texi (Function Attributes, Target Builtins): Document
839 nds32 specific attributes.
841 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
843 * lra-constraints (process_alt_operands): Use the result
844 elimination register for operand when matching constraints.
846 2013-10-31 Jakub Jelinek <jakub@redhat.com>
848 * tree-vrp.c (maybe_set_nonzero_bits): New function.
849 (remove_range_assertions): Call it.
851 * tree.c (tree_ctz): New function.
852 * tree.h (tree_ctz): New prototype.
853 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
854 first argument from tree to const_tree.
855 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
856 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
857 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
859 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
860 * tree-vect-loop.c (vect_analyze_loop_operations,
861 vect_transform_loop): Don't force scalar loop for bound just because
862 number of iterations is unknown, only do it if it is not known to be
863 a multiple of vectorization_factor.
864 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
866 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
867 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
868 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
869 (set_nonzero_bits, get_nonzero_bits): New prototypes.
870 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
871 a default def isn't partially constant.
872 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
873 is known to be partially zero.
874 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
875 to see if a default def isn't partially constant.
876 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
877 creation of a range, if VR_RANGE, try to improve nonzero_bits from
879 (set_nonzero_bits, get_nonzero_bits): New functions.
881 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
882 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
883 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
884 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
885 uses but in the condition and ASSERT_EXPR and the other successor of
886 the predecessor bb is __builtin_unreachable (), set_range_info of the
887 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
889 2013-10-31 Martin Jambor <mjambor@suse.cz>
891 PR rtl-optimization/58934
893 2013-10-30 Martin Jambor <mjambor@suse.cz>
894 PR rtl-optimization/10474
895 * ira.c (find_moveable_pseudos): Do not calculate dominance info
897 (interesting_dest_for_shprep): New function.
898 (split_live_ranges_for_shrink_wrap): Likewise.
899 (ira): Calculate dominance info and df analysis. Call
900 split_live_ranges_for_shrink_wrap.
902 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
903 Yury Gribov <y.gribov@samsung.com>
906 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
907 shadow pointer to avoid clobbering the main one.
909 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
911 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
913 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
916 * ira-color.c (update_costs_from_copies): Add new parameter. Use
917 it for calling update_costs_from_allocno.
918 (assign_hard_reg): Call restore_costs_from_copies only for
919 !retry_p. Pass new argument to update_costs_from_copies.
920 (color_pass): Pass new argument to update_costs_from_copies.
921 (ira_mark_allocation_change): Ditto.
923 2013-10-30 Sharad Singhai <singhai@google.com>
926 * opts.c (common_handle_option): Remove deprecated option
927 -ftree-vectorizer-verbose.
928 * doc/invoke.texi (Debugging Options): Ditto.
929 * opts-global.c (handle_common_deferred_options): Ditto.
930 (dump_remap_tree_vectorizer_verbose): Delete.
931 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
933 2013-10-30 DJ Delorie <dj@redhat.com>
935 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
936 that take no arguments.
938 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
941 * reload1.c (update_eliminables_and_spill): New function, broken
943 (reload): Use it. Check for frame size change after frame size
944 alignment, and call update_eliminables_and_spill first if continue-ing.
946 2013-10-30 Cong Hou <congh@google.com>
949 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
950 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
951 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
953 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
955 PR rtl-optimization/58369
956 * reload1.c (compute_reload_subreg_offset): New function.
957 (choose_reload_regs): Use it to pass endian-correct
958 offset to subreg_regno_offset.
960 2013-10-30 Tobias Burnus <burnus@net-b.de>
963 * tree-cfg.c (replace_loop_annotate): Replace warning by
966 2013-10-30 Jason Merrill <jason@redhat.com>
968 * configure.ac (loose_warn): Add -Wno-format if
969 --disable-build-format-warnings.
971 2013-10-30 David Malcolm <dmalcolm@redhat.com>
973 * cgraphunit.c (analyze_functions): Split symtab_node declarations
974 onto multiple lines to make things easier for rename_symtab.py.
976 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
977 (symtab_semantically_equivalent_p): Likewise.
979 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
982 * lra.c (check_rtl): Remove address check before LRA work.
984 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
986 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
987 POINTER_PLUS_EXPR in the defining statement.
989 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
992 * tree-pass.h (make_pass_regmove): Remove.
993 * timevar.def (TV_REGMOVE): Remove.
994 * passes.def (pass_regmove): Remove.
995 * opts.c (default_options_table): Remove entry for regmove.
996 * doc/passes.texi: Remove regmove pass description.
997 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
999 (-fdump-rtl-regmove): Ditto.
1000 * common.opt (foptimize-register-move, fregmove): Ignore.
1001 * Makefile.in (OBJS): Remove regmove.o.
1002 * regmove.c: Remove.
1003 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
1005 (struct ira_allocno) New member allocno_prefs.
1006 (ALLOCNO_PREFS): New macro.
1007 (ira_prefs, ira_prefs_num): New external vars.
1008 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
1009 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
1010 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
1012 (ira_add_allocno_copy_to_list): Remove prototype.
1013 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
1014 (ira_pref_iterator): New type.
1015 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
1016 (FOR_EACH_PREF): New macro.
1017 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
1018 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
1020 (ira_setup_alts): New function.
1021 (decrease_live_ranges_number): New function.
1022 (ira): Call the above function.
1023 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
1024 (ira_create_allocno): Initialize allocno prefs.
1025 (pref_pool, pref_vec): New static vars.
1026 (initiate_prefs, find_allocno_pref, ira_create_pref): New
1028 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
1029 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
1030 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
1031 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
1032 (ira_add_allocno_copy_to_list): Make static. Rename to
1033 add_allocno_copy_to_list.
1034 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
1035 swap_allocno_copy_ends_if_necessary.
1036 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
1037 ira_remove_allocno_prefs.
1038 (ira_flattening): Ditto.
1039 (ira_build): Call initiate_prefs, print_prefs.
1040 (ira_destroy): Call finish_prefs.
1041 * ira-color.c (struct update_cost_record): New.
1042 (struct allocno_color_data): Add new member update_cost_records.
1043 (update_cost_record_pool): New static var.
1044 (init_update_cost_records, get_update_cost_record): New functions.
1045 (free_update_cost_record_list, finish_update_cost_records): Ditto.
1046 (struct update_cost_queue_elem): Add member from.
1047 (initiate_cost_update): Call init_update_cost_records.
1048 (finish_cost_update): Call finish_update_cost_records.
1049 (queue_update_cost, get_next_update_cost): Add new param from.
1050 (Update_allocno_cost, update_costs_from_allocno): New functions.
1051 (update_costs_from_prefs): Ditto.
1052 (update_copy_costs): Rename to update_costs_from_copies.
1053 (restore_costs_from_copies): New function.
1054 (update_conflict_hard_regno_costs): Don't go back.
1055 (assign_hard_reg): Call restore_costs_from_copies. Add printing
1057 (pop_allocnos): Add priniting more debug info.
1058 (color_allocnos): Remove prefs for conflicting hard regs.
1059 Call update_costs_from_prefs.
1060 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
1061 (get_dup_num): Rename, modify, and move to ira.c
1062 (process_regs_for_copy): Add prefs.
1063 (add_insn_allocno_copies): Put src as first arg of
1064 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
1065 * ira-costs.c (record_reg_classes): Modify and move code into
1066 record_operands_costs.
1067 (find_costs_and_classes): Create prefs for the hard reg of small
1069 (process_bb_node_for_hard_reg_moves): Add prefs.
1071 2013-10-30 Richard Biener <rguenther@suse.de>
1074 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
1075 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
1077 (pre_and_rev_post_order_compute): ... which now wraps it.
1078 * graph.c (draw_cfg_nodes_no_loops): Use
1079 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
1082 2013-10-30 Christian Bruel <christian.bruel@st.com>
1084 * config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
1085 (sh_expand_cmpstr): Handle known align and schedule improvements.
1086 * config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
1087 * config/sh/sh.md (cmpstrnsi): New pattern.
1089 2013-10-30 Martin Jambor <mjambor@suse.cz>
1091 PR rtl-optimization/10474
1092 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1094 (interesting_dest_for_shprep): New function.
1095 (split_live_ranges_for_shrink_wrap): Likewise.
1096 (ira): Calculate dominance info and df analysis. Call
1097 split_live_ranges_for_shrink_wrap.
1099 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1102 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
1104 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1106 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
1107 * tree.h (POINTER_BOUNDS_P): New.
1108 (BOUNDED_TYPE_P): New.
1110 (pointer_bounds_type_node): New.
1111 * tree.c (build_common_tree_nodes): Initialize
1112 pointer_bounds_type_node.
1113 * gimple.h (gimple_call_get_nobnd_arg_index): New.
1114 (gimple_call_num_nobnd_args): New.
1115 (gimple_call_nobnd_arg): New.
1116 (gimple_return_retbnd): New.
1117 (gimple_return_set_retbnd): New
1118 * gimple.c (gimple_build_return): Increase number of ops
1119 for return statement.
1120 (gimple_call_get_nobnd_arg_index): New.
1121 * gimple-pretty-print.c (dump_gimple_return): Print second op.
1123 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1125 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
1126 with "chkp ctor" and "bnd_legacy" attributes.
1127 * gimplify.c (gimplify_init_constructor): Avoid infinite
1128 loop during gimplification of bounds initializer.
1130 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1132 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
1133 (handle_bnd_legacy): New.
1134 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
1135 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
1138 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
1140 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
1141 (BT_FN_PTR_CONST_PTR): New.
1142 (BT_FN_CONST_PTR_CONST_PTR): New.
1143 (BT_FN_PTR_CONST_PTR_SIZE): New.
1144 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
1145 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
1146 (BT_FN_VOID_CONST_PTR_SIZE): New.
1147 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
1148 * chkp-builtins.def: New.
1149 * builtins.def: include chkp-builtins.def.
1150 (DEF_CHKP_BUILTIN): New.
1151 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
1152 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
1153 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
1154 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
1155 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
1156 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
1157 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
1158 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
1159 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
1160 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
1161 * common.opt (fcheck-pointer-bounds): New.
1162 * toplev.c (process_options): Check Pointer Bounds Checker is
1164 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
1166 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1168 * target.def (builtin_chkp_function): New.
1169 (chkp_bound_type): New.
1170 (chkp_bound_mode): New.
1171 (fn_abi_va_list_bounds_size): New.
1172 (load_bounds_for_arg): New.
1173 (store_bounds_for_arg): New.
1174 * targhooks.h (default_load_bounds_for_arg): New.
1175 (default_store_bounds_for_arg): New.
1176 (default_fn_abi_va_list_bounds_size): New.
1177 (default_chkp_bound_type): New.
1178 (default_chkp_bound_mode): New.
1179 (default_builtin_chkp_function): New.
1180 * targhooks.c (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 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
1187 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
1188 (TARGET_STORE_BOUNDS_FOR_ARG): New.
1189 (TARGET_BUILTIN_CHKP_FUNCTION): New.
1190 (TARGET_CHKP_BOUND_TYPE): New.
1191 (TARGET_CHKP_BOUND_MODE): New.
1192 * doc/tm.texi: Regenerated.
1193 * langhooks.h (lang_hooks): Add chkp_supported field.
1194 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
1195 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
1197 2013-10-29 Andrew Pinski <apinski@cavium.com>
1199 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
1200 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
1201 fails, combining the branches anyways.
1202 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
1203 increases the number of combinings.
1204 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
1206 2013-10-29 Mike Stump <mikestump@comcast.net>
1208 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
1209 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
1210 (make_vector_mode): Increase namebuf to 16.
1211 (emit_insn_modes_h): When processing BImode, don't
1212 also match partial int modes.
1213 (emit_class_narrowest_mode): Likewise.
1215 * config/bfin/bfin-modes.def: Add precision to PDI.
1216 * config/m32c/m32c-modes.def: Add precision to PSI.
1217 * config/msp430/msp430-modes.def: Add precision to PSI.
1218 * config/rs6000/rs6000-modes.def: Add precision to PTI.
1219 * config/sh/sh-modes.def: Add precision to PSI and PDI.
1221 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
1224 * config/sh/sh.md (*addc): Rename existing variations to ...
1225 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
1226 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
1227 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
1228 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
1230 2013-10-29 Teresa Johnson <tejohnson@google.com>
1233 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
1234 insanities when updating probabilities.
1236 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1238 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
1239 prettyprinter for cgraph_node to reflect the conversion of the
1240 symtable types to a C++ class hierarchy: it now *is* a
1241 symtab_node_base, rather than having one (named "symbol").
1243 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
1245 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
1246 __cilkrts_pop_frame. If matched, then return true for built-in
1248 (expand_builtin): Added BUILT_IN_CILK_DETACH and
1249 BUILT_IN_CILK_POP_FRAME case.
1250 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
1251 (lhs_cilk_detect_spawn): Likewise.
1252 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
1253 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
1254 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
1255 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
1256 (LANG_HOOKS_CILKPLUS): Likewise.
1257 * tree.h (CILK_SPAWN_FN): Likewise.
1258 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
1259 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
1260 (OBJS): Added cilk-common.o.
1261 (BUILTINS_DEF): Added cilk-builtins.def.
1262 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
1263 (lhd_cilk_detect_spawn): Likewise.
1264 * langhooks.h (lang_hooks_for_cilkplus): New struct.
1265 (struct lang_hooks): Added new field called "cilkplus."
1266 * cilk-common.c: New file.
1268 * cilk-builtins.def: Likewise.
1269 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
1270 "__cilk" macro and set it to 200.
1271 * function.h (struct function::cilk_frame_decl): New field.
1272 (struct function::is_cilk_function): Likewise.
1273 (struct function::calls_cilk_spawn): Likewise.
1274 * gimplify.c (gimplify_call_expr): Added a check if the function call
1275 being gimplified is a spawn detach point. If so, then add pop_frame
1276 and detach function calls.
1277 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
1278 for gimplifying _Cilk_spawn and _Cilk_sync statements.
1279 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
1280 function. If so, added a _Cilk_sync and gimplified it.
1281 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
1282 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
1283 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
1285 (can_inline_edge_p): Prevent inling of spawnee function.
1286 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
1287 for functions that use Cilk keywords.
1288 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
1289 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
1290 CILK_SYNC_STMT cases.
1291 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
1293 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
1294 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
1295 * passes.texi (Cilk Keywords): New section that describes the compiler
1296 code changes for handling Cilk Keywords.
1298 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1300 Patch autogenerated by refactor_symtab.py from
1301 https://github.com/davidmalcolm/gcc-refactoring-scripts
1302 revision 58bb219cc090b2f4516a9297d868c245495ee622
1304 * asan.c (asan_finish_file): Update for conversion of symtab types to
1305 a true class hierarchy.
1306 * cfgexpand.c (estimated_stack_frame_size): Likewise.
1307 * cgraph.c (cgraph_get_body): Likewise.
1308 (cgraph_get_create_real_symbol_node): Likewise.
1309 (verify_cgraph_node): Likewise.
1310 (verify_edge_corresponds_to_fndecl): Likewise.
1311 (verify_edge_count_and_frequency): Likewise.
1312 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
1313 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1314 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
1315 (cgraph_node_cannot_return): Likewise.
1316 (cgraph_set_pure_flag_1): Likewise.
1317 (cgraph_set_const_flag_1): Likewise.
1318 (cgraph_set_nothrow_flag_1): Likewise.
1319 (cgraph_make_node_local_1): Likewise.
1320 (cgraph_for_node_and_aliases): Likewise.
1321 (cgraph_for_node_thunks_and_aliases): Likewise.
1322 (cgraph_node_can_be_local_p): Likewise.
1323 (cgraph_node_cannot_be_local_p_1): Likewise.
1324 (cgraph_function_body_availability): Likewise.
1325 (dump_cgraph_node): Likewise.
1326 (cgraph_rtl_info): Likewise.
1327 (cgraph_mark_address_taken_node): Likewise.
1328 (cgraph_remove_node): Likewise.
1329 (cgraph_release_function_body): Likewise.
1330 (cgraph_update_edges_for_call_stmt_node): Likewise.
1331 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
1332 (cgraph_make_edge_direct): Likewise.
1333 (cgraph_resolve_speculation): Likewise.
1334 (cgraph_speculative_call_info): Likewise.
1335 (cgraph_turn_edge_to_speculative): Likewise.
1336 (cgraph_create_edge_1): Likewise.
1337 (cgraph_set_call_stmt): Likewise.
1338 (cgraph_node_for_asm): Likewise.
1339 (cgraph_add_thunk): Likewise.
1340 (cgraph_same_body_alias): Likewise.
1341 (cgraph_create_function_alias): Likewise.
1342 (cgraph_create_node): Likewise.
1343 (cgraph_create_empty_node): Likewise.
1344 (record_function_versions): Likewise.
1345 (used_from_object_file_p): Likewise.
1346 * cgraph.h (symtab_can_be_discarded): Likewise.
1347 (symtab_real_symbol_p): Likewise.
1348 (cgraph_mark_force_output_node): Likewise.
1349 (cgraph_edge_recursive_p): Likewise.
1350 (symtab_alias_target): Likewise.
1351 (varpool_all_refs_explicit_p): Likewise.
1352 (varpool_can_remove_if_no_refs): Likewise.
1353 (cgraph_only_called_directly_or_aliased_p): Likewise.
1354 (cgraph_next_function_with_gimple_body): Likewise.
1355 (cgraph_first_function_with_gimple_body): Likewise.
1356 (cgraph_function_with_gimple_body_p): Likewise.
1357 (cgraph_next_function): Likewise.
1358 (cgraph_first_function): Likewise.
1359 (cgraph_next_defined_function): Likewise.
1360 (cgraph_first_defined_function): Likewise.
1361 (varpool_next_defined_variable): Likewise.
1362 (varpool_first_defined_variable): Likewise.
1363 (varpool_next_static_initializer): Likewise.
1364 (varpool_first_static_initializer): Likewise.
1365 (varpool_next_variable): Likewise.
1366 (varpool_first_variable): Likewise.
1367 (varpool_node_name): Likewise.
1368 (varpool): Likewise.
1370 (is_a_helper <varpool_node>::test): Likewise.
1371 (is_a_helper <cgraph_node>::test): Likewise.
1372 (varpool_variable_node): Likewise.
1373 (cgraph_function_or_thunk_node): Likewise.
1374 (varpool_alias_target): Likewise.
1375 (cgraph_alias_target): Likewise.
1376 (cgraph_node_name): Likewise.
1377 (varpool_node_asm_name): Likewise.
1378 (cgraph_node_asm_name): Likewise.
1379 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
1380 (cgraph_rebuild_references): Likewise.
1381 (rebuild_cgraph_edges): Likewise.
1382 (record_eh_tables): Likewise.
1383 (build_cgraph_edges): Likewise.
1384 (mark_store): Likewise.
1385 (mark_load): Likewise.
1386 (mark_address): Likewise.
1387 (record_type_list): Likewise.
1388 (record_reference): Likewise.
1389 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
1390 (cgraph_materialize_clone): Likewise.
1391 (cgraph_function_versioning): Likewise.
1392 (cgraph_copy_node_for_versioning): Likewise.
1393 (update_call_expr): Likewise.
1394 (cgraph_find_replacement_node): Likewise.
1395 (cgraph_create_virtual_clone): Likewise.
1396 (cgraph_clone_node): Likewise.
1397 * cgraphunit.c (compile): Likewise.
1398 (output_weakrefs): Likewise.
1399 (output_in_order): Likewise.
1400 (expand_function): Likewise.
1401 (assemble_thunks_and_aliases): Likewise.
1402 (expand_thunk): Likewise.
1403 (mark_functions_to_output): Likewise.
1404 (handle_alias_pairs): Likewise.
1405 (analyze_functions): Likewise.
1406 (walk_polymorphic_call_targets): Likewise.
1407 (varpool_finalize_decl): Likewise.
1408 (process_function_and_variable_attributes): Likewise.
1409 (cgraph_process_same_body_aliases): Likewise.
1410 (analyze_function): Likewise.
1411 (cgraph_add_new_function): Likewise.
1412 (cgraph_finalize_function): Likewise.
1413 (referred_to_p): Likewise.
1414 (cgraph_reset_node): Likewise.
1415 (cgraph_process_new_functions): Likewise.
1416 (enqueue_node): Likewise.
1417 (decide_is_symbol_needed): Likewise.
1418 * coverage.c (coverage_compute_profile_id): Likewise.
1419 * dbxout.c (dbxout_expand_expr): Likewise.
1420 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
1421 (reference_to_unused): Likewise.
1422 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1423 * gimplify.c (unvisit_body): Likewise.
1424 (unshare_body): Likewise.
1425 * ipa-cp.c (ipcp_generate_summary): Likewise.
1426 (ipcp_decision_stage): Likewise.
1427 (identify_dead_nodes): Likewise.
1428 (decide_whether_version_node): Likewise.
1429 (decide_about_value): Likewise.
1430 (perhaps_add_new_callers): Likewise.
1431 (create_specialized_node): Likewise.
1432 (update_profiling_info): Likewise.
1433 (ipcp_propagate_stage): Likewise.
1434 (estimate_local_effects): Likewise.
1435 (good_cloning_opportunity_p): Likewise.
1436 (devirtualization_time_bonus): Likewise.
1437 (propagate_constants_accross_call): Likewise.
1438 (initialize_node_lattices): Likewise.
1439 (ipcp_cloning_candidate_p): Likewise.
1440 (determine_versionability): Likewise.
1441 (print_all_lattices): Likewise.
1442 (print_lattice): Likewise.
1443 (ipcp_discover_new_direct_edges): Likewise.
1444 * ipa-devirt.c (ipa_devirt): Likewise.
1445 (likely_target_p): Likewise.
1446 (update_type_inheritance_graph): Likewise.
1447 (possible_polymorphic_call_target_p): Likewise.
1448 (dump_possible_polymorphic_call_targets): Likewise.
1449 (devirt_variable_node_removal_hook): Likewise.
1450 (record_binfo): Likewise.
1451 (maybe_record_node): Likewise.
1452 (build_type_inheritance_graph): Likewise.
1453 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1454 (inline_generate_summary): Likewise.
1455 (inline_analyze_function): Likewise.
1456 (do_estimate_growth): Likewise.
1457 (simple_edge_hints): Likewise.
1458 (estimate_node_size_and_time): Likewise.
1459 (estimate_edge_devirt_benefit): Likewise.
1460 (compute_inline_parameters): Likewise.
1461 (estimate_function_body_sizes): Likewise.
1462 (compute_bb_predicates): Likewise.
1463 (initialize_inline_failed): Likewise.
1464 (dump_inline_summary): Likewise.
1465 (dump_inline_edge_summary): Likewise.
1466 * ipa-inline-transform.c (inline_transform): Likewise.
1467 (preserve_function_body_p): Likewise.
1468 (save_inline_function_body): Likewise.
1469 (inline_call): Likewise.
1470 (clone_inlined_nodes): Likewise.
1471 (can_remove_node_now_p): Likewise.
1472 (can_remove_node_now_p_1): Likewise.
1473 * ipa-inline.c (early_inliner): Likewise.
1474 (early_inline_small_functions): Likewise.
1475 (inline_always_inline_functions): Likewise.
1476 (ipa_inline): Likewise.
1477 (flatten_function): Likewise.
1478 (inline_small_functions): Likewise.
1479 (speculation_useful_p): Likewise.
1480 (recursive_inlining): Likewise.
1481 (update_caller_keys): Likewise.
1482 (reset_edge_caches): Likewise.
1483 (update_edge_key): Likewise.
1484 (edge_badness): Likewise.
1485 (relative_time_benefit): Likewise.
1486 (want_inline_self_recursive_call_p): Likewise.
1487 (want_inline_small_function_p): Likewise.
1488 (want_early_inline_function_p): Likewise.
1489 (num_calls): Likewise.
1490 (can_early_inline_edge_p): Likewise.
1491 (can_inline_edge_p): Likewise.
1492 (report_inline_failed_reason): Likewise.
1493 * ipa-profile.c (ipa_profile): Likewise.
1494 (ipa_propagate_frequency): Likewise.
1495 (ipa_propagate_frequency_1): Likewise.
1496 (ipa_profile_generate_summary): Likewise.
1497 * ipa-prop.c (ipcp_transform_function): Likewise.
1498 (read_replacements_section): Likewise.
1499 (ipa_prop_read_section): Likewise.
1500 (ipa_modify_call_arguments): Likewise.
1501 (ipa_print_node_params): Likewise.
1502 (propagate_controlled_uses): Likewise.
1503 (update_indirect_edges_after_inlining): Likewise.
1504 (remove_described_reference): Likewise.
1505 (ipa_make_edge_direct_to_target): Likewise.
1506 (ipa_analyze_node): Likewise.
1507 (ipa_analyze_params_uses): Likewise.
1508 (ipa_compute_jump_functions): Likewise.
1509 (ipa_get_callee_param_type): Likewise.
1510 (ipa_print_node_jump_functions): Likewise.
1511 (ipa_initialize_node_params): Likewise.
1512 (ipa_populate_param_decls): Likewise.
1513 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
1514 (write_agg_replacement_chain): Likewise.
1515 (ipa_write_node_info): Likewise.
1516 (ipa_edge_duplication_hook): Likewise.
1517 (try_decrement_rdesc_refcount): Likewise.
1518 * ipa-pure-const.c (propagate_nothrow): Likewise.
1519 (propagate_pure_const): Likewise.
1520 (pure_const_read_summary): Likewise.
1521 (pure_const_write_summary): Likewise.
1522 (analyze_function): Likewise.
1523 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
1524 (ipa_ref_referring_ref_list): Likewise.
1525 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
1526 (ipa_remove_stmt_references): Likewise.
1527 (ipa_find_reference): Likewise.
1528 (ipa_dump_referring): Likewise.
1529 (ipa_dump_references): Likewise.
1530 (ipa_record_reference): Likewise.
1531 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
1532 (ipa_reference_write_optimization_summary): Likewise.
1533 (write_node_summary_p): Likewise.
1534 (propagate): Likewise.
1535 (read_write_all_from_decl): Likewise.
1536 (generate_summary): Likewise.
1537 (analyze_function): Likewise.
1538 (propagate_bits): Likewise.
1539 (ipa_reference_get_not_written_global): Likewise.
1540 (ipa_reference_get_not_read_global): Likewise.
1541 * ipa-split.c (execute_split_functions): Likewise.
1542 (split_function): Likewise.
1543 * ipa-utils.c (ipa_merge_profiles): Likewise.
1544 (dump_cgraph_node_set): Likewise.
1545 (ipa_reverse_postorder): Likewise.
1546 (ipa_edge_within_scc): Likewise.
1547 (ipa_get_nodes_in_cycle): Likewise.
1548 (ipa_free_postorder_info): Likewise.
1549 (ipa_reduced_postorder): Likewise.
1550 (searchc): Likewise.
1551 (recursive_call_p): Likewise.
1552 * ipa.c (ipa_cdtor_merge): Likewise.
1553 (record_cdtor_fn): Likewise.
1554 (function_and_variable_visibility): Likewise.
1555 (varpool_externally_visible_p): Likewise.
1556 (cgraph_externally_visible_p): Likewise.
1557 (comdat_can_be_unshared_p): Likewise.
1558 (comdat_can_be_unshared_p_1): Likewise.
1559 (address_taken_from_non_vtable_p): Likewise.
1560 (ipa_discover_readonly_nonaddressable_vars): Likewise.
1561 (symtab_remove_unreachable_nodes): Likewise.
1562 (walk_polymorphic_call_targets): Likewise.
1563 (process_references): Likewise.
1564 (enqueue_node): Likewise.
1565 (has_addr_references_p): Likewise.
1566 (cgraph_non_local_node_p_1): Likewise.
1567 * is-a.h (varpool_analyze_node): Likewise.
1568 * lto-cgraph.c (input_symtab): Likewise.
1569 (merge_profile_summaries): Likewise.
1570 (input_cgraph_1): Likewise.
1571 (input_edge): Likewise.
1572 (input_varpool_node): Likewise.
1573 (input_node): Likewise.
1574 (input_overwrite_node): Likewise.
1575 (compute_ltrans_boundary): Likewise.
1576 (output_refs): Likewise.
1577 (lto_output_varpool_node): Likewise.
1578 (lto_output_node): Likewise.
1579 (reachable_from_other_partition_p): Likewise.
1580 (referenced_from_other_partition_p): Likewise.
1581 (lto_output_edge): Likewise.
1582 (output_node_opt_summary): Likewise.
1583 (add_node_to): Likewise.
1584 (reachable_from_this_partition_p): Likewise.
1585 (lto_set_symtab_encoder_in_partition): Likewise.
1586 (lto_symtab_encoder_in_partition_p): Likewise.
1587 (lto_set_symtab_encoder_encode_initializer): Likewise.
1588 (lto_symtab_encoder_encode_initializer_p): Likewise.
1589 (lto_set_symtab_encoder_encode_body): Likewise.
1590 (lto_symtab_encoder_encode_body_p): Likewise.
1591 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1593 * lto-streamer-in.c (lto_read_body): Likewise.
1594 (fixup_call_stmt_edges): Likewise.
1595 (fixup_call_stmt_edges_1): Likewise.
1596 * lto-streamer-out.c (produce_symtab): Likewise.
1597 (output_symbol_p): Likewise.
1598 (write_symbol): Likewise.
1599 (lto_output): Likewise.
1600 (copy_function): Likewise.
1601 (output_function): Likewise.
1602 * passes.c (function_called_by_processed_nodes_p): Likewise.
1603 (ipa_write_optimization_summaries): Likewise.
1604 (ipa_write_summaries): Likewise.
1605 (do_per_function_toporder): Likewise.
1606 (do_per_function): Likewise.
1607 (dump_passes): Likewise.
1608 * symtab.c (symtab_semantically_equivalent_p): Likewise.
1609 (symtab_nonoverwritable_alias): Likewise.
1610 (symtab_nonoverwritable_alias_1): Likewise.
1611 (symtab_for_node_and_aliases): Likewise.
1612 (symtab_resolve_alias): Likewise.
1613 (fixup_same_cpp_alias_visibility): Likewise.
1614 (symtab_alias_ultimate_target): Likewise.
1615 (symtab_used_from_object_file_p): Likewise.
1616 (verify_symtab_base): Likewise.
1617 (dump_symtab_base): Likewise.
1618 (symtab_node_name): Likewise.
1619 (symtab_node_asm_name): Likewise.
1620 (symtab_dissolve_same_comdat_group_list): Likewise.
1621 (symtab_add_to_same_comdat_group): Likewise.
1622 (symtab_unregister_node): Likewise.
1623 (symtab_insert_node_to_hashtable): Likewise.
1624 (symtab_register_node): Likewise.
1625 (unlink_from_assembler_name_hash): Likewise.
1626 (insert_to_assembler_name_hash): Likewise.
1627 (eq_assembler_name): Likewise.
1628 (hash_node_by_assembler_name): Likewise.
1629 (eq_node): Likewise.
1630 (hash_node): Likewise.
1631 * toplev.c (wrapup_global_declaration_2): Likewise.
1632 * trans-mem.c (ipa_tm_execute): Likewise.
1633 (ipa_tm_transform_clone): Likewise.
1634 (ipa_tm_transform_transaction): Likewise.
1635 (ipa_tm_transform_calls_redirect): Likewise.
1636 (ipa_tm_insert_gettmclone_call): Likewise.
1637 (ipa_tm_insert_irr_call): Likewise.
1638 (ipa_tm_create_version): Likewise.
1639 (ipa_tm_create_version_alias): Likewise.
1640 (ipa_tm_mark_forced_by_abi_node): Likewise.
1641 (ipa_tm_mark_force_output_node): Likewise.
1642 (ipa_tm_diagnose_tm_safe): Likewise.
1643 (ipa_tm_mayenterirr_function): Likewise.
1644 (ipa_tm_scan_irr_function): Likewise.
1645 (ipa_tm_note_irrevocable): Likewise.
1646 (ipa_tm_scan_calls_clone): Likewise.
1647 (get_cg_data): Likewise.
1648 * tree-eh.c (tree_could_trap_p): Likewise.
1649 * tree-emutls.c (ipa_lower_emutls): Likewise.
1650 (create_emultls_var): Likewise.
1651 (lower_emutls_function_body): Likewise.
1652 (gen_emutls_addr): Likewise.
1653 (emutls_decl): Likewise.
1654 (new_emutls_decl): Likewise.
1655 * tree-inline.c (tree_function_versioning): Likewise.
1656 (optimize_inline_calls): Likewise.
1657 (expand_call_inline): Likewise.
1658 (estimate_num_insns): Likewise.
1659 (copy_bb): Likewise.
1660 (delete_unreachable_blocks_update_callgraph): Likewise.
1661 * tree-nested.c (gimplify_all_functions): Likewise.
1662 (create_nesting_tree): Likewise.
1663 (check_for_nested_with_variably_modified): Likewise.
1664 * tree-pretty-print.c (dump_function_header): Likewise.
1665 * tree-profile.c (tree_profiling): Likewise.
1666 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
1667 (modify_function): Likewise.
1668 (convert_callers): Likewise.
1669 (convert_callers_for_node): Likewise.
1670 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
1671 (associate_varinfo_to_alias): Likewise.
1672 (create_variable_info_for): Likewise.
1673 (get_constraint_for_ssa_var): Likewise.
1674 * tree-vectorizer.c (increase_alignment): Likewise.
1675 * tree.c (find_decls_types_in_var): Likewise.
1676 (find_decls_types_in_node): Likewise.
1677 (free_lang_data_in_decl): Likewise.
1678 * value-prof.c (gimple_ic_transform): Likewise.
1679 (gimple_ic): Likewise.
1680 (check_ic_target): Likewise.
1681 (init_node_map): Likewise.
1682 * varasm.c (decl_binds_to_current_def_p): Likewise.
1683 (default_binds_local_p_1): Likewise.
1684 (dump_tm_clone_pairs): Likewise.
1685 (assemble_alias): Likewise.
1686 (find_decl): Likewise.
1687 (mark_decl_referenced): Likewise.
1688 * varpool.c (varpool_for_node_and_aliases): Likewise.
1689 (varpool_extra_name_alias): Likewise.
1690 (varpool_create_variable_alias): Likewise.
1691 (add_new_static_var): Likewise.
1692 (varpool_finalize_named_section_flags): Likewise.
1693 (varpool_remove_unreferenced_decls): Likewise.
1694 (enqueue_node): Likewise.
1695 (varpool_assemble_decl): Likewise.
1696 (assemble_aliases): Likewise.
1697 (varpool_analyze_node): Likewise.
1698 (cgraph_variable_initializer_availability): Likewise.
1699 (varpool_add_new_variable): Likewise.
1700 (ctor_for_folding): Likewise.
1701 (dump_varpool_node): Likewise.
1702 (varpool_remove_initializer): Likewise.
1703 (varpool_remove_node): Likewise.
1704 (varpool_node_for_decl): Likewise.
1705 (varpool_create_empty_node): Likewise.
1706 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
1707 (ix86_get_function_versions_dispatcher): Likewise.
1709 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1711 * cgraph.h (symtab_node_base): Convert to a class;
1712 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
1713 chain_next/prev from symtab_node_def.
1714 (cgraph_node): Inherit from symtab_node; add GTY option
1715 tag ("SYMTAB_FUNCTION").
1716 (varpool_node): Inherit from symtab_node; add GTY option
1717 tag ("SYMTAB_VARIABLE").
1718 (symtab_node_def): Remove.
1719 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
1720 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
1721 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
1722 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
1724 * ipa-ref.h (symtab_node_def): Drop.
1725 (symtab_node): Change underlying type from symtab_node_def to
1727 (const_symtab_node): Likwise.
1729 * is-a.h: Update examples in comment.
1731 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
1732 (assembler_name_hash): Likewise.
1734 2013-10-29 Martin Liska <marxin.liska@gmail.com>
1736 * doc/tree-ssa.texi (gimple_phi_result): Document.
1737 (gimple_phi_num_args, gimple_phi_arg): Likewise.
1738 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
1739 (PHI_RESULT, PHI_NUM_ARGS): Remove.
1740 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
1742 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1744 * expr.h: Revert change and include tree-core.h.
1745 * rtl.h: Revert change and don't include tree-core.h.
1747 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1749 * config/darwin.c: Include gimple.h.
1750 * config/i386/winnt.c: Likewise.
1752 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
1754 PR tree-optimization/19831
1755 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
1757 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1759 * tree-outof-ssa.h: Remove include files.
1760 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
1762 * tree-ssa-coalesce.c: Likewise.
1763 * cfgexpand.c: Likewise.
1764 * tree-ssa-ter.c: Likewise.
1765 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
1766 * lto-streamer.h: Likewise.
1767 * cgraphbuild.c: Add gimple.h to include list.
1768 * data-streamer-in.c: Likewise.
1769 * ipa-cp.c: Likewise.
1770 * tree-streamer.c: Likewise.
1771 * lto-compress.c: Likewise.
1772 * ipa-reference.c: Likewise.
1773 * data-streamer-out.c: Likewise.
1774 * lto-cgraph.c: Likewise.
1775 * cgraphclones.c: Likewise.
1776 * ipa-utils.c: Likewise.
1777 * data-streamer.c: Likewise.
1778 * ipa-split.c: Likewise.
1779 * lto-section-in.c: Likewise.
1780 * tree-streamer-out.c: Likewise.
1781 * ipa-prop.c: Likewise.
1782 * tree-streamer-in.c: Likewise.
1783 * symtab.c: Likewise.
1784 * opts-global.c: Likewise.
1785 * lto-opts.c: Likewise.
1786 * lto-section-out.c: Likewise.
1787 * lto-streamer.c: Likewise.
1788 * rtl.h: Add tree-core.h to include list.
1789 * expr.h: Remove tree-core.h from include list.
1790 * gimple.h: Likewise.
1791 * ipa-utils.h: Likewise.
1792 * streamer-hooks.h: Likewise.
1793 * streamer-hooks.c: Include input.h.
1795 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1797 * config/arm/arm.c (cortexa7_extra_costs): New table.
1798 (arm_cortex_a7_tune): New.
1799 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
1801 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
1803 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
1805 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1807 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
1808 to use autogenerated markers for a class-hierarchy, every class
1809 must have a GTY marker.
1810 * gengtype.h (struct type): Add linked list of subclasses to
1811 the "s" member of the union.
1812 (add_subclass): New decl.
1813 * gengtype-state.c (read_state_struct_type): Set up subclass
1815 * gengtype.c (get_ultimate_base_class): New.
1816 (add_subclass): New.
1817 (new_structure): Set up subclass linked list.
1818 (set_gc_used_type): Propagate usage information to subclasses.
1819 (output_mangled_typename): Use get_ultimate_base_class.
1820 (walk_subclasses): Use the subclass linked list, avoiding an
1821 O(N^2) when writing out all types.
1822 (walk_type): Issue an error if the base class is missing a tag,
1823 rather than generating bogus C code. Add a gcc_unreachable
1824 default case, in case people omit tags from concrete subclasses,
1825 or get the values wrong.
1826 (write_func_for_structure): Issue an error for subclasses for
1827 which the base doesn't have a "desc", since otherwise the
1828 autogenerated routines for the base would silently fail to visit
1829 any subclass fields.
1830 (write_root): Use get_ultimate_base_class, tweaking constness of
1831 tp to match that function's signature.
1833 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1835 * doc/gty.texi (GTY Options): Add note about inheritance to
1836 description of desc and tag.
1837 (Inheritance and GTY): New.
1839 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1841 * gengtype-parse.c (opts_have): Drop "static" so that
1842 we can use this from gengtype.c.
1843 * gengtype.c (set_gc_used_type): Mark any base class as used;
1844 update field traversal to visit inherited fields.
1845 (output_mangled_typename): Convert references to classes within
1846 an inheritance hierarchy to reference the ultimate base class,
1847 since only it will have gt_ functions.
1848 (get_string_option): New.
1849 (walk_subclasses): New.
1850 (walk_type): Treat GTY structs that have a "desc" as being the
1851 root of an inheritance hierarchy. Generate a switch on it
1852 within the marking function which walks all subclasses, adding
1853 cases for them via walk_subclasses. For subclasses, visit all
1854 fields of the type (including inherited ones).
1855 (write_func_for_structure): Don't write fns for subclasses, only
1856 for the ultimate base class within an inheritance hierarchy.
1857 Subclasses-marking will be handled by the base class marking functions.
1858 (write_types): Likewise.
1859 (write_local_func_for_structure): Likewise.
1860 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
1861 a "tag" option (and are thus concrete subclasses).
1862 (write_root): Use the marker function for the ultimate base class.
1863 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
1864 (opts_have): Add declaration.
1866 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
1868 * lra-spills.c (lra_final_code_change): Remove useless move insns
1869 originated from moves of pseudos.
1871 2013-10-28 Jeff Law <law@redhat.com>
1873 * lower-subreg.c (resolve_simple_move): Fix comment typo.
1875 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
1877 * df-scan.c (df_collection_rec): Adjust.
1878 (copy_defs): New constant.
1879 (copy_uses): Likewise.
1880 (copy_eq_uses): Likewise.
1881 (copy_mw): Likewise.
1882 (copy_all): Likewise.
1883 (df_insn_rescan): Adjust.
1884 (df_notes_rescan): Likewise.
1885 (df_swap_refs): Likewise.
1886 (df_sort_and_compress_refs): Likewise.
1887 (df_sort_and_compress_mws): Likewise.
1888 (df_install_refs): Likewise.
1889 (df_install_mws): Likewise.
1890 (df_refs_add_to_chains): Add flags parameter controlling which vectors
1892 (df_bb_refs_record): Adjust.
1893 (df_record_entry_block_defs): Likewise.
1894 (df_record_exit_block_defs): Likewise.
1895 (df_refs_verify): Likewise.
1896 (df_mws_verify): Likewise.
1897 (df_insn_refs_verify): Likewise.
1898 (df_bb_verify): Likewise.
1899 * ipa-pure-const.c (finish_state): Remove.
1900 (propagate): Adjust.
1901 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
1902 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
1903 var-tracking.c: Adjust.
1904 * vec.c (stack_vecs): Remove.
1905 (register_stack_vec): Likewise.
1906 (stack_vec_register_index): Likewise.
1907 (unregister_stack_vec): Likewise.
1908 * vec.h (struct va_stack): Remove.
1909 (struct vec<T, A, vl_ptr>): Specialize as
1910 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
1911 allocation strategy compatable with the vl_ptr layout.
1912 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
1913 specialization anyway.
1914 (class stack_vec): New class.
1915 (vec_stack_alloc): Remove.
1916 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
1918 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1919 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1920 Sergey Lega <sergey.s.lega@intel.com>
1921 Anna Tikhonova <anna.tikhonova@intel.com>
1922 Ilya Tocar <ilya.tocar@intel.com>
1923 Andrey Turetskiy <andrey.turetskiy@intel.com>
1924 Ilya Verbin <ilya.verbin@intel.com>
1925 Kirill Yukhin <kirill.yukhin@intel.com>
1926 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1928 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
1929 (*prefetch_avx512pf_<mode>): New.
1930 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
1931 (avx512f_maskcmp<mode>3): Ditto.
1932 (vashrv16si3): Ditto.
1934 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1935 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1936 Sergey Lega <sergey.s.lega@intel.com>
1937 Anna Tikhonova <anna.tikhonova@intel.com>
1938 Ilya Tocar <ilya.tocar@intel.com>
1939 Andrey Turetskiy <andrey.turetskiy@intel.com>
1940 Ilya Verbin <ilya.verbin@intel.com>
1941 Kirill Yukhin <kirill.yukhin@intel.com>
1942 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1944 * config/i386/i386.md (any_truncate): New.
1945 (trunsuffix): Ditto.
1946 * config/i386/predicates.md (const_8_to_9_operand): New.
1947 (const_10_to_11_operand): Ditto.
1948 (const_12_to_13_operand): Ditto.
1949 (const_14_to_15_operand): Ditto.
1950 (const_16_to_19_operand): Ditto.
1951 (const_20_to_23_operand): Ditto.
1952 (const_24_to_27_operand): Ditto.
1953 (const_28_to_31_operand): Ditto.
1954 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
1955 (cvtusi2<ssescalarmodesuffix>32): New.
1956 (cvtusi2<ssescalarmodesuffix>64): Ditto.
1957 (ufloatv16siv16sf2): Ditto.
1958 (avx512f_fix_notruncv16sfv16si): Ditto.
1959 (avx512f_ufix_notruncv16sfv16si): Ditto.
1960 (avx512f_vcvtss2usi): Ditto.
1961 (avx512f_vcvtss2usiq): Ditto.
1962 (avx512f_vcvttss2usi): Ditto.
1963 (avx512f_vcvttss2usiq): Ditto.
1964 (avx512f_vcvtsd2usi): Ditto.
1965 (avx512f_vcvtsd2usiq): Ditto.
1966 (avx512f_vcvttsd2usi): Ditto.
1967 (avx512f_vcvttsd2usiq): Ditto.
1968 (ufloatv8siv8df): Ditto.
1969 (avx512f_cvtdq2pd512_2): Ditto.
1970 (avx512f_cvtpd2dq512): Ditto.
1971 (avx512f_ufix_notruncv8dfv8si): Ditto.
1972 (avx512f_cvtpd2ps512): Ditto.
1973 (vec_unpacks_lo_v16sf): Ditto.
1974 (vec_unpacks_hi_v16sf): Ditto.
1975 (vec_unpacks_float_hi_v16si): Ditto.
1976 (vec_unpacks_float_lo_v16si): Ditto.
1977 (avx512f_unpckhps512): Ditto.
1978 (avx512f_unpcklps512): Ditto.
1979 (avx512f_movshdup512): Ditto.
1980 (avx512f_movsldup512): Ditto.
1981 (vec_extract_lo_v32hi): Ditto.
1982 (vec_extract_hi_v32hi): Ditto.
1983 (vec_extract_lo_v64qi): Ditto.
1984 (vec_extract_hi_v64qi): Ditto.
1985 (avx512f_unpckhpd512): Ditto.
1986 (avx512f_movddup512): Ditto.
1987 (avx512f_unpcklpd512): Ditto.
1988 (*avx512f_unpcklpd512): Ditto.
1989 (avx512f_shufps512_1): Ditto.
1990 (avx512f_shufpd512_1): Ditto.
1991 (avx512f_interleave_highv8di): Ditto.
1992 (avx512f_interleave_lowv8di): Ditto.
1993 (PMOV_DST_MODE): Ditto.
1994 (pmov_src_mode): Ditto.
1995 (pmov_src_lower): Ditto.
1997 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
1998 (*avx512f_<code>v8div16qi2): Ditto.
1999 (*avx512f_<code>v8div16qi2_store): Ditto.
2000 (vec_widen_umult_even_v16si): Ditto.
2001 (*vec_widen_umult_even_v16si): Ditto.
2002 (vec_widen_smult_even_v16si): Ditto.
2003 (*vec_widen_smult_even_v16si): Ditto.
2004 (avx512f_interleave_highv16si): Ditto.
2005 (avx512f_interleave_lowv16si): Ditto.
2006 (avx512f_<code>v16qiv16si2): Ditto.
2007 (avx512f_<code>v16hiv16si2): Ditto.
2008 (avx512f_<code>v8qiv8di2): Ditto.
2009 (avx512f_<code>v8hiv8di2): Ditto.
2010 (avx512f_<code>v8siv8di2): Ditto.
2011 (avx512cd_maskb_vec_dupv8di): Ditto.
2012 (avx512cd_maskw_vec_dupv16si): Ditto.
2013 (avx512f_vcvtph2ps512): Ditto.
2014 (avx512f_vcvtps2ph512): Ditto.
2015 (VEC_EXTRACT_MODE): Extened with wider modes.
2016 (VEC_PERM_AVX2): Ditto.
2017 (VEC_PERM_CONST): Ditto.
2019 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
2021 * config/arc/arc.c (arc_ccfsm_post_advance):
2022 Add comment about TYPE_RETURN.
2024 2013-10-28 Bin Cheng <bin.cheng@arm.com>
2026 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
2027 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
2028 (strip_offset): Convert offset to unsigned number.
2030 2013-10-27 Tom de Vries <tom@codesourcery.com>
2032 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
2033 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
2035 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
2037 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
2038 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
2039 (TARGET_BIG_ENDIAN): New macro.
2040 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
2041 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
2042 * config/sh/sh.c: Likewise.
2043 * config/sh/sh.h: Likewise.
2045 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
2047 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
2048 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
2050 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2052 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
2054 (cmpstr_t, cmpstrsi): Fix formatting.
2056 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2059 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
2060 addressing, do not use general_operand for memory operands.
2062 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
2065 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2066 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2068 2013-10-26 Jeff Law <law@redhat.com>
2070 * predict.c (PRED_MUDFLAP): Remove.
2071 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
2073 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
2074 (OBJS): Remove tree-nomudflap.o
2075 (GTFILES): Remove tree-mudflap.c
2076 * builtins.c (expand_builtin_alloc): Remove mudflap support.
2077 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2078 (mfwrap_spec, mflib_spec): Likewise.
2079 (cpp_unique_options, cc1_options, static_specs): Likewise.
2080 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
2081 * passes.def: Likewise.
2082 * toplev.c (compile_file, process_options): Likewise.
2083 * tree-inline.c (copy_tree_r): Likewise.
2084 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
2085 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
2086 (build_constant_desc, output_constant_def_contents): Likewise.
2087 (categorize_decl_for_section): Likewise.
2088 * tree-mudflap.c: Removed.
2089 * tree-mudflap.h: Removed.
2090 * tree-nomudflap.c: Removed.
2091 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
2092 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
2093 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2094 * config/sol2.h (MFLIB_SPEC): Likewise.
2095 * doc/install.texi: Remove mudflap references.
2096 * doc/passes.texi: Similarly.
2097 * doc/sourcebuild.texi: Similarly.
2098 * doc/invoke.texi: Remove mudlfap related options.
2100 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2102 PR rtl-optimization/58759
2103 * lra-constraints.c (lra_constraints): Remove wrong condition to
2104 remove insn setting up an equivalent pseudo.
2106 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2108 * config/rs6000/rs6000-protos.h
2109 (rs6000_secondary_memory_needed_mode): New prototype.
2110 * config/rs6000/rs6000.c: Include ira.h.
2111 (TARGET_LRA_P): Redefine.
2112 (rs6000_legitimate_offset_address_p): Call
2113 legitimate_constant_pool_address_p in strict mode for LRA.
2114 (rs6000_legitimate_address_p): Ditto.
2115 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
2116 (rs6000_emit_move): Add LRA version of code to generate load/store
2118 (rs6000_secondary_memory_needed_mode): New.
2119 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
2120 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
2121 constants, memory, and FP registers.
2122 (rs6000_lra_p): New.
2123 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
2124 * config/rs6000/rs6000.opt (mlra): New option.
2125 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2127 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
2129 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
2130 has_single_use () and not do the conversion if has_single_use ()
2131 returns false for the multiplication result.
2133 2013-10-25 David Malcolm <dmalcolm@redhat.com>
2135 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
2136 to "NODE", since this works on a "tree", not an
2138 (CONSTANT_CLASS_P): Likewise.
2141 (INDIRECT_REF_P): Likewise.
2142 (REFERENCE_CLASS_P): Likewise.
2143 (COMPARISON_CLASS_P): Likewise.
2144 (UNARY_CLASS_P): Likewise.
2145 (BINARY_CLASS_P): Likewise.
2146 (STATEMENT_CLASS_P): Likewise.
2147 (VL_EXP_CLASS_P): Likewise.
2148 (EXPRESSION_CLASS_P): Likewise.
2149 (IS_TYPE_OR_DECL_P): Likewise.
2151 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
2153 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
2154 ADDR_EXPR in the defining statement.
2156 2013-10-25 Richard Biener <rguenther@suse.de>
2158 PR tree-optimization/58626
2159 * tree-loop-distribution.c (enum rdg_dep_type): Remove
2160 anti_dd, output_dd and input_dd.
2161 (struct rdg_edge): Remove level and relation members.
2162 (RDGE_LEVEL, RDGE_RELATION): Remove.
2163 (dot_rdg_1): Adjust.
2164 (create_rdg_edge_for_ddr): Remove.
2165 (create_rdg_edges_for_scalar): Adjust.
2166 (create_edge_for_control_dependence): Likewise.
2167 (create_rdg_edges): Split into ...
2168 (create_rdg_flow_edges): ... this
2169 (create_rdg_cd_edges): ... and this.
2171 (build_rdg): Likewise, do not compute data dependences or
2173 (pg_add_dependence_edges): New function.
2175 (distribute_loop): First apply all non-dependence based
2176 partition mergings. Then compute dependences between partitions
2177 and merge and order partitions according to them.
2179 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2181 PR rtl-optimization/58831
2182 * alias.c (init_alias_analysis): At the beginning of each iteration,
2183 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
2185 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2187 * recog.c (search_ofs): New static variable moved from...
2188 (peep2_find_free_register): ...here.
2189 (peephole2_optimize): Initialize it.
2191 2013-10-25 Tobias Burnus <burnus@net-b.de>
2193 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
2195 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
2197 * config/i386/i386.h (TARGET_MPX): New define.
2198 (TARGET_MPX_P): Ditto.
2200 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2202 * config/i386/constraints.md (B): New.
2205 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
2206 * config/i386/i386-modes.def (BND32): New.
2208 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
2209 * config/i386/i386.c (isa_opts): Add mmpx.
2210 (regclass_map): Add bound registers.
2211 (dbx_register_map): Likewise.
2212 (dbx64_register_map): Likewise.
2213 (svr4_dbx_register_map): Likewise.
2215 (ix86_option_override_internal): Support MPX ISA.
2216 (ix86_conditional_register_usage): Support bound registers.
2217 (print_reg): Likewise.
2218 (ix86_code_end): Add MPX bnd prefix.
2219 (output_set_got): Likewise.
2220 (ix86_output_call_insn): Likewise.
2221 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
2222 (ix86_print_operand_punct_valid_p): Likewise.
2223 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
2225 (ix86_class_likely_spilled_p): Add bound regs support.
2226 (ix86_hard_regno_mode_ok): Likewise.
2227 (x86_order_regs_for_local_alloc): Likewise.
2228 (ix86_bnd_prefixed_insn_p): New.
2229 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
2230 (FIXED_REGISTERS): Add bound registers.
2231 (CALL_USED_REGISTERS): Likewise.
2232 (REG_ALLOC_ORDER): Likewise.
2233 (HARD_REGNO_NREGS): Likewise.
2235 (VALID_BND_REG_MODE): New.
2236 (FIRST_BND_REG): New.
2237 (LAST_BND_REG): New.
2238 (reg_class): Add BND_REGS.
2239 (REG_CLASS_NAMES): Likewise.
2240 (REG_CLASS_CONTENTS): Likewise.
2242 (ANY_BND_REG_P): New.
2244 (HI_REGISTER_NAMES): Add bound registers.
2245 * config/i386/i386.md (UNSPEC_BNDMK): New.
2246 (UNSPEC_BNDMK_ADDR): New.
2247 (UNSPEC_BNDSTX): New.
2248 (UNSPEC_BNDLDX): New.
2249 (UNSPEC_BNDLDX_ADDR): New.
2250 (UNSPEC_BNDCL): New.
2251 (UNSPEC_BNDCU): New.
2252 (UNSPEC_BNDCN): New.
2253 (UNSPEC_MPX_FENCE): New.
2256 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
2257 (length_immediate): Likewise.
2258 (prefix_0f): Likewise.
2260 (prefix_rep): Check for bnd prefix.
2261 (length_nobnd): New.
2262 (length): Use length_nobnd if specified.
2267 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
2270 (simple_return_internal): Likewise.
2271 (simple_return_pop_internal): Likewise.
2272 (*indirect_jump): Add MPX bnd prefix.
2273 (*tablejump_1): Likewise.
2274 (simple_return_internal_long): Likewise.
2275 (simple_return_indirect_internal): Likewise.
2279 (*mov<mode>_internal_mpx): New.
2280 (<mode>_<bndcheck>): New.
2281 (*<mode>_<bndcheck>): New.
2286 * config/i386/predicates.md (lea_address_operand): Rename to...
2287 (address_no_seg_operand): ... this.
2288 (address_mpx_no_base_operand): New.
2289 (address_mpx_no_index_operand): New.
2290 (bnd_mem_operator): New.
2291 * config/i386/i386.opt (mmpx): New.
2292 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
2293 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
2295 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2297 * mode-classes.def (MODE_POINTER_BOUNDS): New.
2298 * tree.def (POINTER_BOUNDS_TYPE): New.
2299 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
2300 (POINTER_BOUNDS_MODE): New.
2301 (make_pointer_bounds_mode): New.
2302 * machmode.h (POINTER_BOUNDS_MODE_P): New.
2303 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
2304 (layout_type): Support POINTER_BOUNDS_TYPE.
2305 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
2306 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
2307 (type_contains_placeholder_1): Likewise.
2308 * tree.h (POINTER_BOUNDS_TYPE_P): New.
2309 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
2310 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
2312 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2314 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
2315 mode of address computation when calling memory_address_addr_space.
2317 2013-08-24 Richard Henderson <rth@twiddle.net>
2320 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
2321 instead of create_convert_operand_to.
2322 (maybe_emit_sync_lock_test_and_set): Likewise.
2323 (expand_atomic_compare_and_swap): Likewise.
2324 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
2326 2013-08-24 Sriraman Tallam <tmsriram@google.com>
2328 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
2329 Change param_is to use the struct and not the pointer to the struct.
2331 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
2333 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
2334 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2335 Move to targhooks.c.
2336 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
2337 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2338 Relocate from gimplify.c.
2339 * targhooks.h: Add 2 prototypes.
2340 * tree.h: Delete 2 prototypes.
2342 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2344 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
2345 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
2347 2013-10-24 Cong Hou <congh@google.com>
2349 * convert.c (convert_to_real): Guard those unsafe math function
2350 convertions with flag_unsafe_math_optimizations. Handle sqrt()
2353 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
2356 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
2358 (cgraph_create_edge): Use the new argument.
2359 (cgraph_create_indirect_edge): Likewise.
2361 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
2363 * config/arc/arc.c (arc_ccfsm_post_advance): Also handle
2365 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
2366 changes statep->state.
2368 2013-10-24 Tobias Burnus <burnus@net-b.de>
2371 * tree-cfg.c (replace_loop_annotate): New function.
2372 (execute_build_cfg): Call it.
2373 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
2374 * internal-fn.c (expand_ANNOTATE): New function.
2375 * internal-fn.def (ANNOTATE): Define as new internal function.
2376 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
2377 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
2378 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
2379 * doc/extend.texi (Pragmas): Document #pragma ivdep.
2380 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
2382 2013-10-17 Ian Bolton <ian.bolton@arm.com>
2383 Marcus Shawcroft <marcus.shawcroft@arm.com>
2385 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
2386 Special case reload SP+C into none GENERAL_REGS.
2388 2013-10-24 Michael Matz <matz@suse.de>
2390 * gengtype.c (is_file_equal): Check that files will be same length.
2392 2013-10-25 Christian Bruel <christian.bruel@st.com>
2394 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
2395 * config/sh/t-sh (sh-mem.o): New rule.
2396 * config/sh/sh-mem.cc (expand_block_move): Moved here.
2397 (sh_expand_cmpstr): New function.
2398 * config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
2399 * config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
2400 * config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
2401 (rotlhi3_8): Rename.
2403 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2405 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
2406 * Makefile.in (CXXDEPMODE): Assign and change users.
2407 (CCDEPMODE): Delete.
2408 * configure: Regenerate.
2410 2013-10-23 David Malcolm <dmalcolm@redhat.com>
2412 * gengtype-parse.c (require_without_advance): New.
2413 (type): For GTY-marked types that are not GTY((user)), parse any
2414 base classes, requiring them to be single-inheritance, and not
2415 be templates. For non-GTY-marked types and GTY((user)),
2416 continue to skip over any C++ inheritance specification.
2417 * gengtype-state.c (state_writer::write_state_struct_type):
2418 Write base class of type (if any).
2419 (read_state_struct_type): Read base class of type (if any).
2420 * gengtype.c (new_structure): Add a "base_class" parameter.
2421 (create_optional_field_): Update for new parameter to new_structure.
2422 (adjust_field_rtx_def): Likewise.
2423 (adjust_field_tree_exp): Likewise.
2424 * gengtype.h (struct type): Add "base_class" field to the s
2426 (new_structure): Add "base" parameter.
2428 2013-10-23 Sriraman Tallam <tmsriram@google.com>
2431 * config/i386/i386.c (ix86_option_override_internal):
2432 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
2433 (ix86_valid_target_attribute_tree):
2434 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
2435 Change TARGET_SSE to TARGET_SSE_P (opts->...)
2437 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2439 * tree-ssa-loop.h: Remove include files.
2440 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2441 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
2442 * graphite-clast-to-gimple.c: Likewise.
2443 * graphite-scop-detection.c: Likewise.
2444 * graphite-sese-to-poly.c: Likewise.
2445 * ipa-inline-analysis.c: Likewise.
2446 * ipa-pure-const.c: Likewise.
2447 * loop-init.c: Likewise.
2448 * passes.c: Likewise.
2449 * predict.c: Likewise.
2450 * tree-cfg.c: Likewise.
2451 * tree-cfgcleanup.c: Likewise.
2452 * tree-chrec.c: Likewise.
2453 * tree-data-ref.c: Likewise.
2454 * tree-loop-distribution.c: Likewise.
2455 * tree-parloops.c: Likewise.
2456 * tree-predcom.c: Likewise.
2457 * tree-scalar-evolution.c: Likewise.
2458 * tree-ssa-address.c: Likewise.
2459 * tree-ssa.c: Likewise.
2460 * tree-ssa-dce.c: Likewise.
2461 * tree-ssa-loop.c: Likewise.
2462 * tree-ssa-loop-im.c: Likewise.
2463 * tree-ssa-loop-ivcanon.c: Likewise.
2464 * tree-ssa-loop-ivopts.c: Likewise.
2465 * tree-ssa-loop-manip.c: Likewise.
2466 * tree-ssa-loop-niter.c: Likewise.
2467 * tree-ssa-loop-prefetch.c: Likewise.
2468 * tree-ssa-loop-unswitch.c: Likewise.
2469 * tree-ssa-reassoc.c: Likewise.
2470 * tree-vect-data-refs.c: Likewise.
2471 * tree-vect-loop.c: Likewise.
2472 * tree-vect-loop-manip.c: Likewise.
2473 * tree-vectorizer.c: Likewise.
2474 * tree-vect-stmts.c: Likewise.
2475 * tree-vrp.c: Likewise.
2477 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2479 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
2481 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2483 PR tree-optimization/58775
2484 PR tree-optimization/58791
2485 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
2486 (insert_stmt_after): Rewritten, don't move the stmt, but really
2488 (get_stmt_uid_with_default): Remove.
2489 (build_and_add_sum): Use insert_stmt_after and
2490 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
2491 (update_range_test): Set uid on stmts added by
2492 force_gimple_operand_gsi. Don't immediately modify statements
2493 in inter-bb optimization, just update oe->op values.
2494 (optimize_range_tests): Return bool whether any changed have been made.
2495 (update_ops): New function.
2496 (struct inter_bb_range_test_entry): New type.
2497 (maybe_optimize_range_tests): Perform statement changes here.
2498 (not_dominated_by, appears_later_in_bb, get_def_stmt,
2499 ensure_ops_are_available): Remove.
2500 (find_insert_point): Rewritten.
2501 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
2502 return LHS of the (new resp. old) stmt. Don't call
2503 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
2504 instead of last, move new stmt at the right place.
2505 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
2506 (negate_value): Likewise. Set uids.
2507 (break_up_subtract_bb): Initialize uids.
2508 (reassociate_bb): Adjust rewrite_expr_tree caller.
2509 (do_reassoc): Don't call renumber_gimple_stmt_uids.
2511 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
2514 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
2515 TARGET_32BIT final condition.
2516 (mulsi3_internal2 and splitter): Same.
2518 2013-10-23 Jeff Law <law@redhat.com>
2520 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
2521 through joiner blocks with abnormal outgoing edges.
2523 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
2524 Add parameter JOINERS, to allow/disallow threading through joiner
2526 (thread_block): New. Call thread_block_1.
2527 (mark_threaded_blocks): Remove code to filter out certain cases
2528 of threading through joiner blocks.
2529 (thread_through_all_blocks): Document how we can have a dangling
2530 edge AUX field and clear it.
2532 2013-10-23 Ian Lance Taylor <iant@google.com>
2534 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
2535 (C++ Dialect Options): Likewise.
2536 (Optimize Options): Likewise.
2538 2013-10-23 Tom de Vries <tom@codesourcery.com>
2540 PR tree-optimization/58805
2541 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
2543 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2545 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
2546 sequence based on get_range_info returned range.
2548 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2550 * tree-ssa.h: Remove all #include's
2551 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2552 * alias.c: Move required includes from tree-ssa.h.
2554 * builtins.c: Likewise.
2555 * calls.c: Likewise.
2556 * cfgexpand.c: Likewise.
2557 * cfghooks.c: Likewise.
2558 * cfgloop.c: Likewise.
2559 * cfgloopmanip.c: Likewise.
2560 * cgraph.c: Likewise.
2561 * cgraphbuild.c: Likewise.
2562 * cgraphclones.c: Likewise.
2563 * cgraphunit.c: Likewise.
2565 * except.c: Likewise.
2567 * final.c: Likewise.
2568 * fold-const.c: Likewise.
2569 * ggc-page.c: Likewise.
2570 * gimple-builder.c: Likewise.
2571 * gimple-fold.c: Likewise.
2572 * gimple-iterator.c: Likewise.
2573 * gimple-low.c: Likewise.
2574 * gimple-pretty-print.c: Likewise.
2575 * gimple-ssa-strength-reduction.c: Likewise.
2576 * gimple-streamer-in.c: Likewise.
2577 * gimple-streamer-out.c: Likewise.
2578 * gimplify.c: Likewise.
2579 * graphite-blocking.c: Likewise.
2580 * graphite-clast-to-gimple.c: Likewise.
2581 * graphite-dependences.c: Likewise.
2582 * graphite-interchange.c: Likewise.
2583 * graphite-optimize-isl.c: Likewise.
2584 * graphite-poly.c: Likewise.
2585 * graphite-scop-detection.c: Likewise.
2586 * graphite-sese-to-poly.c: Likewise.
2587 * graphite.c: Likewise.
2588 * ipa-cp.c: Likewise.
2589 * ipa-inline-analysis.c: Likewise.
2590 * ipa-inline-transform.c: Likewise.
2591 * ipa-inline.c: Likewise.
2592 * ipa-prop.c: Likewise.
2593 * ipa-pure-const.c: Likewise.
2594 * ipa-reference.c: Likewise.
2595 * ipa-split.c: Likewise.
2596 * ipa-utils.c: Likewise.
2597 * loop-init.c: Likewise.
2598 * lto-cgraph.c: Likewise.
2599 * lto-section-in.c: Likewise.
2600 * lto-section-out.c: Likewise.
2601 * lto-streamer-in.c: Likewise.
2602 * lto-streamer-out.c: Likewise.
2603 * lto-streamer.c: Likewise.
2604 * omp-low.c: Likewise.
2605 * passes.c: Likewise.
2606 * predict.c: Likewise.
2607 * print-tree.c: Likewise.
2608 * profile.c: Likewise.
2610 * targhooks.c: Likewise.
2611 * tracer.c: Likewise.
2612 * trans-mem.c: Likewise.
2613 * tree-call-cdce.c: Likewise.
2614 * tree-cfg.c: Likewise.
2615 * tree-cfgcleanup.c: Likewise.
2616 * tree-chrec.c: Likewise.
2617 * tree-complex.c: Likewise.
2618 * tree-data-ref.c: Likewise.
2619 * tree-dfa.c: Likewise.
2620 * tree-eh.c: Likewise.
2621 * tree-emutls.c: Likewise.
2622 * tree-if-conv.c: Likewise.
2623 * tree-inline.c: Likewise.
2624 * tree-into-ssa.c: Likewise.
2625 * tree-loop-distribution.c: Likewise.
2626 * tree-mudflap.c: Likewise.
2627 * tree-nested.c: Likewise.
2628 * tree-nrv.c: Likewise.
2629 * tree-object-size.c: Likewise.
2630 * tree-outof-ssa.c: Likewise.
2631 * tree-parloops.c: Likewise.
2632 * tree-phinodes.c: Likewise.
2633 * tree-predcom.c: Likewise.
2634 * tree-pretty-print.c: Likewise.
2635 * tree-profile.c: Likewise.
2636 * tree-scalar-evolution.c: Likewise.
2637 * tree-sra.c: Likewise.
2638 * tree-ssa-address.c: Likewise.
2639 * tree-ssa-alias.c: Likewise.
2640 * tree-ssa-ccp.c: Likewise.
2641 * tree-ssa-coalesce.c: Likewise.
2642 * tree-ssa-copy.c: Likewise.
2643 * tree-ssa-copyrename.c: Likewise.
2644 * tree-ssa-dce.c: Likewise.
2645 * tree-ssa-dom.c: Likewise.
2646 * tree-ssa-dse.c: Likewise.
2647 * tree-ssa-forwprop.c: Likewise.
2648 * tree-ssa-ifcombine.c: Likewise.
2649 * tree-ssa-live.c: Likewise.
2650 * tree-ssa-loop-ch.c: Likewise.
2651 * tree-ssa-loop-im.c: Likewise.
2652 * tree-ssa-loop-ivcanon.c: Likewise.
2653 * tree-ssa-loop-ivopts.c: Likewise.
2654 * tree-ssa-loop-manip.c: Likewise.
2655 * tree-ssa-loop-niter.c: Likewise.
2656 * tree-ssa-loop-prefetch.c: Likewise.
2657 * tree-ssa-loop-unswitch.c: Likewise.
2658 * tree-ssa-loop.c: Likewise.
2659 * tree-ssa-math-opts.c: Likewise.
2660 * tree-ssa-operands.c: Likewise.
2661 * tree-ssa-phiopt.c: Likewise.
2662 * tree-ssa-phiprop.c: Likewise.
2663 * tree-ssa-pre.c: Likewise.
2664 * tree-ssa-propagate.c: Likewise.
2665 * tree-ssa-reassoc.c: Likewise.
2666 * tree-ssa-sccvn.c: Likewise.
2667 * tree-ssa-sink.c: Likewise.
2668 * tree-ssa-strlen.c: Likewise.
2669 * tree-ssa-structalias.c: Likewise.
2670 * tree-ssa-tail-merge.c: Likewise.
2671 * tree-ssa-ter.c: Likewise.
2672 * tree-ssa-threadedge.c: Likewise.
2673 * tree-ssa-threadupdate.c: Likewise.
2674 * tree-ssa-uncprop.c: Likewise.
2675 * tree-ssa-uninit.c: Likewise.
2676 * tree-ssa.c: Likewise.
2677 * tree-ssanames.c: Likewise.
2678 * tree-stdarg.c: Likewise.
2679 * tree-streamer-in.c: Likewise.
2680 * tree-switch-conversion.c: Likewise.
2681 * tree-tailcall.c: Likewise.
2682 * tree-vect-data-refs.c: Likewise.
2683 * tree-vect-generic.c: Likewise.
2684 * tree-vect-loop-manip.c: Likewise.
2685 * tree-vect-loop.c: Likewise.
2686 * tree-vect-patterns.c: Likewise.
2687 * tree-vect-slp.c: Likewise.
2688 * tree-vect-stmts.c: Likewise.
2689 * tree-vectorizer.c: Likewise.
2690 * tree-vrp.c: Likewise.
2693 * value-prof.c: Likewise.
2694 * var-tracking.c: Likewise.
2695 * varpool.c: Likewise.
2696 * vtable-verify.c: Likewise.
2698 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2700 * config/tilegx/tilegx.c: Include "tree.h".
2702 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2704 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
2705 the info, not after it.
2706 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
2707 dump_ssaname_info call.
2708 (pp_gimple_stmt_1): Adjust caller.
2709 (dump_phi_nodes): Likewise. Don't print "# " here.
2711 2013-10-22 Jan Hubicka <jh@suse.cz>
2713 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
2715 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
2716 * i386.c (expand_small_movmem_or_setmem): New function.
2717 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
2719 (alg_usable_p): Add support for value ranges; cleanup.
2720 (ix86_expand_set_or_movmem): Add support for misaligned moves.
2722 2013-10-22 Sterling Augustine <saugustine@google.com>
2724 * doc/invoke.texi: Document -ggnu-pubnames.
2725 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
2727 * dwarf2out.c: Include gdb/gdb-index.h.
2728 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
2729 debug_generate_pub_sections.
2730 (is_java, output_pubtables, output_pubname): New functions.
2731 (include_pubname_in_output): Handle debug_generate_pub_sections at
2733 (size_of_pubnames): Use new local space_for_flags based on
2734 debug_generate_pub_sections.
2735 (output_pubnames): Unify pubnames and pubtypes output logic.
2736 Genericize comments. Call output_pubname.
2737 (dwarf2out_finish): Move logic to output_pubnames and call it.
2739 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
2742 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
2743 Remove CCCmode handling.
2744 <case LTU>: Return 'c' suffix for CCCmode.
2745 <case GEU>: Return 'nc' suffix for CCCmode.
2746 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
2747 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
2748 (*sub<mode>3_cc_overflow): Ditto.
2749 (*subsi3_zext_cc_overflow): Ditto.
2751 2013-10-22 Steve Ellcey <sellcey@mips.com>
2753 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
2754 (AND (NOT OP1) (NOT OP2)).
2756 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
2758 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2759 meaning of merge-high and merge-low masks for little endian; avoid
2760 use of vector-pack masks for little endian for mismatched modes.
2762 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2764 * config/tilepro/tilepro.c: Include "tree.h".
2766 2013-10-22 Andreas Schwab <schwab@suse.de>
2768 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
2771 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
2773 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
2774 [Ed Smith-Rowland]): New entries.
2775 ([Stephen M. Webb]): Update.
2777 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
2779 * tree-ssa-ter.h: Remove duplicate copy of file contents.
2781 2013-10-21 Marek Polacek <polacek@redhat.com>
2784 * fold-const.c (fold_range_test): Return 0 if the type is not
2787 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
2789 * system.h: Move hwint.h include further down.
2790 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
2791 gcc_checking_asserts.
2792 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
2794 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2796 Fix volatile issues in optimize_bit_field_compare.
2797 * fold-const.c (optimize_bit_field_compare): Bail out if
2798 lvolatilep or rvolatilep.
2800 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2802 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
2803 and get_inner_reference returning different pmode for non-volatile
2804 bit-field members dependent on flag_strict_volatile_bitfields.
2805 * stor-layout.c (layout_decl): Remove special handling of
2806 flag_strict_volatile_bitfields.
2807 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
2808 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
2810 2013-10-21 Paulo Matos <pmatos@broadcom.com>
2812 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
2815 2013-10-21 Jeff Law <law@redhat.com>
2817 * tree-ssa-threadedge.c (thread_through_normal_block): New
2818 argument VISITED. Remove VISISTED as a local variable. When we
2819 have a threadable jump, verify the destination of the jump has not
2821 (thread_across_edge): Allocate VISITED bitmap once at function
2822 scope and use it throughout. Make sure to set appropriate bits in
2823 VISITED for E (start of jump thread path).
2824 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
2825 through a joiner if any edge on the path has a recorded jump thread.
2827 2013-10-21 Ian Lance Taylor <iant@google.com>
2829 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
2830 don't imply that attributes can solve all problems.
2831 (Directory Options): Fix typo.
2833 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2835 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
2836 extend and extend_add.
2838 2013-10-21 Richard Biener <rguenther@suse.de>
2840 PR tree-optimization/58794
2841 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
2842 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
2844 2013-10-21 Richard Biener <rguenther@suse.de>
2847 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
2848 to (T) X for sign-changing conversions (or no conversion).
2850 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
2852 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
2854 2013-10-20 Jan Hubicka <jh@suse.cz>
2856 * config/i386/i386-tune.def: Add comment; organize into categories
2858 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2860 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
2861 argument. Update function comment.
2862 (expand_set_or_movmem_via_rep): New function combining
2863 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
2864 (expand_movmem_via_rep_mov): Remove.
2865 expand_setmem_via_rep_stos): Remove.
2866 (expand_movmem_epilogue): Update calls correspondingly.
2867 (expand_setmem_epilogue_via_loop): Likewise.
2869 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
2870 (expand_set_or_movmem_prologue): New function combining
2871 expand_movmem_prologue and expand_setmem_prologue.
2872 (expand_movmem_prologue): Remove.
2873 (expand_setmem_prologue): Remove.
2874 (expand_set_or_movmem_constant_prologue): New function combining
2875 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
2876 (expand_constant_movmem_prologue): Remove.
2877 (expand_constant_setmem_prologue): Remove.
2878 (promote_duplicated_reg): Allow vector-const0 value.
2879 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
2880 and ix86_expand_setmem.
2881 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
2882 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
2884 2013-10-21 Diego Novillo <dnovillo@google.com>
2886 * asan.c: Include tree.h
2887 * bb-reorder.c: Likewise.
2888 * cfgcleanup.c: Likewise.
2889 * cfgloopmanip.c: Likewise.
2890 * data-streamer-in.c: Likewise.
2891 * data-streamer-out.c: Likewise.
2892 * data-streamer.c: Likewise.
2893 * dwarf2cfi.c: Likewise.
2894 * graphite-blocking.c: Likewise.
2895 * graphite-clast-to-gimple.c: Likewise.
2896 * graphite-dependences.c: Likewise.
2897 * graphite-interchange.c: Likewise.
2898 * graphite-optimize-isl.c: Likewise.
2899 * graphite-poly.c: Likewise.
2900 * graphite-scop-detection.c: Likewise.
2901 * graphite-sese-to-poly.c: Likewise.
2902 * graphite.c: Likewise.
2903 * ipa-devirt.c: Likewise.
2904 * ipa-profile.c: Likewise.
2907 * loop-init.c: Likewise.
2908 * loop-unroll.c: Likewise.
2909 * lower-subreg.c: Likewise.
2910 * lto/lto-object.c: Likewise.
2911 * recog.c: Likewise.
2912 * reginfo.c: Likewise.
2913 * tree-loop-distribution.c: Likewise.
2914 * tree-parloops.c: Likewise.
2915 * tree-ssa-strlen.c: Likewise.
2916 * tree-streamer.c: Likewise.
2917 * value-prof.c: Likewise.
2918 * target-globals.c: Likewise.
2919 * expr.h: Include tree-core.h instead of tree.h.
2920 * gimple.h: Likewise.
2921 * ipa-prop.h: Likewise.
2922 * ipa-utils.h: Likewise.
2923 * lto-streamer.h: Likewise.
2924 * streamer-hooks.h: Likewise.
2925 * ipa-reference.h: Include cgraph.h instead of tree.h.
2926 * cgraph.h: Include basic-block.h instead of tree.h.
2927 * tree-streamer.h: Do not include tree.h.
2928 * genattrtab.c (write_header): Generate inclusion of
2930 * genautomata.c (main): Likewise.
2931 * genemit.c: Likewise.
2932 * genopinit.c: Likewise.
2933 * genoutput.c (output_prologue): Likewise.
2934 * genpeep.c: Likewise.
2936 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2938 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
2940 (vec_unpacku_hi_v8hi): Likewise.
2941 (vec_unpacku_lo_v16qi): Likewise.
2942 (vec_unpacku_lo_v8hi): Likewise.
2944 2013-10-20 Jan Hubicka <jh@suse.cz>
2946 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
2947 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
2948 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
2950 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
2952 * config/mips/mips.h (ISA_HAS_WSBH): Define.
2953 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
2955 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
2957 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
2960 * system.h: Undef m_slot.
2962 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2964 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
2965 all elements for both endian flavors.
2967 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2970 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
2971 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
2972 and SI_REG for 64bit SYSV ABI targets.
2974 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2976 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
2977 to multi_reg_return. Clarify that we are skipping USEs of multiple
2978 return registers. Use bool type where appropriate.
2980 2013-10-18 Jan Hubicka <jh@suse.cz>
2982 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
2984 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
2985 (X86_TUNE_PUSH_MEMORY): Likewise.
2986 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
2987 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
2988 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
2990 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
2991 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
2993 (ix86_option_override_internal): Update to use tune features instead
2996 2013-10-18 Cong Hou <congh@google.com>
2998 PR tree-optimization/58508
2999 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
3000 statement that contains data refs with zero-step.
3002 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3004 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
3005 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
3006 * gimple-low.c (gimple_check_call_arg,
3007 gimple_check_call_matching_types): Move to cgraph.c.
3008 * gimple-low.h: Remove prototype.
3009 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
3010 Relocate from gimple-low.c.
3011 * cgraph.h: Add prototype. Don't include basic-block.h.
3012 * gimplify.c: Add gimple-low to include list.
3013 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
3014 * tree-eh.c: Add gimple-low to include list.
3015 * tree-nested.c: Likewise.
3016 * cfgexpand.c: Add tree-ssa-address.h to include list.
3018 * gimple-fold.c: Likewise.
3019 * gimple-ssa-strength-reduction.c: Likewise.
3020 * trans-mem.c: Likewise.
3021 * tree-mudflap.c: Likewise.
3022 * tree-ssa-loop-ivopts.c: Likewise.
3023 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
3024 (degenerate_phi_result): Move to tree-phinodes.c.
3025 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
3026 * tree-ssa-threadedge.c: Likewise.
3027 * tree-vrp.c: Likewise.
3028 * tree-phinodes.c (degenerate_phi_result): Relocate here.
3029 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
3030 * tree-phinodes.h (degenerate_phi_result): Add prototype.
3031 * tree-ssa-copy.c: Include tree-ssa-dom.h.
3032 * tree-ssa-forwprop.c: Likewise.
3033 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
3034 pass_data_cleanup_cfg_post_optimizing,
3035 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
3036 * tree-optimize.c: Delete File.
3037 * graphite.c: Include tree-cfgcleanup.h.
3038 * passes.c: Likewise.
3039 * tree-cfg.c: Likewise.
3040 * tree-profile.c: Likewise.
3041 * tree-ssa-dse.c: Likewise.
3042 * tree-ssa-loop-ivcanon.c: Likewise.
3043 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
3044 * tree-outof-ssa.c: Include sbitmap.h.
3045 * tree-ssa-live.c: Likewise.
3046 * tree-ssa-propagate.c: Likewise.
3047 * tree-ssa-structalias.c: Likewise.
3048 * tree-stdarg.c: Likewise.
3049 * Makefile.in (OBJS): Delete tree-optimize.o.
3050 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
3051 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
3052 * varasm.c: Include basic-block.h.
3053 * cfgloop.h: Include function.h instead of basic-block.h
3054 (bb_loop_depth): Move to cfgloop.c.
3055 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
3057 2013-10-18 Teresa Johnson <tejohnson@google.com>
3059 * predict.c (probably_never_executed): Compare frequency-based
3060 count to number of training runs.
3061 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
3063 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3065 * config/arm/arm.c (cortexa9_extra_costs): New table.
3066 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
3068 2013-10-18 Jeff Law <law@redhat.com>
3070 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
3072 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
3073 (register_jump_thread): Add "registered_jump_thread" debug
3075 * dbgcnt.def (registered_jump_thread): New debug counter.
3077 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3079 * config/rs6000/rs6000.c: Include cgraph.h.
3081 2013-10-18 Teresa Johnson <tejohnson@google.com>
3083 * tree-ssa-tail-merge.c (replace_block_by): Update edge
3084 weights during merging.
3086 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3088 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
3089 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
3090 * tree-ssa-operands.h: Remove prototype.
3091 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
3092 * gimple.c (virtual_operand_p): Relocate from gimple.c.
3093 * gimple.h: Add prototype.
3094 * gimple-ssa.h: Include tree-ssa-operands.h.
3095 * tree-dump.c: Add tree-cfg.h to include list.
3096 * tree-ssa-alias.c: Add ipa-reference.h to include list.
3097 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
3098 * config/i386/i386.c: Don't include tree-flow.h.
3099 * config/rs6000/rs6000.c: Likewise.
3101 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3103 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
3105 2013-10-18 Richard Biener <rguenther@suse.de>
3107 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
3108 or TYPE_UNSIGNED of integral types.
3109 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
3110 NULL_TREE for zero-precision integral types.
3112 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
3114 * config/aarch64/arm_neon.h
3115 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
3117 2013-10-17 Sriraman Tallam <tmsriram@google.com>
3120 * opth-gen.awk: Define target_flags_explicit.
3122 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3124 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
3125 fields to the reg_addr array that describes the valid addressing
3126 mode for any register, general purpose registers, floating point
3127 registers, and Altivec registers.
3128 (FIRST_RELOAD_REG_CLASS): Likewise.
3129 (LAST_RELOAD_REG_CLASS): Likewise.
3130 (struct reload_reg_map_type): Likewise.
3131 (reload_reg_map_type): Likewise.
3132 (RELOAD_REG_VALID): Likewise.
3133 (RELOAD_REG_MULTIPLE): Likewise.
3134 (RELOAD_REG_INDEXED): Likewise.
3135 (RELOAD_REG_OFFSET): Likewise.
3136 (RELOAD_REG_PRE_INCDEC): Likewise.
3137 (RELOAD_REG_PRE_MODIFY): Likewise.
3138 (reg_addr): Likewise.
3139 (mode_supports_pre_incdec_p): New helper functions to say whether
3140 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
3141 (mode_supports_pre_modify_p): Likewise.
3142 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
3143 print the valid address mode bits for each mode.
3144 (rs6000_debug_print_mode): Likewise.
3145 (rs6000_debug_reg_global): Likewise.
3146 (rs6000_setup_reg_addr_masks): New function to set up the address
3147 mask bits for each type.
3148 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
3149 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
3150 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
3151 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
3152 PRE_MODIFY are supported.
3153 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
3154 registers, instead of {src,dest}_av_p.
3155 (rs6000_print_options_internal): Tweak the debug output slightly.
3157 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3159 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
3162 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
3164 * tree-flow.h (struct omp_region): Move to omp-low.c.
3165 Remove omp_ prototypes and variables.
3166 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
3167 (copy_var_decl): Relocate prototype from tree-flow.h.
3168 * gimple.c (copy_var_decl): Relocate from omp-low.c.
3169 * tree.h: Move prototype to omp-low.h.
3170 * omp-low.h: New File. Relocate prototypes here.
3171 * omp-low.c (struct omp_region): Make local here.
3172 (root_omp_region): Make static.
3173 (copy_var_decl) Move to gimple.c.
3174 (new_omp_region): Make static.
3175 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
3176 * tree-cfg.c: Include omp-low.h.
3177 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
3178 * gimplify.c: Include omp-low.h.
3179 * tree-parloops.c: Likewise.
3181 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3183 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
3184 operands are in memory, prefer to force non-matched operand 1 to
3187 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3190 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
3191 restrict TImode addresses to single indirect registers if both
3192 -mquad-memory and -mvsx-timode are used.
3193 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
3194 we should emit load/store quad. Remove using %y for quad memory
3197 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
3198 constraints to allow load/store quad on machines where TImode is
3199 not allowed in VSX registers. Use 'n' instead of 'F' constraint
3200 for TImode to load integer constants.
3202 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3204 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
3206 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
3208 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
3209 handling of STACK_REG.
3211 2013-10-17 Richard Biener <rguenther@suse.de>
3213 PR tree-optimization/58143
3214 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
3216 (rewrite_to_defined_overflow): Likewise.
3217 (move_computations_dom_walker::before_dom): Rewrite stmts
3218 with undefined signed overflow that are not always executed
3219 into unsigned arithmetic.
3221 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3224 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
3225 explicit isa flag to be an options variable, instead of using
3226 global_options_set. Remove define from rs6000.h.
3227 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
3229 * config/rs6000/rs6000.c (rs6000_option_override_internal):
3230 Initialize rs6000_isa_flags_explicit.
3231 (rs6000_function_specific_save): Add gcc_options* parameter, so
3232 that the powerpc builds after the 2013-10-15 changes.
3233 (rs6000_function_specific_restore): Likewise.
3235 2013-10-16 DJ Delorie <dj@redhat.com>
3237 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
3238 op is a REG before checking REGNO.
3239 (rl78_alloc_physical_registers): Verify pattern is a SET before
3242 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3244 * config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
3246 (vec_unpacks_lo_v4sf): Likewise.
3247 (vec_unpacks_float_hi_v4si): Likewise.
3248 (vec_unpacks_float_lo_v4si): Likewise.
3249 (vec_unpacku_float_hi_v4si): Likewise.
3250 (vec_unpacku_float_lo_v4si): Likewise.
3252 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3254 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
3255 (vsx_concat_v2sf): Likewise.
3257 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
3259 * config/aarch64/aarch64.md
3260 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
3262 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
3264 PR tree-optimization/58697
3265 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
3266 estimated_loop_iterations_int.
3267 (max_stmt_executions_int): Call get_max_loop_iterations_int.
3268 (get_max_loop_iterations_int): New. HWINT version of
3269 get_max_loop_iterations.
3270 * cfgloop.h: Add prototypes.
3271 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
3272 * loop-unroll.c (decide_peel_once_rolling): Call
3273 get_estimated_loop_iterations_int.
3274 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
3275 * tree-ssa-loop-niter.h: Tweak prototypes.
3277 2013-10-16 David Malcolm <dmalcolm@redhat.com>
3279 * gengtype-parse.c (struct_field_seq): Ignore access-control
3280 keywords ("public:" etc).
3282 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
3284 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
3285 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
3287 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
3289 * config/arm/arm.opt (mlra): New option.
3290 * config/arm/arm.c (arm_lra_p): New function.
3291 (TARGET_LRA_P): Define.
3293 2013-10-16 Paulo Matos <pmatos@broadcom.com>
3295 * tree-core.h (tree_code_name): Remove.
3296 * tree.h (get_tree_code_name): New prototype.
3297 * tree.c (tree_code_name): Make static.
3298 (get_tree_code_name): New function.
3299 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
3300 tree_class_check_failed, tree_range_check_failed,
3301 tree_not_class_check_failed, omp_clause_check_failed,
3302 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
3303 wrapper get_tree_code_name instead of calling tree_code_name directly.
3304 * tree-vrp.c (dump_asserts_for): Likewise.
3305 * tree-dump.c (dequeue_and_dump): Likewise.
3306 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
3307 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
3308 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
3309 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
3310 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
3311 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
3312 dump_gimple_omp_for): Likewise.
3313 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
3314 * tree-ssa-pre.c (print_pre_expr): Likewise.
3315 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
3316 * print-tree.c (print_node_brief, print_node): Likewise.
3317 * gimple.c (gimple_check_failed): Likewise.
3318 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
3319 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
3320 * config/mep/mep.c (mep_validate_vliw): Likewise.
3321 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
3322 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
3324 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
3326 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
3329 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
3331 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
3332 (MULTILIB_MATCHES): Add multilib for -march=v8.
3334 2013-10-15 Sriraman Tallam <tmsriram@google.com>
3337 * optc-save-gen.awk: Add extra parameter to the save and restore
3339 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
3340 * tree.c (build_optimization_node): New parameter. Add extra parameter
3341 to call to cl_optimization_save.
3342 (build_target_option_node): New parameter. Add extra parameter
3343 to call to cl_target_option_save.
3344 * tree.h (build_optimization_node): New parameter.
3345 (build_target_option_node): New parameter.
3346 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
3347 build_optimization_node and build_target_option_node.
3348 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
3349 (handle_pragma_push_options): Ditto.
3350 * toplev.c (process_options): Ditto.
3351 * opts.c (init_options_struct): Check for opts_set non-null.
3352 * target.def (target_option.save): New parameter.
3353 (target_option.restore): New parameter.
3354 * tm.texi: Generate.
3355 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
3356 (ix86_pragma_target_parse): Ditto.
3357 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
3359 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
3360 to build_optimization_node and build_target_option_node.
3361 (rs6000_valid_attribute_p): Ditto.
3362 (rs6000_pragma_target_parse): Ditto.
3363 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
3365 * config/i386/i386.h:
3366 TARGET_64BIT_P: New Macro
3367 TARGET_MMX_P: New Macro.
3368 TARGET_3DNOW_P: New Macro.
3369 TARGET_3DNOW_A_P: New Macro.
3370 TARGET_SSE_P: New Macro.
3371 TARGET_SSE2_P: New Macro.
3372 TARGET_SSE3_P: New Macro.
3373 TARGET_SSSE3_P: New Macro.
3374 TARGET_SSE4_1_P: New Macro.
3375 TARGET_SSE4_2_P: New Macro.
3376 TARGET_AVX_P: New Macro.
3377 TARGET_AVX2_P: New Macro.
3378 TARGET_AVX512F_P: New Macro.
3379 TARGET_AVX512PF_P: New Macro.
3380 TARGET_AVX512ER_P: New Macro.
3381 TARGET_AVX512CD_P: New Macro.
3382 TARGET_FMA_P: New Macro.
3383 TARGET_SSE4A_P: New Macro.
3384 TARGET_FMA4_P: New Macro.
3385 TARGET_XOP_P: New Macro.
3386 TARGET_LWP_P: New Macro.
3387 TARGET_ABM_P: New Macro.
3388 TARGET_BMI_P: New Macro.
3389 TARGET_BMI2_P: New Macro.
3390 TARGET_LZCNT_P: New Macro.
3391 TARGET_TBM_P: New Macro.
3392 TARGET_POPCNT_P: New Macro.
3393 TARGET_SAHF_P: New Macro.
3394 TARGET_MOVBE_P: New Macro.
3395 TARGET_CRC32_P: New Macro.
3396 TARGET_AES_P: New Macro.
3397 TARGET_PCLMUL_P: New Macro.
3398 TARGET_CMPXCHG16B_P: New Macro.
3399 TARGET_FSGSBASE_P: New Macro.
3400 TARGET_RDRND_P: New Macro.
3401 TARGET_F16C_P: New Macro.
3402 TARGET_RTM_P: New Macro.
3403 TARGET_HLE_P: New Macro.
3404 TARGET_RDSEED_P: New Macro.
3405 TARGET_PRFCHW_P: New Macro.
3406 TARGET_ADX_P: New Macro.
3407 TARGET_FXSR_P: New Macro.
3408 TARGET_XSAVE_P: New Macro.
3409 TARGET_XSAVEOPT_P: New Macro.
3410 TARGET_LP64_P: New Macro.
3411 TARGET_X32_P: New Macro.
3412 TARGET_FPMATH_DEFAULT_P: New Macro.
3413 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
3414 * config/i386/i386.c (ix86_option_override_internal): New parameters.
3416 Change ix86_tune_string to access opts->x_ix86_tune_string.
3417 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
3418 Change ix86_arch_string to access opts->x_ix86_arch_string.
3419 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
3420 Change ix86_pmode to access opts->x_ix86_pmode.
3421 Change ix86_abi to access opts->x_ix86_abi.
3422 Change ix86_cmodel to access opts->x_ix86_cmodel.
3423 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
3424 Change ix86_isa_flags_explicit to access
3425 opts->x_ix86_isa_flags_explicit.
3426 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
3427 Change ix86_regparm to access opts->x_ix86_regparm.
3428 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
3429 Change ix86_preferred_stack_boundary_arg to access
3430 opts->x_ix86_preferred_stack_boundary_arg.
3431 Change ix86_force_align_arg_pointer to access
3432 opts->x_ix86_force_align_arg_pointer.
3433 Change ix86_incoming_stack_boundar_arg to access
3434 opts->x_ix86_incoming_stack_boundar_arg.
3435 Change ix86_fpmath to access opts->x_ix86_fpmath.
3436 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
3437 Change ix86_recip_name to access opts->x_ix86_recip_name.
3438 Change ix86_stack_protector_guard to access
3439 opts->x_ix86_stack_protector_guard.
3440 Change ix86_tune_memcpy_strategy to access
3441 opts->x_ix86_tune_memcpy_strategy.
3442 Change ix86_tune_memset_strategy to access
3443 opts->x_ix86_tune_memset_strategy.
3444 Change global_options to access opts.
3445 Change global_options_set to access opts_set.
3446 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
3447 Change TARGET_MMX to TARGET_MMX_P (opts->...).
3448 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
3449 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
3450 Change TARGET_SSE to TARGET_SSE_P (opts->...).
3451 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
3452 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
3453 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
3454 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
3455 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
3456 Change TARGET_AVX to TARGET_AVX_P (opts->...).
3457 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
3458 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
3459 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
3460 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
3461 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
3462 Change TARGET_FMA to TARGET_FMA_P (opts->...).
3463 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
3464 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
3465 Change TARGET_XOP to TARGET_XOP_P (opts->...).
3466 Change TARGET_LWP to TARGET_LWP_P (opts->...).
3467 Change TARGET_ABM to TARGET_ABM_P (opts->...).
3468 Change TARGET_BMI to TARGET_BMI_P (opts->...).
3469 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
3470 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
3471 Change TARGET_TBM to TARGET_TBM_P (opts->...).
3472 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
3473 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
3474 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
3475 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
3476 Change TARGET_AES to TARGET_AES_P (opts->...).
3477 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
3478 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
3479 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
3480 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
3481 Change TARGET_F16C to TARGET_F16C_P (opts->...).
3482 Change TARGET_RTM to TARGET_RTM_P (opts->...).
3483 Change TARGET_HLE to TARGET_HLE_P (opts->...).
3484 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
3485 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
3486 Change TARGET_ADX to TARGET_ADX_P (opts->...).
3487 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
3488 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
3489 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
3490 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
3491 Change TARGET_X32 to TARGET_X32_P (opts->...).
3492 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
3493 Change TARGET_FLOAT_RETURNS_IN_80387 to
3494 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
3495 (ix86_function_specific_save): New parameter. Use opts-> fields
3496 to replace global fields.
3497 (ix86_function_specific_restore): Ditto.
3498 (ix86_valid_target_attribute_inner_p): New parameters.
3500 Fix call to ix86_handle_option and set_option.
3501 (ix86_valid_target_attribute_tree): New parameters.
3502 Change global_options to access opts.
3503 Change global_options_set to access opts_set.
3504 Fix call to ix86_valid_target_attribute_inner_p.
3505 Change ix86_tune_string to access opts->x_ix86_tune_string.
3506 Change ix86_arch_string to access opts->x_ix86_arch_string.
3507 Change ix86_fpmath to access opts->x_ix86_fpmath
3508 Fix call to ix86_option_override_internal.
3509 Fix call to ix86_add_new_builtins.
3510 Fix calls to build_optimization_node and build_target_option_node.
3511 (ix86_valid_target_attribute_p): Remove access to global_options.
3512 Use new gcc_options structure func_options.
3513 Fix call to ix86_valid_target_attribute_tree.
3514 Fix call to build_optimization_node.
3515 (get_builtin_code_for_version): Fix call to
3516 ix86_valid_target_attribute_tree.
3518 2013-10-15 David Malcolm <dmalcolm@redhat.com>
3520 * Makefile.in (PICFLAG): New.
3521 (enable_host_shared): New.
3522 (INTERNAL_CFLAGS): Use PICFLAG.
3523 (LIBIBERTY): Use pic build of libiberty.a if configured with
3524 --enable-host-shared.
3525 * configure.ac: Add --enable-host-shared, setting up new
3527 * configure: Regenerate.
3528 * doc/install.texi (--enable-shared): Add note contrasting it with ...
3529 (--enable-host-shared): New option.
3531 2013-10-15 Richard Biener <rguenther@suse.de>
3533 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
3534 for built-in functions.
3536 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
3538 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
3539 (optimize_range_tests_1): New function,
3540 extracted from optimize_range_tests.
3541 (optimize_range_tests_xor): Similarly.
3542 (optimize_range_tests_diff): New function.
3543 (optimize_range_tests): Use optimize_range_tests_1.
3545 2013-10-15 Cong Hou <congh@google.com>
3547 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
3548 requirement of the reduction pattern so that one operand of the
3549 reduction operation can come from outside of the loop.
3551 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3553 * config/arm/neon-schedgen.ml: Remove.
3554 * config/arm/cortex-a9-neon.md: Remove comment regarding
3557 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3559 * config/arm/types: Remove old neon types.
3561 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3563 * config/arm/cortex-a7.md
3564 (cortex_a7_neon_type): New.
3565 (cortex_a7_neon_mul): Update for new types.
3566 (cortex_a7_neon_mla): Likewise.
3567 (cortex_a7_neon): Likewise.
3569 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3571 * config/arm/cortex-a15-neon.md
3572 (cortex_a15_neon_type): New,
3574 (cortex_a15_neon_int_1): Remove.
3575 (cortex_a15_neon_int_2): Likewise.
3576 (cortex_a15_neon_int_3): Likewise.
3577 (cortex_a15_neon_int_4): Likewise.
3578 (cortex_a15_neon_int_5): Likewise.
3579 (cortex_a15_neon_vqneg_vqabs): Likewise.
3580 (cortex_a15_neon_vmov): Likewise.
3581 (cortex_a15_neon_vaba): Likewise.
3582 (cortex_a15_neon_vaba_qqq): Likewise.
3583 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3584 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3585 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3587 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3588 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3589 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
3590 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3591 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3592 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3593 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3594 (cortex_a15_neon_shift_1): Likewise.
3595 (cortex_a15_neon_shift_2): Likewise.
3596 (cortex_a15_neon_shift_3): Likewise.
3597 (cortex_a15_neon_vshl_ddd): Likewise.
3598 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3599 (cortex_a15_neon_vsra_vrsra): Likewise.
3600 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3601 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3602 (cortex_a15_neon_bp_3cycle): Likewise.
3603 (cortex_a15_neon_ldm_2): Likewise.
3604 (cortex_a15_neon_stm_2): Likewise.
3605 (cortex_a15_neon_mcr): Likewise.
3606 (cortex_a15_neon_mrc): Likewise.
3607 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3608 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3609 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3610 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3611 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3612 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3613 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3614 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3615 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3616 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3617 (cortex_a15_neon_bp_simple): Likewise.
3618 (cortex_a15_neon_bp_2cycle): Likewise.
3619 (cortex_a15_neon_bp_3cycle): Likewise.
3620 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3621 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3622 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3623 (cortex_a15_neon_vld2_4_regs): Likewise.
3624 (cortex_a15_neon_vld3_vld4): Likewise.
3625 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3626 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3627 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
3628 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
3629 (cortex_a15_neon_vst3_vst4): Rename to...
3630 (cortex_a15_neon_vst4): This, update for new attributes.
3631 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
3632 (cortex_a15_neon_vld3_vld4_lane): Likewise.
3633 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3634 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3635 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3636 (cortex_a15_neon_ldm_2): Likewise.
3637 (cortex_a15_neon_stm_2): Likewise.
3638 (cortex_a15_neon_mcr): Likewise.
3639 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3640 (cortex_a15_neon_mrc): Likewise.
3641 (cortex_a15_neon_mrrc): Likewise.
3643 (cortex_a15_neon_abd): New.
3644 (cortex_a15_neon_abd_q): Likewise.
3645 (cortex_a15_neon_aba): Likewise.
3646 (cortex_a15_neon_aba_q): Likewise.
3647 (cortex_a15_neon_acc): Likewise.
3648 (cortex_a15_neon_acc_q): Likewise.
3649 (cortex_a15_neon_arith_basic): Likewise.
3650 (cortex_a15_neon_arith_complex): Likewise.
3651 (cortex_a15_neon_multiply): Likewise.
3652 (cortex_a15_neon_multiply_q): Likewise.
3653 (cortex_a15_neon_mla): Likewise.
3654 (cortex_a15_neon_mla_q): Likewise.
3655 (cortex_a15_neon_sat_mla_long): Likewise.
3656 (cortex_a15_neon_shift_acc): Likewise.
3657 (cortex_a15_neon_shift_imm_basic): Likewise.
3658 (cortex_a15_neon_shift_imm_complex): Likewise.
3659 (cortex_a15_neon_shift_reg_basic): Likewise.
3660 (cortex_a15_neon_shift_reg_basic_q): Likewise.
3661 (cortex_a15_neon_shift_reg_complex): Likewise.
3662 (cortex_a15_neon_shift_reg_complex_q): Likewise.
3663 (cortex_a15_neon_fp_negabs): Likewise
3664 (cortex_a15_neon_fp_arith): Likewise
3665 (cortex_a15_neon_fp_arith_q): Likewise
3666 (cortex_a15_neon_fp_cvt_int): Likewise
3667 (cortex_a15_neon_fp_cvt_int_q): Likewise
3668 (cortex_a15_neon_fp_cvt_16): Likewise
3669 (cortex_a15_neon_fp_mul): Likewise
3670 (cortex_a15_neon_fp_mul_q): Likewise
3671 (cortex_a15_neon_fp_mla): Likewise
3672 (cortex_a15_neon_fp_mla_q): Likewise
3673 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
3674 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
3675 (cortex_a15_neon_bitops): Likewise.
3676 (cortex_a15_neon_bitops_q): Likewise.
3677 (cortex_a15_neon_from_gp): Likewise.
3678 (cortex_a15_neon_from_gp_q): Likewise.
3679 (cortex_a15_neon_tbl3_tbl4): Likewise.
3680 (cortex_a15_neon_zip_q): Likewise.
3681 (cortex_a15_neon_to_gp): Likewise.
3682 (cortex_a15_neon_load_a): Likewise.
3683 (cortex_a15_neon_load_b): Likewise.
3684 (cortex_a15_neon_load_c): Likewise.
3685 (cortex_a15_neon_load_d): Likewise.
3686 (cortex_a15_neon_load_e): Likewise.
3687 (cortex_a15_neon_load_f): Likewise.
3688 (cortex_a15_neon_store_a): Likewise.
3689 (cortex_a15_neon_store_b): Likewise.
3690 (cortex_a15_neon_store_c): Likewise.
3691 (cortex_a15_neon_store_d): Likewise.
3692 (cortex_a15_neon_store_e): Likewise.
3693 (cortex_a15_neon_store_f): Likewise.
3694 (cortex_a15_neon_store_g): Likewise.
3695 (cortex_a15_neon_store_h): Likewise.
3696 (cortex_a15_vfp_to_from_gp): Likewise.
3698 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3700 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
3702 (cortex_a9_neon_vshl_ddd): Remove.
3703 (cortex_a9_neon_vst3_vst4): Likewise.
3704 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3706 (cortex_a9_neon_bit_ops_q): New.
3708 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
3709 (cortex_a9_neon_int_2): Likewise.
3710 (cortex_a9_neon_int_3): Likewise.
3711 (cortex_a9_neon_int_4): Likewise.
3712 (cortex_a9_neon_int_5): Likewise.
3713 (cortex_a9_neon_vqneg_vqabs): Likewise.
3714 (cortex_a9_neon_vmov): Likewise.
3715 (cortex_a9_neon_vaba): Likewise.
3716 (cortex_a9_neon_vaba_qqq): Likewise.
3717 (cortex_a9_neon_shift_1): Likewise.
3718 (cortex_a9_neon_shift_2): Likewise.
3719 (cortex_a9_neon_shift_3): Likewise.
3720 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3721 (cortex_a9_neon_vsra_vrsra): Likewise.
3722 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3723 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3724 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3726 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3727 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3728 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3730 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3731 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3732 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3733 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3734 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3735 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3736 (cortex_a9_neon_fp_vsum): Likewise.
3737 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3738 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3739 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3740 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3741 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3742 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3743 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3744 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3745 (cortex_a9_neon_bp_simple): Likewise.
3746 (cortex_a9_neon_bp_2cycle): Likewise.
3747 (cortex_a9_neon_bp_3cycle): Likewise.
3748 (cortex_a9_neon_ldr): Likewise.
3749 (cortex_a9_neon_str): Likewise.
3750 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3751 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3752 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3753 (cortex_a9_neon_vld2_4_regs): Likewise.
3754 (cortex_a9_neon_vld3_vld4): Likewise.
3755 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3756 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3757 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3758 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3759 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3760 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3761 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3762 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3763 (cortex_a9_neon_mcr): Likewise.
3764 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3765 (cortex_a9_neon_mrc): Likewise.
3766 (cortex_a9_neon_mrrc): Likewise.
3768 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3770 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
3772 (cortex_a8_neon_vshl_ddd): Remove.
3773 (cortex_a8_neon_vst3_vst4): Likewise.
3774 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3776 (cortex_a8_neon_bit_ops_q): New.
3778 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
3779 (cortex_a8_neon_int_2): Likewise..
3780 (cortex_a8_neon_int_3): Likewise.
3781 (cortex_a8_neon_int_5): Likewise.
3782 (cortex_a8_neon_vqneg_vqabs): Likewise.
3783 (cortex_a8_neon_int_4): Likewise.
3784 (cortex_a8_neon_vaba): Likewise.
3785 (cortex_a8_neon_vaba_qqq): Likewise.
3786 (cortex_a8_neon_shift_1): Likewise.
3787 (cortex_a8_neon_shift_2): Likewise.
3788 (cortex_a8_neon_shift_3): Likewise.
3789 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3790 (cortex_a8_neon_vsra_vrsra): Likewise.
3791 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3792 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3793 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3795 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3796 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3797 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3799 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3800 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3801 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3802 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3803 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3804 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3805 (cortex_a8_neon_fp_vsum): Likewise.
3806 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3807 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3808 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3809 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3810 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3811 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3812 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3813 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3814 (cortex_a8_neon_bp_simple): Likewise.
3815 (cortex_a8_neon_bp_2cycle): Likewise.
3816 (cortex_a8_neon_bp_3cycle): Likewise.
3817 (cortex_a8_neon_ldr): Likewise.
3818 (cortex_a8_neon_str): Likewise.
3819 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3820 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3821 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3822 (cortex_a8_neon_vld2_4_regs): Likewise.
3823 (cortex_a8_neon_vld3_vld4): Likewise.
3824 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3825 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3826 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3827 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3828 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3829 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3830 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3831 (cortex_a8_neon_mcr): Likewise.
3832 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3833 (cortex_a8_neon_mrc): Likewise.
3834 (cortex_a8_neon_mrrc): Likewise.
3836 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3838 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
3840 * config/aarch64/aarch64-simd.md (neon_type): Remove.
3841 (aarch64_simd_dup<mode>): Add "type" attribute.
3842 (aarch64_dup_lane<mode>): Likewise.
3843 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
3844 (*aarch64_simd_mov<mode>): Likewise.
3845 (aarch64_simd_mov_from_<mode>low): Likewise.
3846 (aarch64_simd_mov_from_<mode>high): Likewise.
3847 (orn<mode>3): Likewise.
3848 (bic<mode>3): Likewise.
3849 (add<mode>3): Likewise.
3850 (sub<mode>3): Likewise.
3851 (mul<mode>3): Likewise.
3852 (*aarch64_mul3_elt<mode>): Likewise.
3853 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
3854 (*aarch64_mul3_elt_to_128df): Likewise.
3855 (*aarch64_mul3_elt_to_64v2df): Likewise.
3856 (neg<mode>2): Likewise.
3857 (abs<mode>2): Likewise.
3858 (abd<mode>_3): Likewise.
3859 (aba<mode>_3): Likewise.
3860 (fabd<mode>_3): Likewise.
3861 (*fabd_scalar<mode>3): Likewise.
3862 (and<mode>3): Likewise.
3863 (ior<mode>3): Likewise.
3864 (xor<mode>3): Likewise.
3865 (one_cmpl<mode>2): Likewise.
3866 (aarch64_simd_vec_set<mode>): Likewise.
3867 (aarch64_simd_lshr<mode>): Likewise.
3868 (aarch64_simd_ashr<mode>): Likewise.
3869 (aarch64_simd_imm_shl<mode>): Likewise.
3870 (aarch64_simd_reg_sshl<mode): Likewise.
3871 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
3872 (aarch64_simd_reg_shl<mode>_signed): Likewise.
3873 (aarch64_simd_vec_setv2di): Likewise.
3874 (aarch64_simd_vec_set<mode>): Likewise.
3875 (aarch64_mla<mode>): Likewise.
3876 (*aarch64_mla_elt<mode>): Likewise.
3877 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
3878 (aarch64_mls<mode>): Likewise.
3879 (*aarch64_mls_elt<mode>): Likewise.
3880 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
3881 (<su><maxmin><mode>3): Likewise.
3882 (move_lo_quad_<mode>): Likewise.
3883 (aarch64_simd_move_hi_quad_<mode>): Likewise.
3884 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
3885 (vec_pack_trunc_<mode>): Likewise.
3886 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
3887 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
3888 (*aarch64_<su>mlal_lo<mode>): Likewise.
3889 (*aarch64_<su>mlal_hi<mode>): Likewise.
3890 (*aarch64_<su>mlsl_lo<mode>): Likewise.
3891 (*aarch64_<su>mlsl_hi<mode>): Likewise.
3892 (*aarch64_<su>mlal<mode>): Likewise.
3893 (*aarch64_<su>mlsl<mode>): Likewise.
3894 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
3895 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
3896 (add<mode>3): Likewise.
3897 (sub<mode>3): Likewise.
3898 (mul<mode>3): Likewise.
3899 (div<mode>3): Likewise.
3900 (neg<mode>2): Likewise.
3901 (abs<mode>2): Likewise.
3902 (fma<mode>4): Likewise.
3903 (*aarch64_fma4_elt<mode>): Likewise.
3904 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
3905 (*aarch64_fma4_elt_to_128df): Likewise.
3906 (*aarch64_fma4_elt_to_64v2df): Likewise.
3907 (fnma<mode>4): Likewise.
3908 (*aarch64_fnma4_elt<mode>): Likewise.
3909 (*aarch64_fnma4_elt_<vswap_width_name><mode>
3910 (*aarch64_fnma4_elt_to_128df): Likewise.
3911 (*aarch64_fnma4_elt_to_64v2df): Likewise.
3912 (<frint_pattern><mode>2): Likewise.
3913 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
3914 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
3915 (vec_unpacks_lo_v4sf): Likewise.
3916 (aarch64_float_extend_lo_v2df): Likewise.
3917 (vec_unpacks_hi_v4sf): Likewise.
3918 (aarch64_float_truncate_lo_v2sf): Likewise.
3919 (aarch64_float_truncate_hi_v4sf): Likewise.
3920 (aarch64_vmls<mode>): Likewise.
3921 (<su><maxmin><mode>3): Likewise.
3922 (<maxmin_uns><mode>3): Likewise.
3923 (reduc_<sur>plus_<mode>): Likewise.
3924 (reduc_<sur>plus_v2di): Likewise.
3925 (reduc_<sur>plus_v2si): Likewise.
3926 (reduc_<sur>plus_<mode>): Likewise.
3927 (aarch64_addpv4sf): Likewise.
3928 (clz<mode>2): Likewise.
3929 (reduc_<maxmin_uns>_<mode>): Likewise.
3930 (reduc_<maxmin_uns>_v2di): Likewise.
3931 (reduc_<maxmin_uns>_v2si): Likewise.
3932 (reduc_<maxmin_uns>_<mode>): Likewise.
3933 (reduc_<maxmin_uns>_v4sf): Likewise.
3934 (aarch64_simd_bsl<mode>_internal): Likewise.
3935 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
3936 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
3937 (aarch64_get_lane<mode>): Likewise.
3938 (*aarch64_combinez<mode>): Likewise.
3939 (aarch64_combine<mode>): Likewise.
3940 (aarch64_simd_combine<mode>): Likewise.
3941 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
3942 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
3943 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
3944 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
3945 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
3946 (aarch64_<sur>h<addsub><mode>): Likewise.
3947 (aarch64_<sur><addsub>hn<mode>): Likewise.
3948 (aarch64_<sur><addsub>hn2<mode>): Likewise.
3949 (aarch64_pmul<mode>): Likewise.
3950 (aarch64_<su_optab><optab><mode>): Likewise.
3951 (aarch64_<sur>qadd<mode>): Likewise.
3952 (aarch64_sqmovun<mode>): Likewise.
3953 (aarch64_<sur>qmovn<mode>): Likewise.
3954 (aarch64_s<optab><mode>): Likewise.
3955 (aarch64_sq<r>dmulh<mode>): Likewise.
3956 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3957 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
3958 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3959 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
3960 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3961 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3962 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
3963 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
3964 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
3965 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
3966 (aarch64_sqdmull<mode>): Likewise.
3967 (aarch64_sqdmull_lane<mode>_internal): Likewise.
3968 (aarch64_sqdmull_n<mode>): Likewise.
3969 (aarch64_sqdmull2<mode>_internal): Likewise.
3970 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
3971 (aarch64_sqdmull2_n<mode>_internal): Likewise.
3972 (aarch64_<sur>shl<mode>): Likewise.
3973 (aarch64_<sur>q<r>shl<mode>
3974 (aarch64_<sur>shll_n<mode>): Likewise.
3975 (aarch64_<sur>shll2_n<mode>): Likewise.
3976 (aarch64_<sur>shr_n<mode>): Likewise.
3977 (aarch64_<sur>sra_n<mode>): Likewise.
3978 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
3979 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
3980 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
3981 (aarch64_cm<optab><mode>): Likewise.
3982 (aarch64_cm<optab>di): Likewise.
3983 (aarch64_cm<optab><mode>): Likewise.
3984 (aarch64_cm<optab>di): Likewise.
3985 (aarch64_cmtst<mode>): Likewise.
3986 (aarch64_cmtstdi): Likewise.
3987 (aarch64_cm<optab><mode>): Likewise.
3988 (*aarch64_fac<optab><mode>): Likewise.
3989 (aarch64_addp<mode>): Likewise.
3990 (aarch64_addpdi): Likewise.
3991 (sqrt<mode>2): Likewise.
3992 (vec_load_lanesoi<mode>): Likewise.
3993 (vec_store_lanesoi<mode>): Likewise.
3994 (vec_load_lanesci<mode>): Likewise.
3995 (vec_store_lanesci<mode>): Likewise.
3996 (vec_load_lanesxi<mode>): Likewise.
3997 (vec_store_lanesxi<mode>): Likewise.
3998 (*aarch64_mov<mode>): Likewise.
3999 (aarch64_ld2<mode>_dreg): Likewise.
4000 (aarch64_ld2<mode>_dreg): Likewise.
4001 (aarch64_ld3<mode>_dreg): Likewise.
4002 (aarch64_ld3<mode>_dreg): Likewise.
4003 (aarch64_ld4<mode>_dreg): Likewise.
4004 (aarch64_ld4<mode>_dreg): Likewise.
4005 (aarch64_tbl1<mode>): Likewise.
4006 (aarch64_tbl2v16qi): Likewise.
4007 (aarch64_combinev16qi): Likewise.
4008 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
4009 (aarch64_st2<mode>_dreg): Likewise.
4010 (aarch64_st2<mode>_dreg): Likewise.
4011 (aarch64_st3<mode>_dreg): Likewise.
4012 (aarch64_st3<mode>_dreg): Likewise.
4013 (aarch64_st4<mode>_dreg): Likewise.
4014 (aarch64_st4<mode>_dreg): Likewise.
4015 (*aarch64_simd_ld1r<mode>): Likewise.
4016 (aarch64_frecpe<mode>): Likewise.
4017 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
4018 (aarch64_frecps<mode>): Likewise.
4020 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4022 * config/arm/iterators.md (V_elem_ch): New.
4024 (VQH_type): Likewise.
4025 * config/arm/arm.md (is_neon_type): New.
4026 (conds): Use is_neon_type.
4027 (anddi3_insn): Update type attribute.
4028 (xordi3_insn): Likewise.
4029 (one_cmpldi2): Likewise.
4030 * config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
4031 * config/arm/neon.md (neon_mov): Update type attribute.
4032 (*movmisalign<mode>_neon_store): Likewise.
4033 (*movmisalign<mode>_neon_load): Likewise.
4034 (vec_set<mode>_internal): Likewise.
4035 (vec_set<mode>_internal): Likewise.
4036 (vec_setv2di_internal): Likewise.
4037 (vec_extract<mode>): Likewise.
4038 (vec_extract<mode>): Likewise.
4039 (vec_extractv2di): Likewise.
4040 (*add<mode>3_neon): Likewise.
4041 (adddi3_neon): Likewise.
4042 (*sub<mode>3_neon): Likewise.
4043 (subdi3_neon): Likewise.
4044 (fma<VCVTF:mode>4): Likewise.
4045 (fma<VCVTF:mode>4_intrinsic): Likewise.
4046 (*fmsub<VCVTF:mode>4): Likewise.
4047 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
4048 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
4049 (ior<mode>3): Likewise.
4050 (and<mode>3): Likewise.
4051 (orn<mode>3_neon): Likewise.
4052 (orndi3_neon): Likewise.
4053 (bic<mode>3_neon): Likewise.
4054 (bicdi3_neon): Likewise.
4055 (xor<mode>3): Likewise.
4056 (one_cmpl<mode>2): Likewise.
4057 (abs<mode>2): Likewise.
4058 (neg<mode>2): Likewise.
4059 (negdi2_neon): Likewise.
4060 (*umin<mode>3_neon): Likewise.
4061 (*umax<mode>3_neon): Likewise.
4062 (*smin<mode>3_neon): Likewise.
4063 (*smax<mode>3_neon): Likewise.
4064 (vashl<mode>3): Likewise.
4065 (vashr<mode>3_imm): Likewise.
4066 (vlshr<mode>3_imm): Likewise.
4067 (ashl<mode>3_signed): Likewise.
4068 (ashl<mode>3_unsigned): Likewise.
4069 (neon_load_count): Likewise.
4070 (ashldi3_neon_noclobber): Likewise.
4071 (ashldi3_neon): Likewise.
4072 (signed_shift_di3_neon): Likewise.
4073 (unsigned_shift_di3_neon): Likewise.
4074 (ashrdi3_neon_imm_noclobber): Likewise.
4075 (lshrdi3_neon_imm_noclobber): Likewise.
4076 (<shift>di3_neon): Likewise.
4077 (widen_ssum<mode>3): Likewise.
4078 (widen_usum<mode>3): Likewise.
4079 (quad_halves_<code>v4si): Likewise.
4080 (quad_halves_<code>v4sf): Likewise.
4081 (quad_halves_<code>v8hi): Likewise.
4082 (quad_halves_<code>v16qi): Likewise.
4083 (reduc_splus_v2di): Likewise.
4084 (neon_vpadd_internal<mode>): Likewise.
4085 (neon_vpsmin<mode>): Likewise.
4086 (neon_vpsmax<mode>): Likewise.
4087 (neon_vpumin<mode>): Likewise.
4088 (neon_vpumax<mode>): Likewise.
4089 (*ss_add<mode>_neon): Likewise.
4090 (*us_add<mode>_neon): Likewise.
4091 (*ss_sub<mode>_neon): Likewise.
4092 (*us_sub<mode>_neon): Likewise.
4093 (neon_vadd<mode>_unspec): Likewise.
4094 (neon_vaddl<mode>): Likewise.
4095 (neon_vaddw<mode>): Likewise.
4096 (neon_vhadd<mode>): Likewise.
4097 (neon_vqadd<mode>): Likewise.
4098 (neon_vaddhn<mode>): Likewise.
4099 (neon_vmul<mode>): Likewise.
4100 (neon_vfms<VCVTF:mode>): Likewise.
4101 (neon_vmlal<mode>): Likewise.
4102 (neon_vmls<mode>): Likewise.
4103 (neon_vmlsl<mode>): Likewise.
4104 (neon_vqdmulh<mode>): Likewise.
4105 (neon_vqdmlal<mode>): Likewise.
4106 (neon_vqdmlsl<mode>): Likewise.
4107 (neon_vmull<mode>): Likewise.
4108 (neon_vqdmull<mode>): Likewise.
4109 (neon_vsub<mode>_unspec): Likewise.
4110 (neon_vsubl<mode>): Likewise.
4111 (neon_vsubw<mode>): Likewise.
4112 (neon_vqsub<mode>): Likewise.
4113 (neon_vhsub<mode>): Likewise.
4114 (neon_vsubhn<mode>): Likewise.
4115 (neon_vceq<mode>): Likewise.
4116 (neon_vcge<mode>): Likewise.
4117 (neon_vcgeu<mode>): Likewise.
4118 (neon_vcgt<mode>): Likewise.
4119 (neon_vcgtu<mode>): Likewise.
4120 (neon_vcle<mode>): Likewise.
4121 (neon_vclt<mode>): Likewise.
4122 (neon_vcage<mode>): Likewise.
4123 (neon_vcagt<mode>): Likewise.
4124 (neon_vtst<mode>): Likewise.
4125 (neon_vabd<mode>): Likewise.
4126 (neon_vabdl<mode>): Likewise.
4127 (neon_vaba<mode>): Likewise.
4128 (neon_vabal<mode>): Likewise.
4129 (neon_vmax<mode>): Likewise.
4130 (neon_vmin<mode>): Likewise.
4131 (neon_vpaddl<mode>): Likewise.
4132 (neon_vpadal<mode>): Likewise.
4133 (neon_vpmax<mode>): Likewise.
4134 (neon_vpmin<mode>): Likewise.
4135 (neon_vrecps<mode>): Likewise.
4136 (neon_vrsqrts<mode>): Likewise.
4137 (neon_vqabs<mode>): Likewise.
4138 (neon_vqneg<mode>): Likewise.
4139 (neon_vcls<mode>): Likewise.
4140 (clz<mode>2): Likewise.
4141 (popcount<mode>2): Likewise.
4142 (neon_vrecpe<mode>): Likewise.
4143 (neon_vrsqrte<mode>): Likewise.
4144 (neon_vget_lane<mode>_sext_internal): Likewise.
4145 (neon_vget_lane<mode>_zext_internal): Likewise.
4146 (neon_vdup_n<mode>): Likewise.
4147 (neon_vdup_n<mode>): Likewise.
4148 (neon_vdup_nv2di): Likewise.
4149 (neon_vdup_lane<mode>_interal): Likewise.
4150 (*neon_vswp<mode>): Likewise.
4151 (neon_vcombine<mode>): Likewise.
4152 (float<mode><V_cvtto>2): Likewise.
4153 (floatuns<mode><V_cvtto>2): Likewise.
4154 (fix_trunc<mode><V_cvtto>2): Likewise.
4155 (fixuns_trunc<mode><V_cvtto>2
4156 (neon_vcvt<mode>): Likewise.
4157 (neon_vcvt<mode>): Likewise.
4158 (neon_vcvtv4sfv4hf): Likewise.
4159 (neon_vcvtv4hfv4sf): Likewise.
4160 (neon_vcvt_n<mode>): Likewise.
4161 (neon_vcvt_n<mode>): Likewise.
4162 (neon_vmovn<mode>): Likewise.
4163 (neon_vqmovn<mode>): Likewise.
4164 (neon_vqmovun<mode>): Likewise.
4165 (neon_vmovl<mode>): Likewise.
4166 (neon_vmul_lane<mode>): Likewise.
4167 (neon_vmul_lane<mode>): Likewise.
4168 (neon_vmull_lane<mode>): Likewise.
4169 (neon_vqdmull_lane<mode>): Likewise.
4170 (neon_vqdmulh_lane<mode>): Likewise.
4171 (neon_vqdmulh_lane<mode>): Likewise.
4172 (neon_vmla_lane<mode>): Likewise.
4173 (neon_vmla_lane<mode>): Likewise.
4174 (neon_vmlal_lane<mode>): Likewise.
4175 (neon_vqdmlal_lane<mode>): Likewise.
4176 (neon_vmls_lane<mode>): Likewise.
4177 (neon_vmls_lane<mode>): Likewise.
4178 (neon_vmlsl_lane<mode>): Likewise.
4179 (neon_vqdmlsl_lane<mode>): Likewise.
4180 (neon_vext<mode>): Likewise.
4181 (neon_vrev64<mode>): Likewise.
4182 (neon_vrev32<mode>): Likewise.
4183 (neon_vrev16<mode>): Likewise.
4184 (neon_vbsl<mode>_internal): Likewise.
4185 (neon_vshl<mode>): Likewise.
4186 (neon_vqshl<mode>): Likewise.
4187 (neon_vshr_n<mode>): Likewise.
4188 (neon_vshrn_n<mode>): Likewise.
4189 (neon_vqshrn_n<mode>): Likewise.
4190 (neon_vqshrun_n<mode>): Likewise.
4191 (neon_vshl_n<mode>): Likewise.
4192 (neon_vqshl_n<mode>): Likewise.
4193 (neon_vqshlu_n<mode>): Likewise.
4194 (neon_vshll_n<mode>): Likewise.
4195 (neon_vsra_n<mode>): Likewise.
4196 (neon_vsri_n<mode>): Likewise.
4197 (neon_vsli_n<mode>): Likewise.
4198 (neon_vtbl1v8qi): Likewise.
4199 (neon_vtbl2v8qi): Likewise.
4200 (neon_vtbl3v8qi): Likewise.
4201 (neon_vtbl4v8qi): Likewise.
4202 (neon_vtbl1v16qi): Likewise.
4203 (neon_vtbl2v16qi): Likewise.
4204 (neon_vcombinev16qi): Likewise.
4205 (neon_vtbx1v8qi): Likewise.
4206 (neon_vtbx2v8qi): Likewise.
4207 (neon_vtbx3v8qi): Likewise.
4208 (neon_vtbx4v8qi): Likewise.
4209 (*neon_vtrn<mode>_insn): Likewise.
4210 (*neon_vzip<mode>_insn): Likewise.
4211 (*neon_vuzp<mode>_insn): Likewise.
4212 (neon_vld1<mode>): Likewise.
4213 (neon_vld1_lane<mode>): Likewise.
4214 (neon_vld1_lane<mode>): Likewise.
4215 (neon_vld1_dup<mode>): Likewise.
4216 (neon_vld1_dup<mode>): Likewise.
4217 (neon_vld1_dupv2di): Likewise.
4218 (neon_vst1<mode>): Likewise.
4219 (neon_vst1_lane<mode>): Likewise.
4220 (neon_vst1_lane<mode>): Likewise.
4221 (neon_vld2<mode>): Likewise.
4222 (neon_vld2<mode>): Likewise.
4223 (neon_vld2_lane<mode>): Likewise.
4224 (neon_vld2_lane<mode>): Likewise.
4225 (neon_vld2_dup<mode>): Likewise.
4226 (neon_vst2<mode>): Likewise.
4227 (neon_vst2<mode>): Likewise.
4228 (neon_vst2_lane<mode>): Likewise.
4229 (neon_vst2_lane<mode>): Likewise.
4230 (neon_vld3<mode>): Likewise.
4231 (neon_vld3qa<mode>): Likewise.
4232 (neon_vld3qb<mode>): Likewise.
4233 (neon_vld3_lane<mode>): Likewise.
4234 (neon_vld3_lane<mode>): Likewise.
4235 (neon_vld3_dup<mode>): Likewise.
4236 (neon_vst3<mode>): Likewise.
4237 (neon_vst3qa<mode>): Likewise.
4238 (neon_vst3qb<mode>): Likewise.
4239 (neon_vst3_lane<mode>): Likewise.
4240 (neon_vst3_lane<mode>): Likewise.
4241 (neon_vld4<mode>): Likewise.
4242 (neon_vld4qa<mode>): Likewise.
4243 (neon_vld4qb<mode>): Likewise.
4244 (neon_vld4_lane<mode>): Likewise.
4245 (neon_vld4_lane<mode>): Likewise.
4246 (neon_vld4_dup<mode>): Likewise.
4247 (neon_vst4<mode>): Likewise.
4248 (neon_vst4qa<mode>): Likewise.
4249 (neon_vst4qb<mode>): Likewise.
4250 (neon_vst4_lane<mode>): Likewise.
4251 (neon_vst4_lane<mode>): Likewise.
4252 (neon_vec_unpack<US>_lo_<mode>): Likewise.
4253 (neon_vec_unpack<US>_hi_<mode>): Likewise.
4254 (neon_vec_<US>mult_lo_<mode>): Likewise.
4255 (neon_vec_<US>mult_hi_<mode>): Likewise.
4256 (neon_vec_<US>shiftl_<mode>): Likewise.
4257 (neon_unpack<US>_<mode>): Likewise.
4258 (neon_vec_<US>mult_<mode>): Likewise.
4259 (vec_pack_trunc_<mode>): Likewise.
4260 (neon_vec_pack_trunc_<mode>): Likewise.
4261 (neon_vabd<mode>_2): Likewise.
4262 (neon_vabd<mode>_3): Likewise.
4264 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4266 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
4267 (load_pair): Update type attribute.
4268 (store_pair): Update type attribute.
4269 * config/aarch64/iterators.md (q): New.
4271 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4273 * config/arm/types.md: Add new types for Neon insns.
4275 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4276 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4277 Sergey Lega <sergey.s.lega@intel.com>
4278 Anna Tikhonova <anna.tikhonova@intel.com>
4279 Ilya Tocar <ilya.tocar@intel.com>
4280 Andrey Turetskiy <andrey.turetskiy@intel.com>
4281 Ilya Verbin <ilya.verbin@intel.com>
4282 Kirill Yukhin <kirill.yukhin@intel.com>
4283 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4285 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
4286 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
4287 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
4289 (srcp14<mode>): Ditto.
4290 (rsqrt14<mode>): Ditto.
4291 (rsqrt14<mode>): Ditto.
4292 (avx512f_vmscalef<mode>): Ditto.
4293 (avx512f_scalef<mode>): Ditto.
4294 (avx512f_getexp<mode>): Ditto.
4295 (avx512f_sgetexp<mode>): Ditto.
4296 (avx512f_fixupimm<mode>): Ditto.
4297 (avx512f_sfixupimm<mode>): Ditto.
4298 (avx512f_rndscale<mode>): Ditto.
4299 (*avx512er_exp2<mode>): Ditto.
4300 (*avx512er_rcp28<mode>): Ditto.
4301 (avx512er_rsqrt28<mode>): Ditto.
4302 (avx512f_getmant<mode>): Ditto.
4303 (avx512f_getmant<mode>): Ditto.
4304 (avx512f_rndscale<mode>): Fix formatting.
4306 2013-10-15 Martin Jambor <mjambor@suse.cz>
4308 * ipa-utils.h (ipa_edge_within_scc): Declare.
4309 * ipa-cp.c (edge_within_scc): Moved...
4310 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
4312 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4313 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4314 Sergey Lega <sergey.s.lega@intel.com>
4315 Anna Tikhonova <anna.tikhonova@intel.com>
4316 Ilya Tocar <ilya.tocar@intel.com>
4317 Andrey Turetskiy <andrey.turetskiy@intel.com>
4318 Ilya Verbin <ilya.verbin@intel.com>
4319 Kirill Yukhin <kirill.yukhin@intel.com>
4320 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4322 * config/i386/predicates.md (const_8_to_15_operand): New.
4323 (const_16_to_31_operand): Ditto.
4324 * config/i386/sse.md (V8FI): New.
4326 (reduc_splus_v8df): Ditto.
4327 (reduc_splus_v16sf): Ditto.
4328 (avx512f_vextract<shuffletype>32x4_1): Ditto.
4329 (vec_extract_hi_<mode>): Ditto.
4330 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
4331 (vec_set_lo_<mode>): Ditto.
4332 (vec_set_hi_<mode>): Ditto.
4333 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
4334 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
4335 (avx512f_pshufd_1): Ditto.
4336 (avx512f_broadcast<mode>): Ditto.
4337 (avx512f_broadcast<mode>): Ditto.
4338 (define_split): Split vec_extract_lo into move.
4339 (ssequartermode): Ditto.
4340 (ssedoublemode): Extened with wider modes.
4341 (vec_extract_lo_<mode>): Ditto.
4343 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4344 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4345 Sergey Lega <sergey.s.lega@intel.com>
4346 Anna Tikhonova <anna.tikhonova@intel.com>
4347 Ilya Tocar <ilya.tocar@intel.com>
4348 Andrey Turetskiy <andrey.turetskiy@intel.com>
4349 Ilya Verbin <ilya.verbin@intel.com>
4350 Kirill Yukhin <kirill.yukhin@intel.com>
4351 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4353 * config/i386/predicates.md (register_or_constm1_operand): New.
4354 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
4355 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
4356 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
4357 UNSPEC_SCATTER_PREFETCH
4359 (avx512f_ucmp<mode>3): Ditto.
4360 (avx512f_vternlog<mode>): Ditto.
4361 (avx512f_align<mode>): Ditto.
4362 (<shift_insn><mode>3): Ditto.
4363 (avx512f_<rotate>v<mode>): Ditto.
4364 (avx512f_<rotate><mode>): Ditto.
4365 (avx512f_eq<mode>3): Ditto.
4366 (avx512f_eq<mode>3_1): Ditto.
4367 (avx512f_gt<mode>3): Ditto.
4368 (avx512f_testm<mode>3): Ditto.
4369 (avx512f_testnm<mode>3): Ditto.
4370 (avx512pf_gatherpf<mode>): Ditto.
4371 (*avx512pf_gatherpf<mode>_mask): Ditto.
4372 (*avx512pf_gatherpf<mode>): Ditto.
4373 (avx512pf_scatterpf<mode>): Ditto.
4374 (*avx512pf_scatterpf<mode>_mask): Ditto.
4375 (*avx512pf_scatterpf<mode>): Ditto.
4376 (avx512f_vec_dup_gpr<mode>): Ditto.
4377 (clz<mode>2): Ditto.
4378 (conflict<mode>): Ditto.
4379 (REDUC_SMINMAX_MODE): Extened with wider modes.
4380 (reduc_<code>_<mode>): Ditto.
4381 (vlshr<mode>3): Ditto.
4382 (vashl<mode>3): Ditto.
4384 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4385 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4386 Sergey Lega <sergey.s.lega@intel.com>
4387 Anna Tikhonova <anna.tikhonova@intel.com>
4388 Ilya Tocar <ilya.tocar@intel.com>
4389 Andrey Turetskiy <andrey.turetskiy@intel.com>
4390 Ilya Verbin <ilya.verbin@intel.com>
4391 Kirill Yukhin <kirill.yukhin@intel.com>
4392 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4394 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
4397 (avx512fmaskmode): Ditto.
4398 (bcstscalarsuff): Ditto.
4399 (avx512f_blendm<mode>): Ditto.
4400 (cmp_imm_predicate): Ditto.
4401 (avx512f_cmp<mode>3): Ditto.
4402 (avx512f_vec_dup<mode>): Ditto.
4403 (avx512f_vec_dup_mem<mode>): Ditto.
4404 (avx512f_vpermi2var<mode>3): Ditto.
4405 (avx512f_vpermt2var<mode>3): Ditto.
4406 (vec_init<mode>): Ditto.
4407 (avx512f_gathersi<mode>): Ditto.
4408 (*avx512f_gathersi<mode>): Ditto.
4409 (*avx512f_gathersi<mode>_2): Ditto.
4410 (avx512f_gatherdi<mode>): Ditto.
4411 (*avx512f_gatherdi<mode>): Ditto.
4412 (*avx512f_gatherdi<mode>_2): Ditto.
4413 (avx512f_scattersi<mode>): Ditto.
4414 (*avx512f_scattersi<mode>): Ditto.
4415 (avx512f_scatterdi<mode>): Ditto.
4416 (*avx512f_scatterdi<mode>): Ditto.
4417 (sseintprefix): Extened with wider modes.
4418 (VEC_GATHER_IDXSI): Ditto.
4419 (VEC_GATHER_IDXDI): Ditto.
4420 (VEC_GATHER_SRCDI): Ditto.
4422 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
4423 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4425 * config/arm/t-aprofile: New file.
4426 * config.gcc: Handle --with-multilib-list option.
4428 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
4430 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
4433 2013-10-15 Richard Biener <rguenther@suse.de>
4435 * tree-loop-distribution.c (build_empty_rdg): Inline into
4437 (rdg_flag_vertex): Inline into single user.
4438 (rdg_flag_vertex_and_dependent): Likewise.
4439 (build_rdg_partition_for_vertex): Remove processed bitmap.
4440 (rdg_build_partitions): Simplify.
4442 2013-10-15 Richard Biener <rguenther@suse.de>
4444 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
4445 Restructure forwarding through conversions and copies to
4446 avoid performing copy-propagation the wrong way. Adjust
4447 recursion invocations.
4448 (forward_propagate_addr_expr): Add argument stating if we
4449 are recursing from a single-use.
4450 (ssa_forward_propagate_and_combine): Adjust.
4452 2013-10-14 David Malcolm <dmalcolm@redhat.com>
4454 * dumpfile.h (gcc::dump_manager): New class, to hold state
4455 relating to dumpfile management.
4456 (get_dump_file_name): Remove in favor of method of dump_manager.
4457 (dump_initialized_p): Likewise.
4458 (dump_start): Likewise.
4459 (dump_finish): Likewise.
4460 (dump_switch_p): Likewise.
4461 (dump_register): Likewise.
4462 (get_dump_file_info): Likewise.
4463 * context.c (gcc::context::context): Construct the dump_manager
4465 * context.h (gcc::context::get_dumps): New.
4466 (gcc::context::m_dumps): New.
4467 * coverage.c (coverage_init): Port to dump_manager API.
4468 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
4469 (extra_dump_files_in_use): Likewise.
4470 (extra_dump_files_alloced): Likewise.
4471 (gcc::dump_manager::dump_manager): New.
4472 (dump_register): Convert to...
4473 (gcc::dump_manager::dump_register): ...method, replacing
4474 function-static next_dump with m_next_dump field.
4475 (get_dump_file_info): Convert to...
4476 (gcc::dump_manager::get_dump_file_info): ...method.
4477 (get_dump_file_name): Convert to...
4478 (gcc::dump_manager::get_dump_file_name): ...method.
4479 (dump_start): Convert to...
4480 (gcc::dump_manager::dump_start): ...method.
4481 (dump_finish): Convert to...
4482 (gcc::dump_manager::dump_finish): ...method.
4483 (dump_begin): Replace body with...
4484 (gcc::dump_manager::dump_begin): ...new method.
4485 (dump_phase_enabled_p): Convert to...
4486 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4487 (dump_phase_enabled_p): Convert to...
4488 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4489 (dump_initialized_p): Convert to...
4490 (gcc::dump_manager::dump_initialized_p): ...method.
4491 (dump_flag_name): Replace body with...
4492 (gcc::dump_manager::dump_flag_name): ...new method.
4493 (dump_enable_all): Convert to...
4494 (gcc::dump_manager::dump_enable_all): ...new method.
4495 (opt_info_enable_passes): Convert to...
4496 (gcc::dump_manager::opt_info_enable_passes): ...new method.
4497 (dump_switch_p_1): Convert to...
4498 (gcc::dump_manager::dump_switch_p_1): ...new method.
4499 (dump_switch_p): Convert to...
4500 (gcc::dump_manager::dump_switch_p): ...new method.
4501 (opt_info_switch_p): Port to dump_manager API.
4502 (enable_rtl_dump_file): Likewise.
4503 * opts-global.c (handle_common_deferred_options): Port to new
4505 * passes.c (pass_manager::finish_optimization_passes): Likewise.
4506 (pass_manager::register_one_dump_file): Likewise.
4507 (pass_manager::register_pass): Likewise.
4508 (pass_init_dump_file): Likewise.
4509 (pass_fini_dump_file): Likewise.
4510 * statistics.c (statistics_early_init): Likewise.
4512 2013-10-14 Richard Biener <rguenther@suse.de>
4514 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
4515 iterative_hash_canonical_type, gimple_canonical_type_hash,
4516 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
4517 gimple_register_canonical_type, print_gimple_types_stats,
4518 free_gimple_type_tables): Move to lto/lto.c
4519 (gt-gimple.h): Do not include.
4520 * gimple.h (gimple_register_canonical_type,
4521 print_gimple_types_stats, free_gimple_type_tables): Remove.
4522 * Makefile.in (GTFILES): Remove gimple.c.
4524 2013-10-14 Travis Snoozy <quandary@remstate.com>
4527 * config/msp430/msp430.c (msp430_option_override): Correct thinko
4528 scanning for msp430x targets.
4530 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
4533 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
4534 (registers_ok_for_ldd_peep): Move around.
4535 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
4536 * config/sparc/sparc.md (widening peepholes): Use it.
4538 2013-10-14 Richard Biener <rguenther@suse.de>
4542 * gimple.c (iterative_hash_canonical_type): Make sure to
4543 record the hash into the correct hashtable slot.
4545 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
4547 PR rtl-optimization/58662
4548 * combine.c (try_combine): Take into account death nodes on I2 when
4549 splitting a PARALLEL of two independent SETs. Fix dump message.
4551 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
4554 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
4555 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
4556 * config/sh/t-sh (sh_treg_combine.o): New entry.
4557 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
4558 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
4559 (register_sh_passes): New function. Register sh_treg_combine pass.
4560 (sh_option_override): Invoke it.
4561 (sh_canonicalize_comparison): Handle op0_preserve_value.
4562 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
4563 opportunities. Canonicalize branch condition.
4564 (nott): Allow only if pseudos can be created for non-SH2A.
4566 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
4569 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
4570 (ix86_expand_movmem): Replace copy_addr_to_reg with
4571 ix86_copy_addr_to_reg.
4572 (ix86_expand_setmem): Likewise.
4574 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
4576 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
4577 provided by ix86_fp_compare_mode instead of CCFPUmode.
4579 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
4581 * config/aarch64/arm_neon.h
4582 (vtbx<1,3>_<psu>8): Fix register constriants.
4584 2013-10-11 Jeff Law <law@redhat.com>
4586 PR tree-optimization/58640
4587 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
4588 threading paths that cross over two loop entry points.
4590 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4592 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
4593 handle vector float as well.
4594 (*vsx_le_perm_load_v4si): Likewise.
4595 (*vsx_le_perm_store_v2di): Likewise.
4596 (*vsx_le_perm_store_v4si): Likewise.
4598 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4600 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
4601 directly to circumvent subtract from splat{31} workaround.
4602 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
4604 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
4605 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
4606 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
4607 (altivec_vperm_<mode>): Convert to define_insn_and_split to
4608 separate big and little endian logic.
4609 (*altivec_vperm_<mode>_internal): New define_insn.
4610 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
4611 separate big and little endian logic.
4612 (*altivec_vperm_<mode>_uns_internal): New define_insn.
4613 (vec_permv16qi): Add little endian logic.
4615 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
4617 * doc/extend.texi (returns_nonnull): Remove arguments.
4619 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4620 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4621 Sergey Lega <sergey.s.lega@intel.com>
4622 Anna Tikhonova <anna.tikhonova@intel.com>
4623 Ilya Tocar <ilya.tocar@intel.com>
4624 Andrey Turetskiy <andrey.turetskiy@intel.com>
4625 Ilya Verbin <ilya.verbin@intel.com>
4626 Kirill Yukhin <kirill.yukhin@intel.com>
4627 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4629 * config/i386/sse.md (VI48F_256_512): New.
4630 (avx2_permvar<mode>): Change to ...
4631 (<avx2_avx512f>_permvar<mode>): This.
4633 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4634 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4635 Sergey Lega <sergey.s.lega@intel.com>
4636 Anna Tikhonova <anna.tikhonova@intel.com>
4637 Ilya Tocar <ilya.tocar@intel.com>
4638 Andrey Turetskiy <andrey.turetskiy@intel.com>
4639 Ilya Verbin <ilya.verbin@intel.com>
4640 Kirill Yukhin <kirill.yukhin@intel.com>
4641 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4643 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
4644 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
4645 * config/i386/sse.md (VI4_AVX): New.
4646 (sf2simodelower): Ditto.
4647 (sse2_cvtps2dq): Change to ...
4648 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
4650 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4651 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4652 Sergey Lega <sergey.s.lega@intel.com>
4653 Anna Tikhonova <anna.tikhonova@intel.com>
4654 Ilya Tocar <ilya.tocar@intel.com>
4655 Andrey Turetskiy <andrey.turetskiy@intel.com>
4656 Ilya Verbin <ilya.verbin@intel.com>
4657 Kirill Yukhin <kirill.yukhin@intel.com>
4658 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4660 * config/i386/sse.md (V_512): New.
4662 (vcond<V_512:mode><VF_512:mode>): Ditto.
4663 (vcond<V_512:mode><VI_512:mode>): Ditto.
4664 (vcondu<V_512:mode><VI_512:mode>): Ditto.
4666 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4667 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4668 Sergey Lega <sergey.s.lega@intel.com>
4669 Anna Tikhonova <anna.tikhonova@intel.com>
4670 Ilya Tocar <ilya.tocar@intel.com>
4671 Andrey Turetskiy <andrey.turetskiy@intel.com>
4672 Ilya Verbin <ilya.verbin@intel.com>
4673 Kirill Yukhin <kirill.yukhin@intel.com>
4674 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4676 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
4677 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
4679 (fma<mode>4): Removed condition.
4680 (fms<mode>4): Ditto.
4681 (fnma<mode>4): Ditto.
4682 (fnms<mode>4): Ditto.
4683 (fma4i_fmadd_<mode>): Ditto.
4684 (*fma_fmadd_<mode>): Ditto.
4685 (*fma_fmsub_<mode>): Ditto.
4686 (*fma_fnmadd_<mode>): Ditto.
4687 (*fma_fnmsub_<mode>): Ditto.
4688 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
4689 (*fma_fmaddsub_<mode>): Ditto.
4690 (*fma_fmsubadd_<mode>): Ditto.
4691 (*fmai_fmadd_<mode>): Ditto.
4692 (*fmai_fmsub_<mode>): Ditto.
4693 (*fmai_fnmadd_<mode>): Ditto.
4694 (*fmai_fnmsub_<mode>): Ditto.
4696 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4697 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4698 Sergey Lega <sergey.s.lega@intel.com>
4699 Anna Tikhonova <anna.tikhonova@intel.com>
4700 Ilya Tocar <ilya.tocar@intel.com>
4701 Andrey Turetskiy <andrey.turetskiy@intel.com>
4702 Ilya Verbin <ilya.verbin@intel.com>
4703 Kirill Yukhin <kirill.yukhin@intel.com>
4704 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4706 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
4707 (VI124_256): Changed to ...
4708 (VI124_256_48_512): This.
4709 (ssepackmode): Extended with wider modes.
4710 (<code><mode>3): Changed iterator.
4711 (*avx2_<code><mode>3): Ditto.
4712 (vec_pack_trunc_<mode>): Ditto.
4714 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4715 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4716 Sergey Lega <sergey.s.lega@intel.com>
4717 Anna Tikhonova <anna.tikhonova@intel.com>
4718 Ilya Tocar <ilya.tocar@intel.com>
4719 Andrey Turetskiy <andrey.turetskiy@intel.com>
4720 Ilya Verbin <ilya.verbin@intel.com>
4721 Kirill Yukhin <kirill.yukhin@intel.com>
4722 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4724 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
4725 (VI8F_256_512): Ditto.
4726 (abs<mode>2): Changed iterator.
4727 (avx2_perm<mode>): Changed to ...
4728 (<avx2_avx512f>_perm<mode>): This.
4729 (avx2_perm<mode>_1): Changed to ...
4730 (<avx2_avx512f>_perm<mode>_1): This.
4732 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4733 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4734 Sergey Lega <sergey.s.lega@intel.com>
4735 Anna Tikhonova <anna.tikhonova@intel.com>
4736 Ilya Tocar <ilya.tocar@intel.com>
4737 Andrey Turetskiy <andrey.turetskiy@intel.com>
4738 Ilya Verbin <ilya.verbin@intel.com>
4739 Kirill Yukhin <kirill.yukhin@intel.com>
4740 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4742 * config/i386/sse.md (VI48_AVX512F): New.
4743 (VI48_AVX2): Changed to ...
4744 (VI48_AVX2_48_AVX512F): This.
4745 (avx2_ashrv<mode>): Changed to ...
4746 (<avx2_avx512f>_ashrv<mode>): This.
4747 (avx2_<shift_insn>v<mode>): Changed to ...
4748 (<avx2_avx512f>_<shift_insn>v<mode>): This.
4750 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4751 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4752 Sergey Lega <sergey.s.lega@intel.com>
4753 Anna Tikhonova <anna.tikhonova@intel.com>
4754 Ilya Tocar <ilya.tocar@intel.com>
4755 Andrey Turetskiy <andrey.turetskiy@intel.com>
4756 Ilya Verbin <ilya.verbin@intel.com>
4757 Kirill Yukhin <kirill.yukhin@intel.com>
4758 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4760 * config/i386/sse.md (VI4_AVX512F): New.
4761 (VI8_AVX2_AVX512F): Ditto.
4762 (mul<mode>3): Extended with wider modes.
4763 (*<sse4_1_avx2>_mul<mode>3): Ditto.
4764 (mul<mode>3): Ditto.
4765 (vec_widen_<s>mult_odd_<mode>): Ditto.
4767 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4768 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4769 Sergey Lega <sergey.s.lega@intel.com>
4770 Anna Tikhonova <anna.tikhonova@intel.com>
4771 Ilya Tocar <ilya.tocar@intel.com>
4772 Andrey Turetskiy <andrey.turetskiy@intel.com>
4773 Ilya Verbin <ilya.verbin@intel.com>
4774 Kirill Yukhin <kirill.yukhin@intel.com>
4775 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4777 * config/i386/sse.md (VI2_AVX512F): New.
4778 (VI124_AVX512F): Ditto.
4779 (sseunpackmode): Extended with wider modes.
4780 (sseunpackfltmode): Ditto.
4781 (vec_unpacks_float_hi_<mode>): Ditto.
4782 (vec_unpacks_float_lo_<mode>): Ditto.
4783 (vec_unpacku_float_hi_<mode>): Ditto.
4784 (vec_unpacku_float_lo_<mode>): Ditto.
4785 (vec_unpacks_lo_<mode>): Ditto.
4786 (vec_unpacks_hi_<mode>): Ditto.
4787 (vec_unpacku_lo_<mode>): Ditto.
4788 (vec_unpacku_hi_<mode>): Ditto.
4790 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4791 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4792 Sergey Lega <sergey.s.lega@intel.com>
4793 Anna Tikhonova <anna.tikhonova@intel.com>
4794 Ilya Tocar <ilya.tocar@intel.com>
4795 Andrey Turetskiy <andrey.turetskiy@intel.com>
4796 Ilya Verbin <ilya.verbin@intel.com>
4797 Kirill Yukhin <kirill.yukhin@intel.com>
4798 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4800 * config/i386/i386.md (multdiv): New.
4801 (multdiv_mnemonic): Ditto.
4802 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
4803 (<sse>_vm<multdiv_mnemonic><mode>3): This.
4804 (<sse>_vmdiv<mode>3): Removed.
4806 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4807 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4808 Sergey Lega <sergey.s.lega@intel.com>
4809 Anna Tikhonova <anna.tikhonova@intel.com>
4810 Ilya Tocar <ilya.tocar@intel.com>
4811 Andrey Turetskiy <andrey.turetskiy@intel.com>
4812 Ilya Verbin <ilya.verbin@intel.com>
4813 Kirill Yukhin <kirill.yukhin@intel.com>
4814 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4816 * config/i386/sse.md (V): Extended with wider modes.
4818 (ssehalfvecmode): Ditto.
4820 (ssepackfltmode): Ditto.
4821 (avx_vec_concat<mode>): Ditto.
4822 (V_256_512): New iterator.
4823 (VF2_512_256): Ditto.
4824 (si2dfmode): New attribute.
4825 (si2dfmodelower): Ditto.
4827 (concat_tg_mode): Ditto.
4828 (floatv4siv4df2): Changed to ...
4829 (float<si2dfmodelower><mode>2): This.
4830 (avx_cvtps2pd256): Changed to ...
4831 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
4832 (vec_pack_trunc_v4df): Changed to ...
4833 (vec_pack_trunc_<mode>): This.
4834 (avx_vpermil<mode>): Changed to ...
4835 (<sse2_avx_avx512f>_vpermil<mode>): This.
4836 (<fixsuffix>fix_truncv8dfv8si2): New.
4837 (vec_pack_sfix_trunc_v8df): Ditto.
4838 (avx512f_rndscale<mode>): Ditto.
4839 (avx512f_roundpd512): Ditto.
4840 (vec_pack_ufix_trunc_<mode>): Updated iterator.
4842 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4843 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4844 Sergey Lega <sergey.s.lega@intel.com>
4845 Anna Tikhonova <anna.tikhonova@intel.com>
4846 Ilya Tocar <ilya.tocar@intel.com>
4847 Andrey Turetskiy <andrey.turetskiy@intel.com>
4848 Ilya Verbin <ilya.verbin@intel.com>
4849 Kirill Yukhin <kirill.yukhin@intel.com>
4850 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4852 * config/i386/i386.md (any_fix): New iterator.
4853 (fixsuffix): New attribute.
4854 * config/i386/sse.md (VF1): Extened with wider modes.
4858 (sseintvecmodelower): Ditto.
4859 (ssescalarmode): Ditto.
4860 (ssescalarnum): Ditto.
4862 (ssexmmmode): Ditto.
4863 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
4864 (<sse>_rcp<mode>2): Change iterator.
4865 (rsqrt<mode>2): Ditto.
4866 (<sse>_rsqrt<mode>2): Ditto.
4867 (avx2_vec_dup<mode>): Ditto.
4868 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
4869 (round<mode>2_sfix): Ditto.
4870 (avx2_pbroadcast<mode>): Ditto.
4871 (*andnot<mode>3): Handle XI mode.
4872 (*<code><mode>3): Ditto.
4873 (AVXTOSSEMODE): Removed.
4874 (avx_vpermil<mode>): Changed to ...
4875 (<sse2_avx_avx512f>_vpermil<mode>): This.
4877 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4878 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4879 Sergey Lega <sergey.s.lega@intel.com>
4880 Anna Tikhonova <anna.tikhonova@intel.com>
4881 Ilya Tocar <ilya.tocar@intel.com>
4882 Andrey Turetskiy <andrey.turetskiy@intel.com>
4883 Ilya Verbin <ilya.verbin@intel.com>
4884 Kirill Yukhin <kirill.yukhin@intel.com>
4885 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4887 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
4888 (<sse>_comi): Ditto.
4889 (<sse>_ucomi): Ditto.
4890 (sse_cvtss2siq_2): Ditto.
4891 (sse2_cvtsd2si): Ditto.
4892 (sse2_cvtsd2siq): Ditto.
4893 (sse2_cvttsd2si): Ditto.
4894 (sse2_cvttsd2siq): Ditto.
4895 (<shift_insn><mode>3): Ditto.
4896 (sse2_cvtsi2sdq): Update constraint and prefix.
4897 (sse_cvtsi2ss): Update prefix.
4898 (sse_cvtsi2ssq): Ditto.
4900 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4902 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
4903 ignore internal calls.
4905 2013-10-11 Richard Biener <rguenther@suse.de>
4907 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
4908 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
4909 TREE_OVERFLOW is set.
4911 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
4913 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
4915 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
4916 * gimple.h: Likewise.
4918 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
4920 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
4923 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
4925 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
4926 * configure: Regenerate.
4928 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4930 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
4931 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
4932 formatting fixes, use pp_colon instead of pp_character (..., ':'),
4933 similarly pp_right_paren.
4934 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
4935 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
4936 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
4937 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
4939 (omp_declare_simd_clauses_equal,
4940 omp_remove_redundant_declare_simd_attrs): New functions.
4941 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
4942 (walk_tree_1): Handle new OpenMP 4.0 clauses.
4943 * tree.h (OMP_LOOP_CHECK): Define.
4944 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
4945 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
4946 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
4947 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
4948 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
4949 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
4950 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
4951 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
4952 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
4953 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
4954 OMP_CLAUSE_SIMDLEN_EXPR): Define.
4955 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
4956 (omp_remove_redundant_declare_simd_attrs): New prototype.
4957 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
4958 GIMPLE_OMP_TEAMS): New codes.
4959 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
4960 * omp-low.c (struct omp_context): Add cancel_label and cancellable
4962 (target_nesting_level): New variable.
4963 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
4964 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
4965 for collapse > 1 static schedule unless ordered.
4966 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
4967 (determine_parallel_type): Adjust get_ws_args_for caller.
4968 (install_var_field): Handle mask & 4 for double indirection.
4969 (scan_sharing_clauses): Ignore shared clause on teams construct.
4970 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
4971 (create_omp_child_function): If inside target or declare target
4972 constructs, set "omp declare target" attribute on the child function.
4973 (find_combined_for): New function.
4974 (scan_omp_parallel): Handle combined loops.
4975 (scan_omp_target, scan_omp_teams): New functions.
4976 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
4977 restrictions and set ctx->cancellable for cancellable constructs.
4978 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
4979 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
4980 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
4981 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
4982 (omp_clause_aligned_alignment): New function.
4983 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
4984 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
4985 on teams constructs. Handle user defined reductions and new
4987 (lower_reduction_clauses): Don't set placeholder to address of ref
4988 if it has already the right type.
4989 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
4990 (expand_parallel_call): Use the new non-_start suffixed builtins,
4991 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
4992 and GOMP_parallel_end after the call.
4993 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
4994 (expand_omp_for_init_counts): Handle combined loops.
4995 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
4997 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
4998 end of cancellable loops.
4999 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
5000 Likewise. Handle collapse > 1 loops.
5001 (expand_omp_simd): Handle combined loops.
5002 (expand_omp_for): Add inner_stmt argument, adjust callers of
5003 expand_omp_for* functions, use expand_omp_for_static*chunk even
5004 for collapse > 1 unless ordered.
5005 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
5006 of cancellable sections.
5007 (expand_omp_single): Remove need_barrier variable, just rely on
5008 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
5009 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
5010 (expand_omp_atomic_load, expand_omp_atomic_store,
5011 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
5012 (expand_omp_target): New function.
5013 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
5014 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
5015 (build_omp_regions_1): Immediately close region for
5016 GF_OMP_TARGET_KIND_UPDATE.
5017 (maybe_add_implicit_barrier_cancel): New function.
5018 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
5020 (lower_omp_single): Likewise. Add clobber after the barrier.
5021 (lower_omp_taskgroup): New function.
5022 (lower_omp_for): Handle combined loops. Adjust
5023 lower_rec_input_clauses caller. Handle cancellation.
5024 (lower_depend_clauses): New function.
5025 (lower_omp_taskreg): Lower depend clauses. Adjust
5026 lower_rec_input_clauses caller. Add clobber after the call. Handle
5028 (lower_omp_target, lower_omp_teams): New functions.
5029 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
5030 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
5031 and GOMP_cancellation_point calls.
5032 (lower_omp): Fold stmts inside of target region.
5033 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
5034 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5035 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
5036 (BT_FN_VOID_OMPFN_PTR_UINT,
5037 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
5038 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
5039 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
5040 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
5041 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
5042 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
5043 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
5044 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
5045 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
5046 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
5047 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
5048 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
5049 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
5050 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
5051 BUILT_IN_GOMP_PARALLEL_END.
5052 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
5053 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5054 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
5055 GF_OMP_FOR_KIND_DISTRIBUTE.
5056 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
5057 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
5058 (dump_gimple_omp_return): Print lhs if it has any.
5059 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
5060 gimple_omp_atomic_seq_cst_p.
5061 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
5062 and GIMPLE_OMP_TEAMS.
5063 * langhooks.c (lhd_omp_mappable_type): New function.
5064 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
5065 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
5067 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
5068 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
5069 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
5070 (struct gimplify_omp_ctx): Add combined_loop field.
5071 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
5072 on stmts inside of target region.
5073 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
5074 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
5075 ORT_TARGET and ORT_TARGET_DATA.
5076 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
5078 (omp_notice_threadprivate_variable): Complain about threadprivate
5079 variables in target region.
5080 (omp_notice_variable): Complain about vars with non-mappable type
5081 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
5082 (omp_check_private): Ignore ORT_TARGET* regions.
5083 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
5084 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
5085 (find_combined_omp_for): New function.
5086 (gimplify_omp_for): Handle gimplification of combined loops.
5087 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
5089 (gimplify_omp_target_update): New function.
5090 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
5091 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
5092 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
5093 (gimplify_body): If fndecl has "omp declare target" attribute, add
5094 implicit ORT_TARGET context around it.
5095 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
5096 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
5097 * tree-nested.c (convert_nonlocal_reference_stmt,
5098 convert_local_reference_stmt, convert_gimple_call): Handle
5099 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5100 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
5101 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
5102 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
5103 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
5104 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
5105 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
5106 BUILT_IN_GOMP_LOOP_END_CANCEL,
5107 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
5108 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
5109 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
5110 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
5111 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
5112 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
5113 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
5114 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
5115 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
5116 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
5117 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
5118 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
5119 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
5120 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
5121 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
5122 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5123 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
5124 gimple_build_omp_teams): New functions.
5125 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
5126 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
5127 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
5128 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5129 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
5130 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
5131 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
5132 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
5133 GF_OMP_ATOMIC_SEQ_CST): New.
5134 (gimple_build_omp_taskgroup, gimple_build_omp_target,
5135 gimple_build_omp_teams): New prototypes.
5136 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
5137 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5138 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
5139 GIMPLE_OMP_SINGLE as end of range.
5140 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
5141 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
5142 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
5143 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
5144 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
5145 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
5146 gimple_omp_target_kind, gimple_omp_target_set_kind,
5147 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
5148 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
5149 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
5150 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
5151 gimple_omp_teams_set_clauses): New inlines.
5152 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
5153 and GIMPLE_OMP_TASKGROUP.
5154 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
5155 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
5156 enum omp_clause_proc_bind_kind): New.
5157 (union omp_clause_subcode): Add depend_kind, map_kind and
5158 proc_bind_kind fields.
5159 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
5160 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5161 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
5162 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
5163 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
5165 2013-10-10 Teresa Johnson <tejohnson@google.com>
5167 * predict.c (tree_estimate_probability): Add new parameter
5168 for estimate_bb_frequencies.
5169 (estimate_bb_frequencies): Add new parameter to force estimation.
5170 (rebuild_frequencies): When max frequency in function is small,
5171 recompute counts from frequencies.
5172 * predict.h (estimate_bb_frequencies): New parameter.
5174 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5176 * ipa-inline.c (ipa_inline): Fix leak of "order" when
5177 optimizations are disabled.
5179 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5181 * coverage.c (coverage_finish): Fix leak of da_file_name.
5183 2013-10-10 Jan Hubicka <jh@suse.cz>
5185 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
5186 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
5189 2013-10-10 Jakub Jelinek <jakub@redhat.com>
5192 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
5193 if any labels are in FALLTHRU_BB, use a special label emitted
5194 immediately after the asm goto insn rather than label_rtx
5196 (expand_asm_stmt): Adjust caller.
5197 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
5198 edge if the last insn in predecessor is a jump with single successor,
5199 but it isn't simplejump_p.
5201 2013-10-10 Richard Biener <rguenther@suse.de>
5203 PR tree-optimization/58656
5204 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
5206 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5208 * gimplify.c: Include expr.h and tm_p.h for targets with special
5209 va-arg padding requirements.
5211 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5213 * tree-flow.h: Move some prototypes to gimple.h.
5214 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
5215 * gimple.h: Relocate some prototypes from tree-flow.h
5216 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
5218 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
5219 (build_va_arg_indirect_ref): Relocate and make static.
5220 (std_gimplify_va_arg_expr): Relocate here.
5221 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
5222 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
5224 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5226 * doc/md.texi: Document the mnemonic attribute.
5228 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5231 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
5232 (set_attr_alternative x ...) when searching for user defined
5235 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5237 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
5239 2013-10-09 Easwaran Raman <eraman@google.com>
5241 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
5242 * cfgexpand.c (defer_stack_allocation): ...use here
5243 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
5245 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
5247 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
5248 (operand_equal_for_value_replacement): New function, extracted from
5249 value_replacement and enhanced to catch more cases.
5250 (value_replacement): Use operand_equal_for_value_replacement.
5252 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5254 * loop-doloop.c (doloop_modify, doloop_optimize): Use
5255 get_max_loop_iterations.
5257 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5259 * config/arm/aarch-common.c (arm_early_load_addr_dep):
5260 Place comment above function.
5262 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5264 * tree-flow.h: Remove all remaining prototypes, enums and structs that
5265 are not related to tree-cfg.c.
5266 * tree-ssa-address.h: New file. Relocate prototypes.
5267 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
5268 (addr_for_mem_ref): New. Combine call to get_address_description and
5269 return addr_for_mem_ref.
5270 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
5271 * tree-ssa-live.h: Adjust prototypes.
5272 * passes.c: Include tree-ssa-live.h.
5273 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
5274 * graphite.c (graphite_transform_loops): Make static.
5275 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5276 make_pass_graphite, pass_data_graphite_transforms,
5277 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
5278 * ipa-pure-const.c (warn_function_noreturn): Make static.
5279 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5280 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5281 Relocate from tree-cfg.c
5282 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
5284 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5285 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5286 Move to ipa-pure-const.c.
5287 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
5288 Relocate from tree-optimize.c.
5289 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
5290 make_pass_fixup_cfg): Move to tree-cfg.c.
5291 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
5292 Relocate some prototypes.
5293 * tree-data-ref.h (tree_check_data_deps) Add prototype.
5294 * tree-dump.c (dump_function_to_file): Remove prototype.
5295 Add tree-flow.h to the include file.
5296 * tree-dump.h: Remove prototype.
5297 * tree-parloops.h: New File. Add prototypes.
5298 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
5299 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
5300 from tree-ssa-loop.c.
5301 * tree-predcom.c (run_tree_predictive_commoning,
5302 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
5303 Relocate here from tree-ssa-loop.c.
5304 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
5305 ssa_name_values.release ().
5306 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
5307 (ssa_name_values): Relocate from tree-flow.h.
5308 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
5309 * tree-ssa-loop.c (run_tree_predictive_commoning,
5310 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
5311 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5312 make_pass_graphite, pass_data_graphite_transforms,
5313 make_pass_graphite_transforms, gate_tree_parallelize_loops,
5314 tree_parallelize_loops, pass_data_parallelize_loops,
5315 make_pass_parallelize_loops): Move to other files.
5316 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
5318 * tree.h: Remove prototypes from tree-address.c.
5320 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5322 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
5323 (struct int_tree_map): Move to tree-hasher.h
5324 (SCALE, LABEL, PERCENT): Move to gimple.h
5325 * tree-flow-inline.h: Delete. Move functions to other files.
5326 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
5327 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
5328 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
5329 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
5330 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
5331 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
5332 * tree-hasher.h: Don't include tree-flow.h.
5333 (struct int_tree_map): Relocate from tree-flow.h.
5334 * tree-sra.c (contains_view_convert_expr_p): Relocate from
5335 tree-flow-inline.h and make static.
5336 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
5338 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
5339 tree-flow-inline.h and make static.
5340 * tree.h (is_global_var, may_be_aliased): Relocate from
5342 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
5343 * value-prof.c: No longer include tree-flow-inline.h.
5344 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
5346 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5348 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
5349 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
5350 (enum move_pos): Move to tree-ssa-loop-im.h
5351 * cfgloop.h: Move some prototypes.
5352 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
5353 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
5354 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
5355 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
5356 (loop_containing_stmt): Relocate from tree-flow-inline.h.
5357 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
5358 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
5359 (enum move_pos): Relocate here.
5360 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
5362 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
5363 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
5364 make_pass_lim): Relocate here from tree-ssa-loop.c.
5365 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
5367 (loop_edge_to_cancel, unloop_loops): Make static.
5368 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5369 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
5370 (tree_complete_unroll, gate_tree_complete_unroll,
5371 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
5372 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5373 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
5374 * tree-ssa-loop-ivopts.c: Remove local prototypes.
5375 (stmt_invariant_in_loop_p): Remove unused function.
5376 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
5377 * tree-ssa-loop-manip.h: New file. Add prototypes.
5378 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
5379 (gcov_type_to_double_int): Move to cfgloop.h.
5380 (double_int_cmp, bound_index,
5381 estimate_numbers_of_iterations_loop): Make static.
5382 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
5383 (max_loop_iterations): Factor out get_max_loop_iterations.
5384 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
5386 * tree-ssa-loop-niter.h: New file. Add prototypes.
5387 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
5388 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
5389 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
5390 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
5391 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
5392 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
5393 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
5394 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
5395 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
5396 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
5397 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5398 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
5399 pass_data_complete_unroll, make_pass_complete_unroll,
5400 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5401 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
5402 tree-ssa-loop-ivcanon.c.
5403 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
5404 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
5405 tree-ssa-loop-prefetch.c.
5406 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
5408 (get_lsm_tmp_name): Relocate and add suffix parameter.
5409 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
5410 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
5411 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
5412 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
5413 (get_estimated_loop_iterations): Factor out accessor from
5414 estimated_loop_iterations in tree-ssa-loop-niter.c.
5415 (get_max_loop_iterations): Factor out accessor from
5416 _max_loop_iterations in tree-ssa-niter.c.
5417 * loop-unroll.c (decide_unroll_constant_iterations,
5418 decide_unroll_runtime_iterations, decide_peel_simple,
5419 decide_unroll_stupid): Use new get_* accessors.
5421 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
5423 PR tree-optimization/20318
5424 * doc/extend.texi (returns_nonnull): New function attribute.
5425 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
5427 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
5428 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
5430 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
5433 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
5434 false if both components are bitfields.
5436 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5438 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
5439 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
5440 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
5441 * config/aarch64/aarch64.h
5442 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
5443 * config/aarch64/aarch64-simd-builtins.def
5444 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
5446 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5448 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
5450 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5452 * config/aarch64/arm_neon.h (vdiv_f64): Added.
5454 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5456 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
5457 (vneg_f64): New intrinsic.
5458 (vneg_s8): Asm replaced with C.
5459 (vneg_s16): Likewise.
5460 (vneg_s32): Likewise.
5461 (vneg_s64): New intrinsic.
5462 (vnegq_f32): Asm replaced with C.
5463 (vnegq_f64): Likewise.
5464 (vnegq_s8): Likewise.
5465 (vnegq_s16): Likewise.
5466 (vnegq_s32): Likewise.
5467 (vnegq_s64): Likewise.
5469 2013-10-09 Renlin Li <Renlin.Li@arm.com>
5471 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
5473 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5475 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
5476 packed stack special handling.
5477 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
5478 back to fixed stack slots for FPRs saved due to stdarg.
5480 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5482 * config/s390/s390.c (s390_frame_info): Restructure function.
5484 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5486 * config/s390/s390.c (struct s390_frame_layout): New field
5488 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
5489 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
5490 regs_ever_clobbered to char*.
5491 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
5492 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
5493 (s390_register_info): Enable FPR save slots. Move/Copy some
5494 functionality into ...
5495 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
5496 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
5498 (s390_frame_info): Do gpr slot allocation here now. stdarg does
5499 not imply a stack frame.
5500 (s390_init_frame_layout): Remove variable clobbered_regs.
5501 (s390_update_register_info): Remove function.
5502 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
5503 cannot be used for register renaming.
5504 (s390_hard_regno_scratch_ok): New function.
5505 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
5506 (s390_initial_elimination_offset): Change offset calculation of
5507 the return address pointer.
5508 (save_gprs): Deal with only r6 being saved from the call-saved regs.
5509 (restore_gprs): Set frame related flag.
5510 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
5511 (s390_emit_prologue): Call s390_register_info instead of
5512 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
5513 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
5514 (s390_optimize_prologue): Call s390_optimize_register_info.
5515 Try to remove also FPR slot save/restore INSNs. Remove frame
5516 related flags from restore INSNs.
5518 2013-10-08 DJ Delorie <dj@redhat.com>
5520 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
5523 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
5524 avoid conflict with the MI use of %c.
5525 * config/rl78/rl78-real.md: change %c to %C throughout.
5526 * config/rl78/rl78-virt.md: Likewise.
5528 2013-10-08 Jan Hubicka <jh@suse.cz>
5530 * config/i386/i386.c (ix86_option_override_internal): Switch
5531 to SSE math for -ffast-math when target ISA supports SSE2.
5533 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
5535 * tree-flow.h: Remove some prototypes.
5536 * tree.h: Remove some protypes, add a couple.
5537 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
5538 using_eh_for_cleanups_p): Add interface routines for front ends.
5539 * tree-eh.h: New file. Add protoptyes.
5540 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
5541 (add_stmt_to_eh_lp_fn): Make static.
5542 (lower_try_finally): Use new using_eh_for_cleanups_p.
5543 * emit-rtl.c: Include tree-eh.h.
5544 * gimple.h: Include tree-eh.h.
5546 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
5548 PR tree-optimization/58480
5549 * tree-vrp.c (infer_nonnull_range): New function.
5550 (infer_value_range): Call infer_nonnull_range.
5552 2013-10-08 Dehao Chen <dehao@google.com>
5554 PR tree-optimization/58619
5555 * tree-inline.c (copy_phis_for_bb): Combine location data
5558 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5561 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
5562 RTX_FRAME_RELATED_P on INSN.
5564 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5566 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
5567 (altivec_expand_vec_perm_const): Call it.
5569 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5571 * config/rs6000/vector.md (mov<mode>): Emit permuted move
5572 sequences for LE VSX loads and stores at expand time.
5573 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
5575 * config/rs6000/rs6000.c (rs6000_const_vec): New.
5576 (rs6000_gen_le_vsx_permute): New.
5577 (rs6000_gen_le_vsx_load): New.
5578 (rs6000_gen_le_vsx_store): New.
5579 (rs6000_gen_le_vsx_move): New.
5580 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
5581 (*vsx_le_perm_load_v4si): New.
5582 (*vsx_le_perm_load_v8hi): New.
5583 (*vsx_le_perm_load_v16qi): New.
5584 (*vsx_le_perm_store_v2di): New.
5585 (*vsx_le_perm_store_v4si): New.
5586 (*vsx_le_perm_store_v8hi): New.
5587 (*vsx_le_perm_store_v16qi): New.
5588 (*vsx_xxpermdi2_le_<mode>): New.
5589 (*vsx_xxpermdi4_le_<mode>): New.
5590 (*vsx_xxpermdi8_le_V8HI): New.
5591 (*vsx_xxpermdi16_le_V16QI): New.
5592 (*vsx_lxvd2x2_le_<mode>): New.
5593 (*vsx_lxvd2x4_le_<mode>): New.
5594 (*vsx_lxvd2x8_le_V8HI): New.
5595 (*vsx_lxvd2x16_le_V16QI): New.
5596 (*vsx_stxvd2x2_le_<mode>): New.
5597 (*vsx_stxvd2x4_le_<mode>): New.
5598 (*vsx_stxvd2x8_le_V8HI): New.
5599 (*vsx_stxvd2x16_le_V16QI): New.
5601 2013-10-07 Renlin Li <Renlin.Li@arm.com>
5603 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
5605 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5607 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
5608 loop to work also for 31bit ABI.
5609 Save the stack pointer for frame_size > 0.
5611 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5613 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
5614 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
5615 constraint letters from expanders.
5616 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
5617 retry count to general_operand.
5618 ("tabort"): Give operand 0 a mode.
5619 ("tabort_1"): Add mode and constraint letter for operand 0.
5620 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
5622 2013-10-04 Jeff Law <law@redhat.com>
5624 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
5626 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
5628 (thread_across_edge): Here. Call it.
5630 2013-10-04 Cary Coutant <ccoutant@google.com>
5632 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
5633 discarding a location list expression (or a piece of one).
5635 2013-10-03 Jan Hubicka <jh@suse.cz>
5637 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
5638 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
5639 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
5641 2013-10-03 Jan Hubicka <jh@suse.cz>
5643 * config/i386/i386.c (ix86_option_override_internal): Do not enable
5644 accumulate-outgoing-args when producing unwind info.
5646 2013-10-03 Wei Mi <wmi@google.com>
5648 * lra-constraints.c (insert_move_for_subreg): New function
5649 extracted from simplify_operand_subreg.
5650 (simplify_operand_subreg): Add reload for paradoxical subreg.
5652 2013-10-03 Rong Xu <xur@google.com>
5654 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
5655 the candidate of builtin_expect such that we should fix the
5656 size/time estimation.
5657 (estimate_function_body_sizes): Do the acutally size/time fix-up
5660 2013-10-03 Rong Xu <xur@google.com>
5662 * predict.c (tree_predict_by_opcode): Get the probability
5663 for builtin_expect from param builtin_expect_probability.
5664 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
5665 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
5666 * doc/invoke.texi: Add documentation for builtin-expect-probability.
5668 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
5671 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
5672 * calls.c (alloca_call_p): Use get_callee_fndecl.
5673 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
5674 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
5676 (vrp_visit_stmt): Remove duplicated code.
5678 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
5680 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
5681 ceildf2, btruncdf2, instead of vsx_* name.
5683 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
5684 iterators to only do V2DF and V4SF here. Move the DF code to
5685 rs6000.md where it is combined with SF mode. Replace <VSv> with
5686 just 'v' since only vector operations are handled with these insns
5687 after moving the DF support to rs6000.md.
5688 (vsx_sub<mode>3): Likewise.
5689 (vsx_mul<mode>3): Likewise.
5690 (vsx_div<mode>3): Likewise.
5691 (vsx_fre<mode>2): Likewise.
5692 (vsx_neg<mode>2): Likewise.
5693 (vsx_abs<mode>2): Likewise.
5694 (vsx_nabs<mode>2): Likewise.
5695 (vsx_smax<mode>3): Likewise.
5696 (vsx_smin<mode>3): Likewise.
5697 (vsx_sqrt<mode>2): Likewise.
5698 (vsx_rsqrte<mode>2): Likewise.
5699 (vsx_fms<mode>4): Likewise.
5700 (vsx_nfma<mode>4): Likewise.
5701 (vsx_copysign<mode>3): Likewise.
5702 (vsx_btrunc<mode>2): Likewise.
5703 (vsx_floor<mode>2): Likewise.
5704 (vsx_ceil<mode>2): Likewise.
5705 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
5706 (vsx_sminsf3): Likewise.
5707 (vsx_fmadf4): Likewise.
5708 (vsx_fmsdf4): Likewise.
5709 (vsx_nfmadf4): Likewise.
5710 (vsx_nfmsdf4): Likewise.
5711 (vsx_cmpdf_internal1): Likewise.
5713 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
5714 simpler to select whether a target has SPE or traditional floating
5715 point support in iterators.
5716 (TARGET_DF_SPE): Likewise.
5717 (TARGET_SF_FPR): Likewise.
5718 (TARGET_DF_FPR): Likewise.
5719 (TARGET_SF_INSN): Macros to say whether floating point support
5720 exists for a given operation for expanders.
5721 (TARGET_DF_INSN): Likewise.
5723 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
5724 combining of SF/DF mode operations, using both traditional and VSX
5732 (abs<mode>2): Combine SF/DF modes using traditional floating point
5733 instructions. Add support for using the upper DF registers with
5734 VSX support, and SF registers with power8-vector support. Update
5735 expanders for operations supported by both the SPE and traditional
5736 floating point units.
5737 (abs<mode>2_fpr): Likewise.
5738 (nabs<mode>2): Likewise.
5739 (nabs<mode>2_fpr): Likewise.
5740 (neg<mode>2): Likewise.
5741 (neg<mode>2_fpr): Likewise.
5742 (add<mode>3): Likewise.
5743 (add<mode>3_fpr): Likewise.
5744 (sub<mode>3): Likewise.
5745 (sub<mode>3_fpr): Likewise.
5746 (mul<mode>3): Likewise.
5747 (mul<mode>3_fpr): Likewise.
5748 (div<mode>3): Likewise.
5749 (div<mode>3_fpr): Likewise.
5750 (sqrt<mode>3): Likewise.
5751 (sqrt<mode>3_fpr): Likewise.
5752 (fre<Fs>): Likewise.
5753 (rsqrt<mode>2): Likewise.
5754 (cmp<mode>_fpr): Likewise.
5755 (smax<mode>3): Likewise.
5756 (smin<mode>3): Likewise.
5757 (smax<mode>3_vsx): Likewise.
5758 (smin<mode>3_vsx): Likewise.
5759 (negsf2): Delete SF operations that are merged with DF.
5766 (fmasf4_fpr): Likewise.
5767 (fmssf4_fpr): Likewise.
5768 (nfmasf4_fpr): Likewise.
5769 (nfmssf4_fpr): Likewise.
5770 (sqrtsf2): Likewise.
5771 (rsqrtsf_internal1): Likewise.
5772 (smaxsf3): Likewise.
5773 (sminsf3): Likewise.
5774 (cmpsf_internal1): Likewise.
5775 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
5776 (negdf2): Delete DF operations that are merged with SF.
5778 (nabsdf2): Likewise.
5784 (rsqrtdf_internal1): Likewise.
5785 (fmadf4_fpr): Likewise.
5786 (fmsdf4_fpr): Likewise.
5787 (nfmadf4_fpr): Likewise.
5788 (nfmsdf4_fpr): Likewise.
5789 (sqrtdf2): Likewise.
5790 (smaxdf3): Likewise.
5791 (smindf3): Likewise.
5792 (cmpdf_internal1): Likewise.
5793 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
5794 (btrunc<mode>2): Delete separate expander, and combine with the
5795 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
5796 (btrunc<mode>2_fpr): Likewise.
5797 (ceil<mode>2): Likewise.
5798 (ceil<mode>2_fpr): Likewise.
5799 (floor<mode>2): Likewise.
5800 (floor<mode>2_fpr): Likewise.
5801 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
5802 Add support for using the upper registers with VSX and
5803 power8-vector. Move insns to be closer to the define_expands. On
5804 VSX systems, prefer the traditional form of FMA over the VSX
5805 version, since the traditional form allows the target not to
5806 overlap with the inputs.
5807 (fms<mode>4_fpr): Likewise.
5808 (nfma<mode>4_fpr): Likewise.
5809 (nfms<mode>4_fpr): Likewise.
5811 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5812 Richard Earnshaw <richard.earnshaw@arm.com>
5814 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
5815 (struct mult_cost_table): Likewise.
5816 (struct mem_cost_table): Likewise.
5817 (struct fp_cost_table): Likewise.
5818 (struct vector_cost_table): Likewise.
5819 (cpu_cost_table): Likewise.
5820 * config/arm/arm.opt (mold-rts-costs): New option.
5821 (mnew-generic-costs): Likewise.
5822 * config/arm/arm.c (generic_extra_costs): New table.
5823 (cortexa15_extra_costs): Likewise.
5824 (arm_slowmul_tune): Use NULL as new costs.
5825 (arm_fastmul_tune): Likewise.
5826 (arm_strongarm_tune): Likewise.
5827 (arm_xscale_tune): Likewise.
5828 (arm_9e_tune): Likewise.
5829 (arm_v6t2_tune): Likewise.
5830 (arm_cortex_a5_tune): Likewise.
5831 (arm_cortex_a9_tune): Likewise.
5832 (arm_v6m_tune): Likewise.
5833 (arm_fa726te_tune): Likewise.
5834 (arm_cortex_a15_tune): Use cortex15_extra_costs.
5835 (arm_cortex_tune): Use generict_extra_costs.
5836 (shifter_op_p): New function.
5837 (arm_unspec_cost): Likewise.
5838 (LIBCALL_COST): Define.
5839 (arm_new_rtx_costs): New function.
5840 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
5841 table is available. Use old costs otherwise unless mnew-generic-costs
5843 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
5844 (cpu_cost_table): Declare.
5846 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
5849 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
5850 (*subs_mul_imm_<mode>)
5851 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
5852 (*sub_<shift>_<mode>)
5853 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
5854 Remove k constraint.
5856 2013-10-03 Ian Bolton <ian.bolton@arm.com>
5858 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
5860 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
5862 2013-10-02 Teresa Johnson <tejohnson@google.com>
5864 * predict.c (probably_never_executed): New function.
5865 (probably_never_executed_bb_p): Invoke probably_never_executed.
5866 (probably_never_executed_edge_p): Ditto.
5867 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
5868 Treat profile insanities conservatively.
5870 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
5872 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
5874 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
5876 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
5877 use it. Use smaller increase for scratch. Don't increase reject
5878 for early clobber scratch.
5879 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
5880 setting eliminated regs except setting fp from hfp.
5881 (lra_eliminate): Check lra_insn_recog_data on NULL.
5883 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5886 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
5887 setting -mvsx-timode by default until the underlying problem is fixed.
5888 (RS6000_CPU, power7 defaults): Likewise.
5890 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5892 * config/x-linux (host-linux.o): Remove header dependencies.
5893 Use $(COMPILE) and $(POSTCOMPILE).
5894 * config/t-linux-android (linux-android.o): Ditto.
5896 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5898 * Makefile.in (expmed.o-warn): Remove.
5900 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5902 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
5903 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
5905 2013-10-02 Teresa Johnson <tejohnson@google.com>
5907 * dojump.c (do_jump_1): Divide probability between
5908 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
5910 2013-10-02 Tom Tromey <tromey@redhat.com>
5912 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
5914 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5916 * tree-flow.h: Remove some prototypes.
5917 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
5918 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
5919 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
5920 mark_virtual_phi_result_for_renaming): Relocate here.
5921 * tree-into-ssa.h: Add prototypes.
5922 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
5923 single_pred_before_succ_order.
5924 (blocks_in_phiopt_order): Rename and move to cfganal.c.
5925 (nonfreeing_call_p) Move to gimple.c.
5926 * cfganal.c (single_pred_before_succ_order): Move and renamed from
5928 * basic-block.h (single_pred_before_succ_order): Add prototype.
5929 * gimple.c (nonfreeing_call_p): Relocate here.
5930 * gimple.h: Add prototype.
5931 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
5932 * tree-ssa-dom.h: New file. Relocate prototypes here.
5933 * tree-ssa.h: Include tree-ssa-dom.h.
5935 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5937 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
5938 Use $(COMPILE) and $(POSTCOMPILE).
5940 * config/alpha/x-alpha (driver-alpha.o): Ditto.
5942 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5944 * tree-flow.h: Remove some prototypes.
5945 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
5946 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5947 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
5948 propagate_tree_value*): Move from here to...
5949 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
5950 propagate_tree_value*): Relocate here.
5951 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5952 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
5953 * gimple-fold.c: Remove gimple-fold.h from include list.
5954 * tree-vrp.c: Remove gimple-fold.h from include list.
5955 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
5956 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
5957 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
5958 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
5959 * sese.c: Add tree-ssa-propagate.h to include list.
5961 2013-10-02 Richard Biener <rguenther@suse.de>
5963 * tree-loop-distribution.c: Include tree-vectorizer.h for
5965 (enum partition_kind): Remove PKIND_REDUCTION.
5966 (struct partition_s): Remove has_writes member, add reduction_p member.
5967 (partition_alloc): Adjust.
5968 (partition_builtin_p): Likewise.
5969 (partition_has_writes): Remove.
5970 (partition_reduction_p): New function.
5971 (partition_merge_into): Likewise.
5972 (generate_code_for_partition): Commonize builtin partition
5974 (rdg_cannot_recompute_vertex_p): Remove.
5975 (already_processed_vertex_p): Likewise.
5976 (rdg_flag_vertex): Do not set has_writes.
5977 (classify_partition): Adjust.
5978 (rdg_build_partitions): Do not set has_writes, treat all
5979 partitions as useful.
5980 (distribute_loop): Record number of library calls generated. Adjust.
5981 (tree_loop_distribution): Report number of loops and library
5982 calls generated as opt-info.
5984 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5986 * tree-flow.h: Include new .h files. Move prototypes.
5987 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
5988 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
5989 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
5990 * tree-pretty-print.h: Add prototypes from tree-flow.h.
5991 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
5992 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
5993 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
5994 * tree.h (get_ref_base_and_extent): Move prototype out.
5995 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
5997 * gimple-low.h: New File. Add prototypes from tree-flow.h.
5998 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
5999 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
6000 * tree-scalar-evolution.c: Include tree.h.
6001 * sese.c: Include tree.h.
6002 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
6003 * dwarf2out.c: Include tree-dfa.h.
6004 * tree-chrec.c: Include tree.h.
6005 * tree-data-ref.c: Include tree.h.
6007 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
6009 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
6012 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6014 * config/t-sol2 (sol2-c.o): Remove header dependencies.
6015 Use $(COMPILE) and $(POSTCOMPILE).
6016 (sol2-cxx.o): Likewise.
6017 (sol2-stubs.o): Likewise.
6019 * config/x-solaris (host-solaris.o): Likewise.
6021 * config/sparc/t-sparc (sparc.o): Remove.
6022 (sparc-c.o): Remove header dependencies.
6023 Use $(COMPILE) and $(POSTCOMPILE).
6024 * config/sparc/x-sparc: Likewise.
6026 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
6028 * config/arc/arc-opts.h: Add 2013 to Copyright years.
6029 * config/arc/arc700.md: Likewise.
6030 * config/arc/arc-modes.def: Likewise.
6031 * config/arc/arc-simd.h: Likewise.
6032 * config/arc/t-arc-uClibc: Likewise.
6033 * config/arc/t-arc-newlib: Likewise.
6035 2013-10-02 Renlin Li <renlin.li@arm.com>
6037 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
6039 (aarch64_expand_epilogue): Likewise.
6041 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6042 Yufeng Zhang <yufeng.zhang@arm.com>
6044 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
6046 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
6047 'base_in' represent a conversion and legal_cast_p_1 holds; set
6048 'base_in' with the returned value from get_unwidened.
6050 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6052 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
6053 plus_constant is not used.
6055 2013-10-01 Wei Mi <wmi@google.com>
6057 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
6058 * config/i386/i386.md: Add define_peephole2 to
6059 break partial reg stall for cvtss2sd/cvtsd2ss.
6061 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6063 * config/arc/arc.c (pass_arc_ifcvt::clone):
6064 Update for ctxt_ -> m_ctxt change.
6066 2013-10-01 Jeff Law <law@redhat.com>
6068 * tree-ssa-threadupdate.c (struct redirection_data): Delete
6069 outgoing_edge and intermediate_edge fields. Instead store the path.
6070 (redirection_data::hash): Hash on the last edge's destination index.
6071 (redirection_data::equal): Check the entire thread path.
6072 (lookup_redirectio_data): Corresponding changes.
6073 (create_edge_and_update_destination_phis): Likewise.
6074 (thread_single_edge): Likewise.
6076 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6077 Diego Novillo <dnovillo@google.com>
6079 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
6080 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
6081 (vld32wh_insn, vld32wl_insn): Delete commented-out old
6082 versions of these patterns.
6084 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
6085 (__builtin_arc_aligned): Likewise.
6087 * config/arc/arc.md: Expand adc_0 comment stating the intended
6088 purpose and why it isn't ready.
6089 Replace commented out call_value_via_label_mixed with a
6090 plain comment about bl_s.
6092 * config/arc/arc.c (stdio.h): Don't include directly.
6093 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
6094 Remove if (1) condition.
6095 (arc_encode_section_info): Fix comment.
6097 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6099 * config/arc/arc.c (arc_conditional_register_usage):
6100 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
6101 Also set reg_alloc_order for DMA config regs.
6103 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6104 Jeremy Bennett <jeremy.bennett@embecosm.com>
6106 * doc/install.texi (--with-cpu): Mention ARC.
6107 (arc-*-elf32): New paragraph.
6108 (arc-linux-uclibc): Likewise.
6109 * doc/md.texi (Machine Constraints): Add ARC part.
6110 * doc/invoke.texi: (menu): Add ARC Options.
6111 (Machine Dependent Options) <ARC Options>: Add synopsis.
6112 (node ARC Options): Add.
6113 * doc/extend.texi (long_call / short_call attribute): Add ARC.
6114 (ARC Built-in Functions): New section defining
6115 generic ARC built-in functions.
6116 (ARC SIMD Built-in Functions): New section defining SIMD specific
6118 (Declaring Attributes of Functions): Extended
6119 description of short_call and long_call attributes for ARC and
6120 added index entries.
6122 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
6123 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
6124 Joern Rennecke <joern.rennecke@embecosm.com>
6125 Muhammad Khurram Riaz <khurram.riaz@arc.com>
6126 Brendan Kehoe <brendan@zen.org>
6127 Michael Eager <eager@eagercon.com>
6128 Simon Cook <simon.cook@embecosm.com>
6129 Jeremy Bennett <jeremy.bennett@embecosm.com>
6131 * config/arc, common/config/arc: New directories.
6133 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6134 Brendan Kehoe <brendan@zen.org>
6135 Simon Cook <simon.cook@embecosm.com>
6137 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
6139 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6141 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
6142 * tree-ssa-coalesce.h: New. Move prototype to here.
6143 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
6144 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
6145 (gimple_can_coalesce_p): Move to...
6146 * gimple.c (gimple_can_coalesce_p): Here.
6148 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6150 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
6151 (dump_decl_set): Move to gimple.c.
6152 * gimple.h: Don't include tree-ssa-operands.h.
6153 (dump_decl_set): Add prototype.
6154 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6155 Move to gimple-ssa.h.
6156 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
6157 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
6158 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
6159 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
6160 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
6162 (dump_decl_set): Relocate here.
6163 * gimple-ssa.h: New file.
6164 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6165 Relocate from gimple.h.
6166 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
6167 * tree-ssa-operands.c (swap_ssa_operands): Rename from
6168 swap_tree_operands and remove non-ssa path.
6169 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
6170 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
6172 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
6173 vect_is_simple_reduction_1): Use swap_ssa_operands.
6174 * tree-flow.h: Move various prototypes to tree-phinodes.h.
6175 (enum need_phi_state): Move to tree-into-ssa.c.
6176 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6177 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
6178 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
6179 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
6180 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6181 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6182 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6183 num_imm_uses): Move to ssa-iterators.h.
6184 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
6185 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
6187 (op_iter_done, op_iter_next_def, op_iter_next_tree,
6188 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
6189 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
6190 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
6191 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
6192 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
6193 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
6194 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
6195 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
6196 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
6197 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
6198 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
6199 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
6200 (set_phi_nodes): Move to tree-phinodes.h.
6201 * tree-ssa-operands.h (enum ssa_op_iter_type,
6202 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
6203 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
6204 (dump_decl_set): Remove prototype.
6205 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
6206 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
6207 (set_phi_nodes): Relocate from tree-flow-inline.h.
6208 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
6210 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
6211 include list. Temporarily add gimple.h to include list.
6212 * ssa-iterators.h: New file.
6213 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6214 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
6215 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
6216 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
6217 Relocate from tree-ssa-operands.h.
6218 (delink_imm_use, link_imm_use_to_list, link_imm_use,
6219 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6220 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6221 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6222 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
6223 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
6224 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
6225 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
6226 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
6227 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
6228 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
6229 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
6230 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
6231 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
6232 Relocate from tree-flow-inline.h.
6233 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
6235 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
6238 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
6239 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
6240 Insert '\t' to output template.
6241 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
6242 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
6243 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
6244 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
6246 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
6248 * doc/install.texi (Host/target specific installation notes for GCC):
6249 Put @anchor before @heading.
6250 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
6251 Use @email for email addresses.
6253 2013-10-01 Jeff Law <law@redhat.com>
6255 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
6256 a vec. Only delete the path if we create one without successfully
6257 registering a jump thread.
6258 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
6260 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
6261 (paths): New vector of jump threading paths.
6262 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
6263 (THREAD_PATH): New accessor macro for the entire thread path.
6264 (lookup_redirection_data): Get intermediate and final outgoing edge
6265 from the thread path.
6266 (create_edge_and_update_destination_phis): Copy the threading path.
6267 (ssa_fix_duplicate_block_edges): Get edges and block types from the
6268 jump threading path.
6269 (ssa_redirect_edges): Get edges and block types from the jump threading
6270 path. Free the path vector.
6271 (thread_block): Get edges from the jump threading path. Look at the
6272 entire path to see if we thread to a loop exit. If we cancel a jump
6273 thread request, then free the path vector.
6274 (thread_single_edge): Get edges and block types from the jump threading
6275 path. Free the path vector.
6276 (thread_through_loop_header): Get edges and block types from the jump
6277 threading path. Free the path vector.
6278 (mark_threaded_blocks): Iterate over the vector of paths and store
6279 the path on the appropriate edge. Get edges and block types from the
6280 jump threading path.
6281 (mark_threaded_blocks): Get edges and block types from the jump
6282 threading path. Free the path vector.
6283 (thread_through_all_blocks): Use the vector of paths rather than
6284 a vector of 3-edge sets.
6285 (register_jump_thread): Accept pointer to a path vector rather
6286 than the path vector itself. Store the path vector for later use.
6289 2013-10-01 Jakub Jelinek <jakub@redhat.com>
6290 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6293 * config/s390/s390.c (s390_split_branches): Modify check for table
6295 (s390_chunkify_start): Rearrange table jump insn check in order to
6296 deal with compare and branch insns correctly.
6298 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
6302 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
6303 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
6304 define_insn_and_split.
6305 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
6306 (shiftsi3_compare): New pattern.
6308 * config/arm/unspecs.md (UNSPEC_RRX): New.
6310 2013-10-01 Alan Modra <amodra@gmail.com>
6312 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
6313 casing inout operands.
6315 2013-10-01 Richard Biener <rguenther@suse.de>
6317 PR tree-optimization/58553
6318 * tree-loop-distribution.c (struct partition_s): Add niter member.
6319 (classify_partition): Populate niter member for the partition
6320 and properly identify whether the relevant store happens before
6321 or after the loop exit.
6322 (generate_memset_builtin): Use niter member from the partition.
6323 (generate_memcpy_builtin): Likewise.
6325 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6327 * vec.h (vec_prefix, vec): Prefix member names with "m_".
6328 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
6330 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6332 * basic-block.h (edge_list): Prefix member names with "m_".
6333 * context.h (context): Likewise.
6334 * domwalk.h (dom_walker): Likewise.
6335 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
6336 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
6337 * hash-table.h (hash_table): Likewise.
6338 * machmode.h (bit_field_mode_iterator): Likewise.
6339 * pass_manager.h (pass_list): Likewise.
6340 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
6341 * tree-pass.h (pass_data): Likewise.
6342 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
6343 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
6344 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
6345 * asan.c (pass_data_asan): Update accordingly.
6346 * cfganal.c (control_dependences::find_control_dependence): Likewise.
6347 (control_dependences::control_dependences): Likewise.
6348 (control_dependences::~control_dependences): Likewise.
6349 (control_dependences::~control_dependences): Likewise.
6350 (control_dependences::get_edges_dependent_on): Likewise.
6351 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
6352 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
6353 * context.c (gcc::context::context): Likewise.
6354 * cprop.c (pass_rtl_cprop::clone): Likewise.
6355 * domwalk.c (dom_walker::walk): Likewise.
6356 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
6357 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
6358 * mode-switching.c (pass_mode_switching::clone): Likewise.
6359 * passes.c (opt_pass::opt_pass): Likewise.
6360 (pass_manager::pass_manager): Likewise.
6361 * predict.c (pass_strip_predict_hints::clone): Likewise.
6362 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
6363 (pass_split_all_insns::clone): Likewise.
6364 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
6366 (bit_field_mode_iterator::next_mode): Likewise.
6367 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
6368 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
6369 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
6370 * tree-complex.c (pass_lower_complex::clone): Likewise.
6371 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
6372 * tree-object-size.c (pass_object_sizes::clone): Likewise.
6373 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
6374 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
6375 (pass_fold_builtins::clone): Likewise.
6376 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
6377 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
6378 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
6379 (pass_cd_dce::clone): Likewise.
6380 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
6381 (pass_phi_only_cprop::clone): Likewise.
6382 * tree-ssa-dse.c (pass_dse::clone): Likewise.
6383 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
6384 * tree-ssa-loop.c (pass_lim::clone): Likewise.
6385 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
6386 * tree-ssa-pre.c (pass_fre::clone): Likewise.
6387 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
6388 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
6389 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
6390 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
6391 * tree-vrp.c (pass_vrp::clone): Likewise.
6392 * tsan.c (pass_tsan::clone): Likewise.
6394 2013-09-30 Jakub Jelinek <jakub@redhat.com>
6397 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
6398 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
6401 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
6402 optimization, punt if sign_bit_p looked through any zero extension.
6404 2013-09-30 Teresa Johnson <tejohnson@google.com>
6406 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
6407 Update redirected out edge count in joiner case.
6408 (ssa_redirect_edges): Common the joiner and non-joiner cases
6409 so that joiner case gets profile updates.
6411 2013-09-30 Richard Biener <rguenther@suse.de>
6413 PR tree-optimization/58554
6414 * tree-loop-distribution.c (classify_partition): Require
6415 unconditionally executed stores for memcpy and memset recognition.
6416 (tree_loop_distribution): Calculate dominance info.
6418 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
6420 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
6421 (NO_PROFILE_COUNTERS): Likewise.
6422 (PROFILE_HOOK): Likewise.
6423 (FUNCTION_PROFILER): Likewise.
6424 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
6426 2013-09-30 Iain Sandoe <iain@codesourcery.com>
6428 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
6429 calls and defines in TARGET_MACHO conditional.
6430 (load_macho_picbase_di): Likewise.
6431 (reload_macho_picbase): Likewise.
6432 (reload_macho_picbase_si): Likewise.
6433 (reload_macho_picbase_di): Likewise.
6434 (nonlocal_goto_receiver): Likewise.
6436 2013-09-30 Nick Clifton <nickc@redhat.com>
6438 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
6439 that use the MSP430X ISA.
6440 (msp430_option_override): Scan -mmcu command line option for any
6441 MCU name that supports the MSP430X ISA.
6442 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
6443 -mmcu options which enable the MSP430X ISA.
6445 2013-09-30 Richard Biener <rguenther@suse.de>
6448 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
6449 before looking for setjmp-like calls.
6451 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6454 * config/darwin-protos.h (machopic_get_function_picbase): New.
6455 * config/darwin.c (machopic_get_function_picbase): New.
6456 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
6457 label for a new func. (load_macho_picbase_di): Likewise.
6458 (reload_macho_picbase): New expand.
6459 (reload_macho_picbase_si): New insn.
6460 (reload_macho_picbase_di): New insn.
6461 (nonlocal_goto_receiver): New define and split.
6462 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
6463 (unspecv enum): Add UNSPECV_NLGR.
6465 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6467 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
6468 that altivec registers are correctly sized on Darwin.
6470 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6472 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
6473 darwin-driver.o): Use COMPILE and POSTCOMPILE.
6474 * config/x-darwin (host-darwin.o): Likewise.
6475 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
6476 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
6477 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
6479 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
6481 * doc/invoke.texi: Fix usage of @tie{} command.
6483 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
6485 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
6487 2013-09-28 Jan Hubicka <jh@suse.cz>
6489 * config/i386/x86-tune.def: Add documentation for each of the options;
6492 2013-09-28 Jan Hubicka <jh@suse.cz>
6494 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
6496 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
6497 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
6498 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
6499 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
6500 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
6502 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
6504 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
6505 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
6506 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
6507 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
6508 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
6509 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
6510 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
6511 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
6512 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
6513 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
6514 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
6515 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
6516 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
6517 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
6518 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
6519 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
6520 gimple.h, godump.c, graphite-clast-to-gimple.c,
6521 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
6522 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
6523 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
6524 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6525 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
6526 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
6527 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
6528 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
6529 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
6530 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
6531 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
6532 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
6533 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
6534 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
6535 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
6536 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
6537 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
6538 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
6539 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
6540 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
6541 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
6542 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
6543 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
6544 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
6545 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6546 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6547 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
6548 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
6549 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
6550 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6551 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6552 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
6553 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
6554 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6555 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6556 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
6557 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
6558 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
6559 whitespace before "(".
6561 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
6563 * expr.h (extract_bit_field): Remove packedp parameter.
6564 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
6565 from forward declaration.
6566 (store_split_bit_field): Remove packedp arg from calls to
6567 extract_fixed_bit_field.
6568 (extract_bit_field_1): Remove packedp parameter and packedp
6569 argument from recursive calls and calls to extract_fixed_bit_field.
6570 (extract_bit_field): Remove packedp parameter and corresponding
6571 arg to extract_bit_field_1.
6572 (extract_fixed_bit_field): Remove packedp parameter. Remove code
6574 (extract_split_bit_field): Remove packedp arg from call to
6575 extract_fixed_bit_field.
6576 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
6577 (copy_blkmode_from_reg): Likewise.
6578 (copy_blkmode_to_reg): Likewise.
6579 (read_complex_part): Likewise.
6580 (store_field): Likewise.
6581 (expand_expr_real_1): Likewise.
6582 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
6583 to extract_bit_field.
6584 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
6585 call to extract_bit_field.
6586 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
6587 call to extract_bit_field.
6588 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
6589 and special packedp behavior from -fstrict-volatile-bitfields
6592 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6594 * lra-eliminations.c (init_elim_table): Guard value_p.
6596 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6598 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
6599 DFmode, DImode, and SFmode in the upper VSX registers based on the
6600 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
6601 if -mpower8-vector. Combine -mvsx-timode handling with the rest
6602 of the VSX register handling.
6604 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
6606 (zero_extendsidi2_lfiwzx): Add support for loading into the
6607 Altivec registers with -mpower8-vector. Use wu/wv constraints to
6608 only do VSX memory options on Altivec registers.
6609 (extendsidi2_lfiwax): Likewise.
6610 (extendsfdf2_fpr): Likewise.
6611 (mov<mode>_hardfloat, SF/SD modes): Likewise.
6612 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
6613 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
6614 (movdi_internal64): Likewise.
6616 2013-09-27 Xinliang David Li <davidxl@google.com>
6618 * opts.c (finish_options): Adjust parameters
6619 according to vect cost model.
6620 (common_handle_option): Set dynamic vect cost
6622 targhooks.c (default_add_stmt_cost): Compute stmt cost
6624 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
6625 Use helper function.
6626 * tree-vectorizer.h (unlimited_cost_model): New function.
6627 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
6628 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
6630 (vect_enhance_data_refs_alignment): Ditto.
6631 * flag-types.h: New enum.
6632 * common/config/i386/i386-common.c (ix86_option_init_struct):
6633 No need to initialize vect_cost_model flag.
6634 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
6637 2013-09-27 Diego Novillo <dnovillo@google.com>
6639 * gimple.h (enum ssa_mode): Remove.
6641 2013-09-27 Paulo Matos <pmatos@broadcom.com>
6643 * cfgloop.h (number_of_loops): Fix typo in check for null.
6645 2013-09-27 Jakub Jelinek <jakub@redhat.com>
6648 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
6649 are children of outermost saved_cfun's loop, and set it up to
6650 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
6651 if loop != loop0 and SESE region contains bbs that belong to loop0.
6653 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6655 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
6656 (binary_scale_code_p, get_base_term, get_index_term): New functions.
6657 (set_address_segment, set_address_base, set_address_index)
6658 (set_address_disp): Accept the argument unconditionally.
6659 (baseness): Remove must_be_base_p and must_be_index_p checks.
6660 (decompose_normal_address): Classify as much as possible in the
6663 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6665 * cse.c (count_reg_usage): Handle INT_LIST.
6666 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
6667 * reginfo.c (reg_scan_mark_refs): Likewise.
6668 * reload1.c (eliminate_regs_1): Likewise.
6670 2013-09-27 Iain Sandoe <iain@codesourcery.com>
6673 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
6676 2013-09-27 Richard Biener <rguenther@suse.de>
6678 PR tree-optimization/58459
6679 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
6680 restriction not propagating into loops.
6682 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
6684 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
6685 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
6686 * doc/tree-ssa.texi (Walking use-def chains): Delete.
6688 2013-09-26 Richard Biener <rguenther@suse.de>
6690 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
6692 2013-09-26 Richard Biener <rguenther@suse.de>
6694 * alias.h (component_uses_parent_alias_set): Rename to ...
6695 (component_uses_parent_alias_set_from): ... this.
6696 * alias.c (component_uses_parent_alias_set): Rename to ...
6697 (component_uses_parent_alias_set_from): ... this and return
6699 (reference_alias_ptr_type_1): Use the result from
6700 component_uses_parent_alias_set_from instead of stripping
6701 components one at a time.
6702 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
6704 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6706 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
6707 Move prototypes to...
6708 * tree-ssa-ter.h: New File. Move prototypes here.
6709 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
6710 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
6711 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
6712 from is_replaceable_p.
6713 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
6714 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
6715 (process_replaceable): Use ter_is_replaceable_p.
6716 (find_replaceable_in_bb): Use ter_is_replaceable_p.
6717 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
6718 newly refactored ssa_is_replaceable_p.
6719 * cfgexpand.c: Include tree-outof-ssa.h.
6720 * ssaexpand.h: Delete.
6722 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6724 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
6725 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
6727 (create_gimple_tmp): Delete.
6728 (get_expr_type, build_assign, build_type_cast): Move to...
6729 * gimple-builder.c: New File.
6730 (get_expr_type): Relocate from gimple.c.
6731 (build_assign, build_type_cast): Change to only create ssanames.
6732 * gimple.h: Move prototypes to...
6733 * gimple-builder.h: New File. Here.
6734 * tree-ssa.h: And here.
6735 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
6736 count_uses_and_derefs): Relocate from gimple.c.
6737 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
6738 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
6739 * tree-ssa-math-opts (execute_cse_reciprocals): Use
6740 gimple_replace_ssa_lhs.
6741 * asan.c: Include gimple-builder.h.
6742 * Makefile.in: Add gimple-builder.o.
6744 2013-09-26 Richard Biener <rguenther@suse.de>
6746 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
6747 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
6748 (loe_visit_block): Use gcc_checking_assert.
6749 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
6750 gimple_assign_ssa_name_copy_p.
6751 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
6754 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
6756 * config/rs6000/t-rs6000 (rs6000.o): Remove.
6757 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
6759 2013-09-26 Richard Biener <rguenther@suse.de>
6761 PR tree-optimization/58539
6762 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
6763 the fact that debug statements are not taking part in loop-closed
6766 2013-09-26 Nick Clifton <nickc@redhat.com>
6768 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
6769 time warning message.
6770 (msp430_print_operand_raw): Delete unused letter parameter.
6771 (TARGET_PRINT_OPERAND_ADDRESS): Define.
6772 (msp430_print_operand_address): New function.
6773 (msp430_print_operand): Move address printing code from here to
6775 * config/msp430/msp430.md (movsipsi2): Add comment in generated
6777 (zero_extendpsisi2): Likewise.
6778 (extendpsisi2): New pattern.
6779 (andneghi3): New pattern.
6781 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
6783 * config/aarch64/aarch64.opt (mlra): New option.
6784 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
6785 (TARGET_LRA_P): Define.
6787 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
6789 * expr.c (expand_assignment): Remove obsolete comment.
6791 2013-09-25 Jeff Law <law@redhat.com>
6793 * tree-flow.h (thread_through_all_blocks): Prototype moved into
6794 tree-ssa-threadupdate.h.
6795 (register_jump_thread): Similarly.
6796 * tree-ssa-threadupdate.h: New header file.
6797 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
6798 * tree-vrp.c: Likewise.
6799 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
6800 (thread_around_empty_blocks): Change type of path vector argument to
6801 an edge,type pair from just an edge. Initialize both elements when
6802 appending to a jump threading path. Tweak references to elements
6804 (thread_across_edge): Similarly. Release memory for the elements
6806 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
6807 (dump_jump_thread_path): New function broken out from
6808 register_jump_thread.
6809 (register_jump_thread): Use dump_jump_thread_path. Change type of
6810 path vector entries. Search the path for NULL edges and dump
6811 the path if one is found. Tweak the conversion of path to 3-edge
6812 form to use the block copy type information embedded in the path.
6814 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6816 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
6818 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6819 Vladimir Makarov <vmakarov@redhat.com>
6821 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
6822 from the least significant bit.
6823 (strip_address_mutations): Add bitfield operations handling.
6824 (must_be_index_p): Add shifting and rotate operations handling.
6825 (set_address_base): Use must_be_base_p predicate.
6826 (set_address_index): Use must_be_index_p predicate.
6828 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
6829 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6830 Sergey Lega <sergey.s.lega@intel.com>
6831 Anna Tikhonova <anna.tikhonova@intel.com>
6832 Ilya Tocar <ilya.tocar@intel.com>
6833 Andrey Turetskiy <andrey.turetskiy@intel.com>
6834 Ilya Verbin <ilya.verbin@intel.com>
6835 Kirill Yukhin <kirill.yukhin@intel.com>
6836 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6838 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
6840 (ix86_expand_vector_move_misalign): Support new unaligned load and
6841 stores and use new names.
6842 (CODE_FOR_sse2_storedqu): Rename to ...
6843 (CODE_FOR_sse2_storedquv16qi): ... this.
6844 (CODE_FOR_sse2_loaddqu): Rename to ...
6845 (CODE_FOR_sse2_loaddquv16qi): ... this.
6846 (CODE_FOR_avx_loaddqu256): Rename to ...
6847 (CODE_FOR_avx_loaddquv32qi): ... this.
6848 (CODE_FOR_avx_storedqu256): Rename to ...
6849 (CODE_FOR_avx_storedquv32qi): ... this.
6850 * config/i386/i386.md (fpint_logic): New.
6851 * config/i386/sse.md (VMOVE): Extend for AVX512.
6855 (VI_UNALIGNED_LOADSTORE): Ditto.
6856 (sse2_avx_avx512f): Ditto.
6857 (sse2_avx2): Extend for AVX512.
6858 (sse4_1_avx2): Ditto.
6859 (avx2_avx512f): New.
6860 (sse): Extend for AVX512.
6863 (avxsizesuffix): Ditto.
6864 (sseintvecmode): Ditto.
6866 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
6867 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
6868 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6869 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
6870 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6871 (<sse2_avx_avx512f>_storedqu<mode): ... this.
6872 (<sse>_movnt<mode>): Replace constraint "x" with "v".
6873 (STORENT_MODE): Extend for AVX512.
6874 (*absneg<mode>2): Replace constraint "x" with "v".
6875 (*mul<mode>3): Ditto.
6876 (*ieee_smin<mode>3): Ditto.
6877 (*ieee_smax<mode>3): Ditto.
6878 (avx_cmp<mode>3): Replace VF with VF_128_256.
6879 (*<sse>_maskcmp<mode>3_comm): Ditto.
6880 (<sse>_maskcmp<mode>3): Ditto.
6881 (<sse>_andnot<mode>3): Extend for AVX512.
6882 (<code><mode>3, anylogic): Replace VF with VF_128_256.
6883 (<code><mode>3, fpint_logic): New.
6884 (*<code><mode>3): Extend for AVX512.
6885 (avx512flogicsuff): New.
6886 (avx512f_<logic><mode>): Ditto.
6887 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
6889 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
6890 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
6891 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
6892 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
6893 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
6894 (xop_vpermil2<mode>3): Ditto.
6895 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
6896 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
6897 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
6898 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
6900 2013-09-25 Tom Tromey <tromey@redhat.com>
6902 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
6903 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
6904 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
6905 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
6906 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
6907 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
6908 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
6909 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
6910 (GRAPHITE_HTAB_H): Remove.
6912 2013-09-25 Tom Tromey <tromey@redhat.com>
6914 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
6916 2013-09-25 Tom Tromey <tromey@redhat.com>
6918 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
6920 2013-09-25 Tom Tromey <tromey@redhat.com>
6922 * config/i386/t-i386 (i386.o): Remove.
6923 (i386-c.o): Use COMPILE and POSTCOMPILE.
6925 2013-09-25 Tom Tromey <tromey@redhat.com>
6927 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
6929 2013-09-25 Tom Tromey <tromey@redhat.com>
6931 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
6932 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
6933 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
6934 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
6935 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
6936 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
6937 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
6938 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
6939 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
6940 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
6941 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
6942 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
6943 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
6944 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
6945 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
6946 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
6947 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
6948 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
6949 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
6950 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
6951 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
6952 (tree-ssa-pre.o, tree-ssa-sccvn.o)
6953 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
6954 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
6955 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
6956 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
6957 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
6958 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
6959 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
6960 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
6961 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
6962 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
6963 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
6964 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
6965 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
6966 (graphite-clast-to-gimple.o, graphite-dependences.o)
6967 (graphite-interchange.o, graphite-poly.o)
6968 (graphite-scop-detection.o, graphite-sese-to-poly.o)
6969 (graphite-optimize-isl.o, tree-vect-loop.o)
6970 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
6971 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
6972 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
6973 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
6974 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
6975 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
6976 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
6977 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
6978 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
6979 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
6980 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
6981 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
6982 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
6983 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
6984 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
6985 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
6986 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
6987 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
6988 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
6989 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
6990 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
6991 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
6992 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
6993 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
6994 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
6995 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
6996 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
6997 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
6998 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
6999 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
7000 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
7001 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
7002 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
7003 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
7004 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
7005 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
7006 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
7007 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
7008 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
7009 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
7010 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
7011 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
7012 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
7013 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
7014 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
7015 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
7016 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
7017 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
7018 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
7019 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
7020 (gcov-dump.o): Remove.
7021 (default-c.o): Use COMPILE and POSTCOMPILE.
7022 (CFLAGS-gcc.o): New variable.
7023 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
7025 2013-09-25 Tom Tromey <tromey@redhat.com>
7027 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
7028 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
7029 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
7030 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
7031 (c-family/c-pragma.o, c-family/c-pretty-print.o)
7032 (c-family/c-semantics.o, c-family/c-ada-spec.o)
7033 (c-family/array-notation-common.o, c-family/stub-objc.o)
7034 (c-family/c-ubsan.o): Remove.
7036 2013-09-25 Tom Tromey <tromey@redhat.com>
7038 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
7040 2013-09-25 Tom Tromey <tromey@redhat.com>
7042 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
7043 to add -DENABLE_SHARED_LIBGCC.
7044 (gcc.o): Don't use subshell.
7046 2013-09-25 Tom Tromey <tromey@redhat.com>
7048 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
7049 * configure.ac: Don't invoke AM_PROG_CC_C_O.
7050 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
7051 * configure, config.in: Rebuild.
7053 2013-09-25 Tom Tromey <tromey@redhat.com>
7055 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
7056 (COMPILE, POSTCOMPILE): New variables.
7057 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
7058 (DEPFILES): New variable.
7059 Include ".Po" files.
7060 * configure.ac: Add checks for dependency checking.
7061 * configure, aclocal.m4: Regenerate.
7063 2013-09-25 Tom Tromey <tromey@redhat.com>
7065 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
7066 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
7068 2013-09-25 Tom Tromey <tromey@redhat.com>
7070 * Makefile.in (generated_files): Add options.h,
7071 target-hooks-def.h, insn-opinit.h,
7072 common/common-target-hooks-def.h, pass-instances.def,
7073 c-family/c-target-hooks-def.h.
7075 2013-09-25 Jeff Law <law@redhat.com>
7077 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
7078 than foo[foo.length () - 1] to access last member in a vec.
7079 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
7081 2013-09-25 Richard Biener <rguenther@suse.de>
7084 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
7086 2013-09-25 Jan Hubicka <jh@suse.cz>
7088 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
7091 2013-09-25 Marek Polacek <polacek@redhat.com>
7094 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
7095 when determining the type name.
7097 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
7099 * config/sh/sh.md: Fix formatting.
7101 2013-09-24 Xinliang David Li <davidxl@google.com>
7103 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
7104 max peel iterations parameter.
7105 * param.def: New parameter.
7106 * doc/invoke.texi: Document New parameter.
7108 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
7110 * gimple-pretty-print.c: Various whitespace tweaks.
7111 * tree-core.h: Likewise.
7112 * tree-pretty-print.c: Likewise.
7113 * tree-ssa-alias.c: Likewise.
7114 * tree-ssa-copy.c: Likewise.
7115 * tree-ssanames.c: Likewise.
7116 * tree-ssanames.h: Likewise.
7117 * tree-vrp.c: Likewise.
7119 2013-09-24 Alan Modra <amodra@gmail.com>
7123 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
7124 for output operands that disallow regs. Don't use EXPAND_WRITE on
7127 2013-09-24 Richard Biener <rguenther@suse.de>
7130 * tree.c (reference_alias_ptr_type): Move ...
7131 * alias.c (reference_alias_ptr_type): ... here and implement
7132 in terms of the new reference_alias_ptr_type_1.
7133 (ref_all_alias_ptr_type_p): New helper.
7134 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
7135 use ref_all_alias_ptr_type_p.
7136 (get_deref_alias_set): Add flag_strict_aliasing check here.
7137 (reference_alias_ptr_type_1): New function, split out from ...
7138 (get_alias_set): ... here.
7139 (alias_ptr_types_compatible_p): New function.
7140 * alias.h (reference_alias_ptr_type): Declare.
7141 (alias_ptr_types_compatible_p): Likewise.
7142 * tree.h (reference_alias_ptr_type): Remove.
7143 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
7144 to compare MEM_REF alias types.
7146 2013-09-24 Richard Biener <rguenther@suse.de>
7148 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
7150 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7152 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
7153 reload helper function arrays into a single array reg_addr.
7154 (reload_fpr_gpr): Likewise.
7155 (reload_gpr_vsx): Likewise.
7156 (reload_vsx_gpr): Likewise.
7157 (struct rs6000_reg_addr): Likewise.
7158 (reg_addr): Likewise.
7159 (rs6000_debug_reg_global): Change rs6000_vector_reload,
7160 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
7161 (rs6000_init_hard_regno_mode_ok): Likewise.
7162 (rs6000_secondary_reload_direct_move): Likewise.
7163 (rs6000_secondary_reload): Likewise.
7165 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
7166 constraints: wu, ww, and wy. Repurpose wv constraint added during
7167 power8 changes. Put wg constraint in alphabetical order.
7169 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
7170 for future work to add ISA 2.07 VSX single precision support.
7171 (-mvsx-scalar-double): Change default from -1 to 1, update
7172 documentation comment.
7173 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
7174 (-mupper-regs-df): New debug switch to control whether DF values
7175 can go in the traditional Altivec registers.
7176 (-mupper-regs-sf): New debug switch to control whether SF values
7177 can go in the traditional Altivec registers.
7179 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
7181 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
7182 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
7183 Add new constraints, wu/ww/wy. Repurpose wv constraint.
7184 (rs6000_debug_legitimate_address_p): Print if we are running
7185 before, during, or after reload.
7186 (rs6000_secondary_reload): Add a comment.
7187 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
7189 * config/rs6000/constraints.md (wa constraint): Sort w<x>
7190 constraints. Update documentation string.
7191 (wd constraint): Likewise.
7192 (wf constraint): Likewise.
7193 (wg constraint): Likewise.
7194 (wn constraint): Likewise.
7195 (ws constraint): Likewise.
7196 (wt constraint): Likewise.
7197 (wx constraint): Likewise.
7198 (wz constraint): Likewise.
7199 (wu constraint): New constraint for ISA 2.07 SFmode scalar
7201 (ww constraint): Likewise.
7202 (wy constraint): Likewise.
7203 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
7204 in the previous submissions.
7205 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
7207 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
7209 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
7210 (REG_BR_PROB): Say that the probability is stored in an int_list.
7211 * reg-notes.def: Update commentary to mention INT_LIST.
7212 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
7213 (INT_LIST): New rtx.
7214 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
7215 * rtlanal.c (int_reg_note_p): New function.
7216 (alloc_reg_note): Assert that the note does not have an int argument.
7217 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
7218 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
7219 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
7220 rather than an INSN_LIST. Handle INT_LIST.
7221 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
7222 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
7223 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
7224 Manipulate them as ints rather than rtxes.
7225 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
7226 * regmove.c (copy_src_to_dest): Likewise.
7227 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
7229 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
7230 into the cases that need it.
7231 * config/arm/arm.c (arm_unwind_emit): Likewise.
7233 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
7234 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
7235 * loop-doloop.c (add_test, doloop_modify): Likewise.
7236 * loop-unswitch.c (compare_and_jump_seq): Likewise.
7237 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
7238 * predict.c (combine_predictions_for_insn): Likewise.
7239 * print-rtl.c (print_rtx): Handle INT_LIST.
7240 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
7241 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
7242 * config/arm/arm.c (emit_unlikely_jump): Likewise.
7243 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
7244 (ix86_split_fp_branch, predict_jump): Likewise.
7245 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
7246 * config/sh/sh.c (expand_cbranchsi4): Likewise.
7247 * config/spu/spu.c (ea_load_store_inline): Likewise.
7249 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
7250 value of a REG_BR_PROB note.
7251 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
7252 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
7253 * emit-rtl.c (try_split): Likewise.
7254 * predict.c (br_prob_note_reliable_p): Likewise.
7255 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
7256 * reorg.c (mostly_true_jump): Likewise.
7257 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
7258 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
7259 * config/i386/i386.c (ix86_print_operand): Likewise.
7260 * config/ia64/ia64.c (ia64_print_operand): Likewise.
7261 * config/mmix/mmix.c (mmix_print_operand): Likewise.
7262 * config/rs6000/rs6000.c (output_cbranch): Likewise.
7263 * config/s390/s390.c (s390_expand_tbegin): Likewise.
7264 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
7265 * config/sparc/sparc.c (output_cbranch): Likewise.
7266 * config/spu/spu.c (get_branch_target): Likewise.
7267 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
7268 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
7270 2013-09-23 Jan Hubicka <jh@suse.cz>
7272 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
7274 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
7275 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
7277 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
7278 (gimple_fold_call): Dump inconsistent devirtualizations; add
7279 sanity check for type based devirtualizations.
7280 * ipa-prop.c: Include ipa-utils.h
7281 (ipa_intraprocedural_devirtualization): Add sanity check.
7282 (try_make_edge_direct_virtual_call): Likewise.
7284 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
7286 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
7287 assignment statements.
7289 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
7291 * gimple-pretty-print.c (dump_ssaname_info): New function.
7292 (dump_gimple_phi): Call it.
7293 (pp_gimple_stmt_1): Likewise.
7294 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
7295 (range_info_def): Declare.
7296 * tree-pretty-print.c (pp_double_int): New function.
7297 (dump_generic_node): Call it.
7298 * tree-pretty-print.h (pp_double_int): Declare.
7299 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
7300 * tree-ssanames.h (range_info_def): New structure.
7301 (value_range_type): Move definition here.
7302 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
7304 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
7306 (set_range_info): New function.
7307 (get_range_info): Likewise.
7308 (duplicate_ssa_name_range_info): Likewise.
7309 (duplicate_ssa_name_fn): Check pointer type and call
7310 duplicate_ssa_name_range_info.
7311 * tree-ssa-copy.c (fini_copy_prop): Likewise.
7312 * tree-vrp.c (value_range_type): Remove definition, now in
7314 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
7315 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
7316 (SSA_NAME_RANGE_INFO): New macro.
7318 2013-09-23 Richard Biener <rguenther@suse.de>
7320 PR tree-optimization/58464
7321 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7322 (phi_trans_add): Change to add conditionally on being not present.
7323 (phi_translate_1): Remove recursion detection here.
7324 (phi_translate): Pre-seed the cache with NULL to catch
7325 recursion here in a more generic way.
7326 (bitmap_find_leader): Adjust comment.
7327 (get_representative_for): Dump value-numbers.
7328 (create_expression_by_pieces): Likewise.
7329 (insert_into_preds_of_block): Likewise.
7331 2013-09-23 Christian Bruel <christian.bruel@st.com>
7334 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
7335 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
7337 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
7340 2013-09-20 Renlin Li <renlin.li@arm.com>
7342 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
7344 (aarch64_expand_epilogue): Likewise.
7345 (aarch64_legitimize_reload_address): Likewise.
7347 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
7349 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
7350 NULL_TREE before pushing them onto the vector. Likewise for labels.
7352 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
7354 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
7355 comparison operators when -fno-trapping-math is in effect.
7356 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
7357 comparison operators in TFmode and assert that unsupported operators
7359 (ia64_print_operand): Likewise.
7361 2013-09-21 Jan Hubicka <jh@suse.cz>
7363 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
7364 (sse_typeless_stores): Enable for core
7365 (sse_load0_by_pxor): Likewise.
7366 (four_jump_limit): Disable for core.
7367 (pad_returns): Likewise.
7368 (avoid_vector_decode): Likewise.
7369 (fuse_cmp_and_branch): Enable for cores.
7370 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
7372 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7375 * config/pa/pa.c (pa_option_override): Disable auto increment and
7376 decrement instructions until reload is completed.
7378 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
7379 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
7380 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
7382 2013-09-20 DJ Delorie <dj@redhat.com>
7383 Nick Clifton <nickc@redhat.com>
7385 * config/rl78/rl78.c: Various whitespace and comment tweaks.
7386 (need_to_save): Save bank 0 on interrupts.
7387 (characterize_address): Strip far address wrappers.
7388 (rl78_as_legitimate_address): Likewise.
7389 (transcode_memory_rtx): Likewise.
7390 (rl78_peep_movhi_p): Disable this peephole after devirt.
7391 (rl78_propogate_register_origins): Forget all origins when a
7393 * config/rl78/rl78-virt.md: Various whitespace tweaks.
7394 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
7396 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
7397 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
7398 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
7400 * config/rl78/constraints.md: Various whitespace and paren tweaks.
7402 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7404 * config/pa/pa.md: In "scc" insn patterns, change output template to
7405 handle const0_rtx in reg_or_0_operand operands.
7407 2013-09-20 Martin Husemann <martin@NetBSD.org>
7410 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
7411 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
7413 2013-09-20 Richard Biener <rguenther@suse.de>
7416 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
7417 remembering SSA name version and block index.
7418 (new_scev_info_str): Adjust.
7419 (hash_scev_info): Likewise. Also hash the block index.
7420 (eq_scev_info): Adjust.
7421 (find_var_scev_info): Likewise.
7422 (struct instantiate_cache_entry): Remove.
7423 (struct instantiate_cache_type): Use a htab to map name, block
7425 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7426 (get_instantiated_value_entry): Likewise.
7427 (hash_idx_scev_info, eq_idx_scev_info): New functions.
7428 (instantiate_scev_name): Adjust.
7430 2013-09-20 Jeff Law <law@redhat.com>
7432 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
7434 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
7436 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
7437 Call aarch64_simd_expand_args to update op[argc].
7439 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7441 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
7444 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7446 * gengtype.c (file_rules): Added rule for *.cc files.
7447 (get_output_file_with_visibility): Give fatal message when no
7450 2013-09-20 Renlin Li <renlin.li@arm.com>
7452 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
7453 (aarch64_expand_epilogue): Likewise.
7454 (aarch64_legitimize_reload_address): Likewise.
7456 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
7459 * expr.c (expand_assignment): Remove misalignp code path.
7461 2013-09-20 Marek Polacek <polacek@redhat.com>
7464 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
7465 TYPE_PRECISION. Add asserts.
7467 2013-09-20 Richard Biener <rguenther@suse.de>
7469 PR tree-optimization/58453
7470 * tree-loop-distribution.c (distribute_loop): Apply the cost
7471 model for -ftree-loop-distribute-patterns, too.
7473 2013-09-20 Richard Biener <rguenther@suse.de>
7476 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
7477 make type comparison less strict.
7479 2013-09-20 Alan Modra <amodra@gmail.com>
7481 * configure: Regenerate.
7482 * aclocal.m4: Regenerate.
7484 2013-09-20 Marek Polacek <polacek@redhat.com>
7487 * doc/extend.texi: Document that attribute used is meant to be used
7488 on variables with static storage duration.
7490 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7492 PR tree-optimization/58472
7493 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
7494 simd_lane_access set inv_p = false.
7495 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
7496 the simduid magic VAR_DECL.
7498 2013-09-19 Jan Hubicka <jh@suse.cz>
7500 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
7502 2013-09-17 Jeff Law <law@redhat.com>
7504 * tree-ssa-dom.c (record_temporary_equivalences): New function
7505 split out of dom_opt_dom_walker::after_dom_children.
7506 (dom_opt_dom_walker::thread_across_edge): Move common code
7507 in here from dom_opt_dom_walker::after_dom_children.
7508 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
7510 2013-09-19 Jan Hubicka <jh@suse.cz>
7512 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
7513 (TARGET_GENERIC): Use PROCESOR_GENERIC
7514 (enum processor_type): Unify generic32 and 64.
7515 * i386.md (cpu): Likewise.
7516 * x86-tune.def (use_leave): Enable for generic32.
7517 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
7518 * athlon.md: Change generic64 to generic in all occurences.
7519 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
7520 (ix86_target_macros_internal): Likewise.
7521 * driver-i386.c (host_detect_local_cpu): Likewise.
7522 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
7524 (generic_memcpy, generic_memset, generic_cost): This one.
7525 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
7526 (m_GENERIC32, m_GENERIC64): Remove.
7527 (m_GENERIC): Turn into one flag.
7528 (processor_target): Unify generic tunnings.
7529 (ix86_option_override_internal): Replace generic32/64 by generic.
7530 (ix86_issue_rate): Likewise.
7531 (ix86_adjust_cost): Likewise.
7533 2013-09-19 Jan Hubicka <jh@suse.cz>
7535 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
7536 of speculative flag.
7538 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7540 * omp-low.c (expand_omp_sections): Always pass len - 1 to
7541 GOMP_sections_start, even if !exit_reachable.
7543 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7545 * lra-constraints.c (need_for_all_save_p): Use macro
7546 HARD_REGNO_CALL_PART_CLOBBERED.
7547 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
7548 set up pseudo conflict hard regs.
7550 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7553 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
7556 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7558 PR rtl-optimization/58438
7559 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
7560 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
7561 from previous subpasses.
7563 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
7565 * arm.c (arm_get_frame_offsets): Validate architecture supports
7566 LDRD/STRD before accepting the tuning preference.
7567 (arm_expand_prologue): Likewise.
7568 (arm_expand_epilogue): Likewise.
7570 2013-09-18 Richard Biener <rguenther@suse.de>
7572 PR tree-optimization/58417
7573 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
7574 have chrecs with symbols defined in the loop as operands.
7575 (chrec_fold_multiply): Likewise.
7576 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
7577 parameters before folding binary operations.
7578 (struct instantiate_cache_entry_hasher): Remove.
7579 (struct instantiate_cache_type): Use a pointer-map.
7580 (instantiate_cache_type::instantiate_cache_type): New function.
7581 (instantiate_cache_type::get): Likewise.
7582 (instantiate_cache_type::set): Likewise.
7583 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7584 (get_instantiated_value_entry): Likewise.
7585 (global_cache): New global.
7586 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
7587 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
7588 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
7589 (instantiate_scev_name): Adjust.
7590 (instantiate_scev): Construct global instead of local cache.
7591 (resolve_mixers): Likewise.
7593 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
7594 Paolo Carlini <paolo.carlini@oracle.com>
7597 * doc/implement-cxx.texi: Fix references to the C++ standards.
7599 2013-09-18 Jakub Jelinek <jakub@redhat.com>
7601 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
7602 * tree-vect-data-refs.c (vect_analyze_data_refs): For
7603 simd_lane_access drs, update also DR_ALIGNED_TO.
7605 2013-09-18 Marek Polacek <polacek@redhat.com>
7608 * doc/extend.texi: Document no_sanitize_undefined attribute.
7609 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
7610 no_sanitize_undefined attribute.
7612 2013-09-18 Nick Clifton <nickc@redhat.com>
7614 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
7615 (ASM_DECLARE_FUNCTION_NAME): Define.
7617 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
7619 * compare-elim.c (find_comparison_dom_walker): New class
7620 (find_comparisons_in_bb): Rename to
7621 find_comparison_dom_walker::before_dom_children
7622 (find_comparisons): Adjust
7623 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
7625 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
7626 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
7627 (init_walk_dominator_tree): Remove declaration.
7628 (fini_walk_dominator_tree): Remove declaration.
7629 * fwprop.c (single_def_use_dom_walker): New class
7630 (single_def_use_enter_block): Convert to
7631 single_def_use_dom_walker::before_dom_children.
7632 (single_def_use_leave_block): Convert to
7633 single_def_use_dom_walker::after_dom_children.
7634 (build_single_def_use_links): Adjust.
7635 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
7637 (find_candidates_in_block): Convert to
7638 find_candidates_dom_walker::before_dom_children.
7639 (execute_strength_reduction): Adjust.
7640 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
7641 (sese_dom_walker): New class.
7642 (sese_dom_walker::sese_dom_walker): New constructor.
7643 (sese_dom_walker::~sese_dom_walker): New destructor.
7644 (build_sese_conditions_before): Convert to
7645 sese_dom_walker::before_dom_children.
7646 (build_sese_conditions_after): Convert to
7647 sese_dom_walker::after_dom_children.
7648 (build_poly_scop): Adjust
7649 * tree-into-ssa.c (rewrite_dom_walker): New class
7650 (rewrite_enter_block): Convert to
7651 rewrite_dom_walker::before_dom_children.
7652 (rewrite_leave_block): Convert to
7653 rewrite_dom_walker::after_dom_children.
7654 (rewrite_update_dom_walker): New class.
7655 (rewrite_update_enter_block): Convert to
7656 rewrite_update_dom_walker::before_dom_children.
7657 (rewrite_update_leave_block): Convert to
7658 rewrite_update_dom_walker::after_dom_children.
7659 (rewrite_blocks, rewrite_into_ssa): Adjust.
7660 (mark_def_dom_walker): New class.
7661 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
7662 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
7663 (mark_def_sites_blocks): Convert to
7664 mark_def_dom_walker::before_dom_children.
7665 (mark_def_site_blocks): Remove.
7666 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
7667 (tree_ssa_dominator_optimize): Adjust.
7668 (dom_thread_across_edge): Convert to method
7669 dom_opt_dom_walker::thread_across_edge.
7670 (dom_opt_enter_block): Convert to member function
7671 dom_opt_dom_walker::before_dom_children.
7672 (dom_opt_leave_block): Convert to member function
7673 dom_opt_dom_walker::after_dom_children.
7674 * tree-ssa-dse.c (dse_dom_walker): New class.
7675 (dse_enter_block): Convert to member function
7676 dse_dom_walker::before_dom_children.
7677 (tree_ssa_dse): Adjust.
7678 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
7679 (determine_invariantness_stmt): Convert to method
7680 invariantness_dom_walker::before_dom_children.
7681 (determine_invariantness): Remove
7682 (move_computations_dom_walker): New class.
7683 (move_computations_stmt): Convert to method
7684 move_computations_dom_walker::before_dom_children.
7685 (move_computations, tree_ssa_lim): Adjust.
7686 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
7687 (nt_init_block): Convert to method
7688 notrappping_dom_walker::before_dom_children.
7689 (nt_fini_block): Convert to method
7690 method nontrapping_dom_walker::after_dom_children.
7691 (get_non_trapping): Adjust.
7692 * tree-ssa-pre.c (eliminate_dom_walker): New class.
7693 (eliminate_bb): Convert to method
7694 eliminate_dom_walker::before_dom_children.
7695 (eliminate_leave_block): Convert to method
7696 eliminate_dom_walker::after_dom_children.
7697 (eliminate): Adjust.
7698 * tree-ssa-strlen.c (strlen_dom_walker): New class.
7699 (strlen_enter_block): Convert to method
7700 strlen_dom_walker::before_dom_children.
7701 (strlen_leave_block): Convert to method
7702 method strlen_dom_walker::after_dom_children.
7703 (tree_ssa_strlen): Adjust.
7704 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
7705 (tree_ssa_uncprop): Adjust.
7706 (uncprop_leave_block): Convert to method
7707 uncprop_dom_walker::after_dom_children.
7708 (uncprop_leave_block): Convert to method
7709 uncprop_dom_walker::before_dom_children.
7711 2013-09-18 Bin Cheng <bin.cheng@arm.com>
7713 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
7714 branch in same basic block. Check both src and dest of the move insn.
7716 2013-09-17 Nick Clifton <nickc@redhat.com>
7718 * config/rl78/rl78-real.md (bf): New pattern.
7720 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
7721 (rl78_print_operand): Do not put a # before a %B.
7722 * config/rl78/rl78.opt: Tweak doc strings.
7724 2013-09-17 DJ Delorie <dj@redhat.com>
7726 * config/rl78/constraints.md (Wcv): Allow up to $r31.
7727 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
7728 (rl78_option_override): Likewise, if -mallregs.
7729 (is_virtual_register): Likewise.
7730 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
7731 (REGNO_OK_FOR_BASE_P): Likewise.
7732 * config/rl78/rl78.opt (-mallregs): New.
7734 2013-09-17 Nick Clifton <nickc@redhat.com>
7736 * config/rl78/rl78.c (need_to_save): Change return type to bool.
7737 For interrupt functions: save all call clobbered registers if the
7738 interrupt handler is not a leaf function.
7739 (rl78_expand_prologue): Always recompute the frame information.
7740 For interrupt functions: only select bank 0 if one of the bank 0
7741 registers is going to be psuhed.
7743 2013-09-17 DJ Delorie <dj@redhat.com>
7745 * config/rl78/constraints.md: For each W* constraint, rename to C*
7746 and create a W* constraint that checks for an optional ES: prefix
7748 * config/rl78/rl78.md (UNS_ES_ADDR): New.
7749 (es_addr): New. Used to wrap far addresses.
7750 * config/rl78/rl78-protos.h (rl78_es_addr): New.
7751 (rl78_es_base): New.
7752 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
7753 wrapped far addresses.
7754 (rl78_print_operand_1): Unwrap far addresses before processing.
7755 (rl78_lo16): Wrap far addresses in unspecs.
7756 (rl78_es_addr): New.
7757 (rl78_es_base): New.
7758 (insn_ok_now): Check for not-yet-wrapped far addresses.
7759 (transcode_memory_rtx): Properly re-wrap far addresses.
7761 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
7763 * config/sparc/t-rtems: Add leon3 multilibs.
7765 2013-09-17 Cong Hou <congh@google.com>
7767 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
7768 when checking the dot production pattern. The type of rhs operand
7769 of multiply is now checked correctly.
7771 2013-09-17 Jeff Law <law@redhat.com>
7773 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
7774 edge implied equivalences into successor phis.
7775 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
7776 here from tree-ssa-threadedge.c.
7777 (mark_threaded_blocks): When threading through a joiner, if both
7778 successors of the joiner's clone reach the same block, verify the
7779 PHI arguments are equal. If not, cancel the jump threading request.
7780 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
7781 tree-ssa-threadupdate.c
7782 (thread_across_edge): Don't check PHI argument equality when
7783 threading through joiner block here.
7785 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
7787 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
7788 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
7789 (warn_uninit, warn_uninitialized_vars,
7790 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7791 Move to tree-ssa-uninit.c.
7792 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
7793 (has_undefined_value_p): New. Pass dependant parts of
7794 ssa_undefined_value_p.
7795 (uninit_undefined_value_p): Use has_undefined_value_p.
7796 (warn_uninit, warn_uninitialized_vars,
7797 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7798 Move from tree-ssa.c.
7799 * tree-ssa.h: Adjust prototypes.
7801 2013-09-17 Jan Hubicka <jh@suse.cz>
7804 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
7805 * ipa-inline.c (can_inline_edge_p): Do not downgrade
7806 FUNCTION_NOT_OPTIMIZED.
7807 * ipa-inline-analysis.c (compute_inline_parameters): Function
7808 not optimized is not inlinable unless it is alwaysinline.
7809 (inline_analyze_function): Force calls in not optimized
7810 function not inlinable.
7812 2013-09-17 Jan Hubicka <jh@suse.cz>
7815 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
7817 * ipa.c (function_and_variable_visibility): Likewise.
7818 * ipa-profile.c (ipa_profile): Likewise.
7820 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
7823 * cgraph.c (cgraph_function_body_availability): Check for ifunc
7824 attribute, and don't inline the resolver in this case.
7826 2013-09-17 Teresa Johnson <tejohnson@google.com>
7828 * coverage.c (get_coverage_counts): Add missing newline.
7830 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7832 PR tree-optimization/58088
7833 * fold-const.c (mask_with_trailing_zeros): New function.
7834 (fold_binary_loc): Make sure we don't recurse infinitely
7835 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
7836 Use mask_with_trailing_zeros where appropriate.
7838 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
7840 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
7841 of 'prev' var to get better distance estimation.
7843 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
7845 * tree-inline.h (struct copy_body_data): Add transform_parameter.
7846 * tree-inline.c (is_parameter_of): New predicate.
7847 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
7848 a parameter has been remapped.
7849 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
7850 (optimize_inline_calls): Initialize transform_parameter.
7851 (copy_gimple_seq_and_replace_locals): Likewise.
7852 (tree_function_versioning): Likewise.
7853 (maybe_inline_call_in_expr): Likewise.
7855 2013-09-17 Nick Clifton <nickc@redhat.com>
7857 * config/msp430/msp430-protos.h: Add prototypes for new functions.
7858 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
7860 (is_attr_func): New function.
7861 (msp430_is_interrupt_func): New function.
7862 (is_naked_func): New function.
7863 (is_reentrant_func): New function.
7864 (is_critical_func): New function.
7865 (msp430_start_function): Add annotations for function attributes.
7866 (msp430_attr): New function.
7867 (msp430_attribute_table): New.
7868 (msp430_function_section): New function.
7869 (TARGET_ASM_FUNCTION_SECTION): Define.
7870 (msp430_builtin): New enum.
7871 (msp430_init_builtins): New function.
7872 (msp430_builtin_devl): New function.
7873 (msp430_expand_builtin): New function.
7874 (TARGET_INIT_BUILTINS): Define.
7875 (TARGET_EXPAND_BUILTINS): Define.
7876 (TARGET_BUILTIN_DECL): Define.
7877 (msp430_expand_prologue): Add support for naked, interrupt,
7878 critical and reentrant functions.
7879 (msp430_expand_epilogue): Likewise.
7880 (msp430_print_operand): Handle 'O' character.
7881 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
7883 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
7884 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
7885 (pushm): Use a 'n' rather than an 'i' constraint.
7886 (msp_return): Add generation of the interrupt return instruction.
7887 (disable_interrupts): New pattern.
7888 (enable_interrupts): New pattern.
7889 (push_intr_state): New pattern.
7890 (pop_intr_state): New pattern.
7891 (bic_SR): New pattern.
7892 (bis_SR): New pattern.
7893 * doc/extend.texi: Document MSP430 function attributes and builtin
7896 2013-09-17 Richard Biener <rguenther@suse.de>
7898 PR tree-optimization/58432
7899 * tree-loop-distribution.c (tree_loop_distribution): Also
7900 scan PHIs for outside loop uses and seed a partition from them.
7902 2013-09-17 Bin Cheng <bin.cheng@arm.com>
7904 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
7905 (restructure_reference): Call backtrace_base_for_ref.
7907 2013-09-17 Alan Modra <amodra@gmail.com>
7910 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
7913 2013-09-16 DJ Delorie <dj@redhat.com>
7915 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
7916 vregs location for RL78/G10.
7917 (rl78_expand_prologue): Avoid SEL on G10.
7918 (rl78_expand_epilogue): Likewise.
7919 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
7920 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
7921 __RL78_G10__ when appropriate.
7922 (ASM_SPEC): Pass -mg10 along to the assembler.
7923 * config/rl78/rl78.md (sel_rb): Disable for G10.
7924 * config/rl78/rl78.opt: Add -mg10 option.
7925 * config/rl78/t-rl78: Add -mg10 multilib.
7927 2013-09-16 Xinliang David Li <davidxl@google.com>
7929 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
7930 * omp-low.c (omp_max_vf): Ditto.
7931 (expand_omp_simd): Ditto.
7932 * tree-vectorizer.c (vectorize_loops): Ditto.
7933 (gate_vect_slp): Ditto.
7934 (gate_increase_alignment): Ditto.
7935 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
7936 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
7937 (gate_tree_vectorize): Name change.
7938 (tree_vectorize): Ditto.
7939 (pass_vectorize::gate): Call new function.
7940 (pass_vectorize::execute): Ditto.
7941 * opts.c: O3 default setting change.
7942 (finish_options): Check new flag.
7943 * doc/invoke.texi: Document new flags.
7944 * common.opt: New flags.
7946 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
7948 * doc/tm.texi.in (Cond Exec Macros): Remove node.
7949 (Condition Code): Don't reference it.
7950 * doc/tm.texi: Regenerate.
7952 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
7955 * lra-constraints.c (undo_optional_reloads): Consider all optional
7956 reload even if it did not get a hard reg.
7958 2013-09-16 Teresa Johnson <tejohnson@google.com>
7960 * dumpfile.c (dump_loc): Remove newline emission.
7961 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
7962 emission to dump_printf_loc calls where missing.
7963 (vect_mark_for_runtime_alias_test): Ditto.
7964 (vect_analyze_data_ref_dependence): Ditto.
7965 (vect_analyze_data_ref_dependences): Ditto.
7966 (vect_slp_analyze_data_ref_dependence): Ditto.
7967 (vect_slp_analyze_data_ref_dependences): Ditto.
7968 (vect_compute_data_ref_alignment): Ditto.
7969 (vect_update_misalignment_for_peel): Ditto.
7970 (vect_verify_datarefs_alignment): Ditto.
7971 (vector_alignment_reachable_p): Ditto.
7972 (vect_get_data_access_cost): Ditto.
7973 (vect_enhance_data_refs_alignment): Ditto.
7974 (vect_find_same_alignment_drs): Ditto.
7975 (vect_analyze_data_refs_alignment): Ditto.
7976 (vect_analyze_group_access): Ditto.
7977 (vect_analyze_data_ref_access): Ditto.
7978 (vect_analyze_data_ref_accesses): Ditto.
7979 (vect_prune_runtime_alias_test_list): Ditto.
7980 (vect_analyze_data_refs): Ditto.
7981 (vect_create_addr_base_for_vector_ref): Ditto.
7982 (vect_create_data_ref_ptr): Ditto.
7983 (vect_grouped_store_supported): Ditto.
7984 (vect_grouped_load_supported): Ditto.
7985 * value-prof.c (check_counter): Ditto.
7986 (check_ic_target): Ditto.
7987 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
7988 (vect_recog_widen_mult_pattern): Ditto.
7989 (vect_recog_widen_sum_pattern): Ditto.
7990 (vect_recog_over_widening_pattern): Ditto.
7991 (vect_recog_widen_shift_pattern): Ditto.
7992 (vect_recog_rotate_pattern): Ditto.
7993 (vect_recog_vector_vector_shift_pattern): Ditto.
7994 (vect_recog_divmod_pattern): Ditto.
7995 (vect_recog_mixed_size_cond_pattern): Ditto.
7996 (vect_recog_bool_pattern): Ditto.
7997 (vect_pattern_recog_1): Ditto.
7998 (vect_pattern_recog): Ditto.
7999 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
8000 (vect_is_simple_iv_evolution): Ditto.
8001 (vect_analyze_scalar_cycles_1): Ditto.
8002 (vect_get_loop_niters): Ditto.
8003 (vect_analyze_loop_1): Ditto.
8004 (vect_analyze_loop_form): Ditto.
8005 (vect_analyze_loop_operations): Ditto.
8006 (vect_analyze_loop_2): Ditto.
8007 (vect_analyze_loop): Ditto.
8008 (report_vect_op): Ditto.
8009 (vect_is_slp_reduction): Ditto.
8010 (vect_is_simple_reduction_1): Ditto.
8011 (vect_get_known_peeling_cost): Ditto.
8012 (vect_estimate_min_profitable_iters): Ditto.
8013 (vect_model_reduction_cost): Ditto.
8014 (vect_model_induction_cost): Ditto.
8015 (get_initial_def_for_induction): Ditto.
8016 (vect_create_epilog_for_reduction): Ditto.
8017 (vectorizable_reduction): Ditto.
8018 (vectorizable_induction): Ditto.
8019 (vectorizable_live_operation): Ditto.
8020 (vect_loop_kill_debug_uses): Ditto.
8021 (vect_transform_loop): Ditto.
8022 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
8023 (vect_stmt_relevant_p): Ditto.
8024 (process_use): Ditto.
8025 (vect_mark_stmts_to_be_vectorized): Ditto.
8026 (vect_model_simple_cost): Ditto.
8027 (vect_model_promotion_demotion_cost): Ditto.
8028 (vect_model_store_cost): Ditto.
8029 (vect_get_store_cost): Ditto.
8030 (vect_model_load_cost): Ditto.
8031 (vect_get_load_cost): Ditto.
8032 (vect_init_vector_1): Ditto.
8033 (vect_get_vec_def_for_operand): Ditto.
8034 (vect_finish_stmt_generation): Ditto.
8035 (vectorizable_call): Ditto.
8036 (vectorizable_conversion): Ditto.
8037 (vectorizable_assignment): Ditto.
8038 (vectorizable_shift): Ditto.
8039 (vectorizable_operation): Ditto.
8040 (vectorizable_store): Ditto.
8041 (vectorizable_load): Ditto.
8042 (vectorizable_condition): Ditto.
8043 (vect_analyze_stmt): Ditto.
8044 (vect_transform_stmt): Ditto.
8045 (vect_is_simple_use): Ditto.
8046 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
8047 (vect_can_advance_ivs_p): Ditto.
8048 (vect_update_ivs_after_vectorizer): Ditto.
8049 (vect_do_peeling_for_loop_bound): Ditto.
8050 (vect_gen_niters_for_prolog_loop): Ditto.
8051 (vect_update_inits_of_drs): Ditto.
8052 (vect_create_cond_for_alias_checks): Ditto.
8053 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
8054 (vect_build_slp_tree_1): Ditto.
8055 (vect_supported_load_permutation_p): Ditto.
8056 (vect_analyze_slp_instance): Ditto.
8057 (vect_analyze_slp): Ditto.
8058 (vect_make_slp_decision): Ditto.
8059 (vect_detect_hybrid_slp): Ditto.
8060 (vect_bb_vectorization_profitable_p): Ditto.
8061 (vect_slp_analyze_bb_1): Ditto.
8062 (vect_update_slp_costs_according_to_vf): Ditto.
8063 (vect_get_mask_element): Ditto.
8064 (vect_transform_slp_perm_load): Ditto.
8065 (vect_schedule_slp_instance): Ditto.
8066 (vect_schedule_slp): Ditto.
8067 (vect_slp_transform_bb): Ditto.
8068 * profile.c (read_profile_edge_counts): Ditto.
8069 (compute_branch_probabilities): Ditto.
8070 * coverage.c (get_coverage_counts): Ditto.
8072 2013-09-16 Diego Novillo <dnovillo@google.com>
8074 * tree-core.h: Add missing comment lines from refactoring of tree.h.
8076 2013-09-16 Jan Hubicka <jh@suse.cz>
8078 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
8079 abstract functions; for static functions check the presence of body.
8081 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8083 * config/aarch64/aarch64-simd-builtins.def (fma): New.
8084 * config/aarch64/aarch64-simd.md
8085 (aarch64_mla_elt<mode>): New.
8086 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8087 (aarch64_mls_elt<mode>): Likewise.
8088 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8089 (aarch64_fma4_elt<mode>): Likewise.
8090 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8091 (aarch64_fma4_elt_to_128v2df): Likewise.
8092 (aarch64_fma4_elt_to_64df): Likewise.
8093 (fnma<mode>4): Likewise.
8094 (aarch64_fnma4_elt<mode>): Likewise.
8095 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8096 (aarch64_fnma4_elt_to_128v2df): Likewise.
8097 (aarch64_fnma4_elt_to_64df): Likewise.
8098 * config/aarch64/iterators.md (VDQSF): New.
8099 * config/aarch64/arm_neon.h
8100 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
8101 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
8103 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8105 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
8106 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
8107 (aarch64_mul3_elt_to_128df): Likewise.
8108 (aarch64_mul3_elt_to_64v2df): Likewise.
8109 * config/aarch64/iterators.md (VEL): Also handle DFmode.
8111 (VMUL_CHANGE_NLANES) Likewise.
8114 * config/aarch64/arm_neon.h
8115 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
8117 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8119 * config/aarch64/arm_neon.h
8120 (vcvtx_high_f32_f64): Fix parameters.
8122 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8123 Uros Bizjak <ubizjak@gmail.com>
8125 * config/alpha.c: Include tree-ssanames.h.
8127 2013-09-16 Richard Biener <rguenther@suse.de>
8129 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
8130 (dot_rdg_1): Handle control_dd.
8131 (create_edge_for_control_dependence): New function.
8132 (create_rdg_edges): Add control dependences if asked for.
8133 (build_rdg): Likewise.
8134 (generate_loops_for_partition): If there are not necessary
8135 control stmts remove all their dependencies.
8136 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
8137 (distribute_loop): Pass on control dependences.
8138 (tree_loop_distribution): Compute control dependences and remove
8139 restriction on number of loop nodes.
8141 2013-09-16 Jakub Jelinek <jakub@redhat.com>
8143 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
8146 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
8148 * cse.c (try_const_anchors): Punt on CC modes.
8150 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8152 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
8154 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
8157 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
8160 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8163 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
8165 (darwin_objc1_section): Likewise.
8166 (darwin_file_end): Emit Image Info section when required.
8168 2013-09-14 Jan Hubicka <jh@suse.cz>
8170 * tree-into-ssa.c (gate_into_ssa): New.
8171 (pass_data_build_ssa): Use it.
8172 * cgraph.h (expand_thunk): Update prototype.
8173 * cgraphunit.c (analyze_function): Expand thunks early.
8174 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
8175 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
8176 set lowered flag; do not add new function.
8177 (assemble_thunks_and_aliases): Update.
8178 * tree-ssa.c (gate_init_datastructures): New gate.
8179 (pass_data_init_datastructures): Use it.
8181 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8184 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
8185 xmm register set described in the psABI.
8187 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
8189 * dwarf2out.c (should_emit_struct_debug): Add check
8190 for type_decl variable is not NULL.
8192 2013-09-13 Jacek Caban <jacek@codeweavers.com>
8194 * config.gcc: Use new winnt-c.c target hooks
8195 * config/t-winnt: New file
8196 * config/winnt-c.c: New file
8197 * doc/tm.texi.in: Document new hook
8198 * doc/tm.texi: Regenerated
8200 2013-09-13 Jan Hubicka <jh@suse.cz>
8203 * ipa-inline.c (check_callers): New function.
8204 (check_caller_edge): Remove.
8205 (want_inline_function_to_all_callers_p): Also permit alises that are
8207 (inline_to_all_callers): Terminate the walk when devirtualization
8208 introduce new calls.
8210 2013-09-13 Jan Hubicka <jh@suse.cz>
8212 * ipa-inline-analysis.c (struct growth_data): Add node.
8213 (do_estimate_growth_1): Fix detection of recursion.
8215 2013-09-13 Jakub Jelinek <jakub@redhat.com>
8217 PR tree-optimization/58392
8218 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
8219 to avoid shadowing of outer loop variable. If
8220 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
8221 replace_by_duplicate_decl simduid of loops that have it set and
8222 set dest_cfun->has_simduid_loops and/or
8223 dest_cfun->has_force_vect_loops.
8224 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
8225 instead of maybe_lookup_decl.
8226 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
8227 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
8228 call. Copy over force_vect and copy and remap simduid. Set
8229 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
8230 (copy_cfg_body): Remove blocks_to_copy argument. Use
8231 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
8232 caller. Don't set cfun->has_simduid_loops and/or
8233 cfun->has_force_vect_loops here.
8234 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
8236 (expand_call_inline, tree_function_versioning): Adjust copy_body
8239 2013-09-13 Martin Jambor <mjambor@suse.cz>
8242 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
8243 the assert if the edge was a speculative one.
8245 2013-09-13 Richard Biener <rguenther@suse.de>
8247 * tree-data-ref.h (known_dependences_p): Move here ...
8248 * tree-loop-distribution.c (known_dependences_p): ... from here.
8249 (dump_rdg_component, debug_rdg_component): Remove.
8251 (generate_loops_for_partition): Use gimple_uid instead of
8252 relying on matching stmt visit order.
8253 (rdg_build_partitions): Take starting stmt vector.
8254 (ldist_gen): Merge into ...
8255 (distribute_loop): ... this function. Do not compute starting
8257 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
8259 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8261 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
8262 Set type attribute correctly. Set predicable_short_it attribute.
8263 (cmpsi_shiftsi): Remove %? from output template.
8265 2013-09-13 Richard Biener <rguenther@suse.de>
8267 * tree-loop-distribution.c (struct rdg_component,
8268 rdg_defs_used_in_other_loops_p, free_rdg_components,
8269 rdg_build_components): Remove.
8270 (stmts_from_loop): Do not record virtual PHIs.
8271 (generate_loops_for_partition): Skip virtual PHIs.
8272 (build_rdg_partition_for_component): Rename to ...
8273 (build_rdg_partition_for_vertex): ... this and adjust.
8274 (rdg_build_partitions): Take a vector of starting vertices
8275 instead of components. Remove unnecessary leftover handling.
8276 (ldist_gen): Do not build components or record other stores.
8277 (distribute_loop): Do not distribute loops containing stmts
8280 2013-09-13 Christian Bruel <christian.bruel@st.com>
8283 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
8285 2013-09-13 Kai Tietz <ktietz@redhat.com>
8287 * config.gcc: Separate cases for mingw and cygwin targets,
8288 and add 64-bit cygwin target case.
8290 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
8291 dll-export inline-functions.
8292 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
8295 2013-09-13 Jeff Law <law@redhat.com>
8299 2013-09-06 Jeff Law <law@redhat.com>
8301 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8302 edge implied equivalences into successor phis.
8304 2013-09-12 DJ Delorie <dj@redhat.com>
8306 * config/rl78/rl78-virt.md: Change from | to \; for asm line
8309 2013-09-12 Brooks Moses <bmoses@google.com>
8312 * Makefile.in: Do not install driver binaries in $(target)/bin.
8314 2013-09-12 DJ Delorie <dj@redhat.com>
8316 * config/rl78/rl78.opt (mrelax): New.
8317 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
8318 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
8320 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
8322 (rl78_expand_epilogue): Likewise.
8324 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
8327 * lra-eliminations.c (remove_reg_equal_offset_note): New.
8328 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
8329 pointer elimination with using remove_reg_equal_offset_note.
8331 2013-09-12 DJ Delorie <dj@redhat.com>
8333 * config/msp430/: New port.
8334 * config.gcc (msp430): Added.
8335 * doc/invoke.texi: Document MSP430 options.
8336 * doc/install.texi: Document msp430-elf
8337 * doc/md.texi: Document msp430-elf
8338 * doc/contrib.texi: Document msp430-elf
8340 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
8343 2013-09-12 Martin Jambor <mjambor@suse.cz>
8346 * ipa-prop.c (remove_described_reference): Give up if the edge in the
8347 reference descriptor is NULL.
8348 (ipa_edge_removal_hook): If owning a reference descriptor, set its
8351 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8353 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
8354 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
8355 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
8356 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
8357 make_temp_ssa_name): move to tree-ssanames.h
8358 * tree-ssa-alias.h: Move prototype.
8359 * tree-ssa.h: Include tree-ssanames.h.
8360 * tree-ssanames.c (FREE_SSANAMES): Move to here.
8361 * tree-ssanames.h: New. Move items from tree-flow*.h
8362 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
8364 2013-09-12 Richard Biener <rguenther@suse.de>
8366 PR tree-optimization/58404
8367 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
8368 propagate non-invariant addresses into dereferences wrapped
8369 in component references.
8371 2013-09-12 Richard Biener <rguenther@suse.de>
8373 PR tree-optimization/58402
8374 * passes.def: Move pass_late_warn_uninitialized later.
8376 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8378 * tree-ssa.h: New. Move content from tree-flow.h and
8380 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
8381 Move prototypes belonging to tree-ssa.c.
8382 * tree-flow-inline.h (redirect_edge_var_map_def,
8383 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
8385 * gimple.h: Adjust prototypes.
8386 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
8388 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
8389 * tree.h: Move prototype to tree-ssa.h.
8390 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
8391 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
8392 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
8393 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
8394 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
8395 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
8396 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
8397 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
8398 graphite-blocking.c, graphite-clast-to-gimple.c,
8399 graphite-dependences.c, graphite-interchange.c,
8400 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
8401 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
8402 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
8403 ipa-reference.c, ipa-split.c, ipa-utils.c,
8404 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
8405 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
8406 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
8407 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
8408 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
8409 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
8410 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
8411 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
8412 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
8413 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
8414 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
8415 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
8416 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
8417 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
8418 value-prof.c, var-tracking.c,
8419 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
8421 2013-09-12 Richard Biener <rguenther@suse.de>
8423 PR tree-optimization/58396
8424 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
8425 (build_rdg): Take a loop-nest parameter, fix memleaks.
8426 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
8428 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8430 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
8433 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
8435 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
8436 The FMA instruction names should have a 'v' prefix.
8438 2013-09-12 Richard Biener <rguenther@suse.de>
8440 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
8441 (dot_rdg): Use popen instead of system in optional code.
8442 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
8443 (already_processed_vertex_p): Adjust.
8444 (has_anti_or_output_dependence, predecessor_has_mem_write,
8445 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
8446 rdg_flag_uses): Remove.
8447 (rdg_flag_vertex): Simplify.
8448 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
8450 (build_rdg_partition_for_component): Process the first vertex
8451 of a component only.
8452 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
8454 2013-09-12 Alan Modra <amodra@gmail.com>
8456 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
8458 2013-09-11 DJ Delorie <dj@redhat.com>
8459 Nick Clifton <nickc@redhat.com>
8461 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
8462 (rl78_stack_based_mem): New.
8463 * config/rl78/constraints.md (Iv08): New.
8472 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
8474 (movsi): Change from expand to insn-and-split.
8475 (ashrsi3): Clobber AX.
8479 * config/rl78/rl78.md (CC_REG): Fix.
8480 (addsi3): Allow memory and immediate operands.
8481 (addsi3_internal): Split into...
8482 (addsi3_internal_virt): ...new, and ...
8483 (addsi3_internal_real): ...new.
8485 (subsi3_internal_virt): New.
8486 (subsi3_internal_real): New.
8487 (mulsi3): Add memory operand.
8488 (mulsi3_rl78): Likewise.
8489 (mulsi3_g13): Likewise.
8490 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
8491 (cbranchqi4_real): Add more constraint options.
8492 (cbranchhi4_real): Expand pattern.
8493 (cbranchhi4_real_signed): New.
8494 (cbranchhi4_real_inverted): New.
8495 (cbranchsi4_real_lt): New.
8496 (cbranchsi4_real_ge): New.
8497 (cbranchsi4_real_signed): New.
8498 (cbranchsi4_real): New.
8500 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
8502 (lshrsi3_virt): Likewise.
8503 (ashlsi3_virt): Likewise.
8504 (cbranchqi4_virt_signed): New.
8505 (cbranchhi4_virt_signed): New.
8506 (cbranchsi4_virt): New.
8507 * config/rl78/rl78.c: Whitespace fixes throughout.
8508 (move_elim_pass): New.
8509 (pass_data_rl78_move_elim): New.
8510 (pass_rl78_move_elim): New.
8511 (make_pass_rl78_move_elim): New.
8512 (rl78_devirt_info): Run devirt earlier.
8513 (rl78_move_elim_info): New.
8514 (rl78_asm_file_start): Register it.
8515 (rl78_split_movsi): New.
8516 (rl78_as_legitimate_address): Allow virtual base registers when
8518 (rl78_addr_space_convert): Remove spurious debug stuff.
8519 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
8520 (rl78_print_operand): More cases for not printing '#'.
8521 (rl78_expand_compare): Remove most of the logic.
8522 (content_memory): New.
8523 (clear_content_memory): New.
8524 (get_content_index): New.
8525 (get_content_name): New.
8526 (display_content_memory): New.
8527 (update_content): New.
8528 (record_content): New.
8529 (already_contains): New.
8530 (insn_ok_now): Re-recog insns with virtual registers.
8531 (add_postponed_content_update): New.
8532 (process_postponed_content_update): New.
8533 (gen_and_emit_move): New.
8534 (transcode_memory_rtx): Record new location content.
8535 Use gen_and_emit_move.
8536 (force_into_acc): New.
8537 (move_to_acc): Use gen_and_emit_move.
8538 (move_from_acc): Likewise.
8539 (move_acc_to_reg): Likewise.
8540 (move_to_x): Likewise.
8541 (move_to_hl): Likewise.
8542 (move_to_de): Likewise.
8543 (rl78_alloc_physical_registers_op1): Record location content.
8544 (has_constraint): New.
8545 (rl78_alloc_physical_registers_op2): Record location content.
8547 (rl78_alloc_physical_registers_ro1): Likewise.
8548 (rl78_alloc_physical_registers_cmp): Likewise.
8549 (rl78_alloc_physical_registers_umul): Likewise.
8550 (rl78_alloc_address_registers_macax): New.
8551 (rl78_alloc_physical_registers): Initialize and set location
8552 content memory as needed.
8553 (rl78_reorg): Make sure split2 is called.
8554 (rl78_rtx_costs): New.
8556 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
8558 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
8559 for (not (neg ...)) and (neg (not ...)) cases.
8561 2013-09-11 Richard Biener <rguenther@suse.de>
8564 * passes.def: Split critical edges before late uninit warning passes.
8565 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
8567 2013-09-11 Jakub Jelinek <jakub@redhat.com>
8569 PR tree-optimization/58385
8570 * fold-const.c (build_range_check): If both low and high are NULL,
8571 use omit_one_operand_loc to preserve exp side-effects.
8573 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8575 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
8577 2013-09-11 Richard Biener <rguenther@suse.de>
8579 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
8580 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
8581 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
8582 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
8583 stmts_from_loop, known_dependences_p, build_empty_rdg,
8584 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
8585 * tree-loop-distribution.c: ... here.
8586 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
8587 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
8588 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
8589 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
8590 * tree-loop-distribution.c: ... here.
8591 * tree-loop-distribution.c: Include gimple-pretty-print.h.
8592 (struct partition_s): Add loops member.
8593 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
8594 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
8595 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
8597 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8598 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8599 Sergey Lega <sergey.s.lega@intel.com>
8600 Anna Tikhonova <anna.tikhonova@intel.com>
8601 Ilya Tocar <ilya.tocar@intel.com>
8602 Andrey Turetskiy <andrey.turetskiy@intel.com>
8603 Ilya Verbin <ilya.verbin@intel.com>
8604 Kirill Yukhin <kirill.yukhin@intel.com>
8605 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8607 * config/i386/constraints.md (k): New.
8609 * config/i386/i386.c (const regclass_map): Add new mask registers.
8610 (dbx_register_map): Ditto.
8611 (dbx64_register_map): Ditto.
8612 (svr4_dbx_register_map): Ditto.
8613 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
8615 (ix86_preferred_reload_class): Disable constants for mask registers.
8616 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
8617 (ix86_hard_regno_mode_ok): Support new mask registers.
8618 (x86_order_regs_for_local_alloc): Ditto.
8619 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
8620 (FIXED_REGISTERS): Add new mask registers.
8621 (CALL_USED_REGISTERS): Ditto.
8622 (REG_ALLOC_ORDER): Ditto.
8623 (VALID_MASK_REG_MODE): New.
8624 (FIRST_MASK_REG): Ditto.
8625 (LAST_MASK_REG): Ditto.
8626 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
8627 (MAYBE_MASK_CLASS_P): New.
8628 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
8629 (REG_CLASS_CONTENTS): Ditto.
8630 (MASK_REGNO_P): New.
8631 (ANY_MASK_REG_P): Ditto.
8632 (HI_REGISTER_NAMES): Add new mask registers.
8633 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
8634 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
8636 (attribute "type"): Add mskmov, msklog.
8637 (attribute "length_immediate"): Support them.
8638 (attribute "memory"): Ditto.
8639 (attribute "prefix_0f"): Ditto.
8640 (*movhi_internal): Support new mask registers.
8641 (*movqi_internal): Ditto.
8642 (define_split): Split out clobber pattern is a logic
8643 insn on mask registers.
8644 (*k<logic><mode>): New.
8645 (*andhi_1): Extend to support mask regs.
8646 (*andqi_1): Extend to support mask regs.
8648 (define_split): Split and-not to and and not if operands
8650 (*<code><mode>_1): Separate HI mode to new pattern...
8651 (*<code>hi_1): This.
8652 (*<code>qi_1): Extend to support mask regs.
8654 (kortestzhi): Ditto.
8655 (kortestchi): Ditto.
8657 (*one_cmpl<mode>2_1): Remove HImode and handle it...
8658 (*one_cmplhi2_1): ...Here, now with mask registers support.
8659 (*one_cmplqi2_1): Support new mask registers.
8660 (HI/QImode arithmetics splitter): Don't split if mask registers
8662 (HI/QImode not splitter): Ditto.
8663 * config/i386/predicated.md (mask_reg_operand): New.
8664 (general_reg_operand): Ditto.
8666 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8668 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
8669 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
8671 2013-09-10 Jeff Law <law@redhat.com>
8673 PR tree-optimization/58380
8674 * tree-ssa-threadupdate.c (thread_block): Recognize another case
8675 of threading through a buried loop header.
8677 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
8678 return value for single successor case.
8680 2013-09-10 Jan Hubicka <jh@suse.cz>
8682 * ipa-devirt.c (ipa_devirt): Enable with LTO.
8684 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
8687 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
8688 support conditional execution.
8689 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
8691 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
8693 * lra.c (lra): Clear lra_optional_reload_pseudos before every
8695 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
8696 Check destination too to check move insn.
8697 (undo_optional_reloads): Add check that the original peudo did not
8698 changed its allocation and the optional reload was inherited on last
8699 inheritance pass. Break loop after deciding to keep optional reload.
8700 (lra_undo_inheritance): Add check that inherited pseudo still in
8703 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
8705 * config/aarch64/aarch64.md (generic_sched): New.
8706 * config/aarch64/aarch64-generic.md (load): Make conditional
8707 on generic_sched attribute.
8708 (nonload): Likewise.
8710 2013-09-10 Jan Hubicka <jh@suse.cz>
8712 * lto-cgraph.c: Include ipa-utils.h.
8713 (compute_ltrans_boundary): Also add possible targets into the boundary.
8715 2013-09-10 Jan Hubicka <jh@suse.cz>
8717 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
8718 VAR_DECL of vtable rather than full expression.
8720 2013-09-10 Jan Hubicka <jh@suse.cz>
8721 Paolo Carlini <paolo.carlini@oracle.com>
8723 * cgraphunit.c (analyze_functions): Save input_location, set it
8724 to UNKNOWN_LOCATION and restore it at the end.
8726 2013-09-10 Martin Jambor <mjambor@suse.cz>
8728 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
8729 represented by a thunk.
8731 2013-09-10 Jeff Law <law@redhat.com>
8733 PR tree-optimization/58343
8734 * tree-ssa-threadupdate.c (thread_block): Identify and disable
8735 jump threading requests through loop headers buried in the middle
8736 of a jump threading path.
8738 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
8739 in return value/type.
8741 2013-09-10 Jakub Jelinek <jakub@redhat.com>
8743 PR rtl-optimization/58365
8744 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
8745 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
8748 2013-09-10 Richard Biener <rguenther@suse.de>
8750 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
8751 * tree-data-ref.c (create_rdg_vertices): Collect all data
8752 references, signal failure to the caller, use data-ref API.
8753 (build_rdg): Compute data references only once. Maintain lifetime
8754 of data references and data dependences from within RDG.
8755 (free_rdg): Free dependence relations.
8756 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
8757 inventing extra dependences.
8758 (distribute_loop): Update for RDG API changes.
8760 2013-09-10 Kai Tietz <ktietz@redhat.com>
8762 * doc/invoke.texi (fms-extensions): Document changed
8763 behavior for ms-abi targets.
8764 * config/i386/i386.c (ix86_option_override_internal):
8765 Set default value of option -fms-extension for ms-abi targets.
8767 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
8769 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
8771 2013-09-10 Alan Modra <amodra@gmail.com>
8774 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
8776 2013-09-10 Alan Modra <amodra@gmail.com>
8778 * config/rs6000/predicates.md (add_cint_operand): New.
8779 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
8780 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
8781 using add_cint_operand.
8782 (largetoc_high_plus_aix): Likewise.
8784 2013-09-09 Jakub Jelinek <jakub@redhat.com>
8786 PR tree-optimization/58364
8787 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
8788 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
8789 the current range can't be an unconditional true or false.
8791 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
8793 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
8795 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
8797 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
8799 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
8802 * doc/invoke.texi (-Wdelete-incomplete): Document it.
8804 2013-09-09 Ian Bolton <ian.bolton@arm.com>
8806 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
8807 NO_REGS for immediate that can't be moved directly into FP_REGS.
8809 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8811 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
8812 comparison with negated operand.
8813 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
8816 2013-09-09 Richard Biener <rguenther@suse.de>
8819 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
8820 of a subloop record all its block as affecting loop-closed SSA form.
8822 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8824 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
8825 of an rtx/bitpos pair.
8826 (store_fixed_bit_field): Update accordingly.
8828 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8830 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
8832 * builtins.c (expand_errno_check): Likewise.
8833 * dwarf2cfi.c (init_return_column_size): Likewise.
8834 * except.c (sjlj_mark_call_sites): Likewise.
8835 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
8836 * lra-constraints.c (emit_inc): Likewise.
8837 * ree.c (combine_set_extension): Likewise.
8838 * regmove.c (fixup_match_2): Likewise.
8839 * reload1.c (inc_for_reload): Likewise.
8841 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8843 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
8844 (force_to_mode, simplify_shift_const_1, simplify_comparison):
8845 Use gen_int_mode with the mode of the associated simplify_* call.
8846 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
8847 * expmed.c (expand_shift_1): Likewise.
8848 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8849 * loop-iv.c (iv_number_of_iterations): Likewise.
8850 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8851 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
8853 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8855 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
8856 of the associated expand_* call.
8857 (asan_emit_stack_protection): Likewise.
8858 * builtins.c (round_trampoline_addr): Likewise.
8859 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
8860 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
8861 (emit_store_flag): Likewise.
8862 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
8863 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
8865 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8866 * ifcvt.c (noce_try_store_flag_constants): Likewise.
8867 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8868 * modulo-sched.c (generate_prolog_epilog): Likewise.
8869 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
8870 (expand_ctz, expand_ffs, expand_unop): Likewise.
8872 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8874 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
8875 of the associated gen_rtx_* call.
8876 * caller-save.c (init_caller_save): Likewise.
8877 * combine.c (find_split_point, make_extraction): Likewise.
8878 (make_compound_operation): Likewise.
8879 * dwarf2out.c (mem_loc_descriptor): Likewise.
8880 * explow.c (plus_constant, probe_stack_range): Likewise.
8881 * expmed.c (expand_mult_const): Likewise.
8882 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
8883 * reload1.c (init_reload): Likewise.
8884 * valtrack.c (cleanup_auto_inc_dec): Likewise.
8885 * var-tracking.c (adjust_mems): Likewise.
8886 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
8887 rather than gen_rtx_fmt_ee.
8889 2013-09-09 Jan Hubicka <jh@suse.cz>
8892 * value-prof.c (gimple_ic): Copy also abnormal edges.
8894 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8896 * asan.c (asan_shadow_cst): Use gen_int_mode.
8898 2013-09-08 Jan Hubicka <jh@suse.cz>
8900 * ipa-profile.c: Add toplevel comment.
8901 (ipa_propagate_frequency_1): Be more conservative when profile is read.
8902 (contains_hot_call_p): New function.
8903 (ipa_propagate_frequency): Set frequencies based on counts when
8905 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
8906 profile; do not tamper with profile after inlining if it is read.
8908 2013-09-08 Jan Hubicka <jh@suse.cz>
8910 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
8913 2013-09-08 Jan Hubicka <jh@suse.cz>
8915 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
8918 2013-09-08 Jeff Law <law@redhat.com>
8921 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
8924 2013-09-08 Andi Kleen <ak@linux.intel.com>
8926 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
8928 2013-09-08 Jan Hubicka <jh@suse.cz>
8930 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
8932 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
8934 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
8935 for non-debug insns.
8936 * lra.c (new_insn_reg): Take the containing insn as a parameter.
8937 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
8938 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
8941 2013-09-08 Jan Hubicka <jh@suse.cz>
8943 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
8944 targets and devirtualize to BUILT_IN_UNREACHABLE.
8945 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
8946 * ipa.c (walk_polymorphic_call_targets): New function.
8947 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
8948 functions; use the new timevar.
8949 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
8950 was removed from the program.
8951 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
8952 not consider it in the walk when its vtable is dead.
8953 (possible_polymorphic_call_targets_1): Pass anonymous flag to
8955 (devirt_variable_node_removal_hook): New function.
8956 (possible_polymorphic_call_targets): Also register
8957 devirt_variable_node_removal_hook.
8958 (ipa_devirt): Do not do non-speculative devirtualization.
8959 (gate_ipa_devirt): One execute if devirtualizing speculatively.
8961 2013-09-08 Jan Hubicka <jh@suse.cz>
8963 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
8964 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
8965 varpool_remove_variable_insertion_hook): Declare.
8966 * varpool.c (varpool_node_hook_list): New structure.
8967 (first_varpool_node_removal_hook,
8968 first_varpool_variable_insertion_hook): New variables.
8969 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
8970 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
8971 varpool_remove_variable_insertion_hook,
8972 varpool_call_variable_insertion_hooks): New functions.
8973 (varpool_remove_node): Use it.
8975 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
8978 * diagnostic.c (diagnostic_build_prefix): When s.file is
8979 "<built-in>" don't output line and column numbers.
8981 2013-09-06 Jan Hubicka <jh@suse.cz>
8983 * cgraphunit.c (expand_thunk): Get body before touching arguments.
8984 * lto-streamer-out.c: Stream thunks, too.
8985 * lto-streamer-in.c (input_function): Pop cfun here
8986 (lto_read_body): Instead of here.
8988 2013-09-06 Caroline Tice <cmtice@google.com>
8990 * doc/install.texi: Add documentation for the --enable-vtable-verify
8991 and the --disable-libvtv configure options.
8993 2013-09-06 Jeff Law <law@redhat.com>
8995 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8996 edge implied equivalences into successor phis.
8998 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
9000 * resource.c (mark_referenced_resources): Handle COND_EXEC.
9002 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
9004 * resource.c (mark_target_live_regs): Compute resources taking
9005 into account if a call is predicated or not.
9007 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9009 * toplev.c (output_stack_usage): Be prepared for suffixes created by
9010 the compiler in the function names.
9012 2013-09-06 Jan Hubicka <jh@suse.cz>
9015 * ipa-inline.c (has_caller_p): New function.
9016 (want_inline_function_to_all_callers_p): Use it.
9017 (sum_callers, inline_to_all_callers): Break out from ...
9018 (ipa_inline): ... here.
9020 2013-09-06 Jan Hubicka <jh@suse.cz>
9022 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
9023 only when AVX is enabled.
9025 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9027 * config/aarch64/aarch64.md
9028 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
9029 is fpsimd_<load/store>2.
9030 (load_pair<mode>): Likewise.
9031 (store_pair<mode>): Likewise.
9033 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9035 * config/arm/types.md (type): Add "mrs" type.
9036 * config/aarch64/aarch64.md
9037 (aarch64_load_tp_hard): Make type "mrs".
9039 (load_tp_hard): Make type "mrs".
9040 * config/arm/cortex-a15.md: Update with new attributes.
9041 * config/arm/cortex-a5.md: Update with new attributes.
9042 * config/arm/cortex-a53.md: Update with new attributes.
9043 * config/arm/cortex-a7.md: Update with new attributes.
9044 * config/arm/cortex-a8.md: Update with new attributes.
9045 * config/arm/cortex-a9.md: Update with new attributes.
9046 * config/arm/cortex-m4.md: Update with new attributes.
9047 * config/arm/cortex-r4.md: Update with new attributes.
9048 * config/arm/fa526.md: Update with new attributes.
9049 * config/arm/fa606te.md: Update with new attributes.
9050 * config/arm/fa626te.md: Update with new attributes.
9051 * config/arm/fa726te.md: Update with new attributes.
9053 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9055 * config/aarch64/aarch64.md
9056 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
9057 (*movtf_aarch64): Likewise.
9059 (thumb1_movdi_insn): Use "multiple" for type where more than one
9060 instruction is used for a move.
9061 (*arm32_movhf): Likewise.
9062 (*thumb_movdf_insn): Likewise.
9064 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9066 * config/arm/types.md (type): Rename fcpys to fmov.
9068 (*arm_movsi_vfp): Rename type fcpys as fmov.
9069 (*thumb2_movsi_vfp): Likewise
9070 (*movhf_vfp_neon): Likewise
9071 (*movhf_vfp): Likewise
9072 (*movsf_vfp): Likewise
9073 (*thumb2_movsf_vfp): Likewise
9074 (*movsfcc_vfp): Likewise
9075 (*thumb2_movsfcc_vfp): Likewise
9076 * config/aarch64/aarch64-simd.md
9077 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
9078 * config/aarch64/aarch64.md
9079 (*movsi_aarch64): Replace type mov_reg with fmovs.
9080 (*movdi_aarch64): Likewise
9081 (*movsf_aarch64): Likewise
9082 (*movdf_aarch64): Likewise
9084 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
9085 * config/arm/iwmmxt.md
9086 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
9087 * config/arm/arm1020e.md: Update with new attributes.
9088 * config/arm/cortex-a15-neon.md: Update with new attributes.
9089 * config/arm/cortex-a5.md: Update with new attributes.
9090 * config/arm/cortex-a53.md: Update with new attributes.
9091 * config/arm/cortex-a7.md: Update with new attributes.
9092 * config/arm/cortex-a8-neon.md: Update with new attributes.
9093 * config/arm/cortex-a9.md: Update with new attributes.
9094 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9095 * config/arm/cortex-r4f.md: Update with new attributes.
9096 * config/arm/marvell-pj4.md: Update with new attributes.
9097 * config/arm/vfp11.md: Update with new attributes.
9099 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9101 * config/aarch64/aarch64.md
9102 (*madd<mode>): Fix type attribute.
9103 (*maddsi_uxtw): Likewise.
9104 (*msub<mode>): Likewise.
9105 (*msubsi_uxtw): Likewise.
9106 (<su_optab>maddsidi4): Likewise.
9107 (<su_optab>msubsidi4): Likewise.
9109 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9111 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
9112 * config/arm/arm.md (core_cycles): Remove fdiv.
9113 * config/arm/vfp.md:
9114 (*sqrtsf2_vfp): Update for attribute changes.
9115 (*sqrtdf2_vfp): Likewise.
9116 * config/aarch64/aarch64.md:
9117 (sqrt<mode>2): Update for attribute changes.
9118 * config/arm/arm1020e.md: Update with new attributes.
9119 * config/arm/cortex-a15-neon.md: Update with new attributes.
9120 * config/arm/cortex-a5.md: Update with new attributes.
9121 * config/arm/cortex-a53.md: Update with new attributes.
9122 * config/arm/cortex-a7.md: Update with new attributes.
9123 * config/arm/cortex-a8-neon.md: Update with new attributes.
9124 * config/arm/cortex-a9.md: Update with new attributes.
9125 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9126 * config/arm/cortex-r4f.md: Update with new attributes.
9127 * config/arm/marvell-pj4.md: Update with new attributes.
9128 * config/arm/vfp11.md: Update with new attributes.
9130 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9132 * config/arm/types.md
9133 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
9134 * config/aarch64/aarch64.md
9135 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
9137 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
9138 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
9139 (float<GPI:mode><GPF:mode>2): Likewise.
9141 (*truncsisf2_vfp): Update with new attributes.
9142 (*truncsidf2_vfp): Likewise.
9143 (fixuns_truncsfsi2): Likewise.
9144 (fixuns_truncdfsi2): Likewise.
9145 (*floatsisf2_vfp): Likewise.
9146 (*floatsidf2_vfp): Likewise.
9147 (floatunssisf2): Likewise.
9148 (floatunssidf2): Likewise.
9149 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
9150 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
9151 * config/arm/arm1020e.md: Update with new attributes.
9152 * config/arm/cortex-a15-neon.md: Update with new attributes.
9153 * config/arm/cortex-a5.md: Update with new attributes.
9154 * config/arm/cortex-a53.md: Update with new attributes.
9155 * config/arm/cortex-a7.md: Update with new attributes.
9156 * config/arm/cortex-a8-neon.md: Update with new attributes.
9157 * config/arm/cortex-a9.md: Update with new attributes.
9158 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9159 * config/arm/cortex-r4f.md: Update with new attributes.
9160 * config/arm/marvell-pj4.md: Update with new attributes.
9161 * config/arm/vfp11.md: Update with new attributes.
9163 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9165 * config/aarch64/arm_neon.h
9166 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
9167 (vqtbx<1,2,3,4><q>_s8): Likewise.
9169 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9171 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
9172 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
9173 (add<mode>3): Add type attribute.
9174 (add<mode>3): Likewise.
9175 (usadd<mode>3): Likewise.
9176 (ssadd<mode>3): Likewise.
9177 (sub<mode>3): Likewise.
9178 (sub<mode>3): Likewise.
9179 (ussub<mode>3): Likewise.
9180 (sssub<mode>3): Likewise.
9181 (ssmulsa3): Likewise.
9182 (usmulusa3): Likewise.
9183 (arm_usatsihi): Likewise.
9185 (*movdi_vfp): Add types for all instructions.
9186 (*movdi_vfp_cortexa8): Likewise.
9187 (*movhf_vfp_neon): Likewise.
9188 (*movhf_vfp): Likewise.
9189 (*movdf_vfp): Likewise.
9190 (*thumb2_movdf_vfp): Likewise.
9191 (*thumb2_movdfcc_vfp): Likewise.
9192 * config/arm/arm.md: Add type attribute to all insn patterns.
9193 (*thumb1_adddi3): Add type attribute.
9194 (*arm_adddi3): Likewise.
9195 (*adddi_sesidi_di): Likewise.
9196 (*adddi_zesidi_di): Likewise.
9197 (*thumb1_addsi3): Likewise.
9198 (addsi3_compare0): Likewise.
9199 (*addsi3_compare0_scratch): Likewise.
9200 (*compare_negsi_si): Likewise.
9201 (cmpsi2_addneg): Likewise.
9202 (*addsi3_carryin_<optab>): Likewise.
9203 (*addsi3_carryin_alt2_<optab>): Likewise.
9204 (*addsi3_carryin_clobercc_<optab>): Likewise.
9205 (*subsi3_carryin): Likewise.
9206 (*subsi3_carryin_const): Likewise.
9207 (*subsi3_carryin_compare): Likewise.
9208 (*subsi3_carryin_compare_const): Likewise.
9209 (*arm_subdi3): Likewise.
9210 (*thumb_subdi3): Likewise.
9211 (*subdi_di_zesidi): Likewise.
9212 (*subdi_di_sesidi): Likewise.
9213 (*subdi_zesidi_di): Likewise.
9214 (*subdi_sesidi_di): Likewise.
9215 (*subdi_zesidi_ze): Likewise.
9216 (thumb1_subsi3_insn): Likewise.
9217 (*arm_subsi3_insn): Likewise.
9218 (*anddi3_insn): Likewise.
9219 (*anddi_zesidi_di): Likewise.
9220 (*anddi_sesdi_di): Likewise.
9221 (*ne_zeroextracts): Likewise.
9222 (*ne_zeroextracts): Likewise.
9223 (*ite_ne_zeroextr): Likewise.
9224 (*ite_ne_zeroextr): Likewise.
9225 (*anddi_notdi_di): Likewise.
9226 (*anddi_notzesidi): Likewise.
9227 (*anddi_notsesidi): Likewise.
9228 (andsi_notsi_si): Likewise.
9229 (thumb1_bicsi3): Likewise.
9230 (*iordi3_insn): Likewise.
9231 (*iordi_zesidi_di): Likewise.
9232 (*iordi_sesidi_di): Likewise.
9233 (*thumb1_iorsi3_insn): Likewise.
9234 (*xordi3_insn): Likewise.
9235 (*xordi_zesidi_di): Likewise.
9236 (*xordi_sesidi_di): Likewise.
9237 (*arm_xorsi3): Likewise.
9238 (*andsi_iorsi3_no): Likewise.
9239 (*smax_0): Likewise.
9240 (*smax_m1): Likewise.
9241 (*arm_smax_insn): Likewise.
9242 (*smin_0): Likewise.
9243 (*arm_smin_insn): Likewise.
9244 (*arm_umaxsi3): Likewise.
9245 (*arm_uminsi3): Likewise.
9246 (*minmax_arithsi): Likewise.
9247 (*minmax_arithsi_): Likewise.
9248 (*satsi_<SAT:code>): Likewise.
9249 (arm_ashldi3_1bit): Likewise.
9250 (arm_ashrdi3_1bit): Likewise.
9251 (arm_lshrdi3_1bit): Likewise.
9252 (*arm_negdi2): Likewise.
9253 (*thumb1_negdi2): Likewise.
9254 (*arm_negsi2): Likewise.
9255 (*thumb1_negsi2): Likewise.
9256 (*negdi_extendsid): Likewise.
9257 (*negdi_zero_extend): Likewise.
9258 (*arm_abssi2): Likewise.
9259 (*thumb1_abssi2): Likewise.
9260 (*arm_neg_abssi2): Likewise.
9261 (*thumb1_neg_abss): Likewise.
9262 (one_cmpldi2): Likewise.
9263 (extend<mode>di2): Likewise.
9264 (*compareqi_eq0): Likewise.
9265 (*arm_extendhisi2addsi): Likewise.
9266 (*arm_movdi): Likewise.
9267 (*thumb1_movdi_insn): Likewise.
9268 (*arm_movt): Likewise.
9269 (*thumb1_movsi_insn): Likewise.
9270 (pic_add_dot_plus_four): Likewise.
9271 (pic_add_dot_plus_eight): Likewise.
9272 (tls_load_dot_plus_eight): Likewise.
9273 (*thumb1_movhi_insn): Likewise.
9274 (*thumb1_movsf_insn): Likewise.
9275 (*movdf_soft_insn): Likewise.
9276 (*thumb_movdf_insn): Likewise.
9277 (cbranchsi4_insn): Likewise.
9278 (cbranchsi4_scratch): Likewise.
9279 (*negated_cbranchsi4): Likewise.
9280 (*tbit_cbranch): Likewise.
9281 (*tlobits_cbranch): Likewise.
9282 (*tstsi3_cbranch): Likewise.
9283 (*cbranchne_decr1): Likewise.
9284 (*addsi3_cbranch): Likewise.
9285 (*addsi3_cbranch_scratch): Likewise.
9286 (*arm_cmpdi_insn): Likewise.
9287 (*arm_cmpdi_unsig): Likewise.
9288 (*arm_cmpdi_zero): Likewise.
9289 (*thumb_cmpdi_zero): Likewise.
9290 (*deleted_compare): Likewise.
9291 (*mov_scc): Likewise.
9292 (*mov_negscc): Likewise.
9293 (*mov_notscc): Likewise.
9294 (*cstoresi_eq0_thumb1_insn): Likewise.
9295 (cstoresi_nltu_thumb1): Likewise.
9296 (cstoresi_ltu_thu): Likewise.
9297 (thumb1_addsi3_addgeu): Likewise.
9298 (*arm_jump): Likewise.
9299 (*thumb_jump): Likewise.
9300 (*check_arch2): Likewise.
9301 (arm_casesi_internal): Likewise.
9302 (thumb1_casesi_dispatch): Likewise.
9303 (*arm_indirect_jump): Likewise.
9304 (*thumb1_indirect_jump): Likewise.
9306 (*and_scc): Likewise.
9307 (*ior_scc): Likewise.
9308 (*compare_scc): Likewise.
9309 (*cond_move): Likewise.
9310 (*cond_arith): Likewise.
9311 (*cond_sub): Likewise.
9312 (*cmp_ite0): Likewise.
9313 (*cmp_ite1): Likewise.
9314 (*cmp_and): Likewise.
9315 (*cmp_ior): Likewise.
9316 (*ior_scc_scc): Likewise.
9317 (*ior_scc_scc_cmp): Likewise.
9318 (*and_scc_scc): Likewise.
9319 (*and_scc_scc_cmp): Likewise.
9320 (*and_scc_scc_nod): Likewise.
9321 (*negscc): Likewise.
9322 (movcond_addsi): Likewise.
9323 (movcond): Likewise.
9324 (*ifcompare_plus_move): Likewise.
9325 (*if_plus_move): Likewise.
9326 (*ifcompare_move_plus): Likewise.
9327 (*if_move_plus): Likewise.
9328 (*ifcompare_arith_arith): Likewise.
9329 (*if_arith_arith): Likewise.
9330 (*ifcompare_arith_move): Likewise.
9331 (*if_arith_move): Likewise.
9332 (*ifcompare_move_arith): Likewise.
9333 (*if_move_arith): Likewise.
9334 (*ifcompare_move_not): Likewise.
9335 (*if_move_not): Likewise.
9336 (*ifcompare_not_move): Likewise.
9337 (*if_not_move): Likewise.
9338 (*ifcompare_shift_move): Likewise.
9339 (*if_shift_move): Likewise.
9340 (*ifcompare_move_shift): Likewise.
9341 (*if_move_shift): Likewise.
9342 (*ifcompare_shift_shift): Likewise.
9343 (*ifcompare_not_arith): Likewise.
9344 (*ifcompare_arith_not): Likewise.
9345 (*if_arith_not): Likewise.
9346 (*ifcompare_neg_move): Likewise.
9347 (*if_neg_move): Likewise.
9348 (*ifcompare_move_neg): Likewise.
9349 (*if_move_neg): Likewise.
9350 (prologue_thumb1_interwork): Likewise.
9351 (*cond_move_not): Likewise.
9352 (*sign_extract_onebit): Likewise.
9353 (*not_signextract_onebit): Likewise.
9354 (stack_tie): Likewise.
9355 (align_4): Likewise.
9356 (align_8): Likewise.
9357 (consttable_end): Likewise.
9358 (consttable_1): Likewise.
9359 (consttable_2): Likewise.
9360 (consttable_4): Likewise.
9361 (consttable_8): Likewise.
9362 (consttable_16): Likewise.
9363 (*thumb1_tablejump): Likewise.
9364 (prefetch): Likewise.
9365 (force_register_use): Likewise.
9366 (thumb_eh_return): Likewise.
9367 (load_tp_hard): Likewise.
9368 (load_tp_soft): Likewise.
9369 (tlscall): Likewise.
9370 (*arm_movtas_ze): Likewise.
9371 (*arm_rev): Likewise.
9372 (*arm_revsh): Likewise.
9373 (*arm_rev16): Likewise.
9374 * config/arm/thumb2.md
9375 (*thumb2_smaxsi3): Likewise.
9376 (*thumb2_sminsi3): Likewise.
9377 (*thumb32_umaxsi3): Likewise.
9378 (*thumb2_uminsi3): Likewise.
9379 (*thumb2_negdi2): Likewise.
9380 (*thumb2_abssi2): Likewise.
9381 (*thumb2_neg_abss): Likewise.
9382 (*thumb2_movsi_insn): Likewise.
9383 (tls_load_dot_plus_four): Likewise.
9384 (*thumb2_movhi_insn): Likewise.
9385 (*thumb2_mov_scc): Likewise.
9386 (*thumb2_mov_negs): Likewise.
9387 (*thumb2_mov_negs): Likewise.
9388 (*thumb2_mov_nots): Likewise.
9389 (*thumb2_mov_nots): Likewise.
9390 (*thumb2_movsicc_): Likewise.
9391 (*thumb2_movsfcc_soft_insn): Likewise.
9392 (*thumb2_indirect_jump): Likewise.
9393 (*thumb2_and_scc): Likewise.
9394 (*thumb2_ior_scc): Likewise.
9395 (*thumb2_ior_scc_strict_it): Likewise.
9396 (*thumb2_cond_move): Likewise.
9397 (*thumb2_cond_arith): Likewise.
9398 (*thumb2_cond_ari): Likewise.
9399 (*thumb2_cond_sub): Likewise.
9400 (*thumb2_negscc): Likewise.
9401 (*thumb2_movcond): Likewise.
9402 (thumb2_casesi_internal): Likewise.
9403 (thumb2_casesi_internal_pic): Likewise.
9404 (*thumb2_alusi3_short): Likewise.
9405 (*thumb2_mov<mode>_shortim): Likewise.
9406 (*thumb2_addsi_short): Likewise.
9407 (*thumb2_subsi_short): Likewise.
9408 (thumb2_addsi3_compare0): Likewise.
9409 (*thumb2_cbz): Likewise.
9410 (*thumb2_cbnz): Likewise.
9411 (*thumb2_one_cmplsi2_short): Likewise.
9412 (*thumb2_negsi2_short): Likewise.
9413 (*orsi_notsi_si): Likewise.
9414 * config/arm/arm1020e.md: Update with new attributes.
9415 * config/arm/arm1026ejs.md: Update with new attributes.
9416 * config/arm/arm1136jfs.md: Update with new attributes.
9417 * config/arm/arm926ejs.md: Update with new attributes.
9418 * config/arm/cortex-a15.md: Update with new attributes.
9419 * config/arm/cortex-a5.md: Update with new attributes.
9420 * config/arm/cortex-a53.md: Update with new attributes.
9421 * config/arm/cortex-a7.md: Update with new attributes.
9422 * config/arm/cortex-a8.md: Update with new attributes.
9423 * config/arm/cortex-a9.md: Update with new attributes.
9424 * config/arm/cortex-m4.md: Update with new attributes.
9425 * config/arm/cortex-r4.md: Update with new attributes.
9426 * config/arm/fa526.md: Update with new attributes.
9427 * config/arm/fa606te.md: Update with new attributes.
9428 * config/arm/fa626te.md: Update with new attributes.
9429 * config/arm/fa726te.md: Update with new attributes.
9431 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9433 * config/aarch64/aarch64-simd.md
9434 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
9435 <vwx> iterator to ensure correct register choice.
9436 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
9437 (aarch64_sqdmull_n<mode>): Likewise.
9438 (aarch64_sqdmull2_n<mode>_internal): Likewise.
9439 * config/aarch64/arm_neon.h
9440 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
9441 (vml<as><q>_n_<su>16): Likewise.
9442 (vml<as>l_high_lane<q>_<su>16): Likewise.
9443 (vml<as>l_high_n_<su>16): Likewise.
9444 (vml<as>l_lane<q>_<su>16): Likewise.
9445 (vml<as>l_n_<su>16): Likewise.
9446 (vmul<q>_lane<q>_<su>16): Likewise.
9447 (vmul<q>_n_<su>16): Likewise.
9448 (vmull_lane<q>_<su>16): Likewise.
9449 (vmull_n_<su>16): Likewise.
9450 (vmull_high_lane<q>_<su>16): Likewise.
9451 (vmull_high_n_<su>16): Likewise.
9452 (vqrdmulh<q>_n_s16): Likewise.
9454 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
9456 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
9457 have the correct lane parameter.
9459 2013-09-06 Richard Biener <rguenther@suse.de>
9461 * cfganal.c (control_dependences::~control_dependences):
9462 Properly free all of the vector.
9464 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
9467 * config/i386/i386.c (ix86_conditional_register_usage):
9468 Proper initialize extended SSE registers.
9470 2013-09-06 Jan Hubicka <jh@suse.cz>
9472 PR tree-optimization/58311
9473 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
9475 2013-09-06 Jan Hubicka <jh@suse.cz>
9477 * Makefile.in (tree-sra.o): Update dependencies.
9478 * tree-sra.c: Include ipa-utils.h
9479 (scan_function): Use recursive_call_p.
9480 (has_caller_p): New function.
9481 (cgraph_for_node_and_aliases): Count also callers of aliases.
9483 2013-09-06 Jan Hubicka <jh@suse.cz>
9486 * cgraph.h (symtab_semantically_equivalent_p): Declare.
9487 * tree-tailcall.c: Include ipa-utils.h.
9488 (find_tail_calls): Use it.
9489 * ipa-pure-const.c (check_call): Likewise.
9490 * ipa-utils.c (recursive_call_p): New function.
9491 * ipa-utils.h (recursive_call_p): Dclare.
9492 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
9493 (symtab_semantically_equivalent_p): New function.
9494 * Makefile.in (tree-tailcall.o): Update dependencies.
9496 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9498 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
9501 2013-09-06 Richard Biener <rguenther@suse.de>
9503 * lto-streamer.h (lto_global_var_decls): Remove.
9504 * Makefile.in (OBJS): Remove lto-symtab.o.
9505 (lto-symtab.o): Remove.
9506 (GTFILES): Remove lto-symtab.c
9507 * lto-symtab.c: Move to lto/
9509 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9511 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
9512 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
9513 (UNSPEC_FPINT_RINT): New constant definitions.
9514 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
9515 definition with 2 attributes.
9516 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
9517 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
9520 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9522 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
9523 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
9524 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
9525 attribute to "z196_alone".
9526 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
9529 2013-09-06 Richard Biener <rguenther@suse.de>
9531 * basic-block.h (class control_dependences): New.
9532 * tree-ssa-dce.c (control_dependence_map): Remove.
9534 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
9535 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
9536 find_pdom, find_control_dependence, find_all_control_dependences):
9538 (mark_control_dependent_edges_necessary,
9539 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
9540 tree_dce_done, perform_tree_ssa_dce): Adjust.
9541 * cfganal.c (set_control_dependence_map_bit,
9542 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
9543 find_all_control_dependences): Move from tree-ssa-dce.c and
9544 implement as methods of control_dependences class.
9545 (control_dependences::control_dependences): New.
9546 (control_dependences::~control_dependences): Likewise.
9547 (control_dependences::get_edges_dependent_on): Likewise.
9548 (control_dependences::get_edge): Likewise.
9550 2013-09-04 Jan Hubicka <jh@suse.cz>
9552 * tree.c (types_same_for_odr): Drop overactive check.
9553 * ipa-devirt.c (hash_type_name): Likewise.
9555 2013-09-04 Jan Hubicka <jh@suse.cz>
9557 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
9558 (analyze_functions): ... here.
9560 2013-09-04 Jan Hubicka <jh@suse.cz>
9563 * cgraphunit.c (analyze_functions): Clear AUX fields
9564 after processing; initialize assembler name has.
9566 2013-09-05 Jeff Law <law@redhat.com>
9568 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
9569 from thread_around_empty_block. Record threading path into PATH.
9570 Recurse if threading through the initial block is successful.
9571 (thread_across_edge): Corresponding changes to slightly simplify.
9573 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9575 * config/aarch64/aarch64.md
9576 (type): Remove frecpe, frecps, frecpx.
9577 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
9578 fix to be a TARGET_SIMD instruction.
9579 (aarch64_frecps): Remove.
9580 * config/aarch64/aarch64-simd.md
9581 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
9582 (aarch64_frecps<mode>): Handle all float/vector of float modes.
9584 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9585 Sofiane Naci <sofiane.naci@arm.com>
9587 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
9588 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
9589 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
9590 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
9591 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
9592 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
9593 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
9594 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
9595 into "clz, "rbit". Rename "shift" to "shift_imm".
9596 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
9597 changes. Update for attribute changes all occurrences of arlo_* and
9599 * config/arm/arm-fixed.md: Update for attribute changes
9600 all occurrences of arlo_* types.
9601 * config/arm/thumb2.md: Update for attribute changes all occurrences
9603 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
9604 (cortexa7_older_only): Likewise.
9605 (cortexa7_younger): Likewise.
9606 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9607 (1020alu_shift_op): Likewise.
9608 (1020alu_shift_reg_op): Likewise.
9609 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9610 (alu_shift_op): Likewise.
9611 (alu_shift_reg_op): Likewise.
9612 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9613 (11_alu_shift_op): Likewise.
9614 (11_alu_shift_reg_op): Likewise.
9615 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9616 (9_alu_shift_reg_op): Likewise.
9617 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
9619 (cortex_a15_alu_shift): Likewise.
9620 (cortex_a15_alu_shift_reg): Likewise.
9621 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
9623 (cortex_a5_alu_shift): Likewise.
9624 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
9626 (cortex_a53_alu_shift): Likewise.
9627 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
9629 (cortex_a7_alu_reg): Likewise.
9630 (cortex_a7_alu_shift): Likewise.
9631 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
9633 (cortex_a8_alu_shift): Likewise.
9634 (cortex_a8_alu_shift_reg): Likewise.
9635 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9636 (cortex_a9_dp_shift): Likewise.
9637 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
9639 * config/arm/cortex-r4.md
9640 (cortex_r4_alu): Update for attribute changes.
9641 (cortex_r4_mov): Likewise.
9642 (cortex_r4_alu_shift_reg): Likewise.
9643 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9644 (526_alu_shift_op): Likewise.
9645 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9646 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9647 (626te_alu_shift_op): Likewise.
9648 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
9649 (726te_alu_shift_op): Likewise.
9650 (726te_alu_shift_reg_op): Likewise.
9651 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9652 (mp626_alu_shift_op): Likewise.
9653 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
9654 (pj4_alu_conds): Likewise.
9655 (pj4_shift): Likewise.
9656 (pj4_shift_conds): Likewise.
9657 (pj4_alu_shift): Likewise.
9658 (pj4_alu_shift_conds): Likewise.
9659 * config/aarch64/aarch64.md: Update for attribute change
9660 all occurrences of arlo_* and shift* types.
9662 2013-09-05 Mike Stump <mikestump@comcast.net>
9664 * tree.h: Move documentation for tree_function_decl to tree-core.h
9665 with the declaration.
9667 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
9670 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
9671 looking for widest mode.
9673 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
9675 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
9677 2013-09-05 Richard Biener <rguenther@suse.de>
9679 PR tree-optimization/58137
9680 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
9681 Do not create vectors of pointers.
9682 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
9683 types for the components of the vector initializer.
9684 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
9685 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
9687 2013-09-05 Martin Jambor <mjambor@suse.cz>
9689 * ipa-prop.c (remove_described_reference): Accept missing references,
9690 return false if that hppens, otherwise return true.
9691 (cgraph_node_for_jfunc): New function.
9692 (try_decrement_rdesc_refcount): Likewise.
9693 (try_make_edge_direct_simple_call): Use them.
9694 (ipa_edge_removal_hook): Remove references from rdescs.
9695 (ipa_edge_duplication_hook): Clone rdescs and their references
9696 when the new edge has the same caller as the old one.
9697 * cgraph.c (cgraph_resolve_speculation): Remove speculative
9698 reference before removing any edges.
9700 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
9702 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
9704 * thumb2.md (thumb2_storewb_parisi): New pattern.
9706 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
9708 * config/aarch64/aarch64-option-extensions.def: Add
9709 AARCH64_OPT_EXTENSION of 'crc'.
9710 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
9711 (AARCH64_ISA_CRC): Ditto.
9712 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
9713 description of the CRC extension.
9715 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
9717 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
9718 * config/rs6000/linux.h: Ditto.
9719 * alpha/linux.h: Ditto.
9720 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
9721 no_c99_libc_has_function.
9722 * config/c6x/uclinux-elf.h: Ditto.
9723 * config/lm32/uclinux-elf.h: Ditto.
9724 * config/m68k/uclinux.h: Ditto.
9725 * config/moxie/uclinux.h: Ditto.
9726 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
9727 (crisv32-*-linux*, cris-*-linux*): Ditto.
9728 * config/bfin/bfin.c: Include "tm_p.h".
9730 2013-09-05 Richard Biener <rguenther@suse.de>
9732 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
9733 check for a definition without a basic-block.
9735 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9736 Sofiane Naci <sofiane.naci@arm.com>
9738 * config/aarch64/aarch64.md
9739 (*movti_aarch64): Rename r_2_f and f_2_r.
9740 (*movsf_aarch64): Likewise.
9741 (*movdf_aarch64): Likewise.
9742 (*movtf_aarch64): Likewise.
9743 (aarch64_movdi_<mode>low): Likewise.
9744 (aarch64_movdi_<mode>high): Likewise.
9745 (aarch64_mov<mode>high_di): Likewise.
9746 (aarch64_mov<mode>low_di): Likewise.
9747 (aarch64_movtilow_tilow): Likewise.
9748 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
9749 values to config/arm/types.md
9750 (attribute "conds"): Update for attribute change.
9751 (anddi3_insn): Likewise.
9752 (iordi3_insn): Likewise.
9753 (xordi3_insn): Likewise.
9754 (one_cmpldi2): Likewise.
9755 * config/arm/types.md (type): Add Neon types.
9756 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
9757 use "type" attribute.
9758 (movmisalign<mode>_neon_store): Likewise.
9759 (movmisalign<mode>_neon_load): Likewise.
9760 (vec_set<mode>_internal): Likewise.
9761 (vec_setv2di_internal): Likewise.
9762 (vec_extract<mode>): Likewise.
9763 (vec_extractv2di): Likewise.
9764 (add<mode>3_neon): Likewise.
9765 (adddi3_neon): Likewise.
9766 (sub<mode>3_neon): Likewise.
9767 (subdi3_neon): Likewise.
9768 (mul<mode>3_neon): Likewise.
9769 (mul<mode>3add<mode>_neon): Likewise.
9770 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
9771 (fma<VCVTF:mode>4)): Likewise.
9772 (fma<VCVTF:mode>4_intrinsic): Likewise.
9773 (fmsub<VCVTF:mode>4)): Likewise.
9774 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
9775 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
9776 (ior<mode>3): Likewise.
9777 (and<mode>3): Likewise.
9778 (anddi3_neon): Likewise.
9779 (orn<mode>3_neon): Likewise.
9780 (orndi3_neon): Likewise.
9781 (bic<mode>3_neon): Likewise.
9782 (bicdi3_neon): Likewise.
9783 (xor<mode>3): Likewise.
9784 (one_cmpl<mode>2): Likewise.
9785 (abs<mode>2): Likewise.
9786 (neg<mode>2): Likewise.
9787 (umin<mode>3_neon): Likewise.
9788 (umax<mode>3_neon): Likewise.
9789 (smin<mode>3_neon): Likewise.
9790 (smax<mode>3_neon): Likewise.
9791 (vashl<mode>3): Likewise.
9792 (vashr<mode>3_imm): Likewise.
9793 (vlshr<mode>3_imm): Likewise.
9794 (ashl<mode>3_signed): Likewise.
9795 (ashl<mode>3_unsigned): Likewise.
9796 (neon_load_count): Likewise.
9797 (ashldi3_neon_noclobber): Likewise.
9798 (signed_shift_di3_neon): Likewise.
9799 (unsigned_shift_di3_neon): Likewise.
9800 (ashrdi3_neon_imm_noclobber): Likewise.
9801 (lshrdi3_neon_imm_noclobber): Likewise.
9802 (widen_ssum<mode>3): Likewise.
9803 (widen_usum<mode>3): Likewise.
9804 (quad_halves_<code>v4si): Likewise.
9805 (quad_halves_<code>v4sf): Likewise.
9806 (quad_halves_<code>v8hi): Likewise.
9807 (quad_halves_<code>v16qi): Likewise.
9808 (reduc_splus_v2di): Likewise.
9809 (neon_vpadd_internal<mode>): Likewise.
9810 (neon_vpsmin<mode>): Likewise.
9811 (neon_vpsmax<mode>): Likewise.
9812 (neon_vpumin<mode>): Likewise.
9813 (neon_vpumax<mode>): Likewise.
9814 (ss_add<mode>_neon): Likewise.
9815 (us_add<mode>_neon): Likewise.
9816 (ss_sub<mode>_neon): Likewise.
9817 (us_sub<mode>_neon): Likewise.
9818 (neon_vadd<mode>_unspec): Likewise.
9819 (neon_vaddl<mode>): Likewise.
9820 (neon_vaddw<mode>): Likewise.
9821 (neon_vhadd<mode>): Likewise.
9822 (neon_vqadd<mode>): Likewise.
9823 (neon_vaddhn<mode>): Likewise.
9824 (neon_vmul<mode>): Likewise.
9825 (neon_vmla<mode>): Likewise.
9826 (neon_vmlal<mode>): Likewise.
9827 (neon_vmls<mode>): Likewise.
9828 (neon_vmlsl<mode>): Likewise.
9829 (neon_vqdmulh<mode>): Likewise.
9830 (neon_vqdmlal<mode>): Likewise.
9831 (neon_vqdmlsl<mode>): Likewise.
9832 (neon_vmull<mode>): Likewise.
9833 (neon_vqdmull<mode>): Likewise.
9834 (neon_vsub<mode>_unspec): Likewise.
9835 (neon_vsubl<mode>): Likewise.
9836 (neon_vsubw<mode>): Likewise.
9837 (neon_vqsub<mode>): Likewise.
9838 (neon_vhsub<mode>): Likewise.
9839 (neon_vsubhn<mode>): Likewise.
9840 (neon_vceq<mode>): Likewise.
9841 (neon_vcge<mode>): Likewise.
9842 (neon_vcgeu<mode>): Likewise.
9843 (neon_vcgt<mode>): Likewise.
9844 (neon_vcgtu<mode>): Likewise.
9845 (neon_vcle<mode>): Likewise.
9846 (neon_vclt<mode>): Likewise.
9847 (neon_vcage<mode>): Likewise.
9848 (neon_vcagt<mode>): Likewise.
9849 (neon_vtst<mode>): Likewise.
9850 (neon_vabd<mode>): Likewise.
9851 (neon_vabdl<mode>): Likewise.
9852 (neon_vaba<mode>): Likewise.
9853 (neon_vabal<mode>): Likewise.
9854 (neon_vmax<mode>): Likewise.
9855 (neon_vmin<mode>): Likewise.
9856 (neon_vpaddl<mode>): Likewise.
9857 (neon_vpadal<mode>): Likewise.
9858 (neon_vpmax<mode>): Likewise.
9859 (neon_vpmin<mode>): Likewise.
9860 (neon_vrecps<mode>): Likewise.
9861 (neon_vrsqrts<mode>): Likewise.
9862 (neon_vqabs<mode>): Likewise.
9863 (neon_vqneg<mode>): Likewise.
9864 (neon_vcls<mode>): Likewise.
9865 (clz<mode>2): Likewise.
9866 (popcount<mode>2): Likewise.
9867 (neon_vrecpe): Likewise.
9868 (neon_vrsqrte): Likewise.
9869 (neon_vget_lane<mode>_sext_internal): Likewise.
9870 (neon_vget_lane<mode>_zext_internal): Likewise.
9871 (neon_vdup_n<mode>): Likewise.
9872 (neon_vdup_nv2di): Likewise.
9873 (neon_vdpu_lane<mode>_internal): Likewise.
9874 (neon_vswp<mode>): Likewise.
9875 (float<mode><V_cvtto>2): Likewise.
9876 (floatuns<mode><V_cvtto>2): Likewise.
9877 (fix_trunc<mode><V_cvtto>)2): Likewise
9878 (fixuns_trunc<mode><V_cvtto)2): Likewise.
9879 (neon_vcvt<mode>): Likewise.
9880 (neon_vcvtv4sfv4hf): Likewise.
9881 (neon_vcvtv4hfv4sf): Likewise.
9882 (neon_vcvt_n<mode>): Likewise.
9883 (neon_vmovn<mode>): Likewise.
9884 (neon_vqmovn<mode>): Likewise.
9885 (neon_vqmovun<mode>): Likewise.
9886 (neon_vmovl<mode>): Likewise.
9887 (neon_vmul_lane<mode>): Likewise.
9888 (neon_vmull_lane<mode>): Likewise.
9889 (neon_vqdmull_lane<mode>): Likewise.
9890 (neon_vqdmulh_lane<mode>): Likewise.
9891 (neon_vmla_lane<mode>): Likewise.
9892 (neon_vmlal_lane<mode>): Likewise.
9893 (neon_vqdmlal_lane<mode>): Likewise.
9894 (neon_vmls_lane<mode>): Likewise.
9895 (neon_vmlsl_lane<mode>): Likewise.
9896 (neon_vqdmlsl_lane<mode>): Likewise.
9897 (neon_vext<mode>): Likewise.
9898 (neon_vrev64<mode>): Likewise.
9899 (neon_vrev32<mode>): Likewise.
9900 (neon_vrev16<mode>): Likewise.
9901 (neon_vbsl<mode>_internal): Likewise.
9902 (neon_vshl<mode>): Likewise.
9903 (neon_vqshl<mode>): Likewise.
9904 (neon_vshr_n<mode>): Likewise.
9905 (neon_vshrn_n<mode>): Likewise.
9906 (neon_vqshrn_n<mode>): Likewise.
9907 (neon_vqshrun_n<mode>): Likewise.
9908 (neon_vshl_n<mode>): Likewise.
9909 (neon_vqshl_n<mode>): Likewise.
9910 (neon_vqshlu_n<mode>): Likewise.
9911 (neon_vshll_n<mode>): Likewise.
9912 (neon_vsra_n<mode>): Likewise.
9913 (neon_vsri_n<mode>): Likewise.
9914 (neon_vsli_n<mode>): Likewise.
9915 (neon_vtbl1v8qi): Likewise.
9916 (neon_vtbl2v8qi): Likewise.
9917 (neon_vtbl3v8qi): Likewise.
9918 (neon_vtbl4v8qi): Likewise.
9919 (neon_vtbx1v8qi): Likewise.
9920 (neon_vtbx2v8qi): Likewise.
9921 (neon_vtbx3v8qi): Likewise.
9922 (neon_vtbx4v8qi): Likewise.
9923 (neon_vtrn<mode>_internal): Likewise.
9924 (neon_vzip<mode>_internal): Likewise.
9925 (neon_vuzp<mode>_internal): Likewise.
9926 (neon_vld1<mode>): Likewise.
9927 (neon_vld1_lane<mode>): Likewise.
9928 (neon_vld1_dup<mode>): Likewise.
9929 (neon_vld1_dupv2di): Likewise.
9930 (neon_vst1<mode>): Likewise.
9931 (neon_vst1_lane<mode>): Likewise.
9932 (neon_vld2<mode>): Likewise.
9933 (neon_vld2_lane<mode>): Likewise.
9934 (neon_vld2_dup<mode>): Likewise.
9935 (neon_vst2<mode>): Likewise.
9936 (neon_vst2_lane<mode>): Likewise.
9937 (neon_vld3<mode>): Likewise.
9938 (neon_vld3qa<mode>): Likewise.
9939 (neon_vld3qb<mode>): Likewise.
9940 (neon_vld3_lane<mode>): Likewise.
9941 (neon_vld3_dup<mode>): Likewise.
9942 (neon_vst3<mode>): Likewise.
9943 (neon_vst3qa<mode>): Likewise.
9944 (neon_vst3qb<mode>): Likewise.
9945 (neon_vst3_lane<mode>): Likewise.
9946 (neon_vld4<mode>): Likewise.
9947 (neon_vld4qa<mode>): Likewise.
9948 (neon_vld4qb<mode>): Likewise.
9949 (neon_vld4_lane<mode>): Likewise.
9950 (neon_vld4_dup<mode>): Likewise.
9951 (neon_vst4<mode>): Likewise.
9952 (neon_vst4qa<mode>): Likewise.
9953 (neon_vst4qb<mode>): Likewise.
9954 (neon_vst4_lane<mode>): Likewise.
9955 (neon_vec_unpack<US>_lo_<mode>): Likewise.
9956 (neon_vec_unpack<US>_hi_<mode>): Likewise.
9957 (neon_vec_<US>mult_lo_<mode>): Likewise.
9958 (neon_vec_<US>mult_hi_<mode>): Likewise.
9959 (neon_vec_<US>shiftl_<mode>): Likewise.
9960 (neon_unpack<US>_<mode>): Likewise.
9961 (neon_vec_<US>mult_<mode>): Likewise.
9962 (vec_pack_trunc_<mode>): Likewise.
9963 (neon_vec_pack_trunk_<mode>): Likewise.
9964 (neon_vabd<mode>_2): Likewise.
9965 (neon_vabd<mode>_3): Likewise.
9966 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
9967 (thumb2_movsi_vfp): Likewise.
9968 (movdi_vfp): Likewise.
9969 (movdi_vfp_cortexa8): Likewise.
9970 (movhf_vfp_neon): Likewise.
9971 (movhf_vfp): Likewiwse.
9972 (movsf_vfp): Likewiwse.
9973 (thumb2_movsf_vfp): Likewiwse.
9974 (movdf_vfp): Likewise.
9975 (thumb2_movdf_vfp): Likewise.
9976 (movsfcc_vfp): Likewise.
9977 (thumb2_movsfcc_vfp): Likewise.
9978 (movdfcc_vfp): Likewise.
9979 (thumb2_movdfcc_vfp): Likewise.
9980 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
9981 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
9982 (v10_v2c): Likewise.
9983 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
9985 (cortex_a15_neon_int_2): Likewise.
9986 (cortex_a15_neon_int_3): Likewise.
9987 (cortex_a15_neon_int_4): Likewise.
9988 (cortex_a15_neon_int_5): Likewise.
9989 (cortex_a15_neon_vqneg_vqabs): Likewise.
9990 (cortex_a15_neon_vmov): Likewise.
9991 (cortex_a15_neon_vaba): Likewise.
9992 (cortex_a15_neon_vaba_qqq): Likewise.
9993 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9994 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9995 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
9996 scalar_64_32_long_scalar): Likewise.
9997 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9998 (cortex_a15_neon_mla_qqq_8_16): Likewise.
9999 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10000 lotype_qdd_64_32_long): Likewise.
10001 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10002 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10003 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
10004 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10005 (cortex_a15_neon_shift_1): Likewise.
10006 (cortex_a15_neon_shift_2): Likewise.
10007 (cortex_a15_neon_shift_3): Likewise.
10008 (cortex_a15_neon_vshl_ddd): Likewise.
10009 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10010 (cortex_a15_neon_vsra_vrsra): Likewise.
10011 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
10012 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
10013 (cortex_a15_neon_fp_vmul_ddd): Likewise.
10014 (cortex_a15_neon_fp_vmul_qqd): Likewise.
10015 (cortex_a15_neon_fp_vmla_ddd): Likewise.
10016 (cortex_a15_neon_fp_vmla_qqq): Likewise.
10017 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
10018 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
10019 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10020 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10021 (cortex_a15_neon_bp_simple): Likewise.
10022 (cortex_a15_neon_bp_2cycle): Likewise.
10023 (cortex_a15_neon_bp_3cycle): Likewise.
10024 (cortex_a15_neon_vld1_1_2_regs): Likewise.
10025 (cortex_a15_neon_vld1_3_4_regs): Likewise.
10026 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10027 (cortex_a15_neon_vld2_4_regs): Likewise.
10028 (cortex_a15_neon_vld3_vld4): Likewise.
10029 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10030 (cortex_a15_neon_vst1_3_4_regs): Likewise.
10031 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
10032 (cortex_a15_neon_vst3_vst4): Likewise.
10033 (cortex_a15_neon_vld1_vld2_lane): Likewise.
10034 (cortex_a15_neon_vld3_vld4_lane" 10
10035 (cortex_a15_neon_vst1_vst2_lane): Likewise.
10036 (cortex_a15_neon_vst3_vst4_lane): Likewise.
10037 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
10038 (cortex_a15_neon_ldm_2): Likewise.0
10039 (cortex_a15_neon_stm_2): Likewise.
10040 (cortex_a15_neon_mcr): Likewise.
10041 (cortex_a15_neon_mcr_2_mcrr): Likewise.
10042 (cortex_a15_neon_mrc): Likewise.
10043 (cortex_a15_neon_mrrc): Likewise.
10044 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
10046 (cortex_a15_alu_shift): Likewise.
10047 (cortex_a15_alu_shift_reg): Likewise.
10048 (cortex_a15_mult32): Likewise.
10049 (cortex_a15_mult64): Likewise.
10050 (cortex_a15_block): Likewise.
10051 (cortex_a15_branch): Likewise.
10052 (cortex_a15_load1): Likewise.
10053 (cortex_a15_load3): Likewise.
10054 (cortex_a15_store1): Likewise.
10055 (cortex_a15_store3): Likewise.
10056 (cortex_a15_call): Likewise.
10057 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
10058 (cortex_a5_f2r): Likewise.
10059 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
10061 (cortex_a53_f2r): Likewise.
10062 * config/arm/cortex-a7.md
10063 (cortex_a7_branch): Update for attribute change.
10064 (cortex_a7_call): Likewise.
10065 (cortex_a7_alu_imm): Likewise.
10066 (cortex_a7_alu_reg): Likewise.
10067 (cortex_a7_alu_shift): Likewise.
10068 (cortex_a7_mul): Likewise.
10069 (cortex_a7_load1): Likewise.
10070 (cortex_a7_store1): Likewise.
10071 (cortex_a7_load2): Likewise.
10072 (cortex_a7_store2): Likewise.
10073 (cortex_a7_load3): Likewise.
10074 (cortex_a7_store3): Likewise.
10075 (cortex_a7_load4): Likewise.
10076 (cortex_a7_store4): Likewise.
10077 (cortex_a7_fpalu): Likewise.
10078 (cortex_a7_fconst): Likewise.
10079 (cortex_a7_fpmuls): Likewise.
10080 (cortex_a7_neon_mul): Likewise.
10081 (cortex_a7_fpmacs): Likewise.
10082 (cortex_a7_neon_mla: Likewise.
10083 (cortex_a7_fpmuld: Likewise.
10084 (cortex_a7_fpmacd: Likewise.
10085 (cortex_a7_fpfmad: Likewise.
10086 (cortex_a7_fdivs: Likewise.
10087 (cortex_a7_fdivd: Likewise.
10088 (cortex_a7_r2f: Likewise.
10089 (cortex_a7_f2r: Likewise.
10090 (cortex_a7_f_flags: Likewise.
10091 (cortex_a7_f_loads: Likewise.
10092 (cortex_a7_f_loadd: Likewise.
10093 (cortex_a7_f_stores: Likewise.
10094 (cortex_a7_f_stored: Likewise.
10095 (cortex_a7_neon): Likewise.
10096 * config/arm/cortex-a8-neon.md
10097 (cortex_a8_neon_mrc): Update for attribute change.
10098 (cortex_a8_neon_mrrc): Likewise.
10099 (cortex_a8_neon_int_1): Likewise.
10100 (cortex_a8_neon_int_2): Likewise.
10101 (cortex_a8_neon_int_3): Likewise.
10102 (cortex_a8_neon_int_4): Likewise.
10103 (cortex_a8_neon_int_5): Likewise.
10104 (cortex_a8_neon_vqneg_vqabs): Likewise.
10105 (cortex_a8_neon_vmov): Likewise.
10106 (cortex_a8_neon_vaba): Likewise.
10107 (cortex_a8_neon_vaba_qqq): Likewise.
10108 (cortex_a8_neon_vsma): Likewise.
10109 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10110 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10111 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10112 long_scalar): Likewise.
10113 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10114 (cortex_a8_neon_mla_qqq_8_16): Likewise.
10115 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10116 long_scalar_qdd_64_32_long): Likewise.
10117 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10118 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10119 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
10120 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10121 (cortex_a8_neon_shift_1): Likewise.
10122 (cortex_a8_neon_shift_2): Likewise.
10123 (cortex_a8_neon_shift_3): Likewise.
10124 (cortex_a8_neon_vshl_ddd): Likewise.
10125 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10126 (cortex_a8_neon_vsra_vrsra): Likewise.
10127 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
10128 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
10129 (cortex_a8_neon_fp_vsum): Likewise.
10130 (cortex_a8_neon_fp_vmul_ddd): Likewise.
10131 (cortex_a8_neon_fp_vmul_qqd): Likewise.
10132 (cortex_a8_neon_fp_vmla_ddd): Likewise.
10133 (cortex_a8_neon_fp_vmla_qqq): Likewise.
10134 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
10135 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
10136 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10137 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10138 (cortex_a8_neon_bp_simple): Likewise.
10139 (cortex_a8_neon_bp_2cycle): Likewise.
10140 (cortex_a8_neon_bp_3cycle): Likewise.
10141 (cortex_a8_neon_ldr): Likewise.
10142 (cortex_a8_neon_str): Likewise.
10143 (cortex_a8_neon_vld1_1_2_regs): Likewise.
10144 (cortex_a8_neon_vld1_3_4_regs): Likewise.
10145 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10146 (cortex_a8_neon_vld2_4_regs): Likewise.
10147 (cortex_a8_neon_vld3_vld4): Likewise.
10148 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10149 (cortex_a8_neon_vst1_3_4_regs): Likewise.
10150 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
10151 (cortex_a8_neon_vst3_vst4): Likewise.
10152 (cortex_a8_neon_vld1_vld2_lane): Likewise.
10153 (cortex_a8_neon_vld3_vld4_lane): Likewise.
10154 (cortex_a8_neon_vst1_vst2_lane): Likewise.
10155 (cortex_a8_neon_vst3_vst4_lane): Likewise.
10156 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
10157 (cortex_a8_neon_mcr): Likewise.
10158 (cortex_a8_neon_mcr_2_mcrr): Likewise.
10159 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10160 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
10162 (ca9_neon_mrrc): Likewise.
10163 (cortex_a9_neon_int_1): Likewise.
10164 (cortex_a9_neon_int_2): Likewise.
10165 (cortex_a9_neon_int_3): Likewise.
10166 (cortex_a9_neon_int_4): Likewise.
10167 (cortex_a9_neon_int_5): Likewise.
10168 (cortex_a9_neon_vqneg_vqabs): Likewise.
10169 (cortex_a9_neon_vmov): Likewise.
10170 (cortex_a9_neon_vaba): Likewise.
10171 (cortex_a9_neon_vaba_qqq): Likewise.
10172 (cortex_a9_neon_vsma): Likewise.
10173 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10174 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10175 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10176 long_scalar): Likewise.
10177 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10178 (cortex_a9_neon_mla_qqq_8_16): Likewise.
10179 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10180 long_scalar_qdd_64_32_long): Likewise.
10181 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10182 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10183 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
10184 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10185 (cortex_a9_neon_shift_1): Likewise.
10186 (cortex_a9_neon_shift_2): Likewise.
10187 (cortex_a9_neon_shift_3): Likewise.
10188 (cortex_a9_neon_vshl_ddd): Likewise.
10189 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10190 (cortex_a9_neon_vsra_vrsra): Likewise.
10191 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
10192 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
10193 (cortex_a9_neon_fp_vsum): Likewise.
10194 (cortex_a9_neon_fp_vmul_ddd): Likewise.
10195 (cortex_a9_neon_fp_vmul_qqd): Likewise.
10196 (cortex_a9_neon_fp_vmla_ddd): Likewise.
10197 (cortex_a9_neon_fp_vmla_qqq): Likewise.
10198 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
10199 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
10200 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10201 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10202 (cortex_a9_neon_bp_simple): Likewise.
10203 (cortex_a9_neon_bp_2cycle): Likewise.
10204 (cortex_a9_neon_bp_3cycle): Likewise.
10205 (cortex_a9_neon_ldr): Likewise.
10206 (cortex_a9_neon_str): Likewise.
10207 (cortex_a9_neon_vld1_1_2_regs): Likewise.
10208 (cortex_a9_neon_vld1_3_4_regs): Likewise.
10209 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10210 (cortex_a9_neon_vld2_4_regs): Likewise.
10211 (cortex_a9_neon_vld3_vld4): Likewise.
10212 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10213 (cortex_a9_neon_vst1_3_4_regs): Likewise.
10214 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
10215 (cortex_a9_neon_vst3_vst4): Likewise.
10216 (cortex_a9_neon_vld1_vld2_lane): Likewise.
10217 (cortex_a9_neon_vld3_vld4_lane): Likewise.
10218 (cortex_a9_neon_vst1_vst2_lane): Likewise.
10219 (cortex_a9_neon_vst3_vst4_lane): Likewise.
10220 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
10221 (cortex_a9_neon_mcr): Likewise.
10222 (cortex_a9_neon_mcr_2_mcrr): Likewise.
10223 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
10224 (cortex_a9_fps): Likewise.
10225 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
10227 (cortex_m4_fmuls): Likewise.
10228 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
10230 (cortex_r4_mrc): Likewise.
10231 * config/arm/iterators.md: Update comment referring to neon_type.
10232 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
10233 (iwmmxt_movsi_insn): Likewise.
10234 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
10236 (pj4_core_to_vfp): Likewise.
10237 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
10239 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
10240 (vfp_fstore): Likewise.
10241 * doc/md.texi: Change references to neon_type to refer to type.
10243 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10245 * tree.h (DECL_BUILT_IN): Fix typo in comment.
10247 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
10249 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
10250 lglobl if not weak.
10252 2013-09-04 Easwaran Raman <eraman@google.com>
10254 PR middle-end/57370
10255 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
10256 (build_and_add_sum): Use it.
10257 (appears_later_in_bb): Simplify code.
10259 2013-09-04 Teresa Johnson <tejohnson@google.com>
10261 * dumpfile.c (dump_finish): Don't close stderr/stdout.
10263 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
10265 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
10267 2013-09-04 Jan Hubicka <jh@suse.cz>
10269 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
10270 * ipa-devirt.c: Include diganostic.h
10271 (odr_type_d): Add types and types_set.
10272 (hash_type_name): Work for types with vtables during LTO.
10273 (odr_hasher::remove): Fix comment; destroy types_set.
10274 (add_type_duplicate): New function,
10275 (get_odr_type): Use it.
10276 (dump_type_inheritance_graph): Dump type duplicates.
10277 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
10279 * tree.c (types_same_for_odr): Give exact answers on types with
10282 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10284 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
10285 explaining their differences.
10287 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
10289 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
10291 2013-09-03 Jeff Law <law@redhat.com>
10293 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
10294 when not threading through a joiner block. Pass joiner/no joiner
10295 state to register_jump_thread.
10296 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
10297 state from argument rather than implying on path length.
10298 Dump the entire jump thread path into debugging dump.
10299 * tree-flow.h (register_jump_thread): Update prototype.
10301 2013-08-29 Xinliang David Li <davidxl@google.com>
10303 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10304 Remove a trivial gcc_assert.
10306 2013-08-29 Xinliang David Li <davidxl@google.com>
10308 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
10309 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
10310 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10311 Delay base decl alignment adjustment.
10312 * tree-vectorizer.c (vect_destroy_datarefs): New function.
10313 * tree-vectorizer.h: New data structure.
10314 (set_dr_misalignment): New function.
10315 (dr_misalignment): Ditto.
10316 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
10317 (vectorizable_load): Ditto.
10318 (ensure_base_align): New function.
10319 (vectorize_loops): Add dbg_cnt support.
10320 (execute_vect_slp): Ditto.
10321 * dbgcnt.def: New debug counter.
10322 * Makefile: New dependency.
10324 2013-09-03 Meador Inge <meadori@codesourcery.com>
10328 2013-08-30 Meador Inge <meadori@codesourcery.com>
10330 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10332 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
10334 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
10335 function descriptor.
10337 2013-09-03 Richard Biener <rguenther@suse.de>
10339 * tree-affine.c (add_elt_to_tree): Fix association issue,
10340 avoid useless converts and make sure to always return a
10341 properly typed result.
10343 2013-09-03 Richard Biener <rguenther@suse.de>
10345 PR middle-end/57656
10346 * fold-const.c (negate_expr_p): Fix division case.
10347 (negate_expr): Likewise.
10349 2013-09-03 Richard Biener <rguenther@suse.de>
10352 * tree-streamer-out.c: Include tm.h.
10353 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
10355 2013-09-03 Jan Hubicka <jh@suse.cz>
10357 * tree-profile.c (tree_profiling): Cleanup CFG when done.
10359 2013-09-03 Alan Modra <amodra@gmail.com>
10361 * config.gcc (powerpc*-*-linux*): Add support for little-endian
10362 multilibs to big-endian target and vice versa.
10363 * config/rs6000/t-linux64: Use := assignment on all vars.
10364 (MULTILIB_EXTRA_OPTS): Remove fPIC.
10365 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
10366 * config/rs6000/t-linux64le: New file.
10367 * config/rs6000/t-linux64bele: New file.
10368 * config/rs6000/t-linux64lebe: New file.
10370 2013-09-02 Jan Hubicka <jh@suse.cz>
10372 * ipa-inline-transform.c (inline_transform): Do not
10373 optimize_inline_calls when not optimizing.
10375 2013-09-02 Jan Hubicka <jh@suse.cz>
10377 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
10378 duplicated nodes for assembler names.
10379 * symtab.c (symtab_unregister_node): Do not attempt to unlink
10380 hard registers from assembler name hash.
10382 2013-09-02 Jan Hubicka <jh@suse.cz>
10384 * ipa-split.c (execute_split_functions): Split externally visible
10385 functions called once.
10387 2013-09-02 Martin Jambor <mjambor@suse.cz>
10390 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
10391 linked list. When finding the correct duplicate, also consider also
10392 the caller in additon to its inlined_to node.
10394 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
10396 * config/aarch64/aarch64-simd-builtins.def
10397 (dup_lane_scalar): Remove.
10398 * config/aarch64/aarch64-simd.md
10399 (aarch64_simd_dup): Add 'w->w' alternative.
10400 (aarch64_dup_lane<mode>): Allow for VALL.
10401 (aarch64_dup_lane_scalar<mode>): Remove.
10402 (aarch64_dup_lane_<vswap_width_name><mode>): New.
10403 (aarch64_get_lane_signed<mode>): Add w->w altenative.
10404 (aarch64_get_lane_unsigned<mode>): Likewise.
10405 (aarch64_get_lane<mode>): Likewise.
10406 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
10407 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
10408 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
10409 (VCON): Change container of V2SF.
10410 (vswap_width_name): Likewise.
10411 * config/aarch64/arm_neon.h
10412 (__aarch64_vdup_lane_any): New.
10413 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10414 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
10415 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10417 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
10419 PR middle-end/56382
10420 * expr.c (emit_move_complex): Do not move complex FP values as parts if
10421 the source or the destination is a single hard register.
10423 2013-09-02 Richard Biener <rguenther@suse.de>
10425 PR middle-end/57511
10426 * tree-scalar-evolution.c (instantiate_scev_name): Allow
10429 2013-09-02 Richard Biener <rguenther@suse.de>
10431 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
10432 arithmetic to sizetype.
10434 2013-09-02 Bin Cheng <bin.cheng@arm.com>
10436 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
10437 Find auto-increment use both before and after candidate.
10439 2013-09-02 Marek Polacek <polacek@redhat.com>
10441 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
10443 2013-09-01 Jan Hubicka <jh@suse.cz>
10445 * Makefile.in: Add ipa-profile.o
10446 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
10447 * cgraph.c (struct cgraph_propagate_frequency_data,
10448 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
10449 ipa-profile.c; replace cgraph_ by ipa_ prefix.
10450 * cgraph.h (cgraph_propagate_frequency): Remove.
10451 * ipa-inline-analysis.c: Include ipa-utils.h;
10452 drop duplicated cfgloop.h.
10453 (inline_update_callee_summaries): Update.
10454 * ipa-profile.c: New file.
10455 * ipa-utils.h (ipa_propagate_frequency): Declare.
10456 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
10457 data-streamer.h, value-prof.h.
10458 (symtab_remove_unreachable_nodes): Update profile.
10459 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
10460 account_time_size, cmp_counts, dump_histogram,
10461 ipa_profile_generate_summary, ipa_profile_write_summary,
10462 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
10463 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
10464 Move to ipa-profile.c.
10466 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
10468 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
10470 2013-09-01 Jan Hubicka <jh@suse.cz>
10472 * common.opt (fdevirtualize-speculatively): New function.
10473 * invoke.texi (fdevirtualize-speculatively): Document.
10474 * ipa-devirt.c: Include ipa-inline.h
10475 (likely_target_p): New function.
10476 (ipa_devirt): New function.
10477 (gate_ipa_devirt): New function.
10478 (pass_data_ipa_devirt): New static var.
10479 (pass_ipa_devirt): Likewise.
10480 (make_pass_ipa_devirt): New function.
10481 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
10482 (common_handle_option): Disable devirtualization when
10483 value range profiling is available.
10484 * passes.def (pass_ipa_devirt): Add.
10485 * timever.def (TV_IPA_DEVIRT): New timevar.
10486 * tree-pass.h (make_pass_ipa_devirt):
10488 2013-09-01 Iain Sandoe <iain@codesourcery.com>
10490 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
10491 include sanitize(undefined).
10493 2013-08-31 Diego Novillo <dnovillo@google.com>
10495 * Makefile.in (TREE_CORE_H): Define.
10497 (GTFILES): Add tree-core.h.
10498 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
10500 * tree-core.h: New file.
10501 Move all data structures, enum, typedefs, global
10502 declarations and constants from ...
10503 * tree.h: ... here.
10505 2013-08-31 Jan Hubicka <jh@suse.cz>
10507 * bulitins.c (expand_builtin): Do not early exit for gcov
10508 instrumented functions.
10510 2013-08-31 Marek Polacek <polacek@redhat.com>
10512 * ubsan.c: Include tm_p.h.
10514 2013-08-31 Jan Hubicka <jh@suse.cz>
10516 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
10519 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
10520 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
10521 * tree-cfg.c (verify_gimple_label): ... here.
10522 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
10523 (ipa_merge_profiles): New function.
10524 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
10525 (lto_input_function_body): Likewise.
10526 * ipa-utils.h (ipa_merge_profiles): Declare.
10527 * lto-streamer.h (lto_input_function_body): Update prototype.
10528 (emit_label_in_global_context_p): Remove.
10529 * lto-symtab.c: Include ipa-utils.h
10530 (lto_cgraph_replace_node): Use ipa_merge_profiles.
10532 2013-08-31 Jan Hubicka <jh@suse.cz>
10534 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
10536 2013-08-31 Jan Hubicka <jh@suse.cz>
10538 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
10540 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
10542 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
10543 "cmp" RTX before signed_comparison_operator check to account
10544 for "code" changes.
10546 2013-08-30 Jan Hubicka <jh@suse.cz>
10548 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
10549 (detect_type_change_1): Rename to ...
10550 (detect_type_change): ... this one; early return on non-polymorphic
10552 (detect_type_change_ssa): Add comp_type parameter; update
10553 use of detect_type_change.
10554 (compute_complex_assign_jump_func): Add param_type parameter;
10555 update use of detect_type_change_ssa.
10556 (compute_complex_ancestor_jump_func): Likewise.
10557 (ipa_get_callee_param_type): New function.
10558 (ipa_compute_jump_functions_for_edge): Compute parameter type;
10559 update calls to the jump function computation functions.
10561 2013-08-30 Teresa Johnson <tejohnson@google.com>
10562 Steven Bosscher <steven@gcc.gnu.org>
10564 * cfgrtl.c (fixup_new_cold_bb): New routine.
10565 (commit_edge_insertions): Invoke fixup_partitions.
10566 (find_partition_fixes): New routine.
10567 (fixup_partitions): Ditto.
10568 (verify_hot_cold_block_grouping): Update comments.
10569 (rtl_verify_edges): Invoke find_partition_fixes.
10570 (rtl_verify_bb_pointers): Update comments.
10571 (rtl_verify_bb_layout): Ditto.
10572 * basic-block.h (probably_never_executed_edge_p): Declare.
10573 (fixup_partitions): Ditto.
10574 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
10575 * bb-reorder.c (sanitize_hot_paths): New function.
10576 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
10577 sanitize_hot_paths.
10578 * predict.c (probably_never_executed_edge_p): New routine.
10579 * cfg.c (check_bb_profile): Add partition insanity warnings.
10581 2013-08-30 Meador Inge <meadori@codesourcery.com>
10583 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10585 2013-08-30 Marek Polacek <polacek@redhat.com>
10587 * Makefile.in (ubsan.o): Add.
10588 (c-family/c-ubsan.o): Add.
10589 (builtins.o): Add ubsan.h dependency.
10590 * ubsan.h: New file.
10591 * ubsan.c: New file.
10592 * common.opt: Add -fsanitize=undefined option.
10593 (flag_sanitize): Add variable.
10594 (fsanitize=): Add option. Add Driver.
10595 (fsanitize=thread): Remove option.
10596 (fsanitize=address): Likewise.
10597 (static-libubsan): New option.
10598 * doc/invoke.texi: Document the new flag and -static-libubsan.
10599 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
10600 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
10601 * builtin-attrs.def (ATTR_COLD): Define.
10602 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10603 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
10604 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
10605 * flag-types.h (sanitize_code): New enum.
10606 * opts.c (common_handle_option): Parse command line arguments
10607 of -fsanitize=. Add -fsanitize=unreachable option.
10608 * varasm.c (get_variable_section): Adjust.
10609 (assemble_noswitch_variable): Likewise.
10610 (assemble_variable): Likewise.
10611 (output_constant_def_contents): Likewise.
10612 (categorize_decl_for_section): Likewise.
10613 (place_block_symbol): Likewise.
10614 (output_object_block): Likewise.
10615 * builtins.def: Likewise.
10616 * toplev.c (compile_file): Likewise.
10617 (process_options): Likewise.
10618 * cppbuiltin.c: Likewise.
10619 * tsan.c (tsan_pass): Likewise.
10620 (tsan_gate): Likewise.
10621 (tsan_gate_O0): Likewise.
10622 * cfgexpand.c (partition_stack_vars): Likewise.
10623 (expand_stack_vars): Likewise.
10624 (defer_stack_allocation): Likewise.
10625 (expand_used_vars): Likewise.
10626 * cfgcleanup.c (old_insns_match_p): Likewise.
10627 * asan.c (asan_finish_file): Likewise.
10628 (asan_instrument): Likewise.
10629 (gate_asan): Likewise.
10630 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
10631 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10632 (asan_global_struct): Use pointer_sized_int_node instead
10633 calling build_nonstandard_integer_type.
10634 (initialize_sanitizer_builtins): Likewise.
10635 (asan_finish_file): Likewise.
10636 * gcc.c: Document %{%:function(args):X}.
10637 (static_spec_functions): Add sanitize.
10638 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
10639 store funcval != NULL there.
10640 (do_spec_1): Adjust handle_spec_function caller.
10641 (handle_braces): Allow %:function(args) as condition.
10642 (sanitize_spec_function): New function.
10643 (ADD_STATIC_LIBUBSAN_LIBS): Define.
10644 (LIBUBSAN_SPEC): Likewise.
10645 (LIBUBSAN_EARLY_SPEC): Likewise.
10646 (SANITIZER_SPEC): Handle libubsan.
10647 (SANITIZER_EARLY_SPEC): Likewise.
10648 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
10649 instead of fsanitize=address.
10650 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
10651 instead of fsanitize=address*.
10652 * builtins.c: Include ubsan.h.
10653 (fold_builtin_0): Instrument __builtin_unreachable.
10654 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
10655 instead of flag_asan.
10656 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
10657 (pointer_sized_int_node): Define.
10658 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
10660 2013-08-30 Mike Stump <mikestump@comcast.net>
10662 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
10665 2013-08-29 Jan Hubicka <jh@suse.cz>
10667 * cgraph.c (cgraph_function_body_availability): Handle weakref
10669 * passes.def: Remove pass_fixup_cfg.
10670 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
10671 track when we need to remove functions.
10672 (gate_ipa_inline): Execute inlining always; add comment why.
10673 (pass_data_ipa_inline): Remove TODO_remove_functions.
10674 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
10675 do not produce summaries.
10676 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
10677 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
10678 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
10679 constant pool and vtable.
10681 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
10683 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
10684 New arm_neon.h's internal macros to specify 64-bit constants.
10685 Avoid using stdint.h's macros.
10687 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
10689 * recog.c (verify_changes): Verify that changes[i].old is non-zero
10690 before applying REG_P.
10692 2013-08-30 Jakub Jelinek <jakub@redhat.com>
10694 PR tree-optimization/58277
10695 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
10696 after seeing too many stmts with vdef in between dombb and current
10697 bb, invalidate everything.
10699 2013-08-30 Richard Biener <rguenther@suse.de>
10701 * fold-const.c (fold_single_bit_test): Fix overflow test.
10703 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
10705 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
10706 and which can live in a register, always retrieve the value on entry.
10707 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
10708 passed by invisible reference specially.
10709 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
10710 (vt_add_function_parameter): Correctly deal with a parameter passed by
10711 invisible reference.
10713 2013-08-30 Jan Hubicka <jh@suse.cz>
10715 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
10717 2013-08-30 Richard Biener <rguenther@suse.de>
10719 PR tree-optimization/58228
10720 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
10721 allow invariant loads in nested loop vectorization.
10723 2013-08-30 Richard Biener <rguenther@suse.de>
10725 PR tree-optimization/58223
10726 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
10727 (has_anti_or_output_dependence): ... this and adjust to also
10728 look for output dependences.
10729 (mark_nodes_having_upstream_mem_writes): Adjust.
10730 (rdg_flag_uses): Likewise.
10732 2013-08-30 Richard Biener <rguenther@suse.de>
10734 PR tree-optimization/58010
10735 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
10736 assert that we have a loop-closed PHI.
10738 2013-08-29 Jan Hubicka <jh@suse.cz>
10740 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
10741 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
10742 * lto-section-in.c (lto_free_function_in_decl_state): New function.
10743 (lto_free_function_in_decl_state_for_node): New function.
10745 2013-08-29 Xinliang David Li <davidxl@google.com>
10747 * loop-unroll.c (report_unroll_peel): Minor message change.
10748 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
10749 Emit alignment peeling message with default -fopt-info.
10750 (vect_loop_versioning): Emit loop version info message.
10751 * tree-vectorizer.c (vectorize_loops): Minor message change.
10752 (execute_vect_slp): Ditto.
10754 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
10756 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
10757 of the clone from the DECL_NAME of the original function.
10759 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
10761 * passes.c (register_pass): Add overload.
10762 * tree-pass.h (register_pass): Forward declare it. Add comment.
10764 2013-08-29 Jan Hubicka <jh@suse.cz>
10766 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
10767 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
10768 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
10769 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10770 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
10771 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
10772 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10773 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
10775 2013-08-29 Teresa Johnson <tejohnson@google.com>
10777 * dumpfile.c (dump_loc): Output column number.
10778 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
10779 * doc/invoke.texi: Document optall -fopt-info flag.
10780 * profile.c (read_profile_edge_counts): Use new dump framework.
10781 (compute_branch_probabilities): Ditto.
10782 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
10783 when pass not in any opt group.
10784 * pass_manager.h (pass_manager::get_pass_profile): New method.
10785 * value-prof.c (check_counter): Use new dump framework.
10786 (check_ic_target): Ditto.
10787 * coverage.c (get_coverage_counts): Ditto.
10788 (coverage_init): Setup new dump framework.
10790 2013-08-29 Richard Biener <rguenther@suse.de>
10792 PR tree-optimization/58246
10793 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
10794 handle the dominance check inside a basic-block.
10796 2013-08-29 Richard Biener <rguenther@suse.de>
10798 PR middle-end/57287
10799 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
10800 of default defs that appear in abnormal PHI nodes.
10802 2013-08-29 Richard Biener <rguenther@suse.de>
10804 PR tree-optimization/57685
10805 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
10806 single-use operands to avoid exponential complexity.
10808 2013-08-28 Dehao Chen <dehao@google.com>
10810 * ipa-inline.c (edge_badness): Fix integer underflow.
10812 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
10814 * gtm-builtins.def (_ITM_free): Declare leaf.
10816 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10819 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
10820 (*tls_local_dynamic_base_64_largepic): Likewise.
10821 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
10822 Remove predicate from call operand.
10823 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
10824 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
10826 2013-08-28 Jeff Law <law@redhat.com>
10828 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
10829 checks for the number of predecessors and successors allowed.
10830 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
10831 which require copying a joiner block if there is a request which
10832 is a subpath that requires no joiner block copying.
10834 2013-08-28 Jan Hubicka <jh@suse.cz>
10836 * lto-streamer-out.c (DFS_write_tree_body): Drop
10837 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
10838 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
10839 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
10840 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
10841 Do not read DECL_ERROR_ISSUED.
10842 (unpack_ts_decl_with_vis_value_fields): Do not read
10844 (lto_input_ts_binfo_tree_pointers): Do not read
10845 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
10846 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
10847 write DECL_ERROR_ISSUED..
10848 (pack_ts_decl_with_vis_value_fields): Do not write
10850 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
10851 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
10852 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
10853 * tree.h (tree_decl_common): Update comment.
10854 (DECL_ERROR_ISSUED): Remove.
10856 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10858 PR middle-end/58257
10859 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
10861 2013-08-28 Jan Hubicka <jh@suse.cz>
10863 * builtins.def (free): Declare leaf.
10865 2013-08-27 David Malcolm <dmalcolm@redhat.com>
10867 * gdbhooks.py: New.
10868 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
10869 * configure: Regenerate.
10871 2013-08-27 Martin Jambor <mjambor@suse.cz>
10873 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
10874 (ipa_ancestor_jf_data): Likewise.
10875 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
10876 (ipa_get_jf_pass_through_type_preserved): New function.
10877 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
10878 (ipa_get_jf_ancestor_type_preserved): New function.
10879 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
10880 (ipa_get_jf_ancestor_result): Likewise.
10881 (propagate_vals_accross_pass_through): Use
10882 ipa_get_jf_pass_through_result to do all the value mappings.
10883 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
10884 type_preserved flag.
10885 (ipa_set_jf_cst_copy): New function.
10886 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
10887 (ipa_set_jf_arith_pass_through): Likewise.
10888 (ipa_set_ancestor_jf): Likewise.
10889 (compute_complex_assign_jump_func): Set type_preserved instead of
10891 (ipa_compute_jump_functions_for_edge): Likewise.
10892 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
10893 (update_jump_functions_after_inlining): Update type_preserved.
10894 Explicitely create jump functions when combining one with pass_through.
10895 (ipa_write_jump_function): Stream the type_preserved flags.
10896 (ipa_read_jump_function): Likewise.
10898 2013-08-27 Jakub Jelinek <jakub@redhat.com>
10899 Aldy Hernandez <aldyh@redhat.com>
10901 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
10902 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
10903 * function.h (struct function): Add has_force_vect_loops and
10905 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
10906 * gimple.c (gimple_build_omp_critical): Add KIND argument and
10908 * gimple.def: Update CLAUSES comments.
10909 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
10910 (gimple_build_omp_for): Add argument to prototype.
10911 (gimple_omp_for_kind): New.
10912 (gimple_omp_for_set_kind): New.
10913 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
10914 GOVD_DATA_SHARE_CLASS.
10915 (enum omp_region_type): Add ORT_SIMD.
10916 (gimple_add_tmp_var): Handle ORT_SIMD.
10917 (gimplify_var_or_parm_decl): Same.
10918 (is_gimple_stmt): Same.
10919 (omp_firstprivatize_variable): Same.
10920 (omp_add_variable): Only use splay_tree_insert if lookup failed.
10921 (omp_notice_variable): Handle ORT_SIMD.
10922 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
10923 (omp_check_private): Handle ORT_SIMD.
10924 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10925 OMP_CLAUSE_SAFELEN.
10926 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
10927 Handle OMP_CLAUSE_LASTPRIVATE.
10928 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10929 OMP_CLAUSE_SAFELEN.
10930 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
10931 (gimplify_expr): Handle OMP_SIMD.
10932 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
10933 (expand_GOMP_SIMD_VF): New.
10934 (expand_GOMP_SIMD_LAST_LANE): New.
10935 * internal-fn.def (GOMP_SIMD_LANE): New.
10936 (GOMP_SIMD_VF): New.
10937 (GOMP_SIMD_LAST_LANE): New.
10938 * omp-low.c: Include target.h.
10939 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
10940 OMP_CLAUSE_SAFELEN.
10941 (check_omp_nesting_restrictions): Same.
10943 (lower_rec_simd_input_clauses): New.
10944 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
10946 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
10947 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
10948 (expand_omp_build_assign): New.
10949 (expand_omp_for_init_counts): New.
10950 (expand_omp_for_init_vars): New.
10951 (extract_omp_for_update_vars): New.
10952 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
10953 and rewrite accordingly.
10954 (expand_omp_simd): New.
10955 (expand_omp_for): Use expand_omp_simd.
10956 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
10957 (lower_omp_for): Do not lower the body.
10958 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
10959 in their own loops.
10960 * tree-flow.h (find_omp_clause): Remove prototype.
10961 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
10962 forcing vectorization of the loop, or if flag_tree_vectorize.
10963 (gate_tree_if_conversion): Similarly.
10964 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
10965 gimple_build_omp_for.
10966 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
10967 * tree-parloops (create_parallel_loop): Pass kind argument to
10968 gimple_build_omp_for.
10969 * tree-pretty-print.c (dump_omp_clause): Add cases for
10970 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
10971 OMP_CLAUSE__SIMDUID_.
10972 (dump_generic_node): Handle OMP_SIMD.
10973 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
10974 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
10975 unroll OMP_SIMD loops here.
10976 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
10977 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
10979 (vect_analyze_data_refs): Handle simd loops.
10980 * tree-vect-loop.c (vectorizable_live_operation): Handle
10982 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
10983 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
10984 (vectorizable_load): Same.
10985 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
10986 (struct simduid_to_vf): New.
10987 (simduid_to_vf::hash): New.
10988 (simduid_to-vf::equal): New.
10989 (struct simd_array_to_simduid): New.
10990 (simd_array_to_simduid::hash): New.
10991 (simd_array_to_simduid::equal): New.
10992 (adjust_simduid_builtins): New.
10993 (struct note_simd_array_uses_struct): New.
10994 (note_simd_array_uses_cb): New.
10995 (note_simd_array_uses): New.
10996 (vectorize_loops): Handle simd hints and adjust simd builtins
10998 * tree-vectorizer.h (struct _stmt_vec_info): Add
10999 simd_lane_access_p field.
11000 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
11001 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
11002 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
11003 (omp_clause_code_name): Same.
11004 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
11005 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
11006 * tree.def (OMP_SIMD): New entry.
11007 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
11008 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
11009 (OMP_CLAUSE_DECL): Adjust range for new clauses.
11010 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
11011 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
11012 (OMP_CLAUSE_LINEAR_STEP): New.
11013 (OMP_CLAUSE_SAFELEN_EXPR): New.
11014 (OMP_CLAUSE__SIMDUID__DECL): New.
11015 (find_omp_clause): New prototype.
11017 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
11019 * config/i386/driver-i386.c (host_detect_local_cpu): Update
11020 Haswell processor detection.
11022 2013-08-27 Christian Widmer <shadow@umbrox.de>
11025 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
11026 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
11027 AVX2 capable processors.
11029 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
11031 * config/aarch64/arm_neon.h: Replace all inline asm implementations
11032 of vget_low_* with implementations in terms of other intrinsics.
11034 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
11036 PR middle-end/57219
11037 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
11038 values to -1, 0 and 1.
11040 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
11042 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
11043 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
11044 (<optab><mode>3_insn): Remove.
11045 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
11046 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
11047 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
11048 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
11049 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
11050 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
11051 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
11052 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
11053 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
11054 (ror<mode>3_insn): Likewise.
11055 * config/aarch64/predicates.md (aarch64_simd_register): New.
11057 2013-08-27 Richard Biener <rguenther@suse.de>
11059 PR tree-optimization/57521
11060 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
11061 one edge is non-critical.
11062 (find_phi_replacement_condition): Make sure to use a non-critical
11063 edge. Cleanup and remove old bug workarounds.
11064 (bb_postdominates_preds): Remove.
11065 (if_convertible_loop_p_1): Do not compute post-dominators.
11066 (combine_blocks): Do not free post-dominators.
11067 (main_tree_if_conversion): Likewise.
11068 (pass_data_if_conversion): Add TODO_verify_ssa.
11070 2013-08-27 DJ Delorie <dj@redhat.com>
11072 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
11074 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
11076 * function.c (assign_parm_find_data_types): Set passed_mode and
11077 nominal_mode to the TYPE_MODE of nominal_type for the built
11078 pointer type in case of the struct-pass-by-reference.
11080 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11082 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
11083 (UINT16_TYPE): Change default to "unsigned int".
11085 * config/avr/avr.opt (mfract-convert-truncate): New option.
11086 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
11087 is set, round negative fractional integers according to n1169
11088 when converting to integer types.
11090 2013-08-26 Jan Hubicka <jh@suse.cz>
11092 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
11093 methods can not be called indirectly when their address is not taken.
11095 2013-08-26 Jan Hubicka <jh@suse.cz>
11097 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
11100 2013-08-26 Jan Hubicka <jh@suse.cz>
11102 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
11105 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11107 * reload.c (find_valid_class): Allow classes that do not include
11108 FIRST_PSEUDO_REGISTER - 1.
11110 2013-08-26 Jan Hubicka <jh@suse.cz>
11112 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
11113 fix edge count/frequency when speculation failed; fix type check
11114 for the direct call.
11116 2013-08-26 Jan Hubicka <jh@suse.cz>
11118 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
11120 2013-08-26 Jan Hubicka <jh@suse.cz>
11122 * ipa-inline-transform.c (inline_transform): Be ready for basic block
11123 to be changed by edge redirection.
11125 2013-08-26 Jan Hubicka <jh@suse.cz>
11127 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
11128 formating; add sanity check.
11129 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
11130 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
11131 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
11132 (resolve_noninline_speculation): Update callee keys, too.
11134 2013-08-26 Jan Hubicka <jh@suse.cz>
11136 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
11137 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
11139 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11141 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
11144 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
11146 * config/i386/i386.c (ix86_debug_options): Remove prototype.
11147 (x86_64_elf_select_section): Ditto.
11148 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
11150 (ix86_pass_by_reference): Ditto.
11151 (output_set_got): Ditto.
11152 (ix86_unary_operator_ok): Ditto.
11153 (ix86_expand_builtin): Ditto.
11155 2013-08-23 Jan Hubicka <jh@suse.cz>
11157 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
11159 2013-08-23 Jan Hubicka <jh@suse.cz>
11161 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
11162 (tree_decl_with_vis): Add FINAL field.
11164 2013-08-23 Jeff Law <law@redhat.com>
11166 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
11167 the debugging dump when the expression is fully redundant.
11169 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11171 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
11172 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
11173 * pretty-print.c (pp_formatted_text_data): Likewise.
11174 (pp_write_text_to_stream): Likewise.
11175 (pp_write_text_as_dot_label_to_stream): Likewise.
11176 (pp_append_r): Likewise.
11177 (pp_format): Likewise.
11178 (pp_flush): Likewise.
11179 (pp_clear_output_area): Likewise.
11180 (pp_append_text): Likewise.
11181 (pp_formatted_text): Likewise.
11182 (pp_remaining_character_count_for_line): Likewise.
11183 (pp_newline): Likewise.
11184 (pp_character): Likewise.
11185 (output_buffer::~output_buffer): Define.
11186 (pretty_printer::~pretty_printer): Destruct output buffer.
11187 * pretty-print.h (output_buffer::~output_buffer): Declare.
11188 (pretty_printer::~pretty_printer): Declare virtual.
11190 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
11193 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
11194 HOST_WIDE_INT_M1U): New macros.
11195 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
11196 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
11197 unsigned -1 for lshift.
11198 * cse.c (cse_insn): Likewise.
11199 * double-int.c (rshift_double, lshift_double): Likewise.
11200 * builtins.c (fold_builtin_bitop): Likewise.
11201 * combine.c (force_to_mode): Likewise.
11202 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
11203 * simplify-rtx.c (simplify_const_unary_operation,
11204 simplify_const_binary_operation): Likewise.
11205 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
11206 check_va_list_escapes): Likewise.
11207 * rtlanal.c (nonzero_bits1): Likewise.
11208 * expmed.c (expand_smod_pow2): Likewise.
11209 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
11211 2013-08-23 Jan Hubicka <jh@suse.cz>
11213 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
11214 as having address taken.
11216 2013-08-23 Jan Hubicka <jh@suse.cz>
11218 * ipa-utils.h (method_class_type): Declare.
11219 * ipa-devirt.c (method_class_type): Export.
11221 * cgraphunit.c (analyze_functions): Do basic devirtualization;
11222 do not walk base classes of anonymous types.
11224 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
11226 PR rtl-optimization/58220
11227 PR regression/58221
11228 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
11229 handle SEQUENCE insns properly.
11231 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11233 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
11235 (pp_newline_and_indent): Likewise.
11236 (pp_separate_with): Likewise.
11237 * pretty-print.c (pp_newline_and_flush): Define.
11238 (pp_newline_and_indent): Likewise.
11239 (pp_separate_with): Likewise.
11241 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11244 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
11245 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
11247 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
11249 * config/i386/predicates.md (ext_sse_reg_operand): New.
11250 * config/i386/i386.md (*movti_internal): Use
11251 predicate to determine if EVEX is needed.
11252 (*movsi_internal): Ditto.
11253 (*movdf_internal): Ditto.
11254 (*movsf_internal): Ditto.
11255 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
11257 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11259 PR tree-optimization/58209
11260 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
11261 (find_tail_calls): Give up for pointer result types if m is non-NULL.
11262 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
11263 emit POINTER_PLUS_EXPR.
11264 (create_tailcall_accumulator): For pointer result type accumulate in
11267 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
11269 * configure.ac: Add backslashes missing from the last change.
11270 * configure: Regenerate.
11272 2013-08-22 Jan Hubicka <jh@susue.cz>
11274 * ipa.c (function_and_variable_visibility): First remember function
11275 was global and then make it local.
11277 2013-08-22 Julian Brown <julian@codesourcery.com>
11279 * configure.ac: Add aarch64 to list of arches which use "nop" in
11281 * configure: Regenerate.
11283 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11285 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
11286 gnu_libc_has_function.
11287 * config/s390/tpf.h: Likewise.
11289 2013-08-22 Jan Hubicka <jh@susue.cz>
11291 * timevar.c (validate_phases): Add cast.
11293 2013-08-22 Jan Hubicka <jh@susue.cz>
11295 * timevar.c (validate_phases): Use size_t for memory.
11296 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
11298 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
11300 * pretty-print.h (output_buffer::output_buffer): Declare.
11301 (pretty_printer::pretty_printer): Likewise.
11302 (pp_construct): Remove.
11303 * pretty-print.c (output_buffer::output_buffer): Define.
11304 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
11305 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
11306 (print_gimple_expr): Likewise.
11307 (print_gimple_seq): Likewise.
11308 (gimple_dump_bb): Likewise.
11309 * sched-vis.c (dump_value_slim): Likewise.
11310 (dump_insn_slim): Likewise.
11311 (dump_rtl_slim): Likewise.
11312 (str_pattern_slim): Likewise.
11313 * tree-mudflap.c (mf_varname_tree): Likewise.
11314 * graph.c (print_graph_cfg): Likewise.
11315 (start_graph_dump): Likewise.
11316 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
11318 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
11319 pretty printer initialization.
11320 * coretypes.h (diagnostic_context): Remove superflous type alias
11322 (pretty_printer): Likewise. Declare directly as a class.
11323 (pretty_print_info): Remove declaration as class.
11324 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
11325 and pp_clear_output_area.
11326 (asan_add_global): Likewise.
11328 2013-08-22 Jan Hubicka <jh@suse.cz>
11330 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
11331 * ipa-utils.h (update_type_inheritance_graph): Declare.
11332 (possible_polymorphic_call_target_p): Declare.
11333 (possible_polymorphic_call_target_p): New.
11334 * ipa-devirt.c: Update toplevel comments.
11335 (cached_polymorphic_call_targets): Move up.
11336 (odr_type_d): Move ID down.
11337 (polymorphic_type_binfo_p): Update comment.
11338 (odr_hasher::remove): Likewise;
11339 (get_odr_type): Set anonymous_namespace.
11340 (dump_odr_type): Dump it.
11341 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
11342 (maybe_record_node): Record node in cached_polymorphic_call_targets.
11343 (record_binfo): Add comment.
11344 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
11346 (devirt_node_removal_hook): Do not iCE when cache is freed.
11347 (possible_polymorphic_call_target_p): New predicate.
11348 (update_type_inheritance_graph): New function.
11350 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
11351 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
11352 Sergey Lega <sergey.s.lega@intel.com>
11353 Anna Tikhonova <anna.tikhonova@intel.com>
11354 Ilya Tocar <ilya.tocar@intel.com>
11355 Andrey Turetskiy <andrey.turetskiy@intel.com>
11356 Ilya Verbin <ilya.verbin@intel.com>
11357 Kirill Yukhin <kirill.yukhin@intel.com>
11358 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11360 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
11361 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
11362 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
11363 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
11364 (OPTION_MASK_ISA_AVX2_UNSET): Update.
11365 (OPTION_MASK_ISA_AVX512F_UNSET): New.
11366 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
11367 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
11368 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
11369 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
11370 OPT_mavx512pf, OPT_mavx512er cases.
11371 * config/i386/constraints.md (v): New constraint.
11372 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
11373 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
11374 (bit_AVX512CD): New.
11375 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11376 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
11377 * config/i386/i386-c.c (ix86_target_macros_internal):
11378 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
11380 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
11381 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
11382 * config/i386/i386.c (regclass_map, dbx_register_map)
11383 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
11384 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
11385 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
11386 -mavx512pf options.
11387 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
11388 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
11389 -mavx512pf options. Fix formatting.
11390 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
11391 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
11392 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
11393 -mavx512cd, -mavx512pf options.
11394 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
11395 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
11396 (ix86_preferred_output_reload_class): Replace SSE_REGS with
11398 (ix86_hard_regno_mode_ok): Support 512-bit registers.
11399 (ix86_set_reg_reg_cost): Ditto.
11400 (x86_order_regs_for_local_alloc): Ditto.
11401 (MAX_VECT_LEN): Extend to 64-byte.
11402 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
11403 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
11404 (TARGET_AVX512ER, TARGET_AVX512CD): New.
11405 (BIGGEST_ALIGNMENT): Extend to 512-bits.
11406 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
11407 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
11408 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
11409 (SSE_REG_MODE_P): Support new modes.
11410 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
11411 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
11412 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
11413 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
11414 (REG_CLASS_CONTENTS): Add new registers.
11415 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
11416 (EXT_REX_SSE_REGNO_P): New.
11417 (HI_REGISTER_NAMES): Add new registers.
11418 * config/i386/i386.md: Define constants for new registers.
11419 (mode): Add new 512-bit modes.
11420 (prefix): Support evex prefix.
11421 (isa): Support avx512f, noavx512f, fma_avx512f.
11422 (ssemodesuffix): Add new 512-bit modes.
11424 (*movxi_internal_avx512f): Ditto.
11425 (*movdi_internal): Replace constraint "x" with the new constraint "v".
11427 (*movsi_internal): Likewise.
11428 (*movdf_internal): Likewise.
11429 (*movsf_internal): Likewise.
11430 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
11431 (<code><mode>3): Likewise.
11432 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
11434 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
11435 with the new constraint "v".
11436 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
11438 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
11439 with the new constraint "v".
11440 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
11441 (<sse2>_storedqu<avxsizesuffix>): Likewise.
11442 (*<plusminus_insn><mode>3): Likewise.
11443 (<sse>_vm<plusminus_insn><mode>3): Likewise.
11444 (*mul<mode>3): Likewise.
11445 (<sse>_vmmul<mode>3): Likewise.
11446 (<sse>_div<mode>3): Likewise.
11447 (<sse>_vmdiv<mode>3): Likewise.
11448 (<sse>_sqrt<mode>2): Likewise.
11449 (<sse>_vmsqrt<mode>2): Likewise.
11450 (*<code><mode>3_finite): Likewise.
11451 (*<code><mode>3) <smaxmin>: Likewise.
11452 (<sse>_vm<code><mode>3): Likewise.
11453 (*<code><mode>3) <any_logic>: Likewise.
11454 (*fma_fmadd_<mode>): Likewise.
11455 (*fma_fmsub_<mode>): Likewise.
11456 (*fma_fnmadd_<mode>): Likewise.
11457 (*fma_fnmsub_<mode>): Likewise.
11458 (*fma_fmaddsub_<mode>): Likewise.
11459 (*fma_fmsubadd_<mode>): Likewise.
11460 (*fmai_fmadd_<mode>): Likewise.
11461 (*fmai_fmsub_<mode>): Likewise.
11462 (*fmai_fnmadd_<mode>): Likewise.
11463 (*fmai_fnmsub_<mode>): Likewise.
11464 (sse_cvtsi2ss): Likewise.
11465 (sse_cvtsi2ssq): Likewise.
11466 (sse_cvtss2si): Likewise.
11467 (sse_cvtss2si_2): Likewise.
11468 (sse_cvtss2siq): Likewise.
11469 (sse_cvtss2siq_2): Likewise.
11470 (sse_cvttss2si): Likewise.
11471 (sse_cvtss2siq_2): Likewise.
11472 (float<sseintvecmodelower><mode>2): Likewise.
11473 (sse2_cvtsd2si_2): Likewise.
11474 (sse2_cvtsd2siq_2): Likewise.
11475 (*<plusminus_insn><mode>3): Likewise.
11476 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
11477 (*<sse4_1_avx2>_mul<mode>3): Likewise.
11478 (ashr<mode>3): Likewise.
11479 (<shift_insn><mode>3): Likewise.
11480 (avx2_<code><mode>3): Likewise.
11481 (*avx2_<code><mode>3): Likewise.
11482 (*andnot<mode>3): Likewise.
11483 (*<code><mode>3) <any_logic>: Likewise.
11484 (abs<mode>2): Likewise.
11485 (avx2_permvar<mode>): Likewise.
11486 (avx2_perm<mode>_1): Likewise.
11487 (*avx_vpermilp<mode>): Likewise.
11488 (avx_vpermilvar<mode>3): Likewise.
11489 (avx2_ashrv<mode>): Likewise.
11490 (avx2_<shift_insn>v<mode>): Likewise.
11491 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
11493 * doc/rtl.texi: Document XImode.
11495 2013-08-21 Jeff Law <law@redhat.com>
11497 * tree-flow.h (register_jump_thread): Pass vector of edges
11498 instead of each important edge.
11499 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
11500 thread path into a vector and pass that to register_jump_thread.
11501 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
11502 passed in edge vector to the current 3-edge form.
11505 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11507 * dce.c (fini_dce): Call df_analyze again just in case
11508 delete_unmarked_insns removed anything.
11510 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
11512 * reload.h (struct reg_equivs): Rename to ..
11513 (struct reg_equivs_s): .. this.
11515 2013-08-20 Martin Liska <marxin.liska@gmail.com>
11517 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
11519 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11521 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
11523 2013-08-21 Jeff Law <law@redhat.com>
11525 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
11526 simplify assignments too. If the RHS collapses to a singleton
11527 range, then return the value for the range.
11529 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
11531 * config/i386/sse.md (V16): Rename to...
11533 (mov<mode>): Update iterator name.
11534 (*mov<mode>_internal): Ditto.
11535 (push<mode>1): Ditto.
11536 (movmisalign<mode>): Ditto.
11538 2013-08-20 Jan Hubicka <jh@suse.cz>
11541 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
11542 entry for direct edges.
11543 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
11545 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11547 Revert my last two changes, r201865 and r201864:
11550 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11552 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11553 instances can own GC refs.
11555 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11556 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11557 (gcc::context::gt_pch_nx): Likewise.
11558 (gcc::context::gt_pch_nx): Likewise.
11559 * ggc.h (gt_ggc_mx <T>): New.
11560 (gt_pch_nx_with_op <T>): New.
11561 (gt_pch_nx <T>): New.
11562 * passes.c (opt_pass::gt_ggc_mx): New.
11563 (opt_pass::gt_pch_nx): New.
11564 (opt_pass::gt_pch_nx_with_op): New.
11565 (pass_manager::gt_ggc_mx): New.
11566 (pass_manager::gt_pch_nx): New.
11567 (pass_manager::gt_pch_nx_with_op): New.
11568 (pass_manager::operator new): Use
11569 ggc_internal_cleared_alloc_stat rather than xcalloc.
11570 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11571 (pass_manager::gt_ggc_mx): New.
11572 (pass_manager::gt_pch_nx): New.
11573 (pass_manager::gt_pch_nx_with_op): New.
11574 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11575 (opt_pass::operator new): New.
11576 (opt_pass::gt_ggc_mx): New.
11577 (opt_pass::gt_pch_nx): New.
11578 (opt_pass::gt_pch_nx_with_op): New.
11581 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11583 * Makefile.in (GTFILES): Add context.h.
11584 * context.c (gcc::context::operator new): New.
11585 (gcc::context::gt_ggc_mx): New.
11586 (gcc::context::gt_pch_nx): New.
11587 (gcc::context::gt_pch_nx): New.
11588 * context.h (gcc::context): Add GTY((user)) marking.
11589 (gcc::context::operator new): New.
11590 (gcc::context::gt_ggc_mx): New.
11591 (gcc::context::gt_pch_nx): New.
11592 (gcc::context::gt_pch_nx): New.
11593 (g): Add GTY marking.
11594 (gt_ggc_mx (gcc::context *)): New.
11595 (gt_pch_nx (gcc::context *)): New.
11596 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11597 void *cookie)): New.
11598 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11600 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11602 * dce.c (fini_dce): Call df_analyze again just in case
11603 delete_unmarked_insns removed anything.
11605 2013-08-20 Teresa Johnson <tejohnson@google.com>
11607 PR rtl-optimizations/57451
11608 * final.c (reemit_insn_block_notes): Prevent lexical blocks
11609 from crossing split section boundaries.
11611 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11613 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
11614 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
11615 with MULTLIB_DEFAULTS.
11617 2013-08-20 Nick Clifton <nickc@redhat.com>
11619 * target.def (narrow_volatile_bitfield): Note that the default
11620 value is false, not !TARGET_STRICT_ALIGN.
11621 * doc/tm.texi: Regenerate.
11623 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
11625 Fix LIB_SPEC for systems without libpthread.
11627 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
11628 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
11630 * config/i386/linux-common.h: Likewise.
11631 * config/mips/linux-common.h: Likewise.
11633 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11635 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
11638 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11640 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11641 instances can own GC refs.
11643 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11644 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11645 (gcc::context::gt_pch_nx): Likewise.
11646 (gcc::context::gt_pch_nx): Likewise.
11647 * ggc.h (gt_ggc_mx <T>): New.
11648 (gt_pch_nx_with_op <T>): New.
11649 (gt_pch_nx <T>): New.
11650 * passes.c (opt_pass::gt_ggc_mx): New.
11651 (opt_pass::gt_pch_nx): New.
11652 (opt_pass::gt_pch_nx_with_op): New.
11653 (pass_manager::gt_ggc_mx): New.
11654 (pass_manager::gt_pch_nx): New.
11655 (pass_manager::gt_pch_nx_with_op): New.
11656 (pass_manager::operator new): Use
11657 ggc_internal_cleared_alloc_stat rather than xcalloc.
11658 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11659 (pass_manager::gt_ggc_mx): New.
11660 (pass_manager::gt_pch_nx): New.
11661 (pass_manager::gt_pch_nx_with_op): New.
11662 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11663 (opt_pass::operator new): New.
11664 (opt_pass::gt_ggc_mx): New.
11665 (opt_pass::gt_pch_nx): New.
11666 (opt_pass::gt_pch_nx_with_op): New.
11668 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11670 * Makefile.in (GTFILES): Add context.h.
11671 * context.c (gcc::context::operator new): New.
11672 (gcc::context::gt_ggc_mx): New.
11673 (gcc::context::gt_pch_nx): New.
11674 (gcc::context::gt_pch_nx): New.
11675 * context.h (gcc::context): Add GTY((user)) marking.
11676 (gcc::context::operator new): New.
11677 (gcc::context::gt_ggc_mx): New.
11678 (gcc::context::gt_pch_nx): New.
11679 (gcc::context::gt_pch_nx): New.
11680 (g): Add GTY marking.
11681 (gt_ggc_mx (gcc::context *)): New.
11682 (gt_pch_nx (gcc::context *)): New.
11683 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11684 void *cookie)): New.
11685 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11687 2013-08-20 Alan Modra <amodra@gmail.com>
11690 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
11691 (rs6000_emit_epilogue): Likewise.
11693 2013-08-19 Dehao Chen <dehao@google.com>
11695 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
11697 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
11698 Jakub Jelinek <jakub@redhat.com>
11700 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
11701 (BUILT_IN_FABSD64): Likewise.
11702 (BUILT_IN_FABSD128): Likewise.
11703 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
11704 (fold_builtin_1): Likewise.
11705 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
11706 destination and source operands.
11707 (*abstd2_fpr): Likewise.
11708 (*nabstd2_fpr): Likewise.
11710 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
11712 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
11715 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
11717 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
11720 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
11722 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
11723 * builtins.c (default_libc_has_function): New.
11724 (gnu_libc_has_function): Ditto.
11725 (no_c99_libc_has_function): Ditto.
11726 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
11727 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
11728 (fold_builtin_sincos): Likewise.
11729 (fold_builtin_cexp): Likewise.
11730 * builtins.def (DEF_C94_BUILTIN): Likewise.
11731 (DEF_C99_BUILTIN): Likewise.
11732 (DEF_C99_C90RES_BUILTIN): Likewise.
11733 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
11734 definitions to using this define.
11735 * config/darwin-protos.h (darwin_libc_has_function): New.
11736 * config/darwin.c (darwin_libc_has_function): Ditto.
11737 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
11738 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
11739 * config/darwin.h: Ditto.
11740 * config/elfos.h: Ditto.
11741 * config/freebsd.h: Ditto.
11742 * config/i386/cygming.h: Ditto.
11743 * config/i386/djgpp.h: Ditto.
11744 * config/i386/i386-interix.h: Ditto.
11745 * config/microblaze/microblaze.h: Ditto.
11746 * config/mmix/mmix.h: Ditto.
11747 * config/gnu-user.h: Ditto.
11748 * config/ia64/hpux.h: Ditto.
11749 * config/pa/pa-hpux.h: Ditto.
11750 * config/pdp11/pdp11.h: Ditto.
11751 * config/picochip/picochip.h: Ditto.
11752 * config/linux.h: Ditto.
11753 * config/netbsd.h: Ditto.
11754 * config/openbsd.h: Ditto.
11755 * config/rs6000/aix43.h: Ditto.
11756 * config/rs6000/aix51.h: Ditto.
11757 * config/rs6000/aix52.h: Ditto.
11758 * config/rs6000/aix53.h: Ditto.
11759 * config/rs6000/aix61.h: Ditto.
11760 * config/rs6000/darwin.h: Ditto.
11761 * config/rs6000/linux.h: Ditto.
11762 * config/rs6000/linux64.h: Ditto.
11763 * config/s390/tpf.h: Ditto.
11764 * config/sol2-10.h: Ditto.
11765 * config/sol2.h: Ditto.
11766 * config/vms/vms.h: Ditto.
11767 * config/vxworks.h: Ditto.
11768 * config/linux-android.c (linux_android_libc_has_function):
11769 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
11770 * config/linux-protos.h (linux_android_libc_has_function):
11772 * config/i386/i386.c (ix86_libc_has_function): New.
11773 * config/i386/i386-protos.h
11774 (ix86_libc_has_function): New declaration.
11775 * config/i386/i386.md
11776 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
11777 ("isinf<mode>2): Likewise.
11778 * convert.c (convert_to_integer): Using new target hook
11779 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
11780 TARGET_C99_FUNCTIONS.
11781 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
11782 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
11783 * coretypes.h (function_class): New enum for different
11784 classes of functions.
11785 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
11786 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
11787 (TARGET_HAS_SINCOS): Likewise.
11788 (TARGET_LIBC_HAS_FUNCTION): New.
11789 * doc/tm.texi: Regenerated.
11790 * targhooks.h (default_libc_has_function): New declaration.
11791 (no_c99_libc_has_function): Ditto.
11792 (gnu_libc_has_function): Ditto.
11793 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
11794 and TARGET_HAS_SINCOS.
11796 2013-08-18 Jan Hubicka <jh@suse.cz>
11798 * Makeifle-in (ipa-devirt.o): New.
11799 (GTFILES): Add ipa-utils.h and ipa-devirt.c
11800 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
11801 (analyze_functions): Look into possible targets of polymorphic call.
11802 * dumpfile.c (dump_files): Add type-inheritance dump.
11803 * dumpfile.h (TDI_inheritance): New.
11804 * ipa-devirt.c: New file.
11805 * ipa-utils.h (odr_type_d): Forward declare.
11806 (odr_type): New type.
11807 (build_type_inheritance_graph): Declare.
11808 (possible_polymorphic_call_targets): Declare and introduce inline
11809 variant when only edge is pased.
11810 (dump_possible_polymorphic_call_targets): Likewise.
11811 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
11812 * tree.c (type_in_anonymous_namespace_p): Break out from ...
11813 (types_same_for_odr): ... here.
11814 * tree.h (type_in_anonymous_namespace_p): Declare.
11816 2013-08-18 Jakub Jelinek <jakub@redhat.com>
11818 PR tree-optimization/58006
11819 * tree-parloops.c (take_address_of): Don't ICE if get_name
11821 (eliminate_local_variables_stmt): Remove clobber stmts.
11823 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
11825 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
11826 error message is issued for an alias to undefined symbol.
11828 2013-08-18 Jan Hubicka <jh@suse.cz>
11830 * cgraph.c (cgraph_create_indirect_edge): Discover
11831 polymorphic calls and record basic info into indirect_info.
11832 * gimple-fold.c (gimple_fold_call): When doing BINFO based
11833 devirtualization, ignore objc function calls.
11834 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
11835 call with no parm index info.
11836 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
11837 * tree.c (virtual_method_call_p): New function.
11838 * tree.h (virtual_method_call_p): Declare.
11840 2013-08-16 Jan Hubicka <jh@suse.cz>
11842 PR middle-end/58179
11843 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
11845 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
11847 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
11850 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11852 * gengtype.c (type_for_name): Add special-case support for
11853 locating types within the "gcc::" namespace.
11854 (open_base_files): Emit a "using namespace gcc" directive.
11856 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11859 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
11860 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
11862 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
11863 array instead of each individual operand as a separate argument.
11864 (emit_fusion_gpr_load): Likewise.
11865 (expand_fusion_gpr_load): Add new function declaration.
11867 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
11868 signature to have the operands passed as an array, instead of as
11869 separate arguments. Allow ZERO_EXTEND to be in the memory
11870 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
11871 depend on the register live/dead flags when peepholes are run.
11872 (expand_fusion_gpr_load): New function to be called from the
11873 peephole2 pass, to change the register that addis sets to be the
11875 (emit_fusion_gpr_load): Change the calling signature to have the
11876 operands passed as an array, instead of as separate arguments.
11877 Allow ZERO_EXTEND to be in the memory address, and also
11878 SIGN_EXTEND if -mpower8-fusion-sign.
11880 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
11881 unspec enumeration.
11882 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
11883 adjust the register addis loads up in the peephole2 pass. Do not
11884 depend on the register live/dead state when the peephole pass is done.
11886 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11888 * gengtype.c (create_user_defined_type): Ensure that the kind
11889 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
11890 declaration is seen before the GTY((user)) marking.
11892 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
11895 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
11897 2013-08-16 Jan Hubicka <jh@suse.cz>
11899 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
11900 arugment expected_type.
11901 (gimple_fold_call): Use it.
11902 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
11903 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
11904 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
11905 (try_make_edge_direct_virtual_call): Likewise.
11906 * tree.c (obj_type_ref_class): New.
11907 * tree.h (obj_type_ref_class): Use it.
11909 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
11911 * sched-vis.c (rtl_slim_pp_initialized): Remove.
11912 (rtl_slim_pp): Likewise.
11913 (init_rtl_slim_pretty_print): Likewise.
11914 (dump_value_slim): Don't call it. Use local pretty printer.
11915 (dump_insn_slim): Likewise.
11916 (dump_rtl_slim): Likewise.
11917 (str_pattern_slim): Likewise.
11918 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
11921 2013-08-16 Jakub Jelinek <jakub@redhat.com>
11923 PR tree-optimization/58164
11924 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
11925 walk gimple_goto_dest of GIMPLE_GOTO.
11927 PR tree-optimization/58165
11928 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
11929 bi_call must be the last stmt in a bb, don't split_block, instead
11930 use fallthru edge from it and give up if there is none.
11931 Release conds vector when returning early.
11933 2013-08-14 Xinliang David Li <davidxl@google.com>
11935 * config/i386/i386.c (ix86_option_override_internal):
11936 Remove unused variable and field.
11938 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11941 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
11942 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
11943 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
11944 and Linux, correct BLKmode alignment when 128-bit alignment is
11945 required and compatibility flag is not set.
11946 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
11947 for zero-size arguments when compatibility flag is not set.
11949 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11951 PR tree-optimization/58145
11952 * tree-sra.c (build_ref_for_offset): If prev_base has
11953 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
11955 2013-08-14 Xinliang David Li <davidxl@google.com>
11957 * config/i386/i386.c (ix86_option_override_internal):
11958 Fix uninitialized variable error.
11960 2013-08-14 Xinliang David Li <davidxl@google.com>
11962 * config/i386/i386.opt: Define two new options.
11963 * config/i386/x86-tune.def: Add arch selector field in macros.
11964 * config/i386/i386.h: Adjust macro definition.
11965 * config/i386/i386.c (ix86_option_override_internal):
11967 (parse_mtune_ctrl_str): New function.
11968 (set_ix86_tune_features): New function.
11969 (ix86_function_specific_restore): Call the new helper function.
11971 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
11973 PR rtl-optimization/57662
11974 * sel-sched.c (code_motion_process_successors): When the current insn
11975 is removed after the recursive traversal, break from the loop.
11976 Add comments and debug printouts.
11978 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11979 Alexandre Oliva <aoliva@redhat.com>
11982 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
11983 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
11984 there also UNSPEC_PLTOFF.
11986 2013-08-14 Marek Polacek <polacek@redhat.com>
11988 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
11989 NUM_CONDITIONS bit positions.
11991 2013-08-13 Cary Coutant <ccoutant@google.com>
11993 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
11994 (attr_checksum): Hash vector contents instead of pointer.
11995 (attr_checksum_ordered): Likewise.
11997 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
11999 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
12000 when Pmode != word_mode. Add length_address attribute.
12001 (sse3_monitor_<mode>): Merge from sse3_monitor and
12002 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
12003 Pmode != word_mode. Update insn length attribute.
12004 * config/i386/i386.c (ix86_option_override_internal): Update
12005 ix86_gen_monitor selection for merged sse3_monitor insn.
12007 2013-08-13 Julian Brown <julian@codesourcery.com>
12009 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
12010 perform invalid legitimization on greater-than-word-size modes for
12011 TARGET_E500_DOUBLE.
12013 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
12015 * ira.c (setup_class_translate_array): Use aclass instead of cl
12016 for classes not fully covered by allocno classes.
12018 2013-08-13 Jakub Jelinek <jakub@redhat.com>
12020 PR tree-optimization/57661
12021 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
12022 * tree-inline.c (tree_function_versioning): Initialize it.
12023 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
12024 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
12025 that is not being copied.
12028 * asan.c (instrument_strlen_call): Fix typo in comment.
12029 Use char * type even for the lhs of POINTER_PLUS_EXPR.
12031 2013-08-13 Steve Ellcey <sellcey@mips.com>
12033 * config/mips/mips.md (prefetch): Use lw instead of ld on
12034 loongson in 32bit mode.
12036 2013-08-13 Nick Clifton <nickc@redhat.com>
12038 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
12040 2013-08-13 Jan Hubicka <jh@suse.cz>
12042 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
12043 introduced edge; fix typo in sanity check.
12044 (cgraph_resolve_speculation): Export; improve diagnostic.
12045 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
12046 speculation at type mismatch.
12047 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
12048 (cgraph_resolve_speculation): Declare.
12049 (symtab_can_be_discarded): New function.
12050 * value-prof.c (gimple_ic_transform): Remove actual transform code.
12051 * ipa-inline-transform.c (speculation_removed): New global var.
12052 (clone_inlined_nodes): See if speculation can be removed.
12053 (inline_call): If speculations was removed, we growths may not match.
12054 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
12055 (speculation_useful_p): New function.
12056 (resolve_noninline_speculation): New function.
12057 (inline_small_functions): Resolve useless speculations.
12058 * ipa-inline.h (speculation_useful_p): Declare
12059 * ipa.c (can_replace_by_local_alias): Simplify.
12060 (ipa_profile): Produce speculative calls in non-lto, too;
12061 add simple cost model; produce local aliases.
12063 2013-08-13 David Malcolm <dmalcolm@redhat.com>
12065 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
12068 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
12070 * config/i386/i386.c (ix86_function_versions): Use error + inform.
12072 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
12074 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
12075 iterator instead of X87MODEF.
12077 2013-08-12 Perez Read <netfirewall@gmail.com>
12080 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
12081 operand 0 for intel asm alternative.
12082 (*movabs<mode>_2): Ditto for operand 1.
12084 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
12086 * config/aarch64/arm_none.h
12087 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
12089 2013-08-12 Nick Clifton <nickc@redhat.com>
12091 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
12093 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
12095 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
12096 expand for QI/HImode operand to produce more effictive code for
12097 unsigned char(short) --> float(double) conversion.
12099 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
12101 * doc/invoke.texi: Mention that -ftls-model does not force the final
12104 2013-08-12 Marek Polacek <polacek@redhat.com>
12105 Marc Glisse <marc.glisse@inria.fr>
12107 PR tree-optimization/57980
12108 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
12109 when creating -1 constant.
12111 2013-08-10 Jan Hubicka <jh@suse.cz>
12113 Workaround binutils PR14342.
12114 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
12115 (gimple_init_edge_profiler): Likewise.
12116 (gimple_gen_ic_func_profiler): Likewise.
12118 2013-08-09 Jan Hubicka <jh@suse.cz>
12120 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
12122 2013-08-09 Xinliang David Li <davidxl@google.com>
12124 * config/i386/stringop.def: New file.
12125 * config/i386/stringop.opt: New file.
12126 * config/i386/i386-opts.h: Include stringopt.def.
12127 * config/i386/i386.opt: Include stringopt.opt.
12128 * config/i386/i386.c (ix86_option_override_internal):
12129 Override default size based stringop inline strategies with options.
12130 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
12133 2013-08-09 Jan Hubicka <jh@suse.cz>
12135 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
12137 2013-08-09 Jan Hubicka <jh@suse.cz>
12139 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
12141 (dump_cgraph_node): Dump profile-id.
12142 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
12143 and common_target_probability.
12144 * lto-cgraph.c (lto_output_edge): Stream common targets.
12145 (lto_output_node): Stream profile ids.
12146 (input_node): Stream profile ids.
12147 (input_edge): Stream common targets.
12148 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
12149 * ipa.c: Include value-prof.h
12150 (ipa_profile_generate_summary): Turn indirect call statement histograms
12151 into common targets.
12152 (ipa_profile): Turn common targets into speculative edges.
12154 2013-08-09 Jan Hubicka <jh@suse.cz>
12156 * cgraph.h (cgraph_node): Add profile_id.
12157 * value-prof.c (cgraph_node_map): Turn into pointer_map.
12158 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
12159 (del_node_map): Update.
12160 (find_func_by_funcdef_no): Replace by ...
12161 (find_func_by_profile_id): ... this one.
12162 (gimple_ic_transform): Do not remove useful histograms when
12163 speculation is not done; dump info when indirect call removal
12165 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
12166 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
12167 (__gcov_indirect_call_profiler_v2): .. this one.
12168 * profile.h (init_node_map): Update.
12169 * coverage.c (coverage_compute_profile_id): New function.
12170 * coverage.h (coverage_compute_profile_id): Declare.
12171 * tree-profile.c (init_ic_make_global_vars): Make
12172 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
12173 (gimple_init_edge_profiler): Update prototype of
12174 __gcov_indirect_call_profiler.
12175 (gimple_gen_ic_func_profiler): Simplify.
12176 (tree_profiling): Use init_node_map
12178 2013-08-09 Jan Hubicka <jh@suse.cz>
12180 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
12181 non-speculative refs.
12182 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
12183 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
12184 (cgraph_set_call_stmt): Likewise.
12185 (cgraph_create_edge_1): Fix release checking compilatoin;
12186 clear lto_stmt_uid.
12187 (cgraph_free_edge): Free indirect info.
12188 (cgraph_turn_edge_to_speculative): New function.
12189 (cgraph_speculative_call_info): New function.
12190 (cgraph_make_edge_direct): Return direct edge; handle speculation.
12191 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
12192 (dump_cgraph_node): Dump speculation.
12193 (verify_edge_count_and_frequency): Accept speculative edges.
12194 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
12195 (verify_cgraph_node): Handle speculation.
12196 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
12197 (cgraph_set_call_stmt): Update prototype.
12198 (cgraph_make_edge_direct): Update prototype.
12199 (cgraph_speculative_call_info): Declare.
12200 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
12201 to change; update call of ipa_find_references.
12202 * ipa-ref.c (ipa_record_reference): Fix return value; clear
12203 lto_stmt_uid and speculative flags.
12204 (ipa_dump_references): Dump speculation.
12205 (ipa_clone_references): Clone speculative flag.
12206 (ipa_clone_referring): Likewise.
12207 (ipa_clone_ref): New function.
12208 (ipa_find_reference): Look into lto_stmt_uids
12209 (ipa_clear_stmts_in_references): Do not clear speculative calls.
12210 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
12211 (ipa_find_reference): Update declaration.
12212 (ipa_clone_ref): Declare.
12213 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
12214 stream speculative flag.
12215 (lto_output_ref): Stream statements uids and speculation.
12216 (input_ref): Likewise.
12217 (input_edge): Stream speuclation.
12218 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
12219 (cgraph_set_call_stmt_including_clones): Handle speculation.
12220 * ipa-inline.c (heap_edge_removal_hook): New function.
12221 (inline_small_functions): Register it.
12222 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
12223 also initialize refs.
12224 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
12226 (try_make_edge_direct_simple_call): Likewise.
12227 (try_make_edge_direct_simple_call): Likewise.
12228 (update_indirect_edges_after_inlining): Likewise.
12229 (remove_described_reference): Look proper lto_stmt_uid.
12230 (propagate_controlled_uses): Likewise.
12231 (propagate_controlled_uses): Liekwise.
12232 * tree-inline.c (copy_bb): Copy speculative edges.
12233 (redirect_all_calls): New function.
12234 (copy_cfg_body): Do redirection after loop info is updated.
12235 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
12237 2013-08-09 Jan Hubicka <jh@suse.cz>
12239 * lto-streamer-out.c (output_function): Renumber PHIs.
12240 * lto-streamer-in.c (input_function): Likewise.
12242 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
12244 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
12245 (get_lane_unsigned): Likewise.
12246 (dup_lane_scalar): Likewise.
12247 (get_lane): enable for VALL.
12248 * config/aarch64/aarch64-simd.md
12249 (aarch64_dup_lane_scalar<mode>): Remove.
12250 (aarch64_get_lane_signed<mode>): Likewise.
12251 (aarch64_get_lane_unsigned<mode>): Likewise.
12252 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
12253 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
12254 (aarch64_get_lane<mode>): Enable for all vector modes.
12255 (aarch64_get_lanedi): Remove misleading constraints.
12256 * config/aarch64/arm_neon.h
12257 (__aarch64_vget_lane_any): Define.
12258 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
12259 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
12260 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
12261 * config/aarch64/iterators.md (VDQQH): New.
12262 (VDQQHS): Likewise.
12263 (vwcore): Likewise.
12265 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
12267 * configure.ac: Add GAS check for LEON instructions on SPARC.
12268 * configure: Regenerate.
12269 * config.in: Likewise.
12270 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
12272 * config/sparc/sparc.opt (LEON, LEON3): New masks.
12273 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
12275 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
12276 (AS_LEON_FLAG): New macro.
12277 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
12278 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
12279 Deal with LEON and LEON3 for the memory model.
12280 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
12281 (atomic_compare_and_swap<mode>_1): Likewise.
12282 (*atomic_compare_and_swap<mode>_1): Likewise.
12284 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12286 * config/arm/neon.md (vcond): Fix floating-point vector
12287 comparisons against 0.
12289 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
12291 * lra-constraints.c (emit_spill_move): Remove assert.
12292 (process_alt_operands): Add more debugging
12293 output. Increase reject for spilling into memory. Decrease
12294 reject for reloading scratch.
12295 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
12297 2013-08-08 Steve Ellcey <sellcey@mips.com>
12299 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
12300 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
12301 micromips incompatible. Add nan2008.
12302 (MULTILIB_DIRNAMES): Add nan2008.
12303 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12304 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
12305 and micromips incompatible. Add nan2008.
12306 (MULTILIB_DIRNAMES): Add nan2008.
12307 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12309 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
12311 PR rtl-optimization/58079
12312 * combine.c (combine_simplify_rtx): Avoid using SUBST if
12313 simplify_comparison has widened a comparison with an integer.
12315 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12317 * config/arm/neon.md (movmisalign<mode>): Disable when we
12318 don't allow unaligned accesses.
12319 (*movmisalign<mode>_neon_store): Likewise.
12320 (*movmisalign<mode>_neon_load): Likewise.
12321 (*movmisalign<mode>_neon_store): Likewise.
12322 (*movmisalign<mode>_neon_load): Likewise.
12324 2013-08-08 Jan Hubicka <jh@suse.cz>
12326 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
12327 (make_pass_rebuild_cgraph_edges): Also clear references.
12328 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
12329 * ipa-inline-transform.c (inline_transform): Remove all references
12331 * cgraphunit.c (expand_function): Remove all references after
12333 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
12334 (ipa_find_reference): Rewrite to iterator.
12335 (remove_stmt_references): Likewise.
12336 (ipa_clear_stmts_in_references): New function.
12337 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
12338 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
12340 * ipa-split.c (split_function): Remove references in split function.
12342 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12345 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
12346 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
12348 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12351 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
12352 suggested mode for the assignment isn't compatible with the
12353 registers required.
12355 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12358 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
12360 2013-08-07 Xinliang David Li <davidxl@google.com>
12362 * config/i386/i386.opt: New option -mtune-ctrl=.
12363 * config/i386/x86-tune.def: New file.
12364 * config/i386/i386.h: include x86-tune.def.
12365 * config/i386/i386.c (ix86_option_override_internal):
12366 Parsing -mtune-ctrl= option and set tune features.
12368 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
12371 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
12372 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
12375 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12377 * rtl.h (update_alignments): Declare.
12378 * final.c (grow_label_align): New function extracted from...
12379 (shorten_branches): ...here. Call it.
12380 (update_alignments): New function.
12381 * reorg.c (sibling_labels): New variable.
12382 (get_label_before): Add SIBLING parameter. If it is non-zero, push
12383 the new label along with it onto the sibling_labels vector.
12384 (fill_simple_delay_slots): Adjust call to get_label_before.
12385 (fill_slots_from_thread): Likewise.
12386 (relax_delay_slots): Likewise.
12387 (make_return_insns): Likewise.
12388 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
12390 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12392 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
12393 document its semantics.
12394 (diagnostic_report_diagnostic): Adjust accordingly.
12396 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12398 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
12399 (sparc_option_override): ...and port to new C++ pass API.
12400 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
12402 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
12404 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
12406 2013-08-06 Caroline Tice <cmtice@google.com>
12408 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
12409 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
12410 * tree-pass.h: Add pass_vtable_verify.
12411 * varasm.c (assemble_variable): Add code to properly set the comdat
12412 section and name for the .vtable_map_vars section.
12413 (assemble_vtyv_preinit_initializer): New function.
12414 (default_sectin_type_flags): Make sure .vtable_map_vars section has
12416 * output.h: Add function decl for assemble_vtv_preinit_initializer.
12417 * vtable-verify.c: New file.
12418 * vtable-verify.h: New file.
12419 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
12420 initialiation levels.
12421 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
12422 * passes.def: Insert pass_vtable_verify.
12423 * aclocal.m4: Reorder includes.
12424 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
12425 -fvtv-counts options.
12426 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
12427 as appropriate, if -fvtable-verify=... is used.
12428 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
12429 -fvtable-verify=... is used.
12430 * Makefile.in (OBJS): Add vtable-verify.o to list.
12431 (vtable-verify.o): Add new build rule.
12432 (GTFILES): Add vtable-verify.c to list.
12433 * common.opt (fvtable-verify=): New flag.
12434 (vtv_priority): Values for fvtable-verify= flag.
12435 (fvtv-counts): New flag.
12436 (fvtv-debug): New flag.
12437 * tree.h (save_vtable_map_decl): New extern function decl.
12439 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12441 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
12442 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
12443 (pass_data_rl78_devirt): ...new pass_data instance and...
12444 (make_pass_rl78_devirt): ...new function.
12445 (rl78_asm_file_start): Port pass registration to new C++ API.
12447 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12449 * coretypes.h (rtl_opt_pass): Add.
12450 (gcc::context): Add.
12451 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
12452 (epiphany_init): Port to new C++ pass API.
12453 (epiphany_optimize_mode_switching): Likewise.
12454 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
12455 (pass_manager::get_pass_mode_switching): New.
12456 (pass_manager::get_pass_peephole2): New.
12457 * mode-switching.c (pass_mode_switching): Add clone method.
12458 * recog.c (pass_peephole2): Add clone method.
12459 (pass_split_all_insns): Add clone method.
12461 2013-08-06 David Malcolm <dmalcolm@redhat.com>
12463 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
12464 (mips_option_override): ...here, porting to new C++ API for passes.
12466 2013-08-06 Jan Hubicka <jh@suse.cz>
12468 * cgraph.c (cgraph_get_body): New function based on lto.c
12470 * cgraph.h (cgraph_get_body): Declare.
12471 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
12473 * cgraphunit.c (expand_function): Get body prior expanding.
12474 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
12475 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
12477 * passes.c (do_per_function_toporder): Get body.
12478 * tree-inline.c (expand_call_inline): Get body prior inlining it.
12479 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
12481 2013-08-06 Martin Jambor <mjambor@suse.cz>
12484 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
12485 is not re-finalized. Rename second parameter to no_collect.
12487 2013-08-06 Martin Jambor <mjambor@suse.cz>
12489 PR middle-end/58041
12490 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
12491 MEM_REF has proper alignment information.
12493 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
12496 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
12498 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
12499 rtx (*) (rtx, ...) with insn_gen_fn.
12500 * genoutput.c (output_insn_data): Cast gen_? function pointers to
12501 insn_gen_fn::stored_funcptr. Add initializer braces.
12503 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12505 Rewrite how instances of passes are cloned to remove assumptions
12506 about their sizes (thus allowing pass subclasses to have
12507 additional data fields, albeit non-GC-managed ones at this point).
12509 * passes.c (make_pass_instance): Now that passes have clone
12510 methods, rewrite this function to eliminate XNEW and memcpy
12511 calls that used hardcoded sizes. Since this function no longer
12512 creates pass instances, rename it to...
12513 (add_pass_instance): ...this. Document the old way that passes were
12514 numbered and flagged, and rework this function to continue using it.
12515 (next_pass_1): Add an initial_pass argument for use by
12517 (position_pass): When adding multiple instances of a pass, use
12518 the pass's clone method, rather than relying on the XNEW/memcpy
12519 within the former make_pass_instance (now add_pass_instance).
12520 (pass_manager::pass_manager): When invoking next_pass_1, also supply
12521 the initial instance of the current pass within the pass manager.
12523 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12525 This is the automated part of the conversion of passes from C
12526 structs to C++ classes.
12528 Patch autogenerated by refactor_passes.py from
12529 https://github.com/davidmalcolm/gcc-refactoring-scripts
12530 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
12532 * asan.c (pass_asan): Convert from a global struct to a subclass of
12533 gimple_opt_pass along with...
12534 (pass_data_asan): ...new pass_data instance and...
12535 (make_pass_asan): ...new function.
12536 (pass_asan_O0): Convert from a global struct to a subclass of
12537 gimple_opt_pass along with...
12538 (pass_data_asan_O0): ...new pass_data instance and...
12539 (make_pass_asan_O0): ...new function.
12540 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
12541 subclass of rtl_opt_pass along with...
12542 (pass_data_inc_dec): ...new pass_data instance and...
12543 (make_pass_inc_dec): ...new function.
12544 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
12545 a subclass of rtl_opt_pass along with...
12546 (pass_data_reorder_blocks): ...new pass_data instance and...
12547 (make_pass_reorder_blocks): ...new function.
12548 (pass_duplicate_computed_gotos): Convert from a global struct to a
12549 subclass of rtl_opt_pass along with...
12550 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
12551 (make_pass_duplicate_computed_gotos): ...new function.
12552 (pass_partition_blocks): Convert from a global struct to a subclass of
12553 rtl_opt_pass along with...
12554 (pass_data_partition_blocks): ...new pass_data instance and...
12555 (make_pass_partition_blocks): ...new function.
12556 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
12557 struct to a subclass of rtl_opt_pass along with...
12558 (pass_data_branch_target_load_optimize1): ...new pass_data instance
12560 (make_pass_branch_target_load_optimize1): ...new function.
12561 (pass_branch_target_load_optimize2): Convert from a global struct to a
12562 subclass of rtl_opt_pass along with...
12563 (pass_data_branch_target_load_optimize2): ...new pass_data instance
12565 (make_pass_branch_target_load_optimize2): ...new function.
12566 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
12567 of rtl_opt_pass along with...
12568 (pass_data_jump): ...new pass_data instance and...
12569 (make_pass_jump): ...new function.
12570 (pass_jump2): Convert from a global struct to a subclass of
12571 rtl_opt_pass along with...
12572 (pass_data_jump2): ...new pass_data instance and...
12573 (make_pass_jump2): ...new function.
12574 * cfgexpand.c (pass_expand): Convert from a global struct to a
12575 subclass of rtl_opt_pass along with...
12576 (pass_data_expand): ...new pass_data instance and...
12577 (make_pass_expand): ...new function.
12578 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
12579 of rtl_opt_pass along with...
12580 (pass_data_free_cfg): ...new pass_data instance and...
12581 (make_pass_free_cfg): ...new function.
12582 (pass_into_cfg_layout_mode): Convert from a global struct to a
12583 subclass of rtl_opt_pass along with...
12584 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
12585 (make_pass_into_cfg_layout_mode): ...new function.
12586 (pass_outof_cfg_layout_mode): Convert from a global struct to a
12587 subclass of rtl_opt_pass along with...
12588 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
12589 (make_pass_outof_cfg_layout_mode): ...new function.
12590 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
12591 struct to a subclass of gimple_opt_pass along with...
12592 (pass_data_build_cgraph_edges): ...new pass_data instance and...
12593 (make_pass_build_cgraph_edges): ...new function.
12594 (pass_rebuild_cgraph_edges): Convert from a global struct to a
12595 subclass of gimple_opt_pass along with...
12596 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
12597 (make_pass_rebuild_cgraph_edges): ...new function.
12598 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
12599 subclass of gimple_opt_pass along with...
12600 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
12602 (make_pass_remove_cgraph_callee_edges): ...new function.
12603 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
12604 struct to a subclass of rtl_opt_pass along with...
12605 (pass_data_stack_adjustments): ...new pass_data instance and...
12606 (make_pass_stack_adjustments): ...new function.
12607 * combine.c (pass_combine): Convert from a global struct to a subclass
12608 of rtl_opt_pass along with...
12609 (pass_data_combine): ...new pass_data instance and...
12610 (make_pass_combine): ...new function.
12611 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
12612 global struct to a subclass of rtl_opt_pass along with...
12613 (pass_data_compare_elim_after_reload): ...new pass_data instance
12615 (make_pass_compare_elim_after_reload): ...new function.
12616 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
12617 of rtl_opt_pass along with...
12618 (pass_data_rtl_cprop): ...new pass_data instance and...
12619 (make_pass_rtl_cprop): ...new function.
12620 * cse.c (pass_cse): Convert from a global struct to a subclass of
12621 rtl_opt_pass along with...
12622 (pass_data_cse): ...new pass_data instance and...
12623 (make_pass_cse): ...new function.
12624 (pass_cse2): Convert from a global struct to a subclass of
12625 rtl_opt_pass along with...
12626 (pass_data_cse2): ...new pass_data instance and...
12627 (make_pass_cse2): ...new function.
12628 (pass_cse_after_global_opts): Convert from a global struct to a
12629 subclass of rtl_opt_pass along with...
12630 (pass_data_cse_after_global_opts): ...new pass_data instance and...
12631 (make_pass_cse_after_global_opts): ...new function.
12632 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
12633 of rtl_opt_pass along with...
12634 (pass_data_ud_rtl_dce): ...new pass_data instance and...
12635 (make_pass_ud_rtl_dce): ...new function.
12636 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
12637 rtl_opt_pass along with...
12638 (pass_data_fast_rtl_dce): ...new pass_data instance and...
12639 (make_pass_fast_rtl_dce): ...new function.
12640 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
12641 a subclass of rtl_opt_pass along with...
12642 (pass_data_df_initialize_opt): ...new pass_data instance and...
12643 (make_pass_df_initialize_opt): ...new function.
12644 (pass_df_initialize_no_opt): Convert from a global struct to a
12645 subclass of rtl_opt_pass along with...
12646 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
12647 (make_pass_df_initialize_no_opt): ...new function.
12648 (pass_df_finish): Convert from a global struct to a subclass of
12649 rtl_opt_pass along with...
12650 (pass_data_df_finish): ...new pass_data instance and...
12651 (make_pass_df_finish): ...new function.
12652 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
12653 rtl_opt_pass along with...
12654 (pass_data_rtl_dse1): ...new pass_data instance and...
12655 (make_pass_rtl_dse1): ...new function.
12656 (pass_rtl_dse2): Convert from a global struct to a subclass of
12657 rtl_opt_pass along with...
12658 (pass_data_rtl_dse2): ...new pass_data instance and...
12659 (make_pass_rtl_dse2): ...new function.
12660 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
12661 subclass of rtl_opt_pass along with...
12662 (pass_data_dwarf2_frame): ...new pass_data instance and...
12663 (make_pass_dwarf2_frame): ...new function.
12664 * except.c (pass_set_nothrow_function_flags): Convert from a global
12665 struct to a subclass of rtl_opt_pass along with...
12666 (pass_data_set_nothrow_function_flags): ...new pass_data instance
12668 (make_pass_set_nothrow_function_flags): ...new function.
12669 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
12670 subclass of rtl_opt_pass along with...
12671 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
12673 (make_pass_convert_to_eh_region_ranges): ...new function.
12674 * final.c (pass_compute_alignments): Convert from a global struct to a
12675 subclass of rtl_opt_pass along with...
12676 (pass_data_compute_alignments): ...new pass_data instance and...
12677 (make_pass_compute_alignments): ...new function.
12678 (pass_final): Convert from a global struct to a subclass of
12679 rtl_opt_pass along with...
12680 (pass_data_final): ...new pass_data instance and...
12681 (make_pass_final): ...new function.
12682 (pass_shorten_branches): Convert from a global struct to a subclass of
12683 rtl_opt_pass along with...
12684 (pass_data_shorten_branches): ...new pass_data instance and...
12685 (make_pass_shorten_branches): ...new function.
12686 (pass_clean_state): Convert from a global struct to a subclass of
12687 rtl_opt_pass along with...
12688 (pass_data_clean_state): ...new pass_data instance and...
12689 (make_pass_clean_state): ...new function.
12690 * function.c (pass_instantiate_virtual_regs): Convert from a global
12691 struct to a subclass of rtl_opt_pass along with...
12692 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
12693 (make_pass_instantiate_virtual_regs): ...new function.
12694 (pass_leaf_regs): Convert from a global struct to a subclass of
12695 rtl_opt_pass along with...
12696 (pass_data_leaf_regs): ...new pass_data instance and...
12697 (make_pass_leaf_regs): ...new function.
12698 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
12699 subclass of rtl_opt_pass along with...
12700 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
12702 (make_pass_thread_prologue_and_epilogue): ...new function.
12703 (pass_match_asm_constraints): Convert from a global struct to a
12704 subclass of rtl_opt_pass along with...
12705 (pass_data_match_asm_constraints): ...new pass_data instance and...
12706 (make_pass_match_asm_constraints): ...new function.
12707 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
12708 subclass of rtl_opt_pass along with...
12709 (pass_data_rtl_fwprop): ...new pass_data instance and...
12710 (make_pass_rtl_fwprop): ...new function.
12711 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
12712 rtl_opt_pass along with...
12713 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
12714 (make_pass_rtl_fwprop_addr): ...new function.
12715 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
12716 rtl_opt_pass along with...
12717 (pass_data_rtl_pre): ...new pass_data instance and...
12718 (make_pass_rtl_pre): ...new function.
12719 (pass_rtl_hoist): Convert from a global struct to a subclass of
12720 rtl_opt_pass along with...
12721 (pass_data_rtl_hoist): ...new pass_data instance and...
12722 (make_pass_rtl_hoist): ...new function.
12723 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
12724 subclass of gimple_opt_pass along with...
12725 (pass_data_lower_cf): ...new pass_data instance and...
12726 (make_pass_lower_cf): ...new function.
12727 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
12728 from a global struct to a subclass of gimple_opt_pass along with...
12729 (pass_data_strength_reduction): ...new pass_data instance and...
12730 (make_pass_strength_reduction): ...new function.
12731 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
12732 of rtl_opt_pass along with...
12733 (pass_data_rtl_ifcvt): ...new pass_data instance and...
12734 (make_pass_rtl_ifcvt): ...new function.
12735 (pass_if_after_combine): Convert from a global struct to a subclass of
12736 rtl_opt_pass along with...
12737 (pass_data_if_after_combine): ...new pass_data instance and...
12738 (make_pass_if_after_combine): ...new function.
12739 (pass_if_after_reload): Convert from a global struct to a subclass of
12740 rtl_opt_pass along with...
12741 (pass_data_if_after_reload): ...new pass_data instance and...
12742 (make_pass_if_after_reload): ...new function.
12743 * init-regs.c (pass_initialize_regs): Convert from a global struct to
12744 a subclass of rtl_opt_pass along with...
12745 (pass_data_initialize_regs): ...new pass_data instance and...
12746 (make_pass_initialize_regs): ...new function.
12747 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
12748 of ipa_opt_pass_d along with...
12749 (pass_data_ipa_cp): ...new pass_data instance and...
12750 (make_pass_ipa_cp): ...new function.
12751 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
12752 global struct to a subclass of gimple_opt_pass along with...
12753 (pass_data_inline_parameters): ...new pass_data instance and...
12754 (make_pass_inline_parameters): ...new function.
12755 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
12756 subclass of gimple_opt_pass along with...
12757 (pass_data_early_inline): ...new pass_data instance and...
12758 (make_pass_early_inline): ...new function.
12759 (pass_ipa_inline): Convert from a global struct to a subclass of
12760 ipa_opt_pass_d along with...
12761 (pass_data_ipa_inline): ...new pass_data instance and...
12762 (make_pass_ipa_inline): ...new function.
12763 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
12764 struct to a subclass of gimple_opt_pass along with...
12765 (pass_data_local_pure_const): ...new pass_data instance and...
12766 (make_pass_local_pure_const): ...new function.
12767 (pass_ipa_pure_const): Convert from a global struct to a subclass of
12768 ipa_opt_pass_d along with...
12769 (pass_data_ipa_pure_const): ...new pass_data instance and...
12770 (make_pass_ipa_pure_const): ...new function.
12771 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
12772 to a subclass of ipa_opt_pass_d along with...
12773 (pass_data_ipa_reference): ...new pass_data instance and...
12774 (make_pass_ipa_reference): ...new function.
12775 * ipa-split.c (pass_split_functions): Convert from a global struct to
12776 a subclass of gimple_opt_pass along with...
12777 (pass_data_split_functions): ...new pass_data instance and...
12778 (make_pass_split_functions): ...new function.
12779 (pass_feedback_split_functions): Convert from a global struct to a
12780 subclass of gimple_opt_pass along with...
12781 (pass_data_feedback_split_functions): ...new pass_data instance and...
12782 (make_pass_feedback_split_functions): ...new function.
12783 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
12784 global struct to a subclass of simple_ipa_opt_pass along with...
12785 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
12787 (make_pass_ipa_function_and_variable_visibility): ...new function.
12788 (pass_ipa_free_inline_summary): Convert from a global struct to a
12789 subclass of simple_ipa_opt_pass along with...
12790 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
12791 (make_pass_ipa_free_inline_summary): ...new function.
12792 (pass_ipa_whole_program_visibility): Convert from a global struct to a
12793 subclass of ipa_opt_pass_d along with...
12794 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
12796 (make_pass_ipa_whole_program_visibility): ...new function.
12797 (pass_ipa_profile): Convert from a global struct to a subclass of
12798 ipa_opt_pass_d along with...
12799 (pass_data_ipa_profile): ...new pass_data instance and...
12800 (make_pass_ipa_profile): ...new function.
12801 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
12802 ipa_opt_pass_d along with...
12803 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
12804 (make_pass_ipa_cdtor_merge): ...new function.
12805 * ira.c (pass_ira): Convert from a global struct to a subclass of
12806 rtl_opt_pass along with...
12807 (pass_data_ira): ...new pass_data instance and...
12808 (make_pass_ira): ...new function.
12809 (pass_reload): Convert from a global struct to a subclass of
12810 rtl_opt_pass along with...
12811 (pass_data_reload): ...new pass_data instance and...
12812 (make_pass_reload): ...new function.
12813 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
12814 subclass of rtl_opt_pass along with...
12815 (pass_data_cleanup_barriers): ...new pass_data instance and...
12816 (make_pass_cleanup_barriers): ...new function.
12817 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
12818 of rtl_opt_pass along with...
12819 (pass_data_loop2): ...new pass_data instance and...
12820 (make_pass_loop2): ...new function.
12821 (pass_rtl_loop_init): Convert from a global struct to a subclass of
12822 rtl_opt_pass along with...
12823 (pass_data_rtl_loop_init): ...new pass_data instance and...
12824 (make_pass_rtl_loop_init): ...new function.
12825 (pass_rtl_loop_done): Convert from a global struct to a subclass of
12826 rtl_opt_pass along with...
12827 (pass_data_rtl_loop_done): ...new pass_data instance and...
12828 (make_pass_rtl_loop_done): ...new function.
12829 (pass_rtl_move_loop_invariants): Convert from a global struct to a
12830 subclass of rtl_opt_pass along with...
12831 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
12832 (make_pass_rtl_move_loop_invariants): ...new function.
12833 (pass_rtl_unswitch): Convert from a global struct to a subclass of
12834 rtl_opt_pass along with...
12835 (pass_data_rtl_unswitch): ...new pass_data instance and...
12836 (make_pass_rtl_unswitch): ...new function.
12837 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
12838 subclass of rtl_opt_pass along with...
12839 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
12841 (make_pass_rtl_unroll_and_peel_loops): ...new function.
12842 (pass_rtl_doloop): Convert from a global struct to a subclass of
12843 rtl_opt_pass along with...
12844 (pass_data_rtl_doloop): ...new pass_data instance and...
12845 (make_pass_rtl_doloop): ...new function.
12846 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
12847 a subclass of rtl_opt_pass along with...
12848 (pass_data_lower_subreg): ...new pass_data instance and...
12849 (make_pass_lower_subreg): ...new function.
12850 (pass_lower_subreg2): Convert from a global struct to a subclass of
12851 rtl_opt_pass along with...
12852 (pass_data_lower_subreg2): ...new pass_data instance and...
12853 (make_pass_lower_subreg2): ...new function.
12854 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
12855 struct to a subclass of ipa_opt_pass_d along with...
12856 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
12857 (make_pass_ipa_lto_gimple_out): ...new function.
12858 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
12859 of ipa_opt_pass_d along with...
12860 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
12861 (make_pass_ipa_lto_finish_out): ...new function.
12862 * mode-switching.c (pass_mode_switching): Convert from a global struct
12863 to a subclass of rtl_opt_pass along with...
12864 (pass_data_mode_switching): ...new pass_data instance and...
12865 (make_pass_mode_switching): ...new function.
12866 * modulo-sched.c (pass_sms): Convert from a global struct to a
12867 subclass of rtl_opt_pass along with...
12868 (pass_data_sms): ...new pass_data instance and...
12869 (make_pass_sms): ...new function.
12870 * omp-low.c (pass_expand_omp): Convert from a global struct to a
12871 subclass of gimple_opt_pass along with...
12872 (pass_data_expand_omp): ...new pass_data instance and...
12873 (make_pass_expand_omp): ...new function.
12874 (pass_lower_omp): Convert from a global struct to a subclass of
12875 gimple_opt_pass along with...
12876 (pass_data_lower_omp): ...new pass_data instance and...
12877 (make_pass_lower_omp): ...new function.
12878 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
12879 of gimple_opt_pass along with...
12880 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
12881 (make_pass_diagnose_omp_blocks): ...new function.
12882 * passes.c (pass_early_local_passes): Convert from a global struct to
12883 a subclass of simple_ipa_opt_pass along with...
12884 (pass_data_early_local_passes): ...new pass_data instance and...
12885 (make_pass_early_local_passes): ...new function.
12886 (pass_all_early_optimizations): Convert from a global struct to a
12887 subclass of gimple_opt_pass along with...
12888 (pass_data_all_early_optimizations): ...new pass_data instance and...
12889 (make_pass_all_early_optimizations): ...new function.
12890 (pass_all_optimizations): Convert from a global struct to a subclass
12891 of gimple_opt_pass along with...
12892 (pass_data_all_optimizations): ...new pass_data instance and...
12893 (make_pass_all_optimizations): ...new function.
12894 (pass_all_optimizations_g): Convert from a global struct to a subclass
12895 of gimple_opt_pass along with...
12896 (pass_data_all_optimizations_g): ...new pass_data instance and...
12897 (make_pass_all_optimizations_g): ...new function.
12898 (pass_rest_of_compilation): Convert from a global struct to a subclass
12899 of rtl_opt_pass along with...
12900 (pass_data_rest_of_compilation): ...new pass_data instance and...
12901 (make_pass_rest_of_compilation): ...new function.
12902 (pass_postreload): Convert from a global struct to a subclass of
12903 rtl_opt_pass along with...
12904 (pass_data_postreload): ...new pass_data instance and...
12905 (make_pass_postreload): ...new function.
12906 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
12907 subclass of rtl_opt_pass along with...
12908 (pass_data_gcse2): ...new pass_data instance and...
12909 (make_pass_gcse2): ...new function.
12910 * postreload.c (pass_postreload_cse): Convert from a global struct to
12911 a subclass of rtl_opt_pass along with...
12912 (pass_data_postreload_cse): ...new pass_data instance and...
12913 (make_pass_postreload_cse): ...new function.
12914 * predict.c (pass_profile): Convert from a global struct to a subclass
12915 of gimple_opt_pass along with...
12916 (pass_data_profile): ...new pass_data instance and...
12917 (make_pass_profile): ...new function.
12918 (pass_strip_predict_hints): Convert from a global struct to a subclass
12919 of gimple_opt_pass along with...
12920 (pass_data_strip_predict_hints): ...new pass_data instance and...
12921 (make_pass_strip_predict_hints): ...new function.
12922 * recog.c (pass_peephole2): Convert from a global struct to a subclass
12923 of rtl_opt_pass along with...
12924 (pass_data_peephole2): ...new pass_data instance and...
12925 (make_pass_peephole2): ...new function.
12926 (pass_split_all_insns): Convert from a global struct to a subclass of
12927 rtl_opt_pass along with...
12928 (pass_data_split_all_insns): ...new pass_data instance and...
12929 (make_pass_split_all_insns): ...new function.
12930 (pass_split_after_reload): Convert from a global struct to a subclass
12931 of rtl_opt_pass along with...
12932 (pass_data_split_after_reload): ...new pass_data instance and...
12933 (make_pass_split_after_reload): ...new function.
12934 (pass_split_before_regstack): Convert from a global struct to a
12935 subclass of rtl_opt_pass along with...
12936 (pass_data_split_before_regstack): ...new pass_data instance and...
12937 (make_pass_split_before_regstack): ...new function.
12938 (pass_split_before_sched2): Convert from a global struct to a subclass
12939 of rtl_opt_pass along with...
12940 (pass_data_split_before_sched2): ...new pass_data instance and...
12941 (make_pass_split_before_sched2): ...new function.
12942 (pass_split_for_shorten_branches): Convert from a global struct to a
12943 subclass of rtl_opt_pass along with...
12944 (pass_data_split_for_shorten_branches): ...new pass_data instance
12946 (make_pass_split_for_shorten_branches): ...new function.
12947 * ree.c (pass_ree): Convert from a global struct to a subclass of
12948 rtl_opt_pass along with...
12949 (pass_data_ree): ...new pass_data instance and...
12950 (make_pass_ree): ...new function.
12951 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
12952 subclass of rtl_opt_pass along with...
12953 (pass_data_stack_regs): ...new pass_data instance and...
12954 (make_pass_stack_regs): ...new function.
12955 (pass_stack_regs_run): Convert from a global struct to a subclass of
12956 rtl_opt_pass along with...
12957 (pass_data_stack_regs_run): ...new pass_data instance and...
12958 (make_pass_stack_regs_run): ...new function.
12959 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
12960 subclass of rtl_opt_pass along with...
12961 (pass_data_cprop_hardreg): ...new pass_data instance and...
12962 (make_pass_cprop_hardreg): ...new function.
12963 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
12964 subclass of rtl_opt_pass along with...
12965 (pass_data_reginfo_init): ...new pass_data instance and...
12966 (make_pass_reginfo_init): ...new function.
12967 * regmove.c (pass_regmove): Convert from a global struct to a subclass
12968 of rtl_opt_pass along with...
12969 (pass_data_regmove): ...new pass_data instance and...
12970 (make_pass_regmove): ...new function.
12971 * regrename.c (pass_regrename): Convert from a global struct to a
12972 subclass of rtl_opt_pass along with...
12973 (pass_data_regrename): ...new pass_data instance and...
12974 (make_pass_regrename): ...new function.
12975 * reorg.c (pass_delay_slots): Convert from a global struct to a
12976 subclass of rtl_opt_pass along with...
12977 (pass_data_delay_slots): ...new pass_data instance and...
12978 (make_pass_delay_slots): ...new function.
12979 (pass_machine_reorg): Convert from a global struct to a subclass of
12980 rtl_opt_pass along with...
12981 (pass_data_machine_reorg): ...new pass_data instance and...
12982 (make_pass_machine_reorg): ...new function.
12983 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
12984 of rtl_opt_pass along with...
12985 (pass_data_sched): ...new pass_data instance and...
12986 (make_pass_sched): ...new function.
12987 (pass_sched2): Convert from a global struct to a subclass of
12988 rtl_opt_pass along with...
12989 (pass_data_sched2): ...new pass_data instance and...
12990 (make_pass_sched2): ...new function.
12991 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
12992 to a subclass of rtl_opt_pass along with...
12993 (pass_data_stack_ptr_mod): ...new pass_data instance and...
12994 (make_pass_stack_ptr_mod): ...new function.
12995 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
12996 to a subclass of rtl_opt_pass along with...
12997 (pass_data_rtl_store_motion): ...new pass_data instance and...
12998 (make_pass_rtl_store_motion): ...new function.
12999 * tracer.c (pass_tracer): Convert from a global struct to a subclass
13000 of gimple_opt_pass along with...
13001 (pass_data_tracer): ...new pass_data instance and...
13002 (make_pass_tracer): ...new function.
13003 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
13004 to a subclass of gimple_opt_pass along with...
13005 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
13006 (make_pass_diagnose_tm_blocks): ...new function.
13007 (pass_lower_tm): Convert from a global struct to a subclass of
13008 gimple_opt_pass along with...
13009 (pass_data_lower_tm): ...new pass_data instance and...
13010 (make_pass_lower_tm): ...new function.
13011 (pass_tm_init): Convert from a global struct to a subclass of
13012 gimple_opt_pass along with...
13013 (pass_data_tm_init): ...new pass_data instance and...
13014 (make_pass_tm_init): ...new function.
13015 (pass_tm_mark): Convert from a global struct to a subclass of
13016 gimple_opt_pass along with...
13017 (pass_data_tm_mark): ...new pass_data instance and...
13018 (make_pass_tm_mark): ...new function.
13019 (pass_tm_edges): Convert from a global struct to a subclass of
13020 gimple_opt_pass along with...
13021 (pass_data_tm_edges): ...new pass_data instance and...
13022 (make_pass_tm_edges): ...new function.
13023 (pass_tm_memopt): Convert from a global struct to a subclass of
13024 gimple_opt_pass along with...
13025 (pass_data_tm_memopt): ...new pass_data instance and...
13026 (make_pass_tm_memopt): ...new function.
13027 (pass_ipa_tm): Convert from a global struct to a subclass of
13028 simple_ipa_opt_pass along with...
13029 (pass_data_ipa_tm): ...new pass_data instance and...
13030 (make_pass_ipa_tm): ...new function.
13031 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
13032 subclass of gimple_opt_pass along with...
13033 (pass_data_call_cdce): ...new pass_data instance and...
13034 (make_pass_call_cdce): ...new function.
13035 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
13036 subclass of gimple_opt_pass along with...
13037 (pass_data_build_cfg): ...new pass_data instance and...
13038 (make_pass_build_cfg): ...new function.
13039 (pass_split_crit_edges): Convert from a global struct to a subclass of
13040 gimple_opt_pass along with...
13041 (pass_data_split_crit_edges): ...new pass_data instance and...
13042 (make_pass_split_crit_edges): ...new function.
13043 (pass_warn_function_return): Convert from a global struct to a
13044 subclass of gimple_opt_pass along with...
13045 (pass_data_warn_function_return): ...new pass_data instance and...
13046 (make_pass_warn_function_return): ...new function.
13047 (pass_warn_function_noreturn): Convert from a global struct to a
13048 subclass of gimple_opt_pass along with...
13049 (pass_data_warn_function_noreturn): ...new pass_data instance and...
13050 (make_pass_warn_function_noreturn): ...new function.
13051 (pass_warn_unused_result): Convert from a global struct to a subclass
13052 of gimple_opt_pass along with...
13053 (pass_data_warn_unused_result): ...new pass_data instance and...
13054 (make_pass_warn_unused_result): ...new function.
13055 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
13056 a subclass of gimple_opt_pass along with...
13057 (pass_data_merge_phi): ...new pass_data instance and...
13058 (make_pass_merge_phi): ...new function.
13059 * tree-complex.c (pass_lower_complex): Convert from a global struct to
13060 a subclass of gimple_opt_pass along with...
13061 (pass_data_lower_complex): ...new pass_data instance and...
13062 (make_pass_lower_complex): ...new function.
13063 (pass_lower_complex_O0): Convert from a global struct to a subclass of
13064 gimple_opt_pass along with...
13065 (pass_data_lower_complex_O0): ...new pass_data instance and...
13066 (make_pass_lower_complex_O0): ...new function.
13067 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
13068 subclass of gimple_opt_pass along with...
13069 (pass_data_lower_eh): ...new pass_data instance and...
13070 (make_pass_lower_eh): ...new function.
13071 (pass_refactor_eh): Convert from a global struct to a subclass of
13072 gimple_opt_pass along with...
13073 (pass_data_refactor_eh): ...new pass_data instance and...
13074 (make_pass_refactor_eh): ...new function.
13075 (pass_lower_resx): Convert from a global struct to a subclass of
13076 gimple_opt_pass along with...
13077 (pass_data_lower_resx): ...new pass_data instance and...
13078 (make_pass_lower_resx): ...new function.
13079 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
13080 of gimple_opt_pass along with...
13081 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
13082 (make_pass_lower_eh_dispatch): ...new function.
13083 (pass_cleanup_eh): Convert from a global struct to a subclass of
13084 gimple_opt_pass along with...
13085 (pass_data_cleanup_eh): ...new pass_data instance and...
13086 (make_pass_cleanup_eh): ...new function.
13087 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
13088 to a subclass of simple_ipa_opt_pass along with...
13089 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
13090 (make_pass_ipa_lower_emutls): ...new function.
13091 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
13092 a subclass of gimple_opt_pass along with...
13093 (pass_data_if_conversion): ...new pass_data instance and...
13094 (make_pass_if_conversion): ...new function.
13095 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
13096 subclass of gimple_opt_pass along with...
13097 (pass_data_build_ssa): ...new pass_data instance and...
13098 (make_pass_build_ssa): ...new function.
13099 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
13100 global struct to a subclass of gimple_opt_pass along with...
13101 (pass_data_loop_distribution): ...new pass_data instance and...
13102 (make_pass_loop_distribution): ...new function.
13103 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
13104 subclass of gimple_opt_pass along with...
13105 (pass_data_mudflap_1): ...new pass_data instance and...
13106 (make_pass_mudflap_1): ...new function.
13107 (pass_mudflap_2): Convert from a global struct to a subclass of
13108 gimple_opt_pass along with...
13109 (pass_data_mudflap_2): ...new pass_data instance and...
13110 (make_pass_mudflap_2): ...new function.
13111 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
13112 subclass of gimple_opt_pass along with...
13113 (pass_data_mudflap_1): ...new pass_data instance and...
13114 (make_pass_mudflap_1): ...new function.
13115 (pass_mudflap_2): Convert from a global struct to a subclass of
13116 gimple_opt_pass along with...
13117 (pass_data_mudflap_2): ...new pass_data instance and...
13118 (make_pass_mudflap_2): ...new function.
13119 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
13120 gimple_opt_pass along with...
13121 (pass_data_nrv): ...new pass_data instance and...
13122 (make_pass_nrv): ...new function.
13123 (pass_return_slot): Convert from a global struct to a subclass of
13124 gimple_opt_pass along with...
13125 (pass_data_return_slot): ...new pass_data instance and...
13126 (make_pass_return_slot): ...new function.
13127 * tree-object-size.c (pass_object_sizes): Convert from a global struct
13128 to a subclass of gimple_opt_pass along with...
13129 (pass_data_object_sizes): ...new pass_data instance and...
13130 (make_pass_object_sizes): ...new function.
13131 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
13132 global struct to a subclass of gimple_opt_pass along with...
13133 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
13135 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13136 (pass_fixup_cfg): Convert from a global struct to a subclass of
13137 gimple_opt_pass along with...
13138 (pass_data_fixup_cfg): ...new pass_data instance and...
13139 (make_pass_fixup_cfg): ...new function.
13140 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
13141 (make_pass_mudflap_1): ...new function.
13142 (pass_mudflap_2): Replace declaration with that of...
13143 (make_pass_mudflap_2): ...new function.
13144 (pass_asan): Replace declaration with that of...
13145 (make_pass_asan): ...new function.
13146 (pass_asan_O0): Replace declaration with that of...
13147 (make_pass_asan_O0): ...new function.
13148 (pass_tsan): Replace declaration with that of...
13149 (make_pass_tsan): ...new function.
13150 (pass_tsan_O0): Replace declaration with that of...
13151 (make_pass_tsan_O0): ...new function.
13152 (pass_lower_cf): Replace declaration with that of...
13153 (make_pass_lower_cf): ...new function.
13154 (pass_refactor_eh): Replace declaration with that of...
13155 (make_pass_refactor_eh): ...new function.
13156 (pass_lower_eh): Replace declaration with that of...
13157 (make_pass_lower_eh): ...new function.
13158 (pass_lower_eh_dispatch): Replace declaration with that of...
13159 (make_pass_lower_eh_dispatch): ...new function.
13160 (pass_lower_resx): Replace declaration with that of...
13161 (make_pass_lower_resx): ...new function.
13162 (pass_build_cfg): Replace declaration with that of...
13163 (make_pass_build_cfg): ...new function.
13164 (pass_early_tree_profile): Replace declaration with that of...
13165 (make_pass_early_tree_profile): ...new function.
13166 (pass_cleanup_eh): Replace declaration with that of...
13167 (make_pass_cleanup_eh): ...new function.
13168 (pass_sra): Replace declaration with that of...
13169 (make_pass_sra): ...new function.
13170 (pass_sra_early): Replace declaration with that of...
13171 (make_pass_sra_early): ...new function.
13172 (pass_early_ipa_sra): Replace declaration with that of...
13173 (make_pass_early_ipa_sra): ...new function.
13174 (pass_tail_recursion): Replace declaration with that of...
13175 (make_pass_tail_recursion): ...new function.
13176 (pass_tail_calls): Replace declaration with that of...
13177 (make_pass_tail_calls): ...new function.
13178 (pass_tree_loop): Replace declaration with that of...
13179 (make_pass_tree_loop): ...new function.
13180 (pass_tree_loop_init): Replace declaration with that of...
13181 (make_pass_tree_loop_init): ...new function.
13182 (pass_lim): Replace declaration with that of...
13183 (make_pass_lim): ...new function.
13184 (pass_tree_unswitch): Replace declaration with that of...
13185 (make_pass_tree_unswitch): ...new function.
13186 (pass_predcom): Replace declaration with that of...
13187 (make_pass_predcom): ...new function.
13188 (pass_iv_canon): Replace declaration with that of...
13189 (make_pass_iv_canon): ...new function.
13190 (pass_scev_cprop): Replace declaration with that of...
13191 (make_pass_scev_cprop): ...new function.
13192 (pass_empty_loop): Replace declaration with that of...
13193 (make_pass_empty_loop): ...new function.
13194 (pass_record_bounds): Replace declaration with that of...
13195 (make_pass_record_bounds): ...new function.
13196 (pass_graphite): Replace declaration with that of...
13197 (make_pass_graphite): ...new function.
13198 (pass_graphite_transforms): Replace declaration with that of...
13199 (make_pass_graphite_transforms): ...new function.
13200 (pass_if_conversion): Replace declaration with that of...
13201 (make_pass_if_conversion): ...new function.
13202 (pass_loop_distribution): Replace declaration with that of...
13203 (make_pass_loop_distribution): ...new function.
13204 (pass_vectorize): Replace declaration with that of...
13205 (make_pass_vectorize): ...new function.
13206 (pass_slp_vectorize): Replace declaration with that of...
13207 (make_pass_slp_vectorize): ...new function.
13208 (pass_complete_unroll): Replace declaration with that of...
13209 (make_pass_complete_unroll): ...new function.
13210 (pass_complete_unrolli): Replace declaration with that of...
13211 (make_pass_complete_unrolli): ...new function.
13212 (pass_parallelize_loops): Replace declaration with that of...
13213 (make_pass_parallelize_loops): ...new function.
13214 (pass_loop_prefetch): Replace declaration with that of...
13215 (make_pass_loop_prefetch): ...new function.
13216 (pass_iv_optimize): Replace declaration with that of...
13217 (make_pass_iv_optimize): ...new function.
13218 (pass_tree_loop_done): Replace declaration with that of...
13219 (make_pass_tree_loop_done): ...new function.
13220 (pass_ch): Replace declaration with that of...
13221 (make_pass_ch): ...new function.
13222 (pass_ccp): Replace declaration with that of...
13223 (make_pass_ccp): ...new function.
13224 (pass_phi_only_cprop): Replace declaration with that of...
13225 (make_pass_phi_only_cprop): ...new function.
13226 (pass_build_ssa): Replace declaration with that of...
13227 (make_pass_build_ssa): ...new function.
13228 (pass_build_alias): Replace declaration with that of...
13229 (make_pass_build_alias): ...new function.
13230 (pass_build_ealias): Replace declaration with that of...
13231 (make_pass_build_ealias): ...new function.
13232 (pass_dominator): Replace declaration with that of...
13233 (make_pass_dominator): ...new function.
13234 (pass_dce): Replace declaration with that of...
13235 (make_pass_dce): ...new function.
13236 (pass_dce_loop): Replace declaration with that of...
13237 (make_pass_dce_loop): ...new function.
13238 (pass_cd_dce): Replace declaration with that of...
13239 (make_pass_cd_dce): ...new function.
13240 (pass_call_cdce): Replace declaration with that of...
13241 (make_pass_call_cdce): ...new function.
13242 (pass_merge_phi): Replace declaration with that of...
13243 (make_pass_merge_phi): ...new function.
13244 (pass_split_crit_edges): Replace declaration with that of...
13245 (make_pass_split_crit_edges): ...new function.
13246 (pass_pre): Replace declaration with that of...
13247 (make_pass_pre): ...new function.
13248 (pass_profile): Replace declaration with that of...
13249 (make_pass_profile): ...new function.
13250 (pass_strip_predict_hints): Replace declaration with that of...
13251 (make_pass_strip_predict_hints): ...new function.
13252 (pass_lower_complex_O0): Replace declaration with that of...
13253 (make_pass_lower_complex_O0): ...new function.
13254 (pass_lower_complex): Replace declaration with that of...
13255 (make_pass_lower_complex): ...new function.
13256 (pass_lower_vector): Replace declaration with that of...
13257 (make_pass_lower_vector): ...new function.
13258 (pass_lower_vector_ssa): Replace declaration with that of...
13259 (make_pass_lower_vector_ssa): ...new function.
13260 (pass_lower_omp): Replace declaration with that of...
13261 (make_pass_lower_omp): ...new function.
13262 (pass_diagnose_omp_blocks): Replace declaration with that of...
13263 (make_pass_diagnose_omp_blocks): ...new function.
13264 (pass_expand_omp): Replace declaration with that of...
13265 (make_pass_expand_omp): ...new function.
13266 (pass_expand_omp_ssa): Replace declaration with that of...
13267 (make_pass_expand_omp_ssa): ...new function.
13268 (pass_object_sizes): Replace declaration with that of...
13269 (make_pass_object_sizes): ...new function.
13270 (pass_strlen): Replace declaration with that of...
13271 (make_pass_strlen): ...new function.
13272 (pass_fold_builtins): Replace declaration with that of...
13273 (make_pass_fold_builtins): ...new function.
13274 (pass_stdarg): Replace declaration with that of...
13275 (make_pass_stdarg): ...new function.
13276 (pass_early_warn_uninitialized): Replace declaration with that of...
13277 (make_pass_early_warn_uninitialized): ...new function.
13278 (pass_late_warn_uninitialized): Replace declaration with that of...
13279 (make_pass_late_warn_uninitialized): ...new function.
13280 (pass_cse_reciprocals): Replace declaration with that of...
13281 (make_pass_cse_reciprocals): ...new function.
13282 (pass_cse_sincos): Replace declaration with that of...
13283 (make_pass_cse_sincos): ...new function.
13284 (pass_optimize_bswap): Replace declaration with that of...
13285 (make_pass_optimize_bswap): ...new function.
13286 (pass_optimize_widening_mul): Replace declaration with that of...
13287 (make_pass_optimize_widening_mul): ...new function.
13288 (pass_warn_function_return): Replace declaration with that of...
13289 (make_pass_warn_function_return): ...new function.
13290 (pass_warn_function_noreturn): Replace declaration with that of...
13291 (make_pass_warn_function_noreturn): ...new function.
13292 (pass_cselim): Replace declaration with that of...
13293 (make_pass_cselim): ...new function.
13294 (pass_phiopt): Replace declaration with that of...
13295 (make_pass_phiopt): ...new function.
13296 (pass_forwprop): Replace declaration with that of...
13297 (make_pass_forwprop): ...new function.
13298 (pass_phiprop): Replace declaration with that of...
13299 (make_pass_phiprop): ...new function.
13300 (pass_tree_ifcombine): Replace declaration with that of...
13301 (make_pass_tree_ifcombine): ...new function.
13302 (pass_dse): Replace declaration with that of...
13303 (make_pass_dse): ...new function.
13304 (pass_nrv): Replace declaration with that of...
13305 (make_pass_nrv): ...new function.
13306 (pass_rename_ssa_copies): Replace declaration with that of...
13307 (make_pass_rename_ssa_copies): ...new function.
13308 (pass_sink_code): Replace declaration with that of...
13309 (make_pass_sink_code): ...new function.
13310 (pass_fre): Replace declaration with that of...
13311 (make_pass_fre): ...new function.
13312 (pass_check_data_deps): Replace declaration with that of...
13313 (make_pass_check_data_deps): ...new function.
13314 (pass_copy_prop): Replace declaration with that of...
13315 (make_pass_copy_prop): ...new function.
13316 (pass_vrp): Replace declaration with that of...
13317 (make_pass_vrp): ...new function.
13318 (pass_uncprop): Replace declaration with that of...
13319 (make_pass_uncprop): ...new function.
13320 (pass_return_slot): Replace declaration with that of...
13321 (make_pass_return_slot): ...new function.
13322 (pass_reassoc): Replace declaration with that of...
13323 (make_pass_reassoc): ...new function.
13324 (pass_rebuild_cgraph_edges): Replace declaration with that of...
13325 (make_pass_rebuild_cgraph_edges): ...new function.
13326 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
13327 (make_pass_remove_cgraph_callee_edges): ...new function.
13328 (pass_build_cgraph_edges): Replace declaration with that of...
13329 (make_pass_build_cgraph_edges): ...new function.
13330 (pass_local_pure_const): Replace declaration with that of...
13331 (make_pass_local_pure_const): ...new function.
13332 (pass_tracer): Replace declaration with that of...
13333 (make_pass_tracer): ...new function.
13334 (pass_warn_unused_result): Replace declaration with that of...
13335 (make_pass_warn_unused_result): ...new function.
13336 (pass_diagnose_tm_blocks): Replace declaration with that of...
13337 (make_pass_diagnose_tm_blocks): ...new function.
13338 (pass_lower_tm): Replace declaration with that of...
13339 (make_pass_lower_tm): ...new function.
13340 (pass_tm_init): Replace declaration with that of...
13341 (make_pass_tm_init): ...new function.
13342 (pass_tm_mark): Replace declaration with that of...
13343 (make_pass_tm_mark): ...new function.
13344 (pass_tm_memopt): Replace declaration with that of...
13345 (make_pass_tm_memopt): ...new function.
13346 (pass_tm_edges): Replace declaration with that of...
13347 (make_pass_tm_edges): ...new function.
13348 (pass_split_functions): Replace declaration with that of...
13349 (make_pass_split_functions): ...new function.
13350 (pass_feedback_split_functions): Replace declaration with that of...
13351 (make_pass_feedback_split_functions): ...new function.
13352 (pass_strength_reduction): Replace declaration with that of...
13353 (make_pass_strength_reduction): ...new function.
13354 (pass_ipa_lower_emutls): Replace declaration with that of...
13355 (make_pass_ipa_lower_emutls): ...new function.
13356 (pass_ipa_function_and_variable_visibility): Replace declaration with
13358 (make_pass_ipa_function_and_variable_visibility): ...new function.
13359 (pass_ipa_tree_profile): Replace declaration with that of...
13360 (make_pass_ipa_tree_profile): ...new function.
13361 (pass_early_local_passes): Replace declaration with that of...
13362 (make_pass_early_local_passes): ...new function.
13363 (pass_ipa_whole_program_visibility): Replace declaration with that
13365 (make_pass_ipa_whole_program_visibility): ...new function.
13366 (pass_ipa_lto_gimple_out): Replace declaration with that of...
13367 (make_pass_ipa_lto_gimple_out): ...new function.
13368 (pass_ipa_increase_alignment): Replace declaration with that of...
13369 (make_pass_ipa_increase_alignment): ...new function.
13370 (pass_ipa_inline): Replace declaration with that of...
13371 (make_pass_ipa_inline): ...new function.
13372 (pass_ipa_free_lang_data): Replace declaration with that of...
13373 (make_pass_ipa_free_lang_data): ...new function.
13374 (pass_ipa_free_inline_summary): Replace declaration with that of...
13375 (make_pass_ipa_free_inline_summary): ...new function.
13376 (pass_ipa_cp): Replace declaration with that of...
13377 (make_pass_ipa_cp): ...new function.
13378 (pass_ipa_reference): Replace declaration with that of...
13379 (make_pass_ipa_reference): ...new function.
13380 (pass_ipa_pure_const): Replace declaration with that of...
13381 (make_pass_ipa_pure_const): ...new function.
13382 (pass_ipa_pta): Replace declaration with that of...
13383 (make_pass_ipa_pta): ...new function.
13384 (pass_ipa_lto_finish_out): Replace declaration with that of...
13385 (make_pass_ipa_lto_finish_out): ...new function.
13386 (pass_ipa_tm): Replace declaration with that of...
13387 (make_pass_ipa_tm): ...new function.
13388 (pass_ipa_profile): Replace declaration with that of...
13389 (make_pass_ipa_profile): ...new function.
13390 (pass_ipa_cdtor_merge): Replace declaration with that of...
13391 (make_pass_ipa_cdtor_merge): ...new function.
13392 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
13394 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13395 (pass_init_datastructures): Replace declaration with that of...
13396 (make_pass_init_datastructures): ...new function.
13397 (pass_fixup_cfg): Replace declaration with that of...
13398 (make_pass_fixup_cfg): ...new function.
13399 (pass_expand): Replace declaration with that of...
13400 (make_pass_expand): ...new function.
13401 (pass_instantiate_virtual_regs): Replace declaration with that of...
13402 (make_pass_instantiate_virtual_regs): ...new function.
13403 (pass_rtl_fwprop): Replace declaration with that of...
13404 (make_pass_rtl_fwprop): ...new function.
13405 (pass_rtl_fwprop_addr): Replace declaration with that of...
13406 (make_pass_rtl_fwprop_addr): ...new function.
13407 (pass_jump): Replace declaration with that of...
13408 (make_pass_jump): ...new function.
13409 (pass_jump2): Replace declaration with that of...
13410 (make_pass_jump2): ...new function.
13411 (pass_lower_subreg): Replace declaration with that of...
13412 (make_pass_lower_subreg): ...new function.
13413 (pass_cse): Replace declaration with that of...
13414 (make_pass_cse): ...new function.
13415 (pass_fast_rtl_dce): Replace declaration with that of...
13416 (make_pass_fast_rtl_dce): ...new function.
13417 (pass_ud_rtl_dce): Replace declaration with that of...
13418 (make_pass_ud_rtl_dce): ...new function.
13419 (pass_rtl_dce): Replace declaration with that of...
13420 (make_pass_rtl_dce): ...new function.
13421 (pass_rtl_dse1): Replace declaration with that of...
13422 (make_pass_rtl_dse1): ...new function.
13423 (pass_rtl_dse2): Replace declaration with that of...
13424 (make_pass_rtl_dse2): ...new function.
13425 (pass_rtl_dse3): Replace declaration with that of...
13426 (make_pass_rtl_dse3): ...new function.
13427 (pass_rtl_cprop): Replace declaration with that of...
13428 (make_pass_rtl_cprop): ...new function.
13429 (pass_rtl_pre): Replace declaration with that of...
13430 (make_pass_rtl_pre): ...new function.
13431 (pass_rtl_hoist): Replace declaration with that of...
13432 (make_pass_rtl_hoist): ...new function.
13433 (pass_rtl_store_motion): Replace declaration with that of...
13434 (make_pass_rtl_store_motion): ...new function.
13435 (pass_cse_after_global_opts): Replace declaration with that of...
13436 (make_pass_cse_after_global_opts): ...new function.
13437 (pass_rtl_ifcvt): Replace declaration with that of...
13438 (make_pass_rtl_ifcvt): ...new function.
13439 (pass_into_cfg_layout_mode): Replace declaration with that of...
13440 (make_pass_into_cfg_layout_mode): ...new function.
13441 (pass_outof_cfg_layout_mode): Replace declaration with that of...
13442 (make_pass_outof_cfg_layout_mode): ...new function.
13443 (pass_loop2): Replace declaration with that of...
13444 (make_pass_loop2): ...new function.
13445 (pass_rtl_loop_init): Replace declaration with that of...
13446 (make_pass_rtl_loop_init): ...new function.
13447 (pass_rtl_move_loop_invariants): Replace declaration with that of...
13448 (make_pass_rtl_move_loop_invariants): ...new function.
13449 (pass_rtl_unswitch): Replace declaration with that of...
13450 (make_pass_rtl_unswitch): ...new function.
13451 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
13452 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13453 (pass_rtl_doloop): Replace declaration with that of...
13454 (make_pass_rtl_doloop): ...new function.
13455 (pass_rtl_loop_done): Replace declaration with that of...
13456 (make_pass_rtl_loop_done): ...new function.
13457 (pass_web): Replace declaration with that of...
13458 (make_pass_web): ...new function.
13459 (pass_cse2): Replace declaration with that of...
13460 (make_pass_cse2): ...new function.
13461 (pass_df_initialize_opt): Replace declaration with that of...
13462 (make_pass_df_initialize_opt): ...new function.
13463 (pass_df_initialize_no_opt): Replace declaration with that of...
13464 (make_pass_df_initialize_no_opt): ...new function.
13465 (pass_reginfo_init): Replace declaration with that of...
13466 (make_pass_reginfo_init): ...new function.
13467 (pass_inc_dec): Replace declaration with that of...
13468 (make_pass_inc_dec): ...new function.
13469 (pass_stack_ptr_mod): Replace declaration with that of...
13470 (make_pass_stack_ptr_mod): ...new function.
13471 (pass_initialize_regs): Replace declaration with that of...
13472 (make_pass_initialize_regs): ...new function.
13473 (pass_combine): Replace declaration with that of...
13474 (make_pass_combine): ...new function.
13475 (pass_if_after_combine): Replace declaration with that of...
13476 (make_pass_if_after_combine): ...new function.
13477 (pass_ree): Replace declaration with that of...
13478 (make_pass_ree): ...new function.
13479 (pass_partition_blocks): Replace declaration with that of...
13480 (make_pass_partition_blocks): ...new function.
13481 (pass_match_asm_constraints): Replace declaration with that of...
13482 (make_pass_match_asm_constraints): ...new function.
13483 (pass_regmove): Replace declaration with that of...
13484 (make_pass_regmove): ...new function.
13485 (pass_split_all_insns): Replace declaration with that of...
13486 (make_pass_split_all_insns): ...new function.
13487 (pass_fast_rtl_byte_dce): Replace declaration with that of...
13488 (make_pass_fast_rtl_byte_dce): ...new function.
13489 (pass_lower_subreg2): Replace declaration with that of...
13490 (make_pass_lower_subreg2): ...new function.
13491 (pass_mode_switching): Replace declaration with that of...
13492 (make_pass_mode_switching): ...new function.
13493 (pass_sms): Replace declaration with that of...
13494 (make_pass_sms): ...new function.
13495 (pass_sched): Replace declaration with that of...
13496 (make_pass_sched): ...new function.
13497 (pass_ira): Replace declaration with that of...
13498 (make_pass_ira): ...new function.
13499 (pass_reload): Replace declaration with that of...
13500 (make_pass_reload): ...new function.
13501 (pass_clean_state): Replace declaration with that of...
13502 (make_pass_clean_state): ...new function.
13503 (pass_branch_prob): Replace declaration with that of...
13504 (make_pass_branch_prob): ...new function.
13505 (pass_value_profile_transformations): Replace declaration with that
13507 (make_pass_value_profile_transformations): ...new function.
13508 (pass_postreload_cse): Replace declaration with that of...
13509 (make_pass_postreload_cse): ...new function.
13510 (pass_gcse2): Replace declaration with that of...
13511 (make_pass_gcse2): ...new function.
13512 (pass_split_after_reload): Replace declaration with that of...
13513 (make_pass_split_after_reload): ...new function.
13514 (pass_branch_target_load_optimize1): Replace declaration with that
13516 (make_pass_branch_target_load_optimize1): ...new function.
13517 (pass_thread_prologue_and_epilogue): Replace declaration with that
13519 (make_pass_thread_prologue_and_epilogue): ...new function.
13520 (pass_stack_adjustments): Replace declaration with that of...
13521 (make_pass_stack_adjustments): ...new function.
13522 (pass_peephole2): Replace declaration with that of...
13523 (make_pass_peephole2): ...new function.
13524 (pass_if_after_reload): Replace declaration with that of...
13525 (make_pass_if_after_reload): ...new function.
13526 (pass_regrename): Replace declaration with that of...
13527 (make_pass_regrename): ...new function.
13528 (pass_cprop_hardreg): Replace declaration with that of...
13529 (make_pass_cprop_hardreg): ...new function.
13530 (pass_reorder_blocks): Replace declaration with that of...
13531 (make_pass_reorder_blocks): ...new function.
13532 (pass_branch_target_load_optimize2): Replace declaration with that
13534 (make_pass_branch_target_load_optimize2): ...new function.
13535 (pass_leaf_regs): Replace declaration with that of...
13536 (make_pass_leaf_regs): ...new function.
13537 (pass_split_before_sched2): Replace declaration with that of...
13538 (make_pass_split_before_sched2): ...new function.
13539 (pass_compare_elim_after_reload): Replace declaration with that of...
13540 (make_pass_compare_elim_after_reload): ...new function.
13541 (pass_sched2): Replace declaration with that of...
13542 (make_pass_sched2): ...new function.
13543 (pass_stack_regs): Replace declaration with that of...
13544 (make_pass_stack_regs): ...new function.
13545 (pass_stack_regs_run): Replace declaration with that of...
13546 (make_pass_stack_regs_run): ...new function.
13547 (pass_df_finish): Replace declaration with that of...
13548 (make_pass_df_finish): ...new function.
13549 (pass_compute_alignments): Replace declaration with that of...
13550 (make_pass_compute_alignments): ...new function.
13551 (pass_duplicate_computed_gotos): Replace declaration with that of...
13552 (make_pass_duplicate_computed_gotos): ...new function.
13553 (pass_variable_tracking): Replace declaration with that of...
13554 (make_pass_variable_tracking): ...new function.
13555 (pass_free_cfg): Replace declaration with that of...
13556 (make_pass_free_cfg): ...new function.
13557 (pass_machine_reorg): Replace declaration with that of...
13558 (make_pass_machine_reorg): ...new function.
13559 (pass_cleanup_barriers): Replace declaration with that of...
13560 (make_pass_cleanup_barriers): ...new function.
13561 (pass_delay_slots): Replace declaration with that of...
13562 (make_pass_delay_slots): ...new function.
13563 (pass_split_for_shorten_branches): Replace declaration with that of...
13564 (make_pass_split_for_shorten_branches): ...new function.
13565 (pass_split_before_regstack): Replace declaration with that of...
13566 (make_pass_split_before_regstack): ...new function.
13567 (pass_convert_to_eh_region_ranges): Replace declaration with that
13569 (make_pass_convert_to_eh_region_ranges): ...new function.
13570 (pass_shorten_branches): Replace declaration with that of...
13571 (make_pass_shorten_branches): ...new function.
13572 (pass_set_nothrow_function_flags): Replace declaration with that of...
13573 (make_pass_set_nothrow_function_flags): ...new function.
13574 (pass_dwarf2_frame): Replace declaration with that of...
13575 (make_pass_dwarf2_frame): ...new function.
13576 (pass_final): Replace declaration with that of...
13577 (make_pass_final): ...new function.
13578 (pass_rtl_seqabstr): Replace declaration with that of...
13579 (make_pass_rtl_seqabstr): ...new function.
13580 (pass_release_ssa_names): Replace declaration with that of...
13581 (make_pass_release_ssa_names): ...new function.
13582 (pass_early_inline): Replace declaration with that of...
13583 (make_pass_early_inline): ...new function.
13584 (pass_inline_parameters): Replace declaration with that of...
13585 (make_pass_inline_parameters): ...new function.
13586 (pass_update_address_taken): Replace declaration with that of...
13587 (make_pass_update_address_taken): ...new function.
13588 (pass_convert_switch): Replace declaration with that of...
13589 (make_pass_convert_switch): ...new function.
13590 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
13591 to a subclass of simple_ipa_opt_pass along with...
13592 (pass_data_ipa_tree_profile): ...new pass_data instance and...
13593 (make_pass_ipa_tree_profile): ...new function.
13594 * tree-sra.c (pass_sra_early): Convert from a global struct to a
13595 subclass of gimple_opt_pass along with...
13596 (pass_data_sra_early): ...new pass_data instance and...
13597 (make_pass_sra_early): ...new function.
13598 (pass_sra): Convert from a global struct to a subclass of
13599 gimple_opt_pass along with...
13600 (pass_data_sra): ...new pass_data instance and...
13601 (make_pass_sra): ...new function.
13602 (pass_early_ipa_sra): Convert from a global struct to a subclass of
13603 gimple_opt_pass along with...
13604 (pass_data_early_ipa_sra): ...new pass_data instance and...
13605 (make_pass_early_ipa_sra): ...new function.
13606 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
13607 subclass of gimple_opt_pass along with...
13608 (pass_data_ccp): ...new pass_data instance and...
13609 (make_pass_ccp): ...new function.
13610 (pass_fold_builtins): Convert from a global struct to a subclass of
13611 gimple_opt_pass along with...
13612 (pass_data_fold_builtins): ...new pass_data instance and...
13613 (make_pass_fold_builtins): ...new function.
13614 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
13615 subclass of gimple_opt_pass along with...
13616 (pass_data_copy_prop): ...new pass_data instance and...
13617 (make_pass_copy_prop): ...new function.
13618 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
13619 global struct to a subclass of gimple_opt_pass along with...
13620 (pass_data_rename_ssa_copies): ...new pass_data instance and...
13621 (make_pass_rename_ssa_copies): ...new function.
13622 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
13623 subclass of gimple_opt_pass along with...
13624 (pass_data_dce): ...new pass_data instance and...
13625 (make_pass_dce): ...new function.
13626 (pass_dce_loop): Convert from a global struct to a subclass of
13627 gimple_opt_pass along with...
13628 (pass_data_dce_loop): ...new pass_data instance and...
13629 (make_pass_dce_loop): ...new function.
13630 (pass_cd_dce): Convert from a global struct to a subclass of
13631 gimple_opt_pass along with...
13632 (pass_data_cd_dce): ...new pass_data instance and...
13633 (make_pass_cd_dce): ...new function.
13634 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
13635 subclass of gimple_opt_pass along with...
13636 (pass_data_dominator): ...new pass_data instance and...
13637 (make_pass_dominator): ...new function.
13638 (pass_phi_only_cprop): Convert from a global struct to a subclass of
13639 gimple_opt_pass along with...
13640 (pass_data_phi_only_cprop): ...new pass_data instance and...
13641 (make_pass_phi_only_cprop): ...new function.
13642 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
13643 subclass of gimple_opt_pass along with...
13644 (pass_data_dse): ...new pass_data instance and...
13645 (make_pass_dse): ...new function.
13646 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
13647 a subclass of gimple_opt_pass along with...
13648 (pass_data_forwprop): ...new pass_data instance and...
13649 (make_pass_forwprop): ...new function.
13650 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
13651 struct to a subclass of gimple_opt_pass along with...
13652 (pass_data_tree_ifcombine): ...new pass_data instance and...
13653 (make_pass_tree_ifcombine): ...new function.
13654 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
13655 subclass of gimple_opt_pass along with...
13656 (pass_data_ch): ...new pass_data instance and...
13657 (make_pass_ch): ...new function.
13658 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
13659 subclass of gimple_opt_pass along with...
13660 (pass_data_tree_loop): ...new pass_data instance and...
13661 (make_pass_tree_loop): ...new function.
13662 (pass_tree_loop_init): Convert from a global struct to a subclass of
13663 gimple_opt_pass along with...
13664 (pass_data_tree_loop_init): ...new pass_data instance and...
13665 (make_pass_tree_loop_init): ...new function.
13666 (pass_lim): Convert from a global struct to a subclass of
13667 gimple_opt_pass along with...
13668 (pass_data_lim): ...new pass_data instance and...
13669 (make_pass_lim): ...new function.
13670 (pass_tree_unswitch): Convert from a global struct to a subclass of
13671 gimple_opt_pass along with...
13672 (pass_data_tree_unswitch): ...new pass_data instance and...
13673 (make_pass_tree_unswitch): ...new function.
13674 (pass_predcom): Convert from a global struct to a subclass of
13675 gimple_opt_pass along with...
13676 (pass_data_predcom): ...new pass_data instance and...
13677 (make_pass_predcom): ...new function.
13678 (pass_vectorize): Convert from a global struct to a subclass of
13679 gimple_opt_pass along with...
13680 (pass_data_vectorize): ...new pass_data instance and...
13681 (make_pass_vectorize): ...new function.
13682 (pass_graphite): Convert from a global struct to a subclass of
13683 gimple_opt_pass along with...
13684 (pass_data_graphite): ...new pass_data instance and...
13685 (make_pass_graphite): ...new function.
13686 (pass_graphite_transforms): Convert from a global struct to a subclass
13687 of gimple_opt_pass along with...
13688 (pass_data_graphite_transforms): ...new pass_data instance and...
13689 (make_pass_graphite_transforms): ...new function.
13690 (pass_check_data_deps): Convert from a global struct to a subclass of
13691 gimple_opt_pass along with...
13692 (pass_data_check_data_deps): ...new pass_data instance and...
13693 (make_pass_check_data_deps): ...new function.
13694 (pass_iv_canon): Convert from a global struct to a subclass of
13695 gimple_opt_pass along with...
13696 (pass_data_iv_canon): ...new pass_data instance and...
13697 (make_pass_iv_canon): ...new function.
13698 (pass_scev_cprop): Convert from a global struct to a subclass of
13699 gimple_opt_pass along with...
13700 (pass_data_scev_cprop): ...new pass_data instance and...
13701 (make_pass_scev_cprop): ...new function.
13702 (pass_record_bounds): Convert from a global struct to a subclass of
13703 gimple_opt_pass along with...
13704 (pass_data_record_bounds): ...new pass_data instance and...
13705 (make_pass_record_bounds): ...new function.
13706 (pass_complete_unroll): Convert from a global struct to a subclass of
13707 gimple_opt_pass along with...
13708 (pass_data_complete_unroll): ...new pass_data instance and...
13709 (make_pass_complete_unroll): ...new function.
13710 (pass_complete_unrolli): Convert from a global struct to a subclass of
13711 gimple_opt_pass along with...
13712 (pass_data_complete_unrolli): ...new pass_data instance and...
13713 (make_pass_complete_unrolli): ...new function.
13714 (pass_parallelize_loops): Convert from a global struct to a subclass
13715 of gimple_opt_pass along with...
13716 (pass_data_parallelize_loops): ...new pass_data instance and...
13717 (make_pass_parallelize_loops): ...new function.
13718 (pass_loop_prefetch): Convert from a global struct to a subclass of
13719 gimple_opt_pass along with...
13720 (pass_data_loop_prefetch): ...new pass_data instance and...
13721 (make_pass_loop_prefetch): ...new function.
13722 (pass_iv_optimize): Convert from a global struct to a subclass of
13723 gimple_opt_pass along with...
13724 (pass_data_iv_optimize): ...new pass_data instance and...
13725 (make_pass_iv_optimize): ...new function.
13726 (pass_tree_loop_done): Convert from a global struct to a subclass of
13727 gimple_opt_pass along with...
13728 (pass_data_tree_loop_done): ...new pass_data instance and...
13729 (make_pass_tree_loop_done): ...new function.
13730 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
13731 struct to a subclass of gimple_opt_pass along with...
13732 (pass_data_cse_reciprocals): ...new pass_data instance and...
13733 (make_pass_cse_reciprocals): ...new function.
13734 (pass_cse_sincos): Convert from a global struct to a subclass of
13735 gimple_opt_pass along with...
13736 (pass_data_cse_sincos): ...new pass_data instance and...
13737 (make_pass_cse_sincos): ...new function.
13738 (pass_optimize_bswap): Convert from a global struct to a subclass of
13739 gimple_opt_pass along with...
13740 (pass_data_optimize_bswap): ...new pass_data instance and...
13741 (make_pass_optimize_bswap): ...new function.
13742 (pass_optimize_widening_mul): Convert from a global struct to a
13743 subclass of gimple_opt_pass along with...
13744 (pass_data_optimize_widening_mul): ...new pass_data instance and...
13745 (make_pass_optimize_widening_mul): ...new function.
13746 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
13747 subclass of gimple_opt_pass along with...
13748 (pass_data_phiopt): ...new pass_data instance and...
13749 (make_pass_phiopt): ...new function.
13750 (pass_cselim): Convert from a global struct to a subclass of
13751 gimple_opt_pass along with...
13752 (pass_data_cselim): ...new pass_data instance and...
13753 (make_pass_cselim): ...new function.
13754 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
13755 subclass of gimple_opt_pass along with...
13756 (pass_data_phiprop): ...new pass_data instance and...
13757 (make_pass_phiprop): ...new function.
13758 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
13759 subclass of gimple_opt_pass along with...
13760 (pass_data_pre): ...new pass_data instance and...
13761 (make_pass_pre): ...new function.
13762 (pass_fre): Convert from a global struct to a subclass of
13763 gimple_opt_pass along with...
13764 (pass_data_fre): ...new pass_data instance and...
13765 (make_pass_fre): ...new function.
13766 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
13767 subclass of gimple_opt_pass along with...
13768 (pass_data_reassoc): ...new pass_data instance and...
13769 (make_pass_reassoc): ...new function.
13770 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
13771 subclass of gimple_opt_pass along with...
13772 (pass_data_sink_code): ...new pass_data instance and...
13773 (make_pass_sink_code): ...new function.
13774 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
13775 subclass of gimple_opt_pass along with...
13776 (pass_data_strlen): ...new pass_data instance and...
13777 (make_pass_strlen): ...new function.
13778 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
13779 struct to a subclass of gimple_opt_pass along with...
13780 (pass_data_build_alias): ...new pass_data instance and...
13781 (make_pass_build_alias): ...new function.
13782 (pass_build_ealias): Convert from a global struct to a subclass of
13783 gimple_opt_pass along with...
13784 (pass_data_build_ealias): ...new pass_data instance and...
13785 (make_pass_build_ealias): ...new function.
13786 (pass_ipa_pta): Convert from a global struct to a subclass of
13787 simple_ipa_opt_pass along with...
13788 (pass_data_ipa_pta): ...new pass_data instance and...
13789 (make_pass_ipa_pta): ...new function.
13790 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
13791 subclass of gimple_opt_pass along with...
13792 (pass_data_uncprop): ...new pass_data instance and...
13793 (make_pass_uncprop): ...new function.
13794 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
13795 global struct to a subclass of gimple_opt_pass along with...
13796 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
13797 (make_pass_late_warn_uninitialized): ...new function.
13798 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
13799 to a subclass of gimple_opt_pass along with...
13800 (pass_data_init_datastructures): ...new pass_data instance and...
13801 (make_pass_init_datastructures): ...new function.
13802 (pass_early_warn_uninitialized): Convert from a global struct to a
13803 subclass of gimple_opt_pass along with...
13804 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
13805 (make_pass_early_warn_uninitialized): ...new function.
13806 (pass_update_address_taken): Convert from a global struct to a
13807 subclass of gimple_opt_pass along with...
13808 (pass_data_update_address_taken): ...new pass_data instance and...
13809 (make_pass_update_address_taken): ...new function.
13810 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
13811 struct to a subclass of gimple_opt_pass along with...
13812 (pass_data_release_ssa_names): ...new pass_data instance and...
13813 (make_pass_release_ssa_names): ...new function.
13814 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
13815 subclass of gimple_opt_pass along with...
13816 (pass_data_stdarg): ...new pass_data instance and...
13817 (make_pass_stdarg): ...new function.
13818 * tree-switch-conversion.c (pass_convert_switch): Convert from a
13819 global struct to a subclass of gimple_opt_pass along with...
13820 (pass_data_convert_switch): ...new pass_data instance and...
13821 (make_pass_convert_switch): ...new function.
13822 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
13823 to a subclass of gimple_opt_pass along with...
13824 (pass_data_tail_recursion): ...new pass_data instance and...
13825 (make_pass_tail_recursion): ...new function.
13826 (pass_tail_calls): Convert from a global struct to a subclass of
13827 gimple_opt_pass along with...
13828 (pass_data_tail_calls): ...new pass_data instance and...
13829 (make_pass_tail_calls): ...new function.
13830 * tree-vect-generic.c (pass_lower_vector): Convert from a global
13831 struct to a subclass of gimple_opt_pass along with...
13832 (pass_data_lower_vector): ...new pass_data instance and...
13833 (make_pass_lower_vector): ...new function.
13834 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
13835 gimple_opt_pass along with...
13836 (pass_data_lower_vector_ssa): ...new pass_data instance and...
13837 (make_pass_lower_vector_ssa): ...new function.
13838 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
13839 to a subclass of gimple_opt_pass along with...
13840 (pass_data_slp_vectorize): ...new pass_data instance and...
13841 (make_pass_slp_vectorize): ...new function.
13842 (pass_ipa_increase_alignment): Convert from a global struct to a
13843 subclass of simple_ipa_opt_pass along with...
13844 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
13845 (make_pass_ipa_increase_alignment): ...new function.
13846 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
13847 gimple_opt_pass along with...
13848 (pass_data_vrp): ...new pass_data instance and...
13849 (make_pass_vrp): ...new function.
13850 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
13851 subclass of simple_ipa_opt_pass along with...
13852 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
13853 (make_pass_ipa_free_lang_data): ...new function.
13854 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
13855 gimple_opt_pass along with...
13856 (pass_data_tsan): ...new pass_data instance and...
13857 (make_pass_tsan): ...new function.
13858 (pass_tsan_O0): Convert from a global struct to a subclass of
13859 gimple_opt_pass along with...
13860 (pass_data_tsan_O0): ...new pass_data instance and...
13861 (make_pass_tsan_O0): ...new function.
13862 * var-tracking.c (pass_variable_tracking): Convert from a global
13863 struct to a subclass of rtl_opt_pass along with...
13864 (pass_data_variable_tracking): ...new pass_data instance and...
13865 (make_pass_variable_tracking): ...new function.
13866 * web.c (pass_web): Convert from a global struct to a subclass of
13867 rtl_opt_pass along with...
13868 (pass_data_web): ...new pass_data instance and...
13869 (make_pass_web): ...new function.
13870 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
13871 declaration with that of...
13872 (make_pass_mode_switch_use): ...new function.
13873 (pass_resolve_sw_modes): Replace declaration with that of...
13874 (make_pass_resolve_sw_modes): ...new function.
13875 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
13876 from a global struct to a subclass of rtl_opt_pass along with...
13877 (pass_data_mode_switch_use): ...new pass_data instance and...
13878 (make_pass_mode_switch_use): ...new function.
13879 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
13880 from a global struct to a subclass of rtl_opt_pass along with...
13881 (pass_data_resolve_sw_modes): ...new pass_data instance and...
13882 (make_pass_resolve_sw_modes): ...new function.
13883 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
13884 struct to a subclass of rtl_opt_pass along with...
13885 (pass_data_insert_vzeroupper): ...new pass_data instance and...
13886 (make_pass_insert_vzeroupper): ...new function.
13887 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
13888 global struct to a subclass of rtl_opt_pass along with...
13889 (pass_data_work_around_errata): ...new pass_data instance and...
13890 (make_pass_work_around_errata): ...new function.
13891 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
13892 struct to a subclass of rtl_opt_pass along with...
13893 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
13894 (make_pass_mips_machine_reorg2): ...new function.
13896 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13898 * passes.c (pass_manager::operator new): New.
13900 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13902 Handwritten part of conversion of passes to C++ classes.
13904 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
13905 (toplev.o): Add dep on PASS_MANAGER_H.
13906 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
13907 of early local pases to reflect this moving from a global to a
13908 member of gcc::pass_manager.
13909 (cgraph_add_new_function): Likewise.
13910 * lto-cgraph.c (lto_output_node): Update for conversion of
13911 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
13912 * passes.c (opt_pass::clone): New.
13913 (opt_pass::gate): New.
13914 (opt_pass::execute): New.
13915 (opt_pass::opt_pass): New.
13916 (pass_manager::execute_early_local_passes): New.
13917 (pass_manager::execute_pass_mode_switching): new.
13918 (finish_optimization_passes): Convert to...
13919 (pass_manager::finish_optimization_passes): ...this.
13920 (finish_optimization_passes): Update for conversion of passes to
13922 (register_dump_files_1): Use has_gate since we cannot portably
13923 check a vtable entry against NULL.
13924 (dump_one_pass): Likewise.
13925 (ipa_write_summaries_2): Likewise.
13926 (ipa_write_optimization_summaries_1): Likewise.
13927 (ipa_read_summaries_1): Likewise.
13928 (ipa_read_optimization_summaries_1): Likewise.
13929 (execute_ipa_stmt_fixups): Likewise.
13930 (pass_manager::pass_manager): Rewrite pass-creation, invoking
13931 pass-creation functions rather than wiring up globals, and
13932 storing the results in fields of pass_manager generated using
13933 pass-instances.def.
13934 (pass_manager::dump_profile_report): Update for conversion of
13935 passes to C++ classes.
13936 (pass_manager::execute_ipa_summary_passes): Likewise.
13937 (execute_one_ipa_transform_pass): Likewise.
13938 (execute_one_pass): Use has_gate and has_execute since we cannot
13939 portably check a vtable entry against NULL.
13940 * pass_manager.h (pass_manager::finish_optimization_passes): New.
13941 (pass_manager): Use pass-instances.def to add fields for the
13942 various pass instances.
13943 * toplev.c (finalize): Update for move of
13944 finish_optimization_passes to a method of gcc::pass_manager.
13945 * toplev.h (finish_optimization_passes): Move to method of class
13947 * tree-pass.h (struct pass_data): New.
13948 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
13949 (opt_pass::gate): Convert to virtual function.
13950 (opt_pass::~opt_pass): New.
13951 (opt_pass::clone): New.
13952 (opt_pass::execute): Convert to virtual function.
13953 (opt_pass::opt_pass): New.
13954 (opt_pass::ctxt_): new.
13955 (gimple_opt_pass): Convert to subclass of opt_pass.
13956 (gimple_opt_pass::gimple_opt_pass): New.
13957 (rtl_opt_pass): Convert to subclass of opt_pass.
13958 (rtl_opt_pass::rtl_opt_pass): New.
13959 (ipa_opt_pass_d): Convert to subclass of opt_pass.
13960 (ipa_opt_pass_d::ipa_opt_pass_d): New.
13961 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
13962 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
13963 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
13964 invocation of pass_mode_switching to reflect this moving from a
13965 global to a member of gcc::pass_manager.
13966 (ix86_option_override): Rework how pass_insert_vzeroupper is
13967 added to the pass_manager to reflect autogenerated changes.
13968 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
13970 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
13972 PR rtl-optimization/57708
13973 * recog.c (peep2_find_free_register): Validate all regs in a
13976 2013-08-05 Jan Hubicka <jh@suse.cz>
13979 * cgraph.c (verify_cgraph_node): Accept local flags from other
13981 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
13982 (function_and_variable_visibility): Likewise.
13983 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
13985 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13987 * graph.c (init_graph_slim_pretty_print): Remove.
13988 (print_graph_cfg): Do not call it. Use local pretty printer.
13989 (start_graph_dump): Likewise.
13991 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13993 * gimple-pretty-print.c (buffer): Remove.
13994 (initialized): Likewise.
13995 (maybe_init_pretty_print): Likewise.
13996 (print_gimple_stmt): Do not call it. Use non-static local
13997 pretty_printer variable.
13998 (print_gimple_expr): Likewise.
13999 (print_gimple_seq): Likewise.
14000 (gimple_dump_bb): Likewise.
14002 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14004 * asan.c (asan_pp): Remove.
14005 (asan_pp_initialized): Likewise.
14006 (asan_pp_initialize): Likewise.
14007 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
14008 (asan_emit_stack_protection): Tidy. Use local pretty printer.
14009 (asan_add_global): Likewise.
14011 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
14013 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
14014 * diagnostic.h (diagnostic_flush_buffer): Adjust.
14015 * pretty-print.c (pp_formatted_text_data): Likewise.
14016 (pp_indent): Rename from pp_base_indent.
14017 (pp_format): Rename from pp_base_format.
14018 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
14019 (pp_format_verbatim): Rename from pp_base_format_verbatim.
14020 (pp_flush): Rename from pp_base_flush.
14021 (pp_set_line_maximum_length): Rename from
14022 pp_base_set_line_maximum_length.
14023 (pp_clear_output_area): Rename from pp_base_clear_output_area.
14024 (pp_set_prefix): Rename from pp_base_set_prefix.
14025 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
14026 (pp_emit_prefix): Rename from pp_base_emit_prefix.
14027 (pp_append_text): Rename from pp_base_append_text.
14028 (pp_formatted_text): Rename from pp_base_formatted_text.
14029 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
14030 (pp_remaining_character_count_for_line): Rename from
14031 pp_base_remaining_character_count_for_line.
14032 (pp_newline): Rename from pp_base_newline.
14033 (pp_character): Rename from pp_base_character.
14034 (pp_string): Rename from pp_base_string.
14035 (pp_maybe_space): Rename from pp_base_maybe_space.
14036 * asan.c (asan_pp_string): Adjust.
14037 (asan_emit_stack_protection): Likewise.
14038 (asan_add_global): Likewise.
14039 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
14040 * tree-mudflap.c (mf_varname_tree): Likewise.
14041 * tree-pretty-print.c (pp_tree_identifier): Rename from
14042 pp_base_tree_identifier.
14043 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
14044 Declare as function.
14046 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14048 * pretty-print.h (pp_bar_bar): New.
14049 (pp_ampersand_ampersand): Likewise.
14050 (pp_less_equal): Likewise.
14051 (pp_greater_equal): Likewise.
14052 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
14053 printer functions instead of pp_string or operators and punctuators.
14054 (dump_gimple_call): Likewise.
14055 (dump_gimple_omp_for): Likewise.
14056 (dump_gimple_transaction): Likewise.
14057 (dump_gimple_phi): Likewise.
14058 (pp_gimple_stmt_1): Likewise.
14059 * sched-vis.c (print_insn): Likewise.
14060 * tree-mudflap.c (mf_varname_tree): Likewise.
14061 * tree-pretty-print.c (dump_block_node): Likewise.
14062 (dump_generic_node): Likewise.
14064 2013-08-02 Jan Hubicka <jh@suse.cz>
14066 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
14068 * lto-streamer-out.c (tree_is_indexable): Results decls and
14069 parm decls are not indexable.
14070 (DFS_write_tree_body): Do not follow args and results.
14071 (hash_tree): Likewise.
14072 (output_functions): Rearrange so struct function is needed
14073 only when real body is output; be able to also ouptut abstract
14074 functions; output DECL_ARGUMENTS and DECL_RESULT.
14075 (lto_output): When not in WPA, ale store abstract functions.
14076 (write_symbol): Do not care about RESULT_DECL.
14077 (output_symbol_p): Handle correctly sbtract decls.
14078 * lto-streamer-in.c (input_function): Rearrange so struct
14079 function can be NULL at entry; allow streaming of
14080 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
14081 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
14082 sanity check during LTO.
14083 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
14084 RESULT_DECl and DECL_ARGUMENTS.
14085 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
14088 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14090 * pretty-print.h (pp_underscore): New.
14092 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
14093 printer functions instead of pp_character.
14094 (dump_binary_rhs): Likewise.
14095 (dump_ternary_rhs): Likewise.
14096 (dump_gimple_call_args): Likewise.
14097 (pp_points_to_solution): Likewise.
14098 (dump_gimple_call): Likewise.
14099 (dump_gimple_switch): Likewise.
14100 (dump_gimple_cond): Likewise.
14101 (dump_gimple_bind): Likewise.
14102 (dump_gimple_try): Likewise.
14103 (dump_gimple_omp_for): Likewise.
14104 (dump_gimple_omp_continue): Likewise.
14105 (dump_gimple_omp_single): Likewise.
14106 (dump_gimple_omp_sections): Likewise.
14107 (dump_gimple_omp_block): Likewise.
14108 (dump_gimple_omp_critical): Likewise.
14109 (dump_gimple_transaction): Likewise.
14110 (dump_gimple_asm): Likewise.
14111 (dump_gimple_phi): Likewise.
14112 (dump_gimple_omp_parallel): Likewise.
14113 (dump_gimple_omp_task): Likewise.
14114 (dump_gimple_omp_atomic_load): Likewise.
14115 (dump_gimple_omp_atomic_store): Likewise.
14116 (dump_gimple_mem_ops): Likewise.
14117 (pp_gimple_stmt_1): Likewise.
14118 (pp_cfg_jump): Likewise.
14119 (dump_implicit_edges): Likewise.
14120 (gimple_dump_bb_for_graph): Likewise.
14121 * graph.c (draw_cfg_node): Likewise.
14122 * langhooks.c (lhd_print_error_function): Likewise.
14123 * sched-vis.c (print_exp): Likewise.
14124 (print_value): Likewise.
14125 (print_pattern): Likewise.
14126 (print_insn): Likewise.
14127 (rtl_dump_bb_for_graph): Likewise.
14128 * tree-pretty-print.c (dump_function_declaration): Likewise.
14129 (dump_array_domain): Likewise.
14130 (dump_omp_clause): Likewise.
14131 (dump_location): Likewise.
14132 (dump_generic_node): Likewise.
14133 (print_struct_decl): Likewise.
14134 * diagnostic.c (diagnostic_show_locus): Use pp_space.
14136 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14138 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
14139 candidate table when replacing a candidate statement.
14140 (replace_rhs_if_not_dup): Likewise.
14141 (replace_one_candidate): Likewise.
14143 2013-08-02 Jan Hubicka <jh@suse.cz>
14144 Martin Liska <marxin.liska@gmail.com>
14146 * cgraphunit.c (add_new_function): Fix logic when adding from
14148 (assemble_thunk): Rename to ...
14149 (expand_thunk); .. this one; export; get it working with
14150 general functions; make produced gimple valid.
14151 * cgraph.h (expand_thunk): Declare.
14153 2013-08-02 Jan Hubicka <jh@suse.cz>
14155 * ipa-cp.c (gather_context_independent_values): Use
14156 ipa_get_param_move_cost.
14157 (get_replacement_map): Remove PARAM; move parameter folding
14159 (create_specialized_node): Update.
14160 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
14161 assert that we have gimple body; update move_cost.
14162 (count_formal_params): Assert that we have gimple body.
14163 (ipa_dump_param): New function.
14164 (ipa_alloc_node_params): Break out from ...
14165 (ipa_initialize_node_params): ... here.
14166 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
14167 (ipa_write_node_info): Stream move costs.
14168 (ipa_read_node_info): Read move costs.
14169 (ipa_update_after_lto_read): Do not recompute node params.
14170 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
14171 (ipa_get_param): Check we are not in WPA.
14172 (ipa_get_param_move_cost): New.
14173 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
14174 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
14175 parm numbers to be present.
14177 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14179 PR rtl-optimization/58048
14180 * lra-constraints.c (process_alt_operands): Don't check asm
14181 operand on register.
14183 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
14185 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
14186 the implied StoreLoad barrier for atomic operations if before.
14188 2013-08-02 Jan Hubicka <jh@suse.cz>
14189 Martin Liska <marxin.liska@gmail.com>
14191 * cgraph.c (cgraph_function_body_availability): Do not check
14193 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
14194 symtab_node_availability): Declare.
14195 * ipa.c (can_replace_by_local_alias): New.
14196 (function_and_variable_visibility): Use it.
14197 * symtab.c (symtab_for_node_and_aliases,
14198 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
14200 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14202 PR rtl-optimization/57963
14203 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
14204 (lra_constraints): Use them.
14206 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
14208 * config/arm/types.md (define_attr "type"): Add "load_acq"
14210 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
14212 (cortex_a53_store1): Likewise.
14214 2013-08-01 Jan Hubicka <jh@suse.cz>
14216 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
14217 partitions are not needed.
14219 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
14221 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
14222 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
14223 MAYBE_NON_Q_CLASS_P where appropriate.
14225 2013-08-01 Jan Hubicka <jh@suse.cz>
14227 * cgraph.h (release_function_body): Declare.
14228 * tree.c (free_lang_data_in_decl): Free, parameters and return values
14229 of unused delcarations.
14231 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14233 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
14234 RTL form when subtracting a constant.
14236 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14238 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
14239 Generate canonical plus rtx with negated immediate instead of minus
14241 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
14243 2013-08-01 Jan Hubicka <jh@suse.cz>
14245 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
14246 (cgraph_release_function_body): Likewise.
14247 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
14248 * cgraph.h (cgrpah_node): Rename abstract_and_needed
14249 to used_as_abstract_origin.
14250 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
14251 symbols used as abstract origins.
14252 * cgraphunit.c (analyze_functions): Update.
14253 * ipa.c (symtab_remove_unreachable_nodes): Recompute
14254 used_as_abstract_origin.
14255 * tree-inline.c (tree_function_versioning): Update
14256 used_as_abstract_origin; be ready for DECL_RESULT and
14257 DECL_ARGUMENTS to be NULL.
14259 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
14260 for abstract functions.
14261 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
14264 2013-08-01 Jan Hubicka <jh@suse.cz>
14266 * profile.c (compute_value_histograms): Fix thinko.
14268 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14270 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
14271 aarch-common-protos.h to extra_headers.
14272 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14273 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
14274 * config/aarch64/t-aarch64 (aarch-common.o): Define.
14276 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14278 * config/aarch64/aarch64.md (define_attr "type"): Delete.
14279 Include "../arm/types.md". Define "type" attribute for all patterns.
14280 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
14283 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
14285 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
14286 to support power8 load fusion.
14287 (fusion_gpr_mem_load): Likewise.
14289 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
14291 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
14292 declarations for power8 load fusion.
14293 (emit_fusion_gpr_load): Likewise.
14295 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
14296 tuning for power8, turn on fusion mode by default. Turn on sign
14297 extending fusion mode if normal fusion mode is on, and we are at
14299 (fusion_gpr_load_p): New function, return true if we can fuse an
14300 addis instruction with a dependent load to a GPR.
14301 (emit_fusion_gpr_load): Emit the instructions for power8 load
14304 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
14305 (VSX load fusion peepholes): New peepholes to fuse together an
14306 addi instruction with a VSX load instruction.
14308 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
14309 peepholes to fuse an addis instruction with a load to a GPR base
14310 register. If we are supporting sign extending fusions, convert
14311 sign extending loads to zero extending loads and add an explicit
14314 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14316 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
14317 aarch-common-protos.h to extra_headers.
14318 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14319 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
14320 (arm_early_store_addr_dep): Likewise.
14321 (arm_no_early_alu_shift_dep): Likewise.
14322 (arm_no_early_alu_shift_value_dep): Likewise.
14323 (arm_no_early_mul_dep): Likewise.
14324 (arm_no_early_store_addr_dep): Likewise.
14325 (arm_mac_accumulator_is_mul_result): Likewise.
14326 (arm_mac_accumulator_is_result): Likewise.
14327 * config/arm/aarch-common.c: ... here. New file.
14328 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
14330 (arm_early_store_addr_dep): Likewise.
14331 (arm_no_early_alu_shift_dep): Likewise.
14332 (arm_no_early_alu_shift_value_dep): Likewise.
14333 (arm_no_early_mul_dep): Likewise.
14334 (arm_no_early_store_addr_dep): Likewise.
14335 (arm_mac_accumulator_is_mul_result): Likewise.
14336 (arm_mac_accumulator_is_result): Likewise.
14337 * config/arm/aarch-common-protos.h: ... here. New file.
14338 * config/arm/t-arm (aarch-common.o): Define.
14340 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14342 * config/arm/arm.md: Include new file "types.md".
14343 (define_attr "type"): Move from here to ...
14344 (define_attr "mul32"): Likewise.
14345 (define_attr "mul64"): Likewise.
14346 * config/arm/types.md: ... here. New file.
14348 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
14350 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
14351 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
14353 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
14355 * gen-pass-instances.awk: Fix offset of substr().
14357 2013-07-31 David Malcolm <dmalcolm@redhat.com>
14359 * Makefile.in (pass-instances.def): New.
14360 (passes.o): Replace dependency on passes.def with one on
14363 * gen-pass-instances.awk: New.
14365 * passes.c (pass_manager::pass_manager): Use pass-instances.def
14366 rather than passes.def, updating local definition of NEXT_PASS
14367 macro to add an extra NUM parameter (currently unused).
14369 2013-07-30 David Malcolm <dmalcolm@redhat.com>
14371 * Makefile.in (PASS_MANAGER_H): New.
14372 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
14373 (passes.o): Likewise.
14374 (statistics.o): Likewise.
14375 (cgraphunit.o): Likewise.
14376 (context.o): Depend on PASS_MANAGER_H.
14378 * pass_manager.h: New.
14380 * cgraphunit.c (cgraph_add_new_function): Update for moves
14381 of globals to fields of pass_manager.
14382 (analyze_function): Likewise.
14383 (expand_function): Likewise.
14384 (ipa_passes): Likewise.
14385 (compile): Likewise.
14387 * context.c (context::context): New.
14388 * context.h (context::context): New.
14389 (context::get_passes): New.
14390 (context::passes_): New.
14392 * lto-cgraph.c (input_node): Update for moves of globals to
14393 fields of pass_manager.
14395 * passes.c (all_passes): Remove, in favor of a field of the
14396 same name within the new class pass_manager.
14397 (all_small_ipa_passes): Likewise.
14398 (all_lowering_passes): Likewise.
14399 (all_regular_ipa_passes): Likewise.
14400 (all_late_ipa_passes): Likewise.
14401 (all_lto_gen_passes): Likewise.
14402 (passes_by_id): Likewise.
14403 (passes_by_id_size): Likewise.
14404 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14405 the new class pass_manager.
14406 (set_pass_for_id): Convert to...
14407 (pass_manager::set_pass_for_id): ...method.
14408 (get_pass_for_id): Convert to...
14409 (pass_manager::get_pass_for_id): ...method.
14410 (register_one_dump_file): Move body of implementation into...
14411 (pass_manager::register_one_dump_file): ...here.
14412 (register_dump_files_1): Convert to...
14413 (pass_manager::register_dump_files_1): ...method.
14414 (register_dump_files): Convert to...
14415 (pass_manager::register_dump_files): ...method.
14416 (create_pass_tab): Update for moves of globals to fields of
14418 (dump_passes): Move body of implementation into...
14419 (pass_manager::dump_passes): ...here.
14420 (register_pass): Move body of implementation into...
14421 (pass_manager::register_pass): ...here.
14422 (init_optimization_passes): Convert into...
14423 (pass_manager::pass_manager): ...constructor for new
14424 pass_manager class, and initialize the pass_lists array.
14425 (check_profile_consistency): Update for moves of globals to
14426 fields of pass_manager.
14427 (dump_profile_report): Move body of implementation into...
14428 (pass_manager::dump_profile_report): ...here.
14429 (ipa_write_summaries_1): Update for moves of pass lists from
14430 being globals to fields of pass_manager.
14431 (ipa_write_optimization_summaries): Likewise.
14432 (ipa_read_summaries): Likewise.
14433 (ipa_read_optimization_summaries): Likewise.
14434 (execute_all_ipa_stmt_fixups): Likewise.
14436 * statistics.c (statistics_fini): Update for moves of globals to
14437 fields of pass_manager.
14439 * toplev.c (general_init): Replace call to
14440 init_optimization_passes with construction of the pass_manager
14443 * tree-pass.h (all_passes): Remove, in favor of a field of the
14444 same name within the new class pass_manager.
14445 (all_small_ipa_passes): Likewise.
14446 (all_lowering_passes): Likewise.
14447 (all_regular_ipa_passes): Likewise.
14448 (all_lto_gen_passes): Likewise.
14449 (all_late_ipa_passes): Likewise.
14450 (passes_by_id): Likewise.
14451 (passes_by_id_size): Likewise.
14452 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14453 the new class pass_manager.
14454 (get_pass_for_id): Remove.
14456 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14458 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
14461 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14463 * arm.md (mulhi3): New expand pattern.
14465 2013-07-30 Jan Hubicka <jh@suse.cz>
14466 Martin Liska <marxin.liska@gmail.com>
14468 * profile.c (compute_value_histograms): Do not ICE when
14469 there is mismatch only on some counters.
14471 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14473 PR rtl-optimization/57637
14474 * function.c (move_insn_for_shrink_wrap): Also check the
14475 GEN set of the LIVE problem for the liveness analysis
14476 if it exists, otherwise give up.
14478 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14480 PR tree-optimization/57993
14481 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
14482 replaced statement in the candidate table.
14483 (phi_add_costs): Return infinite cost when the hidden basis does
14484 not dominate all phis on which the candidate is dependent.
14485 (replace_one_candidate): Record replaced statement in the
14488 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
14490 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
14491 (ashlv2si3): New expander.
14492 (*ashlv2si3_i): New define_insn_and_split.
14493 * predicates.md (float_operation): Allow patterns with three
14494 basic sub-patterns.
14496 PR rtl-optimization/58021
14497 * mode-switching.c (create_pre_exit): Always split off preceding
14498 insns if we are not at the basic block head.
14500 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
14502 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
14503 (UCLIBC_DYNAMIC_LINKER): New macro.
14504 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
14506 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
14507 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
14509 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
14510 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
14511 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
14512 for SF and DF modes. Use ieee_quad_format for TF mode.
14513 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
14514 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
14515 (mips_option_override): Handle `-mnan=legacy'.
14516 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
14517 `-mabs=2008' and `-mnan=2008'.
14518 (OPTION_DEFAULT_SPECS): Add "nan" default.
14519 (ASM_SPEC): Handle `-mnan='.
14520 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
14521 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
14522 comment accordingly.
14523 (neg<mode>2): Likewise.
14524 * config/mips/mips.opt (mabs, mnan): New options.
14525 * doc/install.texi (Configuration): Document `--with-nan=' option.
14526 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
14528 (MIPS Options): Document them.
14529 * config.gcc <mips*-*-*>: Handle `--with-nan='.
14530 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
14531 * configure: Regenerate.
14532 * config.in: Regenerate.
14534 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14536 * config/i386/i386.md (float post-reload splitters): Do not check
14537 for subregs of SSE registers.
14539 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14540 H.J. Lu <hongjiu.lu@intel.com>
14544 * config/i386/i386.md (post-reload splitter
14545 to avoid partial SSE reg dependency stalls): New pattern.
14547 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
14549 * config/s390/s390.md ("movcc"): Swap load and store instructions.
14551 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
14553 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14554 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
14556 2013-07-26 Cary Coutant <ccoutant@google.com>
14558 * dwarf2out.c (die_checksum_ordered): Don't include template
14559 instantiations in signature.
14560 (is_template_parameter): New function.
14561 (is_template_instantiation): New function.
14562 (generate_skeleton_bottom_up): Don't include template instantiations
14564 (generate_skeleton): Likewise.
14565 (break_out_comdat_types): Move recursive call to break out nested
14567 (prune_unused_types_mark_generic_parms_dies): Call
14568 is_template_parameter.
14570 2013-07-26 Ian Bolton <ian.bolton@arm.com>
14572 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
14573 uses vector registers.
14574 * config/aarch64/iterators.md: Add attributes rtn and vas.
14576 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14577 Richard Earnshaw <richard.earnshaw@arm.com>
14579 * combine.c (simplify_comparison): Re-canonicalize operands
14581 * config/arm/arm.md (movcond_addsi): New splitter.
14583 2013-07-25 Sterling Augustine <saugustine@google.com>
14585 * dwarf2out.c (size_of_pubnames): Move code to...
14586 (include_pubname_in_output): ...here. New.
14587 (want_pubnames): Rearrange.
14588 (output_pubnames): Call include_pubname_in_output. Move assertion.
14590 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14592 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
14594 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14597 * doc/extend.texi: Remove obsolete builtins. Fix
14598 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
14600 2013-07-25 Jan Hubicka <jh@suse.cz>
14602 * cgraph.c (release_function_body): Break out from ...
14603 (cgraph_release_function_body): ... this one; also release DECL_RESULT
14604 and DECL_ARGUMENTS.
14605 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
14606 old_tree in the map.
14607 (create_specialized_node): Update.
14608 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
14610 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
14611 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
14612 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
14613 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
14616 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14618 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
14619 addsi3_carryin_alt2_<optab>): Correct output template.
14621 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14623 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
14624 Adjust for arm_restrict_it.
14625 Remove trailing whitespace.
14627 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
14629 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
14630 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
14632 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
14634 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
14636 PR rtl-optimization/57960
14637 * lra-constraints.c (process_alt_operands): Use the right mode
14638 when checking strict_low.
14640 2013-07-25 Jan Hubicka <jh@suse.cz>
14642 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
14643 * cgraph.c (cgraph_remove_node): Do not release function body
14644 when in cgraph streaming.
14645 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
14646 in other partitions are not considered reachable; fix handling of
14649 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14651 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
14653 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14658 * config/arm/arm.md ("*sibcall_insn): Replace use of
14659 Ss with US. Adjust output for v5 and v4t.
14660 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
14662 * config/arm/constraints.md ("Ss"): Rename to US.
14664 2013-07-25 Terry Guo <terry.guo@arm.com>
14666 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
14667 shift_add/shift_sub0/shift_sub1 RTXs.
14669 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
14670 Anton Blanchard <anton@au1.ibm.com>
14672 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
14673 (altivec_vpks<VI_char>ss): Likewise.
14674 (altivec_vpks<VI_char>us): Likewise.
14675 (altivec_vpku<VI_char>us): Likewise.
14676 (altivec_vpku<VI_char>um): Likewise.
14678 2013-07-24 David Malcolm <dmalcolm@redhat.com>
14680 Introduce context class.
14682 * Makefile.in (CONTEXT_H): New.
14683 (OBJS): Add context.o.
14684 (toplev.o): Add CONTEXT_H to dependencies.
14687 * toplev.c (general_init): Create the singleton gcc::context instance.
14693 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
14695 PR rtl-optimization/57968
14696 * mode-switching.c (create_pre_exit): Allow instructions that
14697 don't set a return register to need a non-exit mode.
14699 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14700 Anton Blanchard <anton@au1.ibm.com>
14702 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
14703 operands to vperm for little endian.
14704 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
14705 of lvsl to create the control mask for a vperm for little endian.
14707 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14708 Anton Blanchard <anton@au1.ibm.com>
14710 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
14711 two operands for little-endian.
14713 2013-07-23 Steve Ellcey <sellcey@mips.com>
14715 * config/mips/mips.c (mips_case_values_threshold): New.
14716 (TARGET_CASE_VALUES_THRESHOLD): Define.
14718 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14719 Anton Blanchard <anton@au1.ibm.com>
14721 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
14722 selection of field for vector splat in little endian mode.
14724 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
14726 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
14727 expanders to rs6000.md.
14728 (ior<mode>3): Likewise.
14729 (and<mode>3): Likewise.
14730 (one_cmpl<mode>2): Likewise.
14731 (nor<mode>3): Likewise.
14732 (andc<mode>3): Likewise.
14733 (eqv<mode>3): Likewise.
14734 (nand<mode>3): Likewise.
14735 (orc<mode>3): Likewise.
14737 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
14740 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
14741 to split multi-word logical operations.
14742 (rs6000_split_logical_di): Likewise.
14743 (rs6000_split_logical): Likewise.
14745 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
14746 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
14747 and allow TImode operations in 32-bit.
14748 (vsx_and<mode>3_64bit): Likewise.
14749 (vsx_ior<mode>3_32bit): Likewise.
14750 (vsx_ior<mode>3_64bit): Likewise.
14751 (vsx_xor<mode>3_32bit): Likewise.
14752 (vsx_xor<mode>3_64bit): Likewise.
14753 (vsx_one_cmpl<mode>2_32bit): Likewise.
14754 (vsx_one_cmpl<mode>2_64bit): Likewise.
14755 (vsx_nor<mode>3_32bit): Likewise.
14756 (vsx_nor<mode>3_64bit): Likewise.
14757 (vsx_andc<mode>3_32bit): Likewise.
14758 (vsx_andc<mode>3_64bit): Likewise.
14759 (vsx_eqv<mode>3_32bit): Likewise.
14760 (vsx_eqv<mode>3_64bit): Likewise.
14761 (vsx_nand<mode>3_32bit): Likewise.
14762 (vsx_nand<mode>3_64bit): Likewise.
14763 (vsx_orc<mode>3_32bit): Likewise.
14764 (vsx_orc<mode>3_64bit): Likewise.
14766 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
14767 logical types in GPRs.
14769 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
14770 logical insns to rs6000.md, and allow TImode operations in
14772 (altivec_ior<mode>3): Likewise.
14773 (altivec_xor<mode>3): Likewise.
14774 (altivec_one_cmpl<mode>2): Likewise.
14775 (altivec_nor<mode>3): Likewise.
14776 (altivec_andc<mode>3): Likewise.
14778 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
14779 attributes for moving the 128-bit logical operations into
14781 (BOOL_REGS_OUTPUT): Likewise.
14782 (BOOL_REGS_OP1): Likewise.
14783 (BOOL_REGS_OP2): Likewise.
14784 (BOOL_REGS_UNARY): Likewise.
14785 (BOOL_REGS_AND_CR0): Likewise.
14786 (one_cmpl<mode>2): Add support for DI logical operations on
14787 32-bit, splitting the operations to 32-bit.
14788 (anddi3): Likewise.
14789 (iordi3): Likewise.
14790 (xordi3): Likewise.
14791 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
14792 changes to combine the 32/64-bit code, allow logical operations on
14793 TI mode in 32-bit, and to use similar match_operator patterns like
14794 scalar mode uses. Combine the Altivec and VSX code for logical
14795 operations, and move it here.
14796 (ior<mode>3, 128-bit types): Likewise.
14797 (xor<mode>3, 128-bit types): Likewise.
14798 (one_cmpl<mode>3, 128-bit types): Likewise.
14799 (nor<mode>3, 128-bit types): Likewise.
14800 (andc<mode>3, 128-bit types): Likewise.
14801 (eqv<mode>3, 128-bit types): Likewise.
14802 (nand<mode>3, 128-bit types): Likewise.
14803 (orc<mode>3, 128-bit types): Likewise.
14804 (and<mode>3_internal): Likewise.
14805 (bool<mode>3_internal): Likewise.
14806 (boolc<mode>3_internal1): Likewise.
14807 (boolc<mode>3_internal2): Likewise.
14808 (boolcc<mode>3_internal1): Likewise.
14809 (boolcc<mode>3_internal2): Likewise.
14810 (eqv<mode>3_internal1): Likewise.
14811 (eqv<mode>3_internal2): Likewise.
14812 (one_cmpl1<mode>3_internal): Likewise.
14814 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14816 * config/microblaze/microblaze.c (microblaze_expand_prologue):
14817 Rename flag_stack_usage to flag_stack_usage_info.
14819 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14821 * config/microblaze/sync.md: New file.
14822 * config/microblaze/microblaze.md: Include sync.md
14823 * config/microblaze/microblaze.c: Add print_operand 'y'.
14824 * config/microblaze/constraints.md: Add memory_contraint
14825 'Q' which is a single register.
14827 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
14829 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
14831 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14833 * reload.c (find_reloads): Exit loop once we find this operand
14834 cannot be reloaded somehow for this alternative.
14836 * reload.c (find_reloads): Exit loop once we find a hard register.
14838 * rtlanal.c (computed_jump_p): Exit loop once we find label
14841 * i386.c (ix86_pad_returns): Exit loop after setting replace.
14843 * cfgloopmanip.c (remove_path): Exit loop after setting
14846 * gensupport.c (subst_dup): Avoid loop if code is not
14847 MATCH_DUP nor MATCH_OP_DUP.
14849 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
14851 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
14853 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14855 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
14856 true for SP_REGNUM if mode == ptr_mode.
14857 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
14858 with value R0_REGNUM + 31.
14860 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14862 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
14863 pad pointer-typed argument downward.
14865 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14867 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
14868 and __ILP32__ when the ILP32 model is in use.
14870 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14872 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
14873 (aarch64_load_symref_appropriately): In the case of
14874 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
14875 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
14876 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
14877 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
14878 if 'mode' doesn't equal to 'ptr_mode'.
14879 (aarch64_output_mi_thunk): Add an assertion on the alignment of
14880 'vcall_offset'; change to call aarch64_emit_move differently depending
14881 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
14882 to calculate the upper bound of 'vcall_offset'.
14883 (aarch64_cannot_force_const_mem): Change to also return true if
14885 (aarch64_legitimize_reload_address): In the case of large
14886 displacements, add new local variable 'xmode' and an assertion
14887 based on it; change to use 'xmode' to generate the new rtx and
14889 (aarch64_asm_trampoline_template): Change to generate the template
14890 differently depending on TARGET_ILP32 or not; change to use
14891 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
14892 (aarch64_trampoline_size): Removed.
14893 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
14894 and replace immediate literals with it. Change to use 'ptr_mode'
14895 instead of 'DImode' and call convert_memory_address if the mode
14896 of 'fnaddr' doesn't equal to 'ptr_mode'.
14897 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
14899 (aarch64_elf_asm_destructor): Likewise.
14900 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
14901 on TARGET_ILP32 instead of aarch64_trampoline_size.
14902 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
14903 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
14904 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
14905 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
14906 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
14907 (storewb_pair<GPI:mode>_<P:mode>): ... this.
14908 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
14909 depending on the value of 'mode'.
14910 (add_losym_<mode>): New.
14911 (ldr_got_small_<mode>): New, based on ldr_got_small.
14912 (ldr_got_small): Remove.
14913 (ldr_got_small_sidi): New.
14914 * config/aarch64/iterators.md (P): New.
14915 (PTR): Change to 'ptr_mode' in the condition.
14917 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14919 * config.gcc (aarch64*-*-*): Support --with-abi.
14920 (aarch64*-*-elf): Support --with-multilib-list.
14921 (aarch64*-*-linux*): Likewise.
14922 (supported_defaults): Add abi to aarch64*-*-*.
14923 * configure.ac: Mention AArch64 for --with-multilib-list.
14924 * configure: Re-generated.
14925 * config/aarch64/biarchilp32.h: New file.
14926 * config/aarch64/biarchlp64.h: New file.
14927 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
14929 (MULTILIB_DEFAULTS): Ditto.
14930 (DRIVER_SELF_SPECS): Ditto.
14931 (ASM_SPEC): Update to also substitute -mabi.
14932 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
14933 file whose name depends on -mabi= and -mbig-endian.
14934 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
14936 (POINTER_SIZE): New define.
14937 (POINTERS_EXTEND_UNSIGNED): Ditto.
14938 (enum aarch64_abi_type): New enumeration tag.
14939 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
14940 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
14941 (TARGET_ILP32): New define.
14942 * config/aarch64/aarch64.opt (mabi): New.
14943 (aarch64_abi): New.
14944 (ilp32, lp64): New values for -mabi.
14945 * config/aarch64/t-aarch64 (comma): New define.
14946 (MULTILIB_OPTIONS): Ditto.
14947 (MULTILIB_DIRNAMES): Ditto.
14948 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
14949 * doc/invoke.texi: Document -mabi for AArch64.
14951 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
14953 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
14955 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14956 Anton Blanchard <anton@au1.ibm.com>
14958 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
14959 endianness when selecting field to splat.
14961 2013-07-22 Eric Christopher <echristo@gmail.com>
14963 * dwarf2out.c (die_odr_checksum): New function to use
14964 CHECKSUM_ macros and ULEB128 for DIE tag.
14965 (generate_type_signature): Use.
14967 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
14969 * config.gcc (sparc*-*-*): Accept leon3 processor.
14970 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
14971 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
14972 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
14973 * config/sparc/sparc.opt (enum processor_type): Add leon3.
14974 (mfix-ut699): Adjust comment.
14975 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
14976 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
14977 (CPP_CPU_SPEC): Likewise.
14978 (ASM_CPU_SPEC): Likewise.
14979 * config/sparc/sparc.c (leon3_cost): New constant.
14980 (sparc_option_override): Add leon3 support.
14981 (mem_ref): New function.
14982 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
14983 (sparc_do_work_around_errata): Look into the instruction in the delay
14984 slot and adjust accordingly. Add fix for the data cache nullify issues
14985 of the UT699. Change insertion position for the NOP.
14986 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
14987 (leon3_load): New reservation.
14988 (leon_store): Bump latency to 2.
14989 (grfpu): New automaton.
14990 (grfpu_alu): New unit.
14991 (grfpu_ds): Likewise.
14992 (leon_fp_alu): Adjust.
14993 (leon_fp_mult): Delete.
14994 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
14995 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
14996 * config/sparc/sparc.md (cpu): Add leon3.
14997 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
14998 (swapsi): Likewise.
14999 (atomic_test_and_set): Likewise.
15000 (ldstub): Likewise.
15002 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
15004 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
15005 default for R5900 targets.
15006 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
15007 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
15008 * config/mips/mips.c (mips_option_override): Report an error for
15009 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
15010 for -march=r5900 -mhard-float.
15012 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15014 * df-problems.c (can_move_insns_across): Exit loop once we
15015 find a non-fixed, non-global register.
15017 * ipa-pure-const.c (propagate_nothrow): Exit loop after
15020 * omega.c (omega_eliminate_red): Break after setting red_found.
15021 (omega_problem_has_red_equations): Similarly after setting found.
15022 (omega_query_variable): Similarly after setting coupled.
15024 2013-07-22 Marek Polacek <polacek@redhat.com>
15026 * gimplify.c: Don't include gimple.h twice.
15028 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15030 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
15031 instead of TARGET_THUMB1.
15032 (Pz): New constraint.
15033 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
15035 (compare_negsi_si): Likewise.
15036 (compare_addsi2_op0): Likewise.
15037 (compare_addsi2_op1): Likewise.
15038 (addsi3_carryin_<optab>): Likewise.
15039 (addsi3_carryin_alt2_<optab>): Likewise.
15040 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
15041 for arm_restrict_it.
15042 (subsi3_carryin): Likewise.
15043 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
15044 (minmax_arithsi): Disable for arm_restrict_it.
15045 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
15046 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
15047 (satsi_<SAT:code>_shift): Likewise.
15048 (arm_shiftsi3): Add alternative for 16-bit encoding.
15049 (arm32_movhf): Disable for arm_restrict_it.
15050 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
15051 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
15053 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
15055 * config/arm/arm.md (attribute "insn"): Delete.
15056 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
15057 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
15058 (not_shiftsi): Update for attribute change.
15059 (not_shiftsi_compare0): Likewise.
15060 (not_shiftsi_compare0_scratch): Likewise.
15061 (arm_one_cmplsi2): Likewise.
15062 (thumb1_one_cmplsi2): Likewise.
15063 (notsi_compare0): Likewise.
15064 (notsi_compare0_scratch): Likewise.
15065 (thumb1_movdi_insn): Likewise.
15066 (arm_movsi_insn): Likewise.
15067 (movhi_insn_arch4): Likewise.
15068 (movhi_bytes): Likewise.
15069 (arm_movqi_insn): Likewise.
15070 (thumb1_movqi_insn): Likewise.
15071 (arm32_movhf): Likewise.
15072 (thumb1_movhf): Likewise.
15073 (arm_movsf_soft_insn): Likewise.
15074 (thumb1_movsf_insn): Likewise.
15075 (thumb_movdf_insn): Likewise.
15076 (movsicc_insn): Likewise.
15077 (movsfcc_soft_insn): Likewise.
15078 (and_scc): Likewise.
15079 (cond_move): Likewise.
15080 (if_move_not): Likewise.
15081 (if_not_move): Likewise.
15082 (if_shift_move): Likewise.
15083 (if_move_shift): Likewise.
15084 (if_shift_shift): Likewise.
15085 (if_not_arith): Likewise.
15086 (if_arith_not): Likewise.
15087 (cond_move_not): Likewise.
15088 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
15089 (neon_mov<mode>): Likewise.
15090 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
15091 (thumb2_movsi_vfp): Likewise.
15092 (movsf_vfp): Likewise.
15093 (thumb2_movsf_vfp): Likewise.
15094 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15096 (cortexa7_older_only): Likewise.
15097 (cortexa7_younger): Likewise.
15098 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
15099 (1020alu_shift_op): Likewise.
15100 (1020alu_shift_reg_op): Likewise.
15101 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
15102 (alu_shift_op): Likewise.
15103 (alu_shift_reg_op): Likewise.
15104 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
15105 (11_alu_shift_op): Likewise.
15106 (11_alu_shift_reg_op): Likewise.
15107 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
15108 (9_alu_shift_reg_op): Likewise.
15109 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15111 (cortex_a15_alu_shift): Likewise.
15112 (cortex_a15_alu_shift_reg): Likewise.
15113 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
15114 (cortex_a5_alu_shift): Likewise.
15115 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
15117 (cortex_a53_alu_shift): Likewise.
15118 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15120 (cortex_a7_alu_reg): Likewise.
15121 (cortex_a7_alu_shift): Likewise.
15122 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15123 (cortex_a8_alu_shift): Likewise.
15124 (cortex_a8_alu_shift_reg): Likewise.
15125 (cortex_a8_mov): Likewise.
15126 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
15127 (cortex_a9_dp_shift): Likewise.
15128 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
15129 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
15130 (cortex_r4_mov): Likewise.
15131 (cortex_r4_alu_shift): Likewise.
15132 (cortex_r4_alu_shift_reg): Likewise.
15133 * config/arm/fa526.md (526_alu_op): Update for attribute change.
15134 (526_alu_shift_op): Likewise.
15135 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
15136 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
15137 (626te_alu_shift_op): Likewise.
15138 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
15139 (726te_alu_op): Likewise.
15140 (726te_alu_shift_op): Likewise.
15141 (726te_alu_shift_reg_op): Likewise.
15142 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
15143 (mp626_alu_shift_op): Likewise.
15144 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
15145 (pj4_alu_e1_conds): Likewise.
15146 (pj4_alu): Likewise.
15147 (pj4_alu_conds): Likewise.
15148 (pj4_shift): Likewise.
15149 (pj4_shift_conds): Likewise.
15150 (pj4_alu_shift): Likewise.
15151 (pj4_alu_shift_conds): Likewise.
15153 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15155 * config/arm/predicates.md (shiftable_operator_strict_it):
15157 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
15158 Disable cond_exec version for arm_restrict_it.
15159 (thumb2_smaxsi3): Convert to generate cond_exec.
15160 (thumb2_sminsi3): Likewise.
15161 (thumb32_umaxsi3): Likewise.
15162 (thumb2_uminsi3): Likewise.
15163 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
15164 (thumb2_neg_abssi2): Likewise.
15165 (thumb2_mov_scc): Add alternative for 16-bit encoding.
15166 (thumb2_movsicc_insn): Adjust alternatives.
15167 (thumb2_mov_negscc): Disable for arm_restrict_it.
15168 (thumb2_mov_negscc_strict_it): New pattern.
15169 (thumb2_mov_notscc_strict_it): New pattern.
15170 (thumb2_mov_notscc): Disable for arm_restrict_it.
15171 (thumb2_ior_scc): Likewise.
15172 (thumb2_ior_scc_strict_it): New pattern.
15173 (thumb2_cond_move): Adjust for arm_restrict_it.
15174 (thumb2_cond_arith): Disable for arm_restrict_it.
15175 (thumb2_cond_arith_strict_it): New pattern.
15176 (thumb2_cond_sub): Adjust for arm_restrict_it.
15177 (thumb2_movcond): Likewise.
15178 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
15179 (thumb2_zero_extendhisi2_v6): Likewise.
15180 (thumb2_zero_extendqisi2_v6): Likewise.
15181 (orsi_notsi_si): Likewise.
15182 (orsi_not_shiftsi_si): Likewise.
15184 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
15186 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
15187 instruction sequence is 1 byte shorter.
15189 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
15191 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
15192 it is not needed after split.
15194 2013-07-20 Iain Sandoe <iain@codesourcery.com>
15197 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
15198 second label for nonlocal goto receivers. Don't output pic base labels
15199 unless we're producing PIC; mark that action unreachable().
15200 (ix86_save_reg): If the function contains a nonlocal label, save the
15202 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
15203 * config/darwin.c (emitted_pic_label_num): New GTY.
15204 (update_pic_label_number_if_needed): New.
15205 (machopic_output_function_base_name): Adjust for nonlocal receiver
15207 (machopic_should_output_picbase_label): New.
15208 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
15209 (nonlocal_goto_receiver): New insn and split.
15211 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
15213 * config/aarch64/aarch64-builtins.c
15214 (aarch64_fold_builtin): Fold abs in all modes.
15215 * config/aarch64/aarch64-simd-builtins.def
15216 (abs): Enable for all modes.
15217 * config/aarch64/arm_neon.h
15218 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
15219 (vabs_f64): Add missing intrinsic.
15221 2013-07-19 Ian Bolton <ian.bolton@arm.com>
15223 * config/aarch64/arm_neon.h (vabs_s64): New function
15225 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
15228 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
15229 * config/avr/avr.md (adjust_len): Add `round'.
15230 * config/avr/avr-protos.h (avr_out_round): New prototype.
15231 (avr_out_plus): Add `out_label' argument.
15232 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
15233 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
15234 Handle the case where `insn' is just a pattern.
15235 (avr_out_bitop): Handle the case where `insn' is just a pattern.
15236 (avr_out_round): New function.
15237 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
15239 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
15241 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15242 Add check for flag_stack_usage to handle -fstack-usage support
15244 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
15246 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
15247 interaction for new Power8 flags and VSX.
15249 2013-07-18 Sriraman Tallam <tmsriram@google.com>
15251 PR middle-end/57698
15252 * tree-inline.c (expand_call_inline): Emit errors during
15253 early_inlining only if optimization is not turned on.
15255 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15259 * passes.c (init_optimization_passes): Move the construction of
15260 the pass hierarchy into a new passes.def file.
15262 * Makefile.in (passes.o): Add dependency on passes.def.
15264 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15266 * passes.c (init_optimization_passes): Introduce macros for
15267 constructing the tree of passes (INSERT_PASSES_AFTER,
15268 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
15269 TERMINATE_PASS_LIST).
15271 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
15272 Wei Mi <wmi@google.com>
15274 PR rtl-optimization/57878
15275 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
15277 (reload_pseudo_compare_func): Check nregs first for reload
15280 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15282 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
15284 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
15286 * read-rtl.c (validate_const_int): Once an invalid character is
15287 seen, quit the loop.
15289 * gengtype.c (write_roots): Similarly once we find the "deletable"
15290 or "if_marked" option.
15292 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15294 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
15295 "xtab" and "sat". Move value "clz" from here to ...
15296 (attriubte "type"): ... here.
15297 (satsi_<SAT:code>): Delete "insn" attribute.
15298 (satsi_<SAT:code>_shift): Likewise.
15299 (arm_zero_extendqisi2addsi): Likewise.
15300 (arm_extendqisi2addsi): Likewise.
15301 (clzsi2): Update for attribute changes.
15302 (rbitsi2): Likewise.
15303 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
15305 (arm_usatsihi): Likewise.
15306 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15308 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15310 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
15311 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
15312 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
15313 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
15314 in alphabetical order.
15315 (attribute "core_cycles"): Update for attribute changes.
15316 (arm_addsi3): Likewise.
15317 (addsi3_compare0): Likewise.
15318 (addsi3_compare0_scratch): Likewise.
15319 (addsi3_compare_op1): Likewise.
15320 (addsi3_compare_op2): Likewise.
15321 (compare_addsi2_op0): Likewise.
15322 (compare_addsi2_op1): Likewise.
15323 (addsi3_carryin_shift_<optab>): Likewise.
15324 (subsi3_carryin_shift): Likewise.
15325 (rsbsi3_carryin_shift): Likewise.
15326 (arm_subsi3_insn): Likewise.
15327 (subsi3_compare0): Likewise.
15328 (subsi3_compare): Likewise.
15329 (arm_andsi3_insn): Likewise.
15330 (thumb1_andsi3_insn): Likewise.
15331 (andsi3_compare0): Likewise.
15332 (andsi3_compare0_scratch): Likewise.
15333 (zeroextractsi_compare0_scratch
15334 (andsi_not_shiftsi_si): Likewise.
15335 (iorsi3_insn): Likewise.
15336 (iorsi3_compare0): Likewise.
15337 (iorsi3_compare0_scratch): Likewise.
15338 (arm_xorsi3): Likewise.
15339 (thumb1_xorsi3_insn): Likewise.
15340 (xorsi3_compare0): Likewise.
15341 (xorsi3_compare0_scratch): Likewise.
15342 (satsi_<SAT:code>_shift): Likewise.
15344 (arm_shiftsi3): Likewise.
15345 (shiftsi3_compare0): Likewise.
15346 (not_shiftsi): Likewise.
15347 (not_shiftsi_compare0): Likewise.
15348 (not_shiftsi_compare0_scratch): Likewise.
15349 (arm_one_cmplsi2): Likewise.
15350 (thumb_one_complsi2): Likewise.
15351 (notsi_compare0): Likewise.
15352 (notsi_compare0_scratch): Likewise.
15353 (thumb1_zero_extendhisi2): Likewise.
15354 (arm_zero_extendhisi2): Likewise.
15355 (arm_zero_extendhisi2_v6): Likewise.
15356 (arm_zero_extendhisi2addsi): Likewise.
15357 (thumb1_zero_extendqisi2): Likewise.
15358 (thumb1_zero_extendqisi2_v6): Likewise.
15359 (arm_zero_extendqisi2): Likewise.
15360 (arm_zero_extendqisi2_v6): Likewise.
15361 (arm_zero_extendqisi2addsi): Likewise.
15362 (thumb1_extendhisi2): Likewise.
15363 (arm_extendhisi2): Likewise.
15364 (arm_extendhisi2_v6): Likewise.
15365 (arm_extendqisi): Likewise.
15366 (arm_extendqisi_v6): Likewise.
15367 (arm_extendqisi2addsi): Likewise.
15368 (thumb1_extendqisi2): Likewise.
15369 (thumb1_movdi_insn): Likewise.
15370 (arm_movsi_insn): Likewise.
15371 (movsi_compare0): Likewise.
15372 (movhi_insn_arch4): Likewise.
15373 (movhi_bytes): Likewise.
15374 (arm_movqi_insn): Likewise.
15375 (thumb1_movqi_insn): Likewise.
15376 (arm32_movhf): Likewise.
15377 (thumb1_movhf): Likewise.
15378 (arm_movsf_soft_insn): Likewise.
15379 (thumb1_movsf_insn): Likewise.
15380 (movdf_soft_insn): Likewise.
15381 (thumb_movdf_insn): Likewise.
15382 (arm_cmpsi_insn): Likewise.
15383 (cmpsi_shiftsi): Likewise.
15384 (cmpsi_shiftsi_swp): Likewise.
15385 (arm_cmpsi_negshiftsi_si): Likewise.
15386 (movsicc_insn): Likewise.
15387 (movsfcc_soft_insn): Likewise.
15388 (arith_shiftsi): Likewise.
15389 (arith_shiftsi_compare0
15390 (arith_shiftsi_compare0_scratch
15391 (sub_shiftsi): Likewise.
15392 (sub_shiftsi_compare0
15393 (sub_shiftsi_compare0_scratch
15394 (and_scc): Likewise.
15395 (cond_move): Likewise.
15396 (if_plus_move): Likewise.
15397 (if_move_plus): Likewise.
15398 (if_move_not): Likewise.
15399 (if_not_move): Likewise.
15400 (if_shift_move): Likewise.
15401 (if_move_shift): Likewise.
15402 (if_shift_shift): Likewise.
15403 (if_not_arith): Likewise.
15404 (if_arith_not): Likewise.
15405 (cond_move_not): Likewise.
15406 (thumb1_ashlsi3): Set type attribute.
15407 (thumb1_ashrsi3): Likewise.
15408 (thumb1_lshrsi3): Likewise.
15409 (thumb1_rotrsi3): Likewise.
15410 (shiftsi3_compare0_scratch): Likewise.
15411 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
15412 (neon_mov<mode>): Likewise.
15413 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
15415 (thumb2_movsi_insn): Likewise.
15416 (thumb2_cmpsi_neg_shiftsi): Likewise.
15417 (thumb2_extendqisi_v6): Likewise.
15418 (thumb2_zero_extendhisi2_v6): Likewise.
15419 (thumb2_zero_extendqisi2_v6): Likewise.
15420 (thumb2_shiftsi3_short): Likewise.
15421 (thumb2_addsi3_compare0_scratch): Likewise.
15422 (orsi_not_shiftsi_si): Likewise.
15423 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
15424 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
15426 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
15427 (1020alu_shift_op): Likewise.
15428 (1020alu_shift_reg_op): Likewise.
15429 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
15430 (alu_shift_op): Likewise.
15431 (alu_shift_reg_op): Likewise.
15432 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
15433 (11_alu_shift_op): Likewise.
15434 (11_alu_shift_reg_op): Likewise.
15435 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
15436 (9_alu_shift_reg_op): Likewise.
15437 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15439 (cortex_a15_alu_shift): Likewise.
15440 (cortex_a15_alu_shift_reg): Likewise.
15441 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
15443 (cortex_a5_alu_shift): Likewise.
15444 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
15446 (cortex_a53_alu_shift): Likewise.
15447 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15449 (cortex_a7_alu_reg): Likewise.
15450 (cortex_a7_alu_shift): Likewise.
15451 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
15453 (cortex_a8_alu_shift): Likewise.
15454 (cortex_a8_alu_shift_reg): Likewise.
15455 (cortex_a8_mov): Likewise.
15456 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
15457 (cortex_a9_dp_shift): Likewise.
15458 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
15460 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
15462 (cortex_r4_mov): Likewise.
15463 (cortex_r4_alu_shift): Likewise.
15464 (cortex_r4_alu_shift_reg): Likewise.
15465 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
15466 (526_alu_shift_op): Likewise.
15467 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
15468 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
15469 (626te_alu_shift_op): Likewise.
15470 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
15471 (726te_alu_op): Likewise.
15472 (726te_alu_shift_op): Likewise.
15473 (726te_alu_shift_reg_op): Likewise.
15474 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
15475 (mp626_alu_shift_op): Likewise.
15476 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
15477 (pj4_alu_e1_conds): Likewise.
15478 (pj4_alu): Likewise.
15479 (pj4_alu_conds): Likewise.
15480 (pj4_shift): Likewise.
15481 (pj4_shift_conds): Likewise.
15482 (pj4_alu_shift): Likewise.
15483 (pj4_alu_shift_conds): Likewise.
15484 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15486 (cortexa7_older_only): Likewise.
15487 (cortexa7_younger): Likewise.
15489 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15491 * ipa-pure-const.c (generate_summary): Rename to...
15492 (pure_const_generate_summary): ... this.
15494 2013-07-17 Iain Sandoe <iain@codesourcery.com>
15496 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
15498 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
15501 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
15504 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15506 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
15507 enabled without -march=zEC12.
15508 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
15511 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15513 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
15514 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
15515 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
15516 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
15517 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
15518 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
15519 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
15521 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
15522 and ISA_HAS_NMADD3_NMSUB3.
15523 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
15524 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
15525 (nmsub4<mode>, nmsub3<mode>): Likewise.
15526 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
15528 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15530 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
15531 TARGET_MIPS5400 checking.
15533 2013-07-16 Jakub Jelinek <jakub@redhat.com>
15534 Peter Bergner <bergner@vnet.ibm.com>
15536 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
15537 registers in the comment.
15538 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
15539 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
15540 rather than FIRST_PSEUDO_REGISTERS.
15542 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
15544 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
15545 enable extra ISA flags with TARGET_HTM.
15547 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15549 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
15552 2013-07-15 Cong Hou <congh@google.com>
15554 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
15555 in compare function for sorting.
15557 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
15559 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
15560 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
15561 * config/rs6000/rs6000.opt: Add -mhtm option.
15562 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
15563 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
15564 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
15565 __HTM__ if the HTM instructions are available.
15566 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
15567 htm_spr_reg_operand): New define_predicates.
15568 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
15569 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
15571 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
15572 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
15573 HTM builtin functions.
15574 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
15575 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
15576 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
15577 (rs6000_builtin_mask_calculate): Likewise.
15578 (rs6000_option_override_internal): Likewise.
15579 (bdesc_htm): Add new HTM builtin support.
15580 (htm_spr_num): New function.
15581 (htm_spr_regno): Likewise.
15582 (rs6000_htm_spr_icode): Likewise.
15583 (htm_expand_builtin): Likewise.
15584 (htm_init_builtins): Likewise.
15585 (rs6000_expand_builtin): Add support for HTM builtin functions.
15586 (rs6000_init_builtins): Likewise.
15587 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
15589 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
15590 (TARGET_HTM, MASK_HTM): Define macros.
15591 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
15592 (FIXED_REGISTERS): Likewise.
15593 (CALL_USED_REGISTERS): Likewise.
15594 (CALL_REALLY_USED_REGISTERS): Likewise.
15595 (REG_ALLOC_ORDER): Likewise.
15596 (enum reg_class): Likewise.
15597 (REG_CLASS_NAMES): Likewise.
15598 (REG_CLASS_CONTENTS): Likewise.
15599 (REGISTER_NAMES): Likewise.
15600 (ADDITIONAL_REGISTER_NAMES): Likewise.
15601 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
15602 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
15603 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
15604 * config/rs6000/htm.md: New file.
15605 * config/rs6000/htmintrin.h: New file.
15606 * config/rs6000/htmxlintrin.h: New file.
15608 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
15610 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15611 Define SYMBOL_TINY_GOT, update comment.
15612 * config/aarch64/aarch64.c
15613 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
15614 (aarch64_expand_mov_immediate): Likewise.
15615 (aarch64_print_operand): Likewise.
15616 (aarch64_classify_symbol): Likewise.
15617 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
15618 (ldr_got_tiny): Define.
15620 2013-07-13 Tobias Grosser <tobias@grosser.es>
15622 PR tree-optimization/54094
15623 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
15624 scheduling dimension for the parallelism check from the polyhedral
15625 information in the AST.
15626 * graphite-dependences.c (carries_deps): Do not assume the schedule is
15629 2013-07-13 Jason Merrill <jason@redhat.com>
15631 * print-tree.c (debug_vec_tree): Use debug_raw.
15632 (debug_raw (vec<tree, va_gc> &)): New.
15633 (debug_raw (vec<tree, va_gc> *)): New.
15634 * tree.h: Declare them.
15636 2013-07-13 Bin Cheng <bin.cheng@arm.com>
15638 * ifcvt.c (ifcvt_after_combine): New static variable.
15639 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
15641 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
15642 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
15643 rest_of_handle_if_after_reload): Pass new argument for if_convert.
15645 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
15647 * config/mips/mips.c (mips_expand_call): Remove empty statement.
15649 2013-07-12 Michael Matz <matz@suse.de>
15651 PR middle-end/55771
15652 * convert.c (convert_to_real): Reject non-float inner types.
15654 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
15656 * config/aarch64/aarch64-protos.h
15657 (aarch64_simd_immediate_valid_for_move): Remove.
15658 * config/aarch64/aarch64.c (simd_immediate_info): New member.
15659 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
15661 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
15663 2013-07-11 Steve Ellcey <sellcey@mips.com>
15665 * config/mips/mips.c (mips_conditional_register_usage): Do not
15666 use t[0-7] registers in MIPS16 mode when optimizing for size.
15668 2013-07-11 Sriraman Tallam <tmsriram@google.com>
15670 * config/i386/i386.c (dispatch_function_versions): Fix array
15671 indexing of function_version_info to match actual_versions.
15673 2013-07-11 Teresa Johnson <tejohnson@google.com>
15675 * vec.h (struct va_gc): Move release out-of-line.
15676 (va_gc::release): Call ggc_free on released vec.
15678 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15680 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
15681 Require GOT register as additional operand in UNSPEC.
15682 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
15683 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15684 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15685 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
15686 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
15687 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
15688 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
15690 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
15693 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
15694 name seen by assembler/linker rather if available.
15696 2013-07-11 Andreas Schwab <schwab@suse.de>
15698 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
15700 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
15702 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
15704 2013-07-10 Joseph Myers <joseph@codesourcery.com>
15706 * doc/tm.texi.in: Move hook documentation to ....
15707 * target.def: ... here.
15709 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
15710 text on @hook line.
15711 * doc/tm.texi: Regenerate.
15713 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
15716 * doc/invoke.texi: Document -Wconditionally-supported.
15718 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15721 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
15724 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15727 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
15728 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
15729 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
15730 Remove duplicate devices.
15731 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
15732 * config/avr/t-multilib: Regenerate.
15733 * config/avr/avr-tables.opt: Regenerate.
15734 * doc/avr-mmcu.texi: Regenerate.
15736 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15739 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
15741 2013-07-10 Graham Stott <graham.stott@btinternet.com>
15743 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
15744 the cost of MULT when optimizing for size.
15746 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
15748 * config/cr16/cr16-protos.h: Don't include target.h.
15750 2013-07-09 Joseph Myers <joseph@codesourcery.com>
15752 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
15753 adjust register size for TDmode and TFmode for VSX registers.
15755 2013-07-08 Kai Tietz <ktietz@redhat.com>
15758 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
15759 hook_bool_const_tree_true.
15761 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15763 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
15764 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
15765 * config/s390/s390.md: Define FPR*_REGNUM constants.
15766 Fix FPR2_REGNUM constant (18 -> 17).
15767 ("*trunc<BFP:mode><DFP_ALL:mode>2")
15768 ("*trunc<DFP_ALL:mode><BFP:mode>2")
15769 ("trunc<BFP:mode><DFP_ALL:mode>2")
15770 ("trunc<DFP_ALL:mode><BFP:mode>2")
15771 ("*extend<BFP:mode><DFP_ALL:mode>2")
15772 ("*extend<DFP_ALL:mode><BFP:mode>2")
15773 ("extend<BFP:mode><DFP_ALL:mode>2")
15774 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
15777 2013-07-08 Graham Stott <graham.stott@btinternet.com>
15779 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
15781 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15783 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
15784 and cfun_fpr_bit_p to cfun_fpr_save_p.
15785 (s390_frame_area, s390_register_info, s390_frame_info)
15786 (s390_emit_prologue, s390_emit_epilogue)
15787 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
15789 * config/s390/s390.h: Define *_REGNUM macros for floating point
15792 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
15794 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
15796 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
15798 PR rtl-optimization/57786
15799 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
15800 and break out of the loop when it is set to false.
15802 2013-07-08 Jakub Jelinek <jakub@redhat.com>
15805 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
15806 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
15807 (const_int 63)) 0)).
15808 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
15809 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
15810 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
15812 PR rtl-optimization/57829
15813 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
15814 mask bits outside of mode are just sign-extension from mode to HWI.
15816 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
15818 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
15819 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
15820 adjust_address instead of change_address to keep info about alignment.
15821 (emit_strmov): Remove.
15822 (emit_memmov): New function.
15823 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
15824 (expand_movmem_epilogue): Likewise and return updated rtx for
15826 (expand_constant_movmem_prologue): Likewise and return updated rtx for
15827 destination and source.
15828 (decide_alignment): Refactor, handle vector_loop.
15829 (ix86_expand_movmem): Likewise.
15830 (ix86_expand_setmem): Likewise.
15831 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
15833 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
15835 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
15836 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
15838 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15840 * config/i386/sse.md (sse_movlhps): Change alternative 3
15841 of operand 2 to "m".
15843 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15846 * config/i386/sse.md (iptr): New mode attribute.
15847 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
15848 (<sse>_vm<plusminus_insn><mode>3): Ditto.
15849 (<sse>_vmmul<mode>3): Ditto.
15850 (<sse>_vmdiv<mode>3): Ditto.
15851 (sse_vmrcpv4sf2): Ditto.
15852 (<sse>_vmsqrt<mode>2): Ditto.
15853 (sse_vmrsqrtv4sf2): Ditto.
15854 (<sse>_vm<code><mode>3): Ditto.
15855 (avx_vmcmp<mode>3): Ditto.
15856 (<sse>_vmmaskcmp<mode>3): Ditto.
15857 (<sse>_comi): Ditto.
15858 (<sse>_ucomi): Ditto.
15859 (*xop_vmfrcz_<mode>): Ditto.
15860 (*fmai_fmadd_<mode>): Ditto.
15861 (*fmai_fmsub_<mode>): Ditto.
15862 (*fmai_fnmadd_<mode>): Ditto.
15863 (*fmai_fnmsub_<mode>): Ditto.
15864 (*fma4i_vmfmadd_<mode>): Ditto.
15865 (*fma4i_vmfmsub_<mode>): Ditto.
15866 (*fma4i_vmfnmadd_<mode>): Ditto.
15867 (*fma4i_vmfnmsub_<mode>): Ditto.
15868 (*xop_vmfrcz_<mode>): Ditto.
15869 (sse_cvtps2pi): Ditto.
15870 (sse_cvttps2pi): Ditto.
15871 (sse_cvtss2si): Ditto.
15872 (sse_cvtss2si_2): Ditto.
15873 (sse_cvtss2siq_2): Ditto.
15874 (sse_cvttss2si): Ditto.
15875 (sse_cvttss2siq): Ditto.
15876 (sse_cvtsd2si): Ditto.
15877 (sse_cvtsd2si_2): Ditto.
15878 (sse_cvtsd2siq_2): Ditto.
15879 (sse_cvttsd2si): Ditto.
15880 (sse_cvttsd2siq): Ditto.
15881 (sse_cvtsd2ss): Ditto.
15882 (sse_cvtss2sd): Ditto.
15883 (avx2_pbroadcast<mode>): Ditto.
15884 (avx2_pbroadcast<mode>_1): Ditto.
15885 (*avx_vperm_broadcast_v4sf): Ditto.
15887 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
15888 (sse_movlhps): Ditto.
15889 (sse_storehps): Ditto.
15890 (sse_loadhps): Ditto.
15891 (sse_storelps): Ditto.
15892 (sse_loadlps): Ditto.
15893 (*vec_concatv4sf): Ditto.
15894 (*vec_interleave_highv2df): Ditto.
15895 (*vec_interleave_lowv2df): Ditto.
15896 (*vec_extractv2df_1_sse): Ditto.
15897 (*vec_extractv2df_0_sse): Ditto.
15898 (sse2_storelpd): Ditto.
15899 (sse2_loadlpd): Ditto.
15900 (sse2_movsd): Ditto.
15901 (*vec_concatv4si): Ditto.
15902 (vec_concatv2di): Ditto.
15904 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
15905 for Intel asm dialect.
15906 (mmx_punpcklwd): Ditto.
15907 (mmx_punpckldq): Ditto.
15909 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
15910 for intel assembler dialect.
15912 2013-07-06 Jakub Jelinek <jakub@redhat.com>
15915 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
15916 for BUILT_IN_C{LZ,LRSB}*.
15917 * tree.h (CASE_INT_FN): Add FN##IMAX case.
15918 * tree-vrp.c (extract_range_basic): Handle
15919 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
15920 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
15921 fall thru to code calling set_value*.
15922 * builtins.c (expand_builtin): Remove *IMAX cases.
15923 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
15924 if width is bigger than 2*HWI.
15926 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
15928 PR rtl-optimization/55342
15929 * lra-int.h (lra_subreg_reload_pseudos): New.
15930 * lra.c: Add undoing optional reloads to the block diagram.
15931 (lra_subreg_reload_pseudos): New.
15932 (lra_optional_reload_pseudos): Change comments.
15933 (lra): Init and clear lra_subreg_reload_pseudos. Clear
15934 lra_optional_reload_pseudos after undo transformations.
15935 * lra-assigns.c (pseudo_prefix_title): New.
15936 (lra_setup_reg_renumber): Use it.
15937 (spill_for): Ditto. Check subreg reload pseudos too.
15938 (assign_by_spills): Consider subreg reload pseudos too.
15939 * lra-constraints.c (simplify_operand_subreg): Use
15940 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
15941 (curr_insn_transform): Recognize and do optional reloads.
15942 (undo_optional_reloads): New.
15943 (lra_undo_inheritance): Call undo_optional_reloads.
15945 2013-07-05 Thomas Quinot <quinot@adacore.com>
15947 * tree-complex.c (expand_complex_operations_1): Fix typo.
15949 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
15951 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
15952 (tune_params): New member 'const vec_costs'.
15953 * config/aarch64/aarch64.c (generic_vector_cost): New.
15954 (generic_tunings): New member 'generic_vector_cost'.
15955 (aarch64_builtin_vectorization_cost): New.
15956 (aarch64_add_stmt_cost): New.
15957 (TARGET_VECTORIZE_ADD_STMT_COST): New.
15958 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
15960 2013-07-03 Jakub Jelinek <jakub@redhat.com>
15963 * config/i386/predicates.md (vsib_address_operand): Disallow
15964 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
15966 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
15968 PR middle-end/55030
15969 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
15970 expand_builtin_setjmp_receiver.
15971 (expand_label): Adjust, call expand_builtin_setjmp_receiver
15972 with NULL for the label parameter.
15973 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
15974 the frame-pointer. Adjust comments.
15975 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
15976 only if LABEL is non-NULL.
15978 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
15980 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
15981 (ARM_ABI_AAPCS64): Ditto.
15983 (ARM_DEFAULT_ABI): Ditto.
15985 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
15987 * config/aarch64/aarch64-builtins.c
15988 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
15989 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
15991 * config/aarch64/aarch64-simd.md
15992 (aarch64_ld1<VALL:mode>): New.
15993 (aarch64_st1<VALL:mode>): Likewise.
15994 * config/aarch64/arm_neon.h
15995 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
15997 2013-07-02 Sriraman Tallam <tmsriram@google.com>
15999 * config/i386/i386.c (gate_insert_vzeroupper): Check if
16001 (ix86_option_override_internal):Turn on all -mavx target flags by
16002 default as they are dependent on AVX anyway.
16004 2013-07-02 Cary Coutant <ccoutant@google.com>
16006 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
16007 deterministic hash.
16008 (loc_checksum_ordered): Likewise.
16009 (hash_loc_operands): Remove inline keyword.
16011 2013-07-02 Jakub Jelinek <jakub@redhat.com>
16013 PR tree-optimization/57741
16014 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
16015 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
16016 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
16017 Allow REAL_CST step_exprs if flag_associative_math.
16018 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
16020 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16022 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
16024 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16026 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
16028 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16030 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
16032 (iorsi3_insn): Likewise.
16033 (arm_xorsi3): Likewise.
16035 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
16037 * arm.md (attribute "wtype"): Delete. Move attribute values from here
16039 (attribute "type"): ... here, and prefix with "wmmx_".
16040 (attribute "core_cycles"): Update for attribute changes.
16041 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
16042 (tbcstv4hi): Likewise.
16043 (tbcstv2si): Likewise.
16044 (iwmmxt_iordi3): Likewise.
16045 (iwmmxt_xordi3): Likewise.
16046 (iwmmxt_anddi3): Likewise.
16047 (iwmmxt_nanddi3): Likewise.
16048 (iwmmxt_arm_movdi): Likewise.
16049 (iwmmxt_movsi_insn): Likewise.
16050 (mov<mode>_internal): Likewise.
16051 (and<mode>3_iwmmxt): Likewise.
16052 (ior<mode>3_iwmmxt): Likewise.
16053 (xor<mode>3_iwmmxt): Likewise.
16054 (add<mode>3_iwmmxt): Likewise.
16055 (ssaddv8qi3): Likewise.
16056 (ssaddv4hi3): Likewise.
16057 (ssaddv2si3): Likewise.
16058 (usaddv8qi3): Likewise.
16059 (usaddv4hi3): Likewise.
16060 (usaddv2si3): Likewise.
16061 (sub<mode>3_iwmmxt): Likewise.
16062 (sssubv8qi3): Likewise.
16063 (sssubv4hi3): Likewise.
16064 (sssubv2si3): Likewise.
16065 (ussubv8qi3): Likewise.
16066 (ussubv4hi3): Likewise.
16067 (ussubv2si3): Likewise.
16068 (mulv4hi3_iwmmxt): Likewise.
16069 (smulv4hi3_highpart): Likewise.
16070 (umulv4hi3_highpart): Likewise.
16071 (iwmmxt_wmacs): Likewise.
16072 (iwmmxt_wmacsz): Likewise.
16073 (iwmmxt_wmacu): Likewise.
16074 (iwmmxt_wmacuz): Likewise.
16075 (iwmmxt_clrdi): Likewise.
16076 (iwmmxt_clrv8qi): Likewise.
16077 (iwmmxt_clr4hi): Likewise.
16078 (iwmmxt_clr2si): Likewise.
16079 (iwmmxt_uavgrndv8qi3): Likewise.
16080 (iwmmxt_uavgrndv4hi3): Likewise.
16081 (iwmmxt_uavgv8qi3): Likewise.
16082 (iwmmxt_uavgv4hi3): Likewise.
16083 (iwmmxt_tinsrb): Likewise.
16084 (iwmmxt_tinsrh): Likewise.
16085 (iwmmxt_tinsrw): Likewise.
16086 (iwmmxt_textrmub): Likewise.
16087 (iwmmxt_textrmsb): Likewise.
16088 (iwmmxt_textrmuh): Likewise.
16089 (iwmmxt_textrmsh): Likewise.
16090 (iwmmxt_textrmw): Likewise.
16091 (iwmxxt_wshufh): Likewise.
16092 (eqv8qi3): Likewise.
16093 (eqv4hi3): Likewise.
16094 (eqv2si3): Likewise.
16095 (gtuv8qi3): Likewise.
16096 (gtuv4hi3): Likewise.
16097 (gtuv2si3): Likewise.
16098 (gtv8qi3): Likewise.
16099 (gtv4hi3): Likewise.
16100 (gtv2si3): Likewise.
16101 (smax<mode>3_iwmmxt): Likewise.
16102 (umax<mode>3_iwmmxt): Likewise.
16103 (smin<mode>3_iwmmxt): Likewise.
16104 (umin<mode>3_iwmmxt): Likewise.
16105 (iwmmxt_wpackhss): Likewise.
16106 (iwmmxt_wpackwss): Likewise.
16107 (iwmmxt_wpackdss): Likewise.
16108 (iwmmxt_wpackhus): Likewise.
16109 (iwmmxt_wpackwus): Likewise.
16110 (iwmmxt_wpackdus): Likewise.
16111 (iwmmxt_wunpckihb): Likewise.
16112 (iwmmxt_wunpckihh): Likewise.
16113 (iwmmxt_wunpckihw): Likewise.
16114 (iwmmxt_wunpckilb): Likewise.
16115 (iwmmxt_wunpckilh): Likewise.
16116 (iwmmxt_wunpckilw): Likewise.
16117 (iwmmxt_wunpckehub): Likewise.
16118 (iwmmxt_wunpckehuh): Likewise.
16119 (iwmmxt_wunpckehuw): Likewise.
16120 (iwmmxt_wunpckehsb): Likewise.
16121 (iwmmxt_wunpckehsh): Likewise.
16122 (iwmmxt_wunpckehsw): Likewise.
16123 (iwmmxt_wunpckelub): Likewise.
16124 (iwmmxt_wunpckeluh): Likewise.
16125 (iwmmxt_wunpckeluw): Likewise.
16126 (iwmmxt_wunpckelsb): Likewise.
16127 (iwmmxt_wunpckelsh): Likewise.
16128 (iwmmxt_wunpckelsw): Likewise.
16129 (ror<mode>3): Likewise.
16130 (ashr<mode>3_iwmmxt): Likewise.
16131 (lshr<mode>3_iwmmxt): Likewise.
16132 (ashl<mode>3_iwmmxt): Likewise.
16133 (ror<mode>3_di): Likewise.
16134 (ashr<mode>3_di): Likewise.
16135 (lshr<mode>3_di): Likewise.
16136 (ashl<mode>3_di): Likewise.
16137 (iwmmxt_wmadds): Likewise.
16138 (iwmmxt_wmaddu): Likewise.
16139 (iwmmxt_tmia): Likewise.
16140 (iwmmxt_tmiaph): Likewise.
16141 (iwmmxt_tmiabb): Likewise.
16142 (iwmmxt_tmiatb): Likewise.
16143 (iwmmxt_tmiabt): Likewise.
16144 (iwmmxt_tmiatt): Likewise.
16145 (iwmmxt_tmovmskb): Likewise.
16146 (iwmmxt_tmovmskh): Likewise.
16147 (iwmmxt_tmovmskw): Likewise.
16148 (iwmmxt_waccb): Likewise.
16149 (iwmmxt_wacch): Likewise.
16150 (iwmmxt_waccw): Likewise.
16151 (iwmmxt_waligni): Likewise.
16152 (iwmmxt_walignr): Likewise.
16153 (iwmmxt_walignr0): Likewise.
16154 (iwmmxt_walignr1): Likewise.
16155 (iwmmxt_walignr2): Likewise.
16156 (iwmmxt_walignr3): Likewise.
16157 (iwmmxt_wsadb): Likewise.
16158 (iwmmxt_wsadh): Likewise.
16159 (iwmmxt_wsadbz): Likewise.
16160 (iwmmxt_wsadhz): Likewise.
16161 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
16162 (iwmmxt_wabsdiffb): Likewise.
16163 (iwmmxt_wabsdiffh): Likewise.
16164 (iwmmxt_wabsdiffw): Likewise.
16165 (iwmmxt_waddsubhx): Likewise
16166 (iwmmxt_wsubaddhx): Likewise.
16167 (addc<mode>3): Likewise.
16168 (iwmmxt_avg4): Likewise.
16169 (iwmmxt_avg4r): Likewise.
16170 (iwmmxt_wmaddsx): Likewise.
16171 (iwmmxt_wmaddux): Likewise.
16172 (iwmmxt_wmaddsn): Likewise.
16173 (iwmmxt_wmaddun): Likewise.
16174 (iwmmxt_wmulwsm): Likewise.
16175 (iwmmxt_wmulwum): Likewise.
16176 (iwmmxt_wmulsmr): Likewise.
16177 (iwmmxt_wmulumr): Likewise.
16178 (iwmmxt_wmulwsmr): Likewise.
16179 (iwmmxt_wmulwumr): Likewise.
16180 (iwmmxt_wmulwl): Likewise.
16181 (iwmmxt_wqmulm): Likewise.
16182 (iwmmxt_wqmulwm): Likewise.
16183 (iwmmxt_wqmulmr): Likewise.
16184 (iwmmxt_wqmulwmr): Likewise.
16185 (iwmmxt_waddbhusm): Likewise.
16186 (iwmmxt_waddbhusl): Likewise.
16187 (iwmmxt_wqmiabb): Likewise.
16188 (iwmmxt_wqmiabt): Likewise.
16189 (iwmmxt_wqmiatb): Likewise.
16190 (iwmmxt_wqmiatt): Likewise.
16191 (iwmmxt_wqmiabbn): Likewise.
16192 (iwmmxt_wqmiabtn): Likewise.
16193 (iwmmxt_wqmiatbn): Likewise.
16194 (iwmmxt_wqmiattn): Likewise.
16195 (iwmmxt_wmiabb): Likewise.
16196 (iwmmxt_wmiabt): Likewise.
16197 (iwmmxt_wmiatb): Likewise.
16198 (iwmmxt_wmiatt): Likewise.
16199 (iwmmxt_wmiabbn): Likewise.
16200 (iwmmxt_wmiabtn): Likewise.
16201 (iwmmxt_wmiatbn): Likewise.
16202 (iwmmxt_wmiattn): Likewise.
16203 (iwmmxt_wmiawbb): Likewise.
16204 (iwmmxt_wmiawbt): Likewise.
16205 (iwmmxt_wmiawtb): Likewise.
16206 (iwmmxt_wmiawtt): Likewise.
16207 (iwmmxt_wmiawbbn): Likewise.
16208 (iwmmxt_wmiawbtn): Likewise.
16209 (iwmmxt_wmiawtbn): Likewise.
16210 (iwmmxt_wmiawttn): Likewise.
16211 (iwmmxt_wmerge): Likewise.
16212 (iwmmxt_tandc<mode>3): Likewise.
16213 (iwmmxt_torc<mode>3): Likewise.
16214 (iwmmxt_torvsc<mode>3): Likewise.
16215 (iwmmxt_textrc<mode>3): Likewise.
16216 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
16217 (wmmxt_pack): Likewise.
16218 (wmmxt_mult_c1): Likewise.
16219 (wmmxt_mult_c2): Likewise.
16220 (wmmxt_alu_c1): Likewise.
16221 (wmmxt_alu_c2): Likewise.
16222 (wmmxt_alu_c3): Likewise.
16223 (wmmxt_transfer_c1): Likewise.
16224 (wmmxt_transfer_c2): Likewise.
16225 (wmmxt_transfer_c3): Likewise.
16226 (marvell_f_iwmmxt_wstr): Likewise.
16227 (marvell_f_iwmmxt_wldr): Likewise.
16229 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
16231 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
16233 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16236 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16237 * lra-constraints.c (need_for_split_p): Check call used hard regs
16238 living through calls.
16240 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
16241 call used regs for call insn.
16243 2013-06-28 Jakub Jelinek <jakub@redhat.com>
16246 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
16247 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
16249 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16251 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
16253 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16255 * lra-constraints.c (need_for_split_p): Check call used hard regs
16256 living through calls.
16258 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
16261 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
16262 to tie with any other modes. Eliminate Altivec vector mode tests,
16263 since these are a subset of ALTIVEC or VSX vector modes. Simplify
16264 code, to return 0 if testing MODE2 for a condition, if we've
16265 already tested MODE1 for the same condition.
16267 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16269 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
16272 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16274 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
16275 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
16277 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16279 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
16281 (aarch64_symbolic_constant_p): Remove.
16282 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
16283 static. Fix line length and white space.
16284 (aarch64_symbolic_constant_p): Remove.
16285 * config/aarch64/predicates.md (aarch64_valid_symref):
16286 Use aarch64_classify_symbol_expression.
16288 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16290 * config/arm/constraints.md (Ts): New constraint.
16291 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
16293 (compare_scc): Use "Ts" constraint for operand 0.
16294 (ior_scc_scc): Likewise.
16295 (and_scc_scc): Likewise.
16296 (and_scc_scc_nodom): Likewise.
16297 (ior_scc_scc_cmp): Likewise for operand 7.
16298 (and_scc_scc_cmp): Likewise.
16299 * config/arm/thumb2.md (thumb2_movsi_insn):
16300 Add alternatives for 16-bit encodings.
16301 (thumb2_movhi_insn): Likewise.
16302 (thumb2_movsicc_insn): Likewise.
16303 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
16304 (thumb2_negscc): Use "Ts" constraint.
16305 Move mvn instruction outside cond_exec block.
16306 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
16307 for 16-bit encodings.
16309 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16311 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
16313 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
16314 (mulsi3subsi): Likewise.
16315 (mulsidi3adddi): Likewise.
16316 (mulsidi3_v6): Likewise.
16317 (umulsidi3_v6): Likewise.
16318 (umulsidi3adddi_v6): Likewise.
16319 (smulsi3_highpart_v6): Likewise.
16320 (umulsi3_highpart_v6): Likewise.
16321 (mulhisi3tb): Likewise.
16322 (mulhisi3bt): Likewise.
16323 (mulhisi3tt): Likewise.
16324 (maddhisi4): Likewise.
16325 (maddhisi4tb): Likewise.
16326 (maddhisi4tt): Likewise.
16327 (maddhidi4): Likewise.
16328 (maddhidi4tb): Likewise.
16329 (maddhidi4tt): Likewise.
16330 (zeroextractsi_compare0_scratch): Likewise.
16331 (insv_zero): Likewise.
16332 (insv_t2): Likewise.
16333 (anddi_notzesidi_di): Likewise.
16334 (anddi_notsesidi_di): Likewise.
16335 (andsi_notsi_si): Likewise.
16336 (iordi_zesidi_di): Likewise.
16337 (xordi_zesidi_di): Likewise.
16338 (andsi_iorsi3_notsi): Likewise.
16339 (smax_0): Likewise.
16340 (smax_m1): Likewise.
16341 (smin_0): Likewise.
16342 (not_shiftsi): Likewise.
16343 (unaligned_loadsi): Likewise.
16344 (unaligned_loadhis): Likewise.
16345 (unaligned_loadhiu): Likewise.
16346 (unaligned_storesi): Likewise.
16347 (unaligned_storehi): Likewise.
16348 (extv_reg): Likewise.
16349 (extzv_t2): Likewise.
16350 (divsi3): Likewise.
16351 (udivsi3): Likewise.
16352 (arm_zero_extendhisi2addsi): Likewise.
16353 (arm_zero_extendqisi2addsi): Likewise.
16354 (compareqi_eq0): Likewise.
16355 (arm_extendhisi2_v6): Likewise.
16356 (arm_extendqisi2addsi): Likewise.
16357 (arm_movt): Likewise.
16358 (thumb2_ldrd): Likewise.
16359 (thumb2_ldrd_base): Likewise.
16360 (thumb2_ldrd_base_neg): Likewise.
16361 (thumb2_strd): Likewise.
16362 (thumb2_strd_base): Likewise.
16363 (thumb2_strd_base_neg): Likewise.
16364 (arm_negsi2): Add alternative for 16-bit encoding.
16365 (arm_one_cmplsi2): Likewise.
16367 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16369 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
16370 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
16371 (movdfcc): Likewise.
16372 * config/arm/vfp.md (*thumb2_movsf_vfp):
16373 Disable predication for arm_restrict_it.
16374 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
16375 (*thumb2_movdfcc_vfp): Likewise.
16376 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
16377 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
16378 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
16379 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
16380 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
16381 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
16382 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
16383 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
16384 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
16385 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
16386 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
16387 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
16388 Disable predication for arm_restrict_it.
16390 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
16392 * config/i386/bmiintrin.h (_bextr_u32): New.
16393 (_bextr_u64): Ditto.
16395 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
16397 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
16398 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
16399 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
16400 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
16401 * config/mips/n32-elf.h: ...this new file.
16403 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
16406 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
16407 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
16409 2013-06-27 Catherine Moore <clm@codesourcery.com>
16411 * config/mips/mips-tables.opt: Regenerate.
16412 * config/mips/mips-cpus.def: Add m14ke and m14kec.
16413 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
16414 * doc/invoke.texi: Add -m14kc.
16416 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16419 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
16420 constraints of operand 1 and 2.
16423 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
16424 to match RTL canonicalization. Swap predicates and
16425 constraints of operand 1 and 2.
16427 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
16429 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
16430 Process OP_INOUT regs for splitting too.
16432 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16434 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
16435 decl before the loop, initialize to NULL.
16436 (vectorizable_load): Initialize ptr_incr to NULL.
16438 2013-06-27 Martin Jambor <mjambor@suse.cz>
16441 * ipa-ref.h (ipa_maybe_record_reference): Declare.
16442 * ipa-ref.c (ipa_maybe_record_reference): New function.
16443 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
16444 * ipa-cp.c (create_specialized_node): Record potential references from
16446 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
16448 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16450 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
16451 parameter 'mode' of type 'enum machine_mode mode'; change to pass
16452 'mode' to force_reg.
16453 (aarch64_add_offset): Update calls to aarch64_force_temporary.
16454 (aarch64_expand_mov_immediate): Likewise.
16456 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16458 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
16459 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
16461 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16463 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
16464 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16465 (struct machine_function): Add tbegin_p.
16466 (s390_canonicalize_comparison): Fold CC mode compares to
16467 conditional jump if possible.
16468 (s390_emit_jump): Return the emitted jump.
16469 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
16470 Handle CCRAWmode compares.
16471 (s390_option_override): Default to -mhtm if available.
16472 (s390_reg_clobbered_rtx): Handle floating point regs as well.
16473 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
16474 FPRs instead of df_regs_ever_live_p.
16475 (s390_optimize_nonescaping_tx): New function.
16476 (s390_init_frame_layout): Extend clobbered_regs array to cover
16478 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
16479 (s390_expand_tbegin): New function.
16480 (enum s390_builtin): New enum definition.
16481 (code_for_builtin): New array definition.
16482 (s390_init_builtins): New function.
16483 (s390_expand_builtin): New function.
16484 (TARGET_INIT_BUILTINS): Define.
16485 (TARGET_EXPAND_BUILTIN): Define.
16486 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
16487 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
16488 (s390_alc_comparison): Likewise.
16489 * config/s390/s390-modes.def: Add CCRAWmode.
16490 * config/s390/s390.h (processor_flags): Add PF_TX.
16491 (TARGET_CPU_HTM): Define macro.
16492 (TARGET_HTM): Define macro.
16493 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
16494 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
16495 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16496 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
16497 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
16498 (TBEGIN_MASK, TBEGINC_MASK): New constants.
16499 ("*cc_to_int"): Move up.
16500 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
16501 constants other than 0.
16502 ("*ccraw_to_int"): New insn and splitter definition.
16503 ("tbegin", "tbegin_nofloat", "tbegin_retry")
16504 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
16505 ("tx_assist"): New expander.
16506 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
16507 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
16508 * config/s390/s390.opt: Add -mhtm option.
16509 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
16510 * config/s390/htmxlintrin.h: New file.
16511 * config/s390/htmintrin.h: New file.
16512 * config/s390/s390intrin.h: New file.
16513 * doc/extend.texi: Document htm builtins.
16514 * config.gcc: Add the new header files to extra_headers.
16516 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
16518 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
16519 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
16521 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
16522 Pat Haugen <pthaugen@us.ibm.com>
16523 Peter Bergner <bergner@vnet.ibm.com>
16525 * config/rs6000/power8.md: New.
16526 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
16527 setting for power8 entry.
16528 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
16529 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
16530 test for Power4/Power5 only.
16531 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
16533 (force_new_group): Adjust comment.
16534 * config/rs6000/rs6000.md: Include power8.md.
16536 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
16538 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
16539 * config/arm/arm-protos.h (arm_max_conditional_execute): New
16541 (tune_params): Update comment.
16542 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
16543 (arm_max_conditional_execute): New function.
16544 (thumb2_final_prescan_insn): Use max_insn_skipped and
16545 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
16547 2013-06-25 Jakub Jelinek <jakub@redhat.com>
16549 PR tree-optimization/57705
16550 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
16551 SSA_NAME step, provided that it is not defined inside the loop.
16552 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
16553 (get_initial_def_for_induction): Handle SSA_NAME IV step.
16555 2013-06-25 Martin Jambor <mjambor@suse.cz>
16557 PR middle-end/57670
16558 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
16559 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
16561 (ipa_note_param_call): Initialize member_ptr flag.
16562 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
16563 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
16564 (ipa_write_indirect_edge_info): Stream member_ptr flag.
16565 (ipa_read_indirect_edge_info): Likewise.
16567 2013-06-25 Richard Biener <rguenther@suse.de>
16569 PR middle-end/56977
16570 * passes.c (init_optimization_passes): Move pass_fold_builtins
16571 and pass_dce earlier with -Og.
16573 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
16575 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
16576 <BIT_FIELD_REF>: Remove trailing TAB.
16577 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
16580 2013-06-24 Martin Jambor <mjambor@suse.cz>
16582 PR tree-optimization/57358
16583 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
16584 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
16585 (ipa_analyze_params_uses): Generate pessimistic info when true.
16587 2013-06-24 Martin Jambor <mjambor@suse.cz>
16589 PR tree-optimization/57539
16590 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
16591 global.inlined_to of the new node to it. All callers changed.
16592 * ipa-inline-transform.c (clone_inlined_nodes): New variable
16593 inlining_into, pass it to cgraph_clone_node.
16594 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
16595 ipa_free_edge_args_substructures.
16596 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
16597 rdesc linked list. Do not assert rdesc edges have inlined caller.
16598 Assert we have found an rdesc in the rdesc list.
16600 2013-06-24 Richard Biener <rguenther@suse.de>
16602 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
16603 (pointer_set_lookup): Declare.
16604 (class pointer_map): New template class implementing a
16605 generic pointer to T map.
16606 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
16607 pointer_map<T>::contains, pointer_map<T>::insert,
16608 pointer_map<T>::traverse): New functions.
16609 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
16610 (pointer_set_lookup): New function.
16611 (pointer_set_contains): Use pointer_set_lookup.
16612 (pointer_set_insert): Likewise.
16613 (insert_aux): Remove.
16614 (struct pointer_map_t): Embed a pointer_set_t.
16615 (pointer_map_create): Adjust.
16616 (pointer_map_destroy): Likewise.
16617 (pointer_map_contains): Likewise.
16618 (pointer_map_insert): Likewise.
16619 (pointer_map_traverse): Likewise.
16620 * tree-streamer.h (struct streamer_tree_cache_d): Use a
16621 pointer_map<unsigned> instead of a pointer_map_t.
16622 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
16623 (streamer_tree_cache_lookup): Likewise.
16624 (streamer_tree_cache_create): Likewise.
16625 (streamer_tree_cache_delete): Likewise.
16626 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
16627 pointer_map<unsigned> instead of a pointer_map_t.
16628 (lto_init_tree_ref_encoder): Adjust.
16629 (lto_destroy_tree_ref_encoder): Likewise.
16630 * lto-section-out.c (lto_output_decl_index): Likewise.
16631 (lto_record_function_out_decl_state): Likewise.
16632 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
16634 2013-06-24 Richard Biener <rguenther@suse.de>
16636 PR tree-optimization/57488
16637 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
16639 2013-06-24 Alan Modra <amodra@gmail.com>
16641 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
16642 (gen_easy_altivec_constant): Likewise.
16643 * config/rs6000/predicates.md (easy_vector_constant_add_self,
16644 easy_vector_constant_msb): Likewise.
16646 2013-06-23 Jakub Jelinek <jakub@redhat.com>
16649 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
16650 add missing return true.
16652 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
16655 * config/sh/predicates.md (general_extend_operand): Invoke
16656 general_movsrc_operand for memory operands.
16657 (general_movsrc_operand): Allow reg+reg addressing, do not use
16658 general_operand for memory operands.
16660 2013-06-23 Sriraman Tallam <tmsriram@google.com>
16662 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
16663 when current target options does not apply.
16664 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
16665 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
16666 * config/i386/bmiintrin.h: Pass appropriate target
16667 attributes to header.
16668 * config/i386/mmintrin.h: Ditto.
16669 * config/i386/nmmintrin.h: Ditto.
16670 * config/i386/avx2intrin.h: Ditto.
16671 * config/i386/fxsrintrin.h: Ditto.
16672 * config/i386/tbmintrin.h: Ditto.
16673 * config/i386/xsaveintrin.h: Ditto.
16674 * config/i386/f16cintrin.h: Ditto.
16675 * config/i386/xtestintrin.h: Ditto.
16676 * config/i386/xsaveoptintrin.h: Ditto.
16677 * config/i386/bmi2intrin.h: Ditto.
16678 * config/i386/lzcntintrin.h: Ditto.
16679 * config/i386/smmintrin.h: Ditto.
16680 * config/i386/wmmintrin.h: Ditto.
16681 * config/i386/x86intrin.h: Remove all header include guards.
16682 * config/i386/prfchwintrin.h: Ditto.
16683 * config/i386/pmmintrin.h: Ditto.
16684 * config/i386/tmmintrin.h: Ditto.
16685 * config/i386/xmmintrin.h: Ditto.
16686 * config/i386/popcntintrin.h: Ditto.
16687 * config/i386/rdseedintrin.h: Ditto.
16688 * config/i386/ammintrin.h: Ditto.
16689 * config/i386/emmintrin.h: Ditto.
16690 * config/i386/immintrin.h: Remove all header include guards.
16691 * config/i386/fma4intrin.h: Ditto.
16692 * config/i386/lwpintrin.h: Ditto.
16693 * config/i386/xopintrin.h: Ditto.
16694 * config/i386/ia32intrin.h: Ditto.
16695 * config/i386/avxintrin.h: Ditto.
16696 * config/i386/rtmintrin.h: Ditto.
16697 * config/i386/fmaintrin.h: Ditto.
16698 * config/i386/mm3dnow.h: Ditto.
16700 2013-06-22 Sriraman Tallam <tmsriram@google.com>
16702 * common/config/i386/i386-common.c: Handle LZCNT.
16704 2013-06-22 Andi Kleen <ak@linux.intel.com>
16706 * doc/extend.texi: Use __atomic_store_n instead of
16707 __atomic_store in HLE example.
16709 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
16711 * config/sh/sh.c: Remove <cstdlib> workaround.
16713 2013-06-21 Andi Kleen <ak@linux.intel.com>
16715 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
16717 2013-06-21 Andi Kleen <ak@linux.intel.com>
16719 * doc/extend.texi: Document that __atomic_clear and
16720 __atomic_test_and_set should only be used with bool.
16722 2013-06-20 Jan Hubicka <jh@suse.cz>
16724 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
16725 types_same_for_odr.
16726 * tree.c (decls_same_for_odr): New function.
16727 (same_for_edr): New function.
16728 (types_same_for_odr): New function.
16729 (get_binfo_at_offset): Use it.
16730 * tree.h (types_same_for_odr): Declare.
16732 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
16733 Jason Merrill <jason@redhat.com>
16735 * system.h: Include <cstdlib> as well as <stdlib.h>.
16737 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
16740 * config/i386/i386.c (construct_container): Report error if
16741 long double is used with disabled x87 float returns.
16743 2013-06-20 Jan Hubicka <jh@suse.cz>
16745 * lto-cgraph.c (input_symtab): Do not set cgraph state.
16747 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
16749 PR rtl-optimization/57425
16750 PR rtl-optimization/57569
16751 * alias.c (write_dependence_p): Remove parameters mem_mode and
16752 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
16753 Changed all callers.
16754 (canon_anti_dependence): Get comments and semantics in sync.
16755 Add parameter mem_canonicalized. Changed all callers.
16756 * rtl.h (canon_anti_dependence): Update prototype.
16758 2013-06-20 Richard Biener <rguenther@suse.de>
16760 * data-streamer-in.c (streamer_read_uhwi): Optimize single
16761 byte case, inline streamer_read_uchar and defer section
16764 2013-06-20 Richard Biener <rguenther@suse.de>
16766 PR tree-optimization/57584
16767 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
16768 SSA names into the expanded expression that take part in
16769 abnormal coalescing.
16771 2013-06-19 Sharad Singhai <singhai@google.com>
16773 * gcov.c (print_usage): Handle new option.
16774 (process_args): Ditto.
16775 (get_gcov_intermediate_filename): New function.
16776 (output_intermediate_file): New function.
16777 (output_gcov_file): New function
16778 (generate_results): Handle new option.
16779 (release_function): Relase demangled name.
16780 (read_graph_file): Handle demangled name.
16781 (output_lines): Ditto.
16782 * doc/gcov.texi: Document gcov intermediate format.
16784 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
16787 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
16788 (lra_emit_add): Use the functions. Add comment about Y as an
16791 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
16794 * collect2.c (collect_atexit): New.
16795 (collect_exit): Delete.
16796 (main): Register collect_atexit with atexit.
16797 (collect_wait): Change collect_exit to exit.
16799 * collect2.h (collect_exit): Delete.
16800 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
16802 2013-06-19 Wei Mi <wmi@google.com>
16804 PR rtl-optimization/57518
16805 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
16806 if regno is used in paradoxical subreg.
16807 (update_equiv_regs): Check pdx_subregs[regno] before
16808 set a reg to be equivalent with a mem.
16810 2013-06-19 Matthias Klose <doko@ubuntu.com>
16813 * file-find.h (find_a_file): Add a mode parameter.
16814 * file-find.c (find_a_file): Likewise.
16815 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
16816 with X_OK for the executables.
16817 * collect2.c (main): Call find_a_file with X_OK.
16819 2013-06-19 Steve Ellcey <sellcey@mips.com>
16822 * config/mips/mips.md (casesi_internal_mips16_<mode>):
16823 Use NEXT_INSN instead of next_real_insn.
16825 2013-06-19 Jan Hubicka <jh@suse.cz>
16827 * cgraph.h (const_value_known_p): Replace by ...
16828 (ctor_for_folding): .. this one.
16829 * cgraphunit.c (process_function_and_variable_attributes): Use it.
16830 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
16831 * expr.c (expand_expr_real_1): Likewise.
16832 (string_constant): Likewise.
16833 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
16834 * ipa.c (process_references): Likewise.
16835 (symtab_remove_unreachable_nodes): Likewise.
16836 * ipa-inline-analysis.c (param_change_prob): Likewise.
16837 * gimple-fold.c (canonicalize_constructor_val): Likewise.
16838 (get_base_constructor): Likwise.
16839 * varpool.c (varpool_remove_node): Likewise.
16840 (varpool_remove_initializer): LIkewise.
16841 (dump_varpool_node): LIkwise.
16842 (const_value_known_p): Rewrite to ...
16843 (ctor_for_folding): ... this one.
16845 2013-06-19 Jakub Jelinek <jakub@redhat.com>
16848 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
16849 PERSONALITY in $PATH derived prefixes.
16851 2013-06-19 Jeff Law <law@redhat.com>
16853 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
16856 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
16857 (simplify_bitwise_binary): Use it to simpify certain binary ops on
16860 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
16862 * config/arm/vfp.md: Move VFP instruction classification documentation
16864 * config/arm/arm.md: ... here. Update instruction classification
16867 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
16869 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
16870 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
16871 pattern. Use more efficient sequences on ARMv5 and Thumb2.
16873 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
16876 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16877 with NEXT_INSN. Use tablejump_p to check for jump table data
16880 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
16883 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
16884 that now in C++ the value is correct per the C++ standards.
16886 2013-06-19 Richard Biener <rguenther@suse.de>
16888 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
16889 for global context.
16891 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16894 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16897 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16898 with next_active_insn.
16900 2013-06-18 Sriraman Tallam <tmsriram@google.com>
16902 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
16903 functions are inlined during failures to flag an error.
16904 * tree-inline.c (expand_call_inline): Allow the error to be flagged
16905 in early inline pass.
16907 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
16909 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
16912 2013-06-18 Julian Brown <julian@codesourcery.com>
16914 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
16915 Permit virtual register pre-reload if !strict.
16916 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
16918 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
16920 * config/arm/neon.md (movmisalign<mode>): Use
16921 neon_perm_struct_or_reg_operand instead of
16922 neon_struct_or_register_operand.
16923 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
16924 neon_permissive_struct_operand instead of neon_struct_operand.
16925 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
16926 neon_vector_mem_operand.
16927 * config/arm/predicates.md (neon_struct_operand): Adjust call to
16928 neon_vector_mem_operand.
16929 (neon_permissive_struct_operand): New.
16930 (neon_struct_or_register_operand): Rename to...
16931 (neon_perm_struct_or_reg_operand): This. Adjust call to
16932 neon_vector_mem_operand.
16934 2013-06-18 Richard Biener <rguenther@suse.de>
16936 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
16937 * lto-streamer.h: Include pointer-set.h.
16938 (struct lto_decl_slot): Remove.
16939 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
16940 Remove next_index entry.
16941 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16942 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16943 (lto_init_tree_ref_encoder): Adjust.
16944 (lto_destroy_tree_ref_encoder): Likewise.
16945 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16946 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16947 (lto_output_decl_index): Adjust.
16948 (lto_new_out_decl_state): Likewise.
16949 (lto_record_function_out_decl_state): Likewise.
16950 * lto-streamer-out.c (copy_function): Likewise.
16952 2013-06-18 Richard Biener <rguenther@suse.de>
16954 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
16955 * cgraphunit.c: Include cfgloop.h.
16956 (init_lowered_empty_function): Initialize the loop tree.
16957 (assemble_thunk): Insert new BBs into loops.
16959 2013-06-18 Richard Biener <rguenther@suse.de>
16961 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
16962 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
16963 the map from cache entry to cache index optional.
16964 (streamer_tree_cache_replace_tree): Adjust accordingly.
16965 (streamer_tree_cache_append): Likewise.
16966 (streamer_tree_cache_delete): Likewise.
16967 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
16968 streamer cache map from cache entry to cache index.
16969 * lto-streamer-out.c (create_output_block): Adjust.
16971 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
16973 * config/arm/arm.md (attribute "insn"): Move multiplication and
16974 division attributes to...
16975 (attribute "type"): ... here. Remove mult.
16976 (attribute "mul32"): New attribute.
16977 (attribute "mul64"): Add umaal.
16978 (*arm_mulsi3): Update attributes.
16979 (*arm_mulsi3_v6): Likewise.
16980 (*thumb_mulsi3): Likewise.
16981 (*thumb_mulsi3_v6): Likewise.
16982 (*mulsi3_compare0): Likewise.
16983 (*mulsi3_compare0_v6): Likewise.
16984 (*mulsi_compare0_scratch): Likewise.
16985 (*mulsi_compare0_scratch_v6): Likewise.
16986 (*mulsi3addsi): Likewise.
16987 (*mulsi3addsi_v6): Likewise.
16988 (*mulsi3addsi_compare0): Likewise.
16989 (*mulsi3addsi_compare0_v6): Likewise.
16990 (*mulsi3addsi_compare0_scratch): Likewise.
16991 (*mulsi3addsi_compare0_scratch_v6): Likewise.
16992 (*mulsi3subsi): Likewise.
16993 (*mulsidi3adddi): Likewise.
16994 (*mulsi3addsi_v6): Likewise.
16995 (*mulsidi3adddi_v6): Likewise.
16996 (*mulsidi3_nov6): Likewise.
16997 (*mulsidi3_v6): Likewise.
16998 (*umulsidi3_nov6): Likewise.
16999 (*umulsidi3_v6): Likewise.
17000 (*umulsidi3adddi): Likewise.
17001 (*umulsidi3adddi_v6): Likewise.
17002 (*smulsi3_highpart_nov6): Likewise.
17003 (*smulsi3_highpart_v6): Likewise.
17004 (*umulsi3_highpart_nov6): Likewise.
17005 (*umulsi3_highpart_v6): Likewise.
17006 (mulhisi3): Likewise.
17007 (*mulhisi3tb): Likewise.
17008 (*mulhisi3bt): Likewise.
17009 (*mulhisi3tt): Likewise.
17010 (maddhisi4): Likewise.
17011 (*maddhisi4tb): Likewise.
17012 (*maddhisi4tt): Likewise.
17013 (maddhidi4): Likewise.
17014 (*maddhidi4tb): Likewise.
17015 (*maddhidi4tt): Likewise.
17016 (divsi3): Likewise.
17017 (udivsi3): Likewise.
17018 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
17019 (thumb2_mulsi_short_compare0): Likewise.
17020 (thumb2_mulsi_short_compare0_scratch): Likewise.
17021 * config/arm/arm1020e.md (1020mult1): Update attribute change.
17022 (1020mult2): Likewise.
17023 (1020mult3): Likewise.
17024 (1020mult4): Likewise.
17025 (1020mult5): Likewise.
17026 (1020mult6): Likewise.
17027 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
17029 (cortex_a15_mult64): Likewise.
17030 (cortex_a15_sdiv): Likewise.
17031 (cortex_a15_udiv): Likewise.
17032 * config/arm/arm1026ejs.md (mult1): Update attribute change.
17038 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
17039 (pj4_ir_div): Likewise.
17040 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
17041 (11_mult2): Likewise.
17042 (11_mult3): Likewise.
17043 (11_mult4): Likewise.
17044 (11_mult5): Likewise.
17045 (11_mult6): Likewise.
17046 (11_mult7): Likewise.
17047 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
17048 (cortex_a8_mla): Likewise.
17049 (cortex_a8_mull): Likewise.
17050 (cortex_a8_smulwy): Likewise.
17051 (cortex_a8_smlald): Likewise.
17052 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
17053 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
17054 (cortex_r4_mul_3): Likewise.
17055 (cortex_r4_mla_4): Likewise.
17056 (cortex_r4_mla_3): Likewise.
17057 (cortex_r4_smlald): Likewise.
17058 (cortex_r4_mull): Likewise.
17059 (cortex_r4_sdiv): Likewise.
17060 (cortex_r4_udiv): Likewise.
17061 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
17062 (cortex_a7_idiv): Likewise.
17063 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
17064 (9_mult2): Likewise.
17065 (9_mult3): Likewise.
17066 (9_mult4): Likewise.
17067 (9_mult5): Likewise.
17068 (9_mult6): Likewise.
17069 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
17070 (cortex_a53_sdiv): Likewise.
17071 (cortex_a53_udiv): Likewise.
17072 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
17073 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
17074 (mp626_mult2): Likewise.
17075 (mp626_mult3): Likewise.
17076 (mp626_mult4): Likewise.
17077 * config/arm/fa526.md (526_mult1): Update attribute change.
17078 (526_mult2): Likewise.
17079 * config/arm/arm-generic.md (mult): Update attribute change.
17080 (mult_ldsched_strongarm): Likewise.
17081 (mult_ldsched): Likewise.
17082 (multi_cycle): Likewise.
17083 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
17084 * config/arm/fa606te.md (606te_mult1): Update attribute change.
17085 (606te_mult2): Likewise.
17086 (606te_mult3): Likewise.
17087 (606te_mult4): Likewise.
17088 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
17089 (cortex_a9_mac16): Likewise.
17090 (cortex_a9_multiply): Likewise.
17091 (cortex_a9_mac): Likewise.
17092 (cortex_a9_multiply_long): Likewise.
17093 * config/arm/fa626te.md (626te_mult1): Update attribute change.
17094 (626te_mult2): Likewise.
17095 (626te_mult3): Likewise.
17096 (626te_mult4): Likewise.
17098 2013-06-18 Richard Biener <rguenther@suse.de>
17101 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
17103 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17106 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17107 with next_active_insn.
17109 2013-06-18 Alan Modra <amodra@gmail.com>
17111 * config/rs6000/rs6000.h (enum data_align): New.
17112 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
17113 (DATA_ABI_ALIGNMENT): Define.
17114 (CONSTANT_ALIGNMENT): Correct comment.
17115 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
17116 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
17118 2013-06-17 David Malcolm <dmalcolm@redhat.com>
17120 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
17121 ATTRIBUTE_UNUSED marking.
17123 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
17125 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
17126 alternative and update.
17127 (aarch64_dup_lanedi): Delete.
17128 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
17129 * config/aarch64/aarch64-simd-builtins.def: Update.
17131 2013-06-17 Richard Biener <rguenther@suse.de>
17133 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
17134 (lto_input_scc): Declare.
17135 (lto_input_tree_1): Likewise.
17136 (struct lto_stats_d): Add num_tree_bodies_output and
17137 num_pickle_refs_output.
17138 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
17139 (lto_read_tree_1): Split out from ...
17140 (lto_read_tree): ... this.
17141 (lto_input_scc): New function.
17142 (lto_input_tree_1): Split out from ...
17143 (lto_input_tree): ... this. Handle LTO_tree_scc.
17144 (lto_data_in_create): Create the streamer cache without hashes.
17145 * lto-streamer-out.c (create_output_block): Create the streamer
17146 cache with hashes when not doing WPA.
17147 (lto_write_tree_1): Split out from ...
17148 (lto_write_tree): ... this.
17149 (get_symbol_initial_value): New function.
17150 (lto_output_tree_1): Split out from ...
17151 (lto_output_tree): ... this. Write trees as series of SCCs
17152 using a DFS walk via DFS_write_tree.
17153 (struct sccs, struct scc_entry): New types.
17154 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
17155 (DFS_write_tree_body): New function.
17156 (DFS_write_tree): Likewise.
17157 (hash_tree): Likewise.
17158 (scc_entry_compare): Likewise.
17159 (hash_scc): Likewise.
17160 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
17161 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
17162 TREE_CHAIN as regular reference.
17163 (streamer_read_integer_cst): Remove.
17164 (streamer_get_pickled_tree): Adjust.
17165 * tree-streamer-out.c (streamer_write_chain): Disable streaming
17166 of DECL_EXTERNALs in BLOCK_VARS for now.
17167 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
17169 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
17170 Add hash value argument and record that if hashes are recorded
17172 (streamer_tree_cache_insert_1): Adjust.
17173 (streamer_tree_cache_insert): Likewise.
17174 (streamer_tree_cache_insert_at): Rename to ...
17175 (streamer_tree_cache_replace_tree): ... this and adjust.
17176 (streamer_tree_cache_append): Adjust.
17177 (record_common_node): Likewise.
17178 (streamer_tree_cache_create): Add argument whether to
17179 record hash values together with trees.
17180 (streamer_tree_cache_delete): Adjust.
17181 * tree-streamer.h (struct streamer_tree_cache_d): Add
17183 (streamer_read_integer_cst): Remove.
17184 (streamer_tree_cache_insert): Adjust.
17185 (streamer_tree_cache_append): Likewise.
17186 (streamer_tree_cache_insert_at): Rename to ...
17187 (streamer_tree_cache_replace_tree): ... this and adjust.
17188 (streamer_tree_cache_create): Add argument whether to record hashes.
17189 (streamer_tree_cache_get): Rename to ...
17190 (streamer_tree_cache_get_tree): ... this.
17191 (streamer_tree_cache_get_hash): New function.
17192 * tree.c (cache_integer_cst): New function.
17193 * tree.h (cache_integer_cst): Declare.
17194 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
17195 * lto-symtab.c (lto_varpool_replace_node): Only release
17196 DECL_INITIAL of non-prevailing decls.
17197 * varpool.c (varpool_remove_initializer): Do not release
17198 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
17200 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17202 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
17203 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
17204 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
17205 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
17206 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
17207 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
17208 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
17209 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
17210 instead of TARGET_64BIT.
17211 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
17212 Require ISA_HAS_<D>DIV.
17214 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
17216 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
17217 (mips*-*-linux*): Move default with_llsc setting to where other
17219 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
17221 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
17222 Likewise. Remove default with_tune setting. Move default float
17223 setting to its own block. Handle with_llsc in the same block as above.
17225 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
17227 PR rtl-optimization/57425
17228 PR rtl-optimization/57569
17229 * alias.c (write_dependence_p): Add new parameters mem_mode,
17230 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
17231 Changed all callers.
17232 (canon_anti_dependence): New function.
17233 * cse.c (check_dependence): Use canon_anti_dependence.
17234 * cselib.c (cselib_invalidate_mem): Likewise.
17235 * rtl.h (canon_anti_dependence): Declare.
17237 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17239 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
17240 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
17241 ".set mips3" for 64-bit targets.
17243 2013-06-15 Dehao Chen <dehao@google.com>
17245 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
17246 * gimple-low.c (gimple_check_call_matching_types): Likewise.
17247 (gimple_check_call_args): Likewise.
17248 * value-prof.c (check_ic_target): Likewise.
17249 * ipa-inline.c (early_inliner): Likewise.
17250 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
17251 * cgraph.c (cgraph_create_edge_1): Likewise.
17252 (cgraph_make_edge_direct): Likewise.
17254 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
17257 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
17258 rs6000_output_move_128bit to handle emitting quad memory
17259 operations. Set attribute length to 8 bytes.
17261 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
17263 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
17265 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
17266 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
17267 (aarch64_<su>mlsl<mode>): Likewise.
17269 2013-06-14 Mike Stump <mikestump@comcast.net>
17271 * Makefile.in (TARGET_H): Add insn-codes.h.
17273 2013-06-14 Alan Modra <amodra@gmail.com>
17275 PR middle-end/57134
17276 PR middle-end/57586
17277 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
17278 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
17279 bitfield expansion when EXPAND_MEMORY.
17280 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
17282 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
17284 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
17285 test for clearing quad memory on 32-bit later.
17287 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17289 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
17290 (fold_negate_expr): Likewise.
17291 (fold_real_zero_addition_p): Handle vectors.
17292 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
17294 2013-06-14 Alan Modra <amodra@gmail.com>
17296 * varasm.c (force_const_mem): Revert 2013-06-07 change.
17298 2013-06-13 Jan Hubicka <jh@suse.cz>
17300 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17301 Local comdats are not externally visible.
17302 * symtab.c (dump_symtab_base): Dump externally visible.
17303 (verify_symtab_base): Verify back links in the symtab hash.
17305 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17307 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
17308 CONVERT_EXPR as equal nodes.
17310 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17312 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
17314 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17316 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
17317 Generalize to complex and vector.
17318 * tree.c (build_all_ones_cst): New function.
17319 * tree.h (build_all_ones_cst): Declare it.
17321 2013-06-13 Alan Modra <amodra@gmail.com>
17323 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
17324 * config/rs6000/rs6000.md (signbittf2): New insn.
17325 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
17326 (abstf2_internal, cmptf_internal2): Likewise.
17327 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
17329 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
17330 Pat Haugen <pthaugen@us.ibm.com>
17331 Peter Bergner <bergner@vnet.ibm.com>
17333 * config/rs6000/rs6000.c (emit_load_locked): Add support for
17334 power8 byte, half-word, and quad-word atomic instructions.
17335 (emit_store_conditional): Likewise.
17336 (rs6000_expand_atomic_compare_and_swap): Likewise.
17337 (rs6000_expand_atomic_op): Likewise.
17339 * config/rs6000/sync.md (larx): Add new modes for power8.
17341 (AINT): New mode iterator to include TImode as well as normal
17342 integer modes on power8.
17343 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
17344 that VSX registers are not considered. Use AINT mode iterator
17345 instead of INT1 to allow inclusion of quad word atomic operations
17347 (load_locked<mode>): Likewise.
17348 (store_conditional<mode>): Likewise.
17349 (atomic_compare_and_swap<mode>): Likewise.
17350 (atomic_exchange<mode>): Likewise.
17351 (atomic_nand<mode>): Likewise.
17352 (atomic_fetch_<fetchop_name><mode>): Likewise.
17353 (atomic_nand_fetch<mode>): Likewise.
17354 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
17356 (ATOMIC): On power8, add QImode, HImode modes.
17357 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
17358 modes that promote to SImode.
17359 (load_lockedti): Convert TImode arguments to PTImode, so that we
17360 get a guaranteed even/odd register pair.
17361 (load_lockedpti): Likewise.
17362 (store_conditionalti): Likewise.
17363 (store_conditionalpti): Likewise.
17365 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
17366 atomic load/store instructions.
17369 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
17371 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
17373 (insn_count): New attribute, with most cases extracted from...
17374 (length): ...here. Redefine most cases in terms of insn_count.
17375 (single_insn): Delete.
17376 (can_delay): Use insn_count to check for single instructions.
17377 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
17378 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
17379 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
17380 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
17381 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
17382 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
17383 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
17384 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
17385 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
17386 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
17387 rather than "length".
17388 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
17389 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
17390 Use "insn_count" rather than "length".
17391 * config/mips/mips-dsp.md
17392 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
17393 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
17396 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
17398 PR tree-optimization/57361
17399 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
17401 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
17403 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
17405 (aarch64_simd_combine<mode>): New instruction expansion.
17406 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
17407 function prototype.
17408 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
17409 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
17411 2013-06-12 Jan Hubicka <jh@suse.cz>
17413 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
17414 decl has when in streaming stage.
17415 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
17416 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
17418 2013-06-12 Roland Stigge <stigge@antcom.de>
17421 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
17423 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17425 PR tree-optimization/57537
17426 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
17427 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
17429 2013-06-12 Richard Biener <rguenther@suse.de>
17431 * data-streamer.h (streamer_write_char_stream): CSE
17432 obs->current_pointer.
17433 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
17434 streamer_write_char_stream manually and optimize the resulting loop.
17435 (streamer_write_hwi_stream): Likewise.
17437 2013-06-12 Jan Hubicka <jh@suse.cz>
17439 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
17440 * cgraph.h (varpool_create_empty_node): Declare.
17441 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
17443 * symtab.c (symtab_unregister_node): Be lax about missin entries
17445 (symtab_get_node): Update comment.
17446 * varpool.c (varpool_create_empty_node): Break out from ...
17447 (varpool_node_for_decl): ... here.
17448 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
17450 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
17452 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
17453 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
17454 part. Use straight-line flow at the end.
17455 <COMPONENT_REF>: Remove superfluous else.
17456 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
17458 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17461 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
17462 target hook even for !TREE_PUBLIC decls. If no resolution info
17463 is available, return false for common and external decls.
17465 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17467 * config/rl78/constraints.md (U): New constraint.
17468 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
17471 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17474 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
17475 to allow returning address to AT_PLATFORM name.
17477 2013-06-11 Jan Hubicka <jh@suse.cz>
17479 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
17480 * cgraph.h (symtab_node_base): Add weakref flag.
17481 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
17482 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
17483 (output_weakrefs): Use weakref flag.
17484 * fold-const.c (simple_operand_p): Handle WEAK.
17485 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
17486 * ipa.c (varpool_externally_visible_p): Drop weakref.
17487 (function_and_variable_visibility): Update comment; fix weakref
17488 sanity checks; do not clear DECL_WEAK on them.
17489 * lto-cgraph.c (lto_output_node): update.
17490 (lto_output_varpool_node): Update.
17491 (input_overwrite_node): Update.
17492 (input_node): Update.
17493 (input_varpool_node): Update.
17494 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
17495 (lto_symtab_merge_symbols): Add sanity check.
17496 (lto_symtab_prevailing_decl): Do not special case weakrefs.
17497 * passes.c (rest_of_decl_compilation): Set static flag, too.
17498 * symtab.c (dump_symtab_base): Dump weakref.
17499 (verify_symtab_base): Sanity check weakrefs.
17500 (symtab_make_decl_local): Remove duplicated code.
17501 (symtab_alias_ultimate_target): Simplify.
17502 * varpool.c (varpool_create_variable_alias): Set weakref flag.
17504 2013-06-11 Tom de Vries <tom@codesourcery.com>
17506 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
17507 sequence_vect == NULL.
17509 2013-06-11 DJ Delorie <dj@redhat.com>
17511 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
17512 (rl78_unwind_word_mode): New.
17514 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17516 * final.c (debug_prefix_maps): Make static.
17518 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17520 * function.c (initial_trampoline): Remove stray copy.
17522 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
17524 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
17526 2013-06-11 Martin Jambor <mjambor@suse.cz>
17528 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
17529 within bounds at the beginning of the function.
17531 2013-06-11 Alan Modra <amodra@gmail.com>
17533 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
17535 (get_named_section): Don't NULL !DECL_P decl.
17537 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
17539 * doc/invoke.texi (core-avx2): Document.
17541 (atom): Updated with MOVBE.
17543 2013-06-11 Richard Biener <rguenther@suse.de>
17545 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
17547 2013-06-11 Anton Blanchard <anton@samba.org>
17549 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
17550 correct shift value in little-endian mode.
17552 2013-06-11 Jakub Jelinek <jakub@redhat.com>
17555 * varasm.c (get_variable_align): Move #endif to the right place.
17557 2013-06-10 Cary Coutant <ccoutant@google.com>
17559 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
17560 for hash so that hash table traversal order is deterministic.
17562 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
17563 Pat Haugen <pthaugen@us.ibm.com>
17564 Peter Bergner <bergner@vnet.ibm.com>
17566 * config/rs6000/vector.md (GPR move splitter): Do not split moves
17567 of vectors in GPRS if they are direct moves or quad word load or
17570 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
17572 (direct_move_p): Likewise.
17573 (quad_load_store_p): Likewise.
17575 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
17576 classes into bins based on the physical register type.
17577 (reg_class_to_reg_type): Likewise.
17578 (IS_STD_REG_TYPE): Likewise.
17579 (IS_FP_VECT_REG_TYPE): Likewise.
17580 (reload_fpr_gpr): Arrays to determine what insn to use if we can
17581 use direct move instructions.
17582 (reload_gpr_vsx): Likewise.
17583 (reload_vsx_gpr): Likewise.
17584 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
17585 information that is a simplification of register classes. Also
17586 precalculate direct move reload helpers.
17587 (direct_move_p): New function to return true if the operation can
17588 be done as a direct move instruciton.
17589 (quad_load_store_p): New function to return true if the operation
17590 is a quad memory operation.
17591 (rs6000_legitimize_address): If quad memory, only allow register
17592 indirect for TImode addresses.
17593 (rs6000_legitimate_address_p): Likewise.
17594 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
17595 (rs6000_reload_register_type): Likewise.
17596 (register_to_reg_type): Return register type.
17597 (rs6000_secondary_reload_simple_move): New helper function for
17598 secondary reload and secondary memory needed to identify anything
17599 that is a simple move, and does not need reloading.
17600 (rs6000_secondary_reload_direct_move): New helper function for
17601 secondary reload to identify cases that can be done with several
17602 instructions via the direct move instructions.
17603 (rs6000_secondary_reload_move): New helper function for secondary
17604 reload to identify moves between register types that can be done.
17605 (rs6000_secondary_reload): Add support for quad memory operations
17606 and for direct move.
17607 (rs6000_secondary_memory_needed): Likewise.
17608 (rs6000_debug_secondary_memory_needed): Change argument names.
17609 (rs6000_output_move_128bit): New function to return the move to
17610 use for 128-bit moves, including knowing about the various
17611 limitations of quad memory operations.
17613 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
17614 memory operations. call rs6000_output_move_128bit for the actual
17615 instruciton(s) to generate.
17616 (vsx_movti_64bit): Likewise.
17618 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
17619 (UNSPEC_P8V_MTVSRWZ): Likewise.
17620 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
17621 (UNSPEC_P8V_MTVSRD): Likewise.
17622 (UNSPEC_P8V_XXPERMDI): Likewise.
17623 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
17624 (UNSPEC_FUSION_GPR): Likewise.
17625 (FMOVE128_GPR): New iterator for direct move.
17626 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
17627 (f32_sv): Likewise.
17628 (f32_dm): Likewise.
17629 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
17630 loads and direct move instructions.
17631 (zero_extendsidi2_lfiwzx): Likewise.
17632 (extendsidi2_lfiwax): Likewise.
17633 (extendsidi2_nocell): Likewise.
17634 (floatsi<mode>2_lfiwax): Likewise.
17635 (lfiwax): Likewise.
17636 (floatunssi<mode>2_lfiwzx): Likewise.
17637 (lfiwzx): Likewise.
17638 (fix_trunc<mode>_stfiwx): Likewise.
17639 (fixuns_trunc<mode>_stfiwx): Likewise.
17640 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
17641 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
17642 (parity<mode>2_cmpb): Set length/type attr.
17643 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
17644 for 'mr.' to fast_compare.
17645 (bpermd_<mode>): Change type attr to popcnt.
17646 (p8_fmrgow_<mode>): New insns for power8 direct move support.
17647 (p8_mtvsrwz_1): Likewise.
17648 (p8_mtvsrwz_2): Likewise.
17649 (reload_fpr_from_gpr<mode>): Likewise.
17650 (p8_mtvsrd_1): Likewise.
17651 (p8_mtvsrd_2): Likewise.
17652 (p8_xxpermdi_<mode>): Likewise.
17653 (reload_vsx_from_gpr<mode>): Likewise.
17654 (reload_vsx_from_gprsf): Likewise.
17655 (p8_mfvsrd_3_<mode>): LIkewise.
17656 (reload_gpr_from_vsx<mode>): Likewise.
17657 (reload_gpr_from_vsxsf): Likewise.
17658 (p8_mfvsrd_4_disf): Likewise.
17659 (multi-word GPR splits): Do not split direct moves or quad memory
17662 2013-06-10 David Malcolm <dmalcolm@redhat.com>
17664 * tree-into-ssa.c (interesting_blocks): Make static.
17666 2013-06-10 Jakub Jelinek <jakub@redhat.com>
17669 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
17670 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
17671 Use DATA_ABI_ALIGNMENT for that case instead if defined.
17672 (get_variable_align): New function.
17673 (get_variable_section, emit_bss, emit_common,
17674 assemble_variable_contents, place_block_symbol): Use
17675 get_variable_align instead of DECL_ALIGN.
17676 (assemble_noswitch_variable): Add align argument, use it
17677 instead of DECL_ALIGN.
17678 (assemble_variable): Adjust caller. Use get_variable_align
17679 instead of DECL_ALIGN.
17680 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
17682 (DATA_ABI_ALIGNMENT): Define.
17683 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
17684 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
17685 opt is false, only return the psABI mandated alignment increase.
17686 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
17687 (DATA_ABI_ALIGNMENT): ... this.
17688 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
17689 (DATA_ABI_ALIGNMENT): ... this.
17690 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
17691 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
17692 (DATA_ABI_ALIGNMENT): ... this.
17693 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
17694 * doc/tm.texi: Regenerated.
17696 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
17698 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
17699 cmp_code to construct REG_EQUAL note.
17701 2013-06-09 Jakub Jelinek <jakub@redhat.com>
17704 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
17705 that operands[2] doesn't overlap with operands[0].
17707 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
17708 Jan Hubicka <jh@suse.cz>
17710 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
17711 hack to mark symbols as used.
17713 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
17715 PR rtl-optimization/57559
17716 * lra-constraints.c (process_alt_operands): Don't discourage
17717 memory with known offset for offsetable memory constraint.
17718 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
17720 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
17722 * varasm.c (struct oc_local_state): Reorder fields.
17723 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
17724 and adjust accordingly.
17725 (output_constructor): Reorder initialization code and adjust call to
17726 output_constructor_bitfield.
17728 2013-06-07 Jan Hubicka <jh@suse.cz>
17730 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
17732 2013-06-07 David Malcolm <dmalcolm@redhat.com>
17734 * tree-object-size.c (unknown): Make const.
17736 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17738 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
17739 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
17740 for last alternative in the cpu_facility attribute.
17742 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17745 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
17746 (xordi3): Change operand 2 constraint to arm_xordi_operand.
17747 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
17748 * config/arm/constraints.md (Dg): New constraint.
17749 * config/arm/neon.md (xordi3_neon): Remove.
17750 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
17751 * config/arm/predicates.md (arm_xordi_operand): New predicate.
17753 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17755 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
17756 Clean up alternatives.
17758 2013-06-07 Alan Modra <amodra@gmail.com>
17760 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
17763 2013-06-07 Alan Modra <amodra@gmail.com>
17765 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
17767 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17769 * config/arm/constraints.md (Df): New constraint.
17770 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
17771 Correct length attribute for last two alternatives.
17773 2013-06-07 Alan Modra <amodra@gmail.com>
17775 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
17776 override user -mfp-in-toc.
17777 (offsettable_ok_by_alignment): Consider just the current access
17778 rather than the whole object, unless BLKmode. Handle
17779 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
17780 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
17781 for -mcmodel=medium.
17782 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
17783 override user -mfp-in-toc or -msum-in-toc. Default to
17784 -mno-fp-in-toc for -mcmodel=medium.
17786 2013-06-06 DJ Delorie <dj@redhat.com>
17788 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
17789 TARGET_VALID_POINTER_MODE.
17791 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
17792 Pat Haugen <pthaugen@us.ibm.com>
17793 Peter Bergner <bergner@vnet.ibm.com>
17795 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
17796 Document new power8 builtins.
17798 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
17799 condition code register, to allow 128-bit logical operations to be
17800 done in the VSX or GPR registers.
17801 (nor<mode>3): Use the canonical form for nor.
17802 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
17803 vclz*, and vpopcnt* vector instructions.
17804 (nand<mode>3): Likewise.
17805 (orc<mode>3): Likewise.
17806 (clz<mode>2): LIkewise.
17807 (popcount<mode>2): Likewise.
17809 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
17810 that only the GPRs are recognized.
17812 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
17813 support for new power8 builtins.
17815 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
17817 (xscvdpspn): Likewise.
17823 (vpopcnt): Likewise.
17824 (vpopcntb): Likewise.
17825 (vpopcnth): Likewise.
17826 (vpopcntw): Likewise.
17827 (vpopcntd): Likewise.
17829 (vmrgew): Likewise.
17830 (vmrgow): Likewise.
17832 (eqv_v16qi3): Likewise.
17833 (eqv_v8hi3): Likewise.
17834 (eqv_v4si3): Likewise.
17835 (eqv_v2di3): Likewise.
17836 (eqv_v4sf3): Likewise.
17837 (eqv_v2df3): Likewise.
17839 (nand_v16qi3): Likewise.
17840 (nand_v8hi3): Likewise.
17841 (nand_v4si3): Likewise.
17842 (nand_v2di3): Likewise.
17843 (nand_v4sf3): Likewise.
17844 (nand_v2df3): Likewise.
17846 (orc_v16qi3): Likewise.
17847 (orc_v8hi3): Likewise.
17848 (orc_v4si3): Likewise.
17849 (orc_v2di3): Likewise.
17850 (orc_v4sf3): Likewise.
17851 (orc_v2df3): Likewise.
17853 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
17854 allow power8 quad mode in 64-bit.
17855 (rs6000_builtin_vectorized_function): Add support to vectorize
17856 ISA 2.07 count leading zeros, population count builtins.
17857 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
17858 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
17859 (builtin_function_type): Add vgbbd builtin function which takes an
17861 (altivec_expand_vec_perm_const): Add support for new power8 merge
17864 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
17865 that does not include TImdoe for use with 32-bit.
17866 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
17868 (UNSPEC_VSX_CVDPSPN): Likewise.
17869 (vsx_xscvdpspn): Likewise.
17870 (vsx_xscvspdpn): Likewise.
17871 (vsx_xscvdpspn_scalar): Likewise.
17872 (vsx_xscvspdpn_directmove): Likewise.
17873 (vsx_and<mode>3): Split logical operations into 32-bit and
17874 64-bit. Add support to do logical operations on TImode as well as
17875 VSX vector types. Allow logical operations to be done in either
17876 VSX registers or in general purpose registers in 64-bit mode. Add
17877 splitters if GPRs were used. For AND, add clobber of CCmode to
17878 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
17880 (vsx_and<mode>3_32bit): Likewise.
17881 (vsx_and<mode>3_64bit): Likewise.
17882 (vsx_ior<mode>3): Likewise.
17883 (vsx_ior<mode>3_32bit): Likewise.
17884 (vsx_ior<mode>3_64bit): Likewise.
17885 (vsx_xor<mode>3): Likewise.
17886 (vsx_xor<mode>3_32bit): Likewise.
17887 (vsx_xor<mode>3_64bit): Likewise.
17888 (vsx_one_cmpl<mode>2): Likewise.
17889 (vsx_one_cmpl<mode>2_32bit): Likewise.
17890 (vsx_one_cmpl<mode>2_64bit): Likewise.
17891 (vsx_nor<mode>3): Likewise.
17892 (vsx_nor<mode>3_32bit): Likewise.
17893 (vsx_nor<mode>3_64bit): Likewise.
17894 (vsx_andc<mode>3): Likewise.
17895 (vsx_andc<mode>3_32bit): Likewise.
17896 (vsx_andc<mode>3_64bit): Likewise.
17897 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
17898 and xxlorc instructions.
17899 (vsx_eqv<mode>3_64bit): Likewise.
17900 (vsx_nand<mode>3_32bit): Likewise.
17901 (vsx_nand<mode>3_64bit): Likewise.
17902 (vsx_orc<mode>3_32bit): Likewise.
17903 (vsx_orc<mode>3_64bit): Likewise.
17905 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
17907 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
17909 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
17910 (p8_vmrgow): Likewise.
17911 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
17912 GPRs to be split under VSX.
17913 (p8v_clz<mode>2): Add power8 count leading zero support.
17914 (p8v_popcount<mode>2): Add power8 population count support.
17915 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
17918 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
17921 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
17923 (vec_nand): Likewise.
17924 (vec_vclz): Likewise.
17925 (vec_vclzb): Likewise.
17926 (vec_vclzd): Likewise.
17927 (vec_vclzh): Likewise.
17928 (vec_vclzw): Likewise.
17929 (vec_vgbbd): Likewise.
17930 (vec_vmrgew): Likewise.
17931 (vec_vmrgow): Likewise.
17932 (vec_vpopcnt): Likewise.
17933 (vec_vpopcntb): Likewise.
17934 (vec_vpopcntd): Likewise.
17935 (vec_vpopcnth): Likewise.
17936 (vec_vpopcntw): Likewise.
17938 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17940 PR rtl-optimization/57468
17941 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
17944 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17946 PR rtl-optimization/57459
17947 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
17948 type when setting live regs.
17950 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17952 * config/s390/s390.opt (mlra): New option.
17953 * config/s390/s390.c (s390_decompose_address): Check displacement
17954 for all registers for LRA.
17955 (s390_secondary_reload): Don't used secondary reloads for LRA.
17956 (s390_lra_p): New function.
17957 (TARGET_LRA_P): Define.
17958 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
17959 of attribute cpu_facility to zarch for the last alternative.
17960 (*cmpmem_short): Ditto.
17962 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
17964 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
17965 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
17966 (arm_expand_prologue): Likewise.
17968 2013-06-06 Teresa Johnson <tejohnson@google.com>
17971 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
17972 as this is now done by redirect_edge_and_branch_force.
17973 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
17974 barriers, and fix interaction with splitting.
17975 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
17976 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
17977 reflect changes made in the routine.
17978 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
17979 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
17980 since this is called in cfglayout mode, and replace partition fixup
17981 with assert as that is now done by force_nonfallthru_and_redirect.
17982 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
17983 already be marked with region crossing note.
17984 (insert_section_boundary_note): Make non-static, gate on flag
17985 has_bb_partition, rewrite to also check for multiple partitions.
17986 (rest_of_handle_reorder_blocks): Remove call to
17987 insert_section_boundary_note, now done later during free_cfg.
17988 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
17989 * bb-reorder.h (insert_section_boundary_note): Declare.
17990 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
17991 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
17992 invoke insert_section_boundary_note.
17993 (try_redirect_by_replacing_jump): Remove unnecessary
17994 check for region crossing note.
17995 (fixup_partition_crossing): New function.
17996 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
17997 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
17998 in non-cfglayout mode.
17999 (force_nonfallthru_and_redirect): Fixup partition boundaries,
18000 remove old code that tried to do this. Emit barrier correctly
18001 when we are in cfglayout mode.
18002 (last_bb_in_partition): New function.
18003 (rtl_split_edge): Correctly fixup partition boundaries.
18004 (commit_one_edge_insertion): Remove old code that tried to
18005 fixup region crossing edge since this is now handled in
18006 split_block, and set up insertion point correctly since
18007 block may now end in a jump.
18008 (verify_hot_cold_block_grouping): Guard against checking when not in
18009 linearized RTL mode.
18010 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
18012 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
18013 rtl_verify_flow_info, so not called in cfglayout mode.
18014 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
18015 (fixup_reorder_chain): Remove old code that attempted to fixup region
18016 crossing note as this is now handled in force_nonfallthru_and_redirect.
18017 (duplicate_insn_chain): Don't duplicate switch section notes.
18018 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
18020 * basic-block.h (emit_barrier_after_bb): Declare.
18022 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18024 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
18025 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
18026 arm_usatsihi): Adjust alternatives for arm_restrict_it.
18028 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18030 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
18032 * config/arm/ldmstm.md: Regenerate.
18034 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18036 * config/arm/sync.md (atomic_loaddi_1):
18037 Disable predication for arm_restrict_it.
18038 (arm_load_exclusive<mode>): Likewise.
18039 (arm_load_exclusivesi): Likewise.
18040 (arm_load_exclusivedi): Likewise.
18041 (arm_load_acquire_exclusive<mode>): Likewise.
18042 (arm_load_acquire_exclusivesi): Likewise.
18043 (arm_load_acquire_exclusivedi): Likewise.
18044 (arm_store_exclusive<mode>): Likewise.
18045 (arm_store_exclusive<mode>): Likewise.
18046 (arm_store_release_exclusivedi): Likewise.
18047 (arm_store_release_exclusive<mode>): Likewise.
18049 2013-06-06 Richard Biener <rguenther@suse.de>
18051 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
18053 (lto_tag_is_tree_code_p): Adjust.
18054 (lto_tag_is_gimple_code_p): Likewise.
18055 (lto_gimple_code_to_tag): Likewise.
18056 (lto_tag_to_gimple_code): Likewise.
18057 (lto_tree_code_to_tag): Likewise.
18058 (lto_tag_to_tree_code): Likewise.
18059 * data-streamer.h (streamer_write_hwi_in_range): Use
18060 uhwi streaming to stream the normalized range.
18061 (streamer_read_hwi_in_range): Likewise.
18063 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18065 * config/arm/arm.md (enabled_for_depr_it): New attribute.
18066 (predicable_short_it): Likewise.
18067 (predicated): Likewise.
18068 (enabled): Handle above.
18069 (define_cond_exec): Set predicated attribute to yes.
18071 2013-06-05 Mike Stump <mikestump@comcast.net>
18073 * gdbinit.in (__FUNCTION__): Add.
18075 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
18077 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
18078 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
18080 2013-06-05 Jan Hubicka <jh@suse.cz>
18082 * varasm.c (mark_decl_referenced): Revert the removal until targets
18085 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
18087 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
18088 instead of mark_decl_referenced.
18090 2013-06-05 Jan Hubicka <jh@suse.cz>
18092 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
18093 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
18094 and symtab_used_from_object_file_p.
18095 (cgraph_make_node_local_1): Clear forced_by_abi.
18096 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
18097 * cgraph.h (symtab_node_base): Add forced_by_abi.
18098 (decide_is_variable_needed): Remove.
18099 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
18100 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
18101 (decide_is_symbol_needed): ... this one; handle symbols in general;
18102 always analyze virtuals; honnor forced_by_abi.
18103 (cgraph_finalize_function): Update.
18104 (varpool_finalize_decl): Update.
18105 (symbol_defined_and_needed): Remove.
18106 (analyze_functions): Update.
18107 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18108 output_refs, input_overwrite_node): Handle forced_by_abi.
18109 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
18110 (address_taken_from_non_vtable_p): ... this one.
18111 (comdat_can_be_unshared_p_1): New function.
18112 (cgraph_comdat_can_be_unshared_p): Rename to ...
18113 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
18114 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
18115 (function_and_variable_visibility): Clear forced_by_abi as needed.
18116 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
18117 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
18118 * symtab.c (dump_symtab_base): Dump forced_by_abi.
18119 * varpool.c (decide_is_variable_needed): Remove.
18121 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18123 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
18124 (arm_option_override): Override arm_restrict_it where appropriate.
18125 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
18126 * config/arm/arm.opt (mrestrict-it): New command-line option.
18127 * doc/invoke.texi: Document -mrestrict-it.
18129 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18131 * tsan.c (tsan_atomic_table): Make const.
18133 2013-06-05 Richard Biener <rguenther@suse.de>
18135 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
18136 index associated with the tree we are supposed to replace.
18137 * tree-streamer-out.c (pack_ts_base_value_fields): Output
18138 TREE_ASM_WRITTEN as zero for everything but SSA names.
18140 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18142 * tree-ssa-structalias.c (call_stmt_vars): Make static.
18144 2013-06-04 Jan Hubicka <jh@suse.cz>
18146 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
18147 (input_node, input_varpool_node): Handle correctly external same
18149 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
18150 nodes at ltrans stage.
18152 2013-06-04 Jan Hubicka <jh@suse.cz>
18154 * ipa-inline.c (update_caller_keys): Fix availability test.
18155 (update_callee_keys): Likewise.
18156 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
18157 to follow ELF standard.
18159 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
18161 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
18162 (mips64r5900el-*-elf*): New configurations.
18163 * config/mips/mips-cpus.def (r5900): New processor.
18164 * config/mips/mips-tables.opt: Regenerate.
18165 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
18166 (mips_issue_rate): Handle PROCESSOR_R5900.
18167 (mips_reorg_process_insns): Force reorder mode for the R5900.
18168 * config/mips/mips.h (TARGET_MIPS5900): Define.
18169 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
18171 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
18173 * config/mips/mips.md (processor): Add r5900.
18174 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
18176 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18178 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
18179 into function to generate MOVI instruction.
18180 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
18181 (aarch64_preferred_simd_mode): Turn into wrapper.
18182 (aarch64_output_scalar_simd_mov_immediate): New function.
18183 * config/aarch64/aarch64-protos.h: Add prototype for above.
18185 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18187 * config/aarch64/aarch64.c (simd_immediate_info): Remove
18188 element_char member.
18189 (sizetochar): Return signed char.
18190 (aarch64_simd_valid_immediate): Remove elchar and other
18191 unnecessary variables.
18192 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
18193 Calculate element_char as required.
18194 * config/aarch64/aarch64-protos.h: Update and move prototype
18195 for aarch64_output_simd_mov_immediate.
18196 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
18199 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18201 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
18202 information completed by aarch64_simd_valid_immediate.
18203 (aarch64_legitimate_constant_p): Update arguments.
18204 (aarch64_simd_valid_immediate): Work with struct rather than many
18206 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
18207 (aarch64_simd_make_constant): Update arguments.
18208 (aarch64_output_simd_mov_immediate): Work with struct rather than
18209 many pointers. Output immediate directly rather than as operand.
18210 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18212 * config/aarch64/constraints.md (Dn): Update arguments.
18214 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18216 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
18218 (aarch64_simd_immediate_valid_for_move): Remove.
18219 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
18220 (aarch64_simd_make_constant): Update call.
18221 (aarch64_output_simd_mov_immediate): Update call.
18222 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18224 * config/aarch64/constraints.md (Dn): Update call.
18226 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18228 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
18229 return type to bool for prototype.
18230 (aarch64_legitimate_constant_p): Check for true instead of not -1.
18231 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
18232 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
18234 2013-06-04 Catherine Moore <clm@codesourcery.com>
18236 * config/mips/mips.opt (meva): New.
18237 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
18238 (ASM_SPEC): Handle -meva.
18239 * doc/invoke.texi (meva): Document.
18241 2013-06-04 Alan Modra <amodra@gmail.com>
18243 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
18246 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18248 * rtl.def: Add extra fourth optional field to define_cond_exec.
18249 * gensupport.c (process_one_cond_exec): Process attributes from
18251 * doc/md.texi: Document fourth field in define_cond_exec.
18253 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
18255 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
18256 out the processing order as in store_bit_field_1.
18258 2013-06-04 Jan Hubicka <jh@suse.cz>
18260 PR middle-end/57500
18261 * cgraphunit.c (cgraph_process_same_body_aliases): Create
18262 non-VAR_DECL node if it does not exist yet.
18264 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
18266 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
18267 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
18268 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
18269 target_cpu_default setting.
18271 2013-06-03 Teresa Johnson <tejohnson@google.com>
18273 * dumpfile.c (opt_info_switch_p): Change -fopt-info
18274 default to -fopt-info=optimized instead of all.
18275 * doc/invoke.texi: Ditto.
18276 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
18277 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
18278 (execute_vect_slp): Emit BB vectorization success under
18279 MSG_OPTIMIZED_LOCATIONS.
18280 * tree-vect-slp.c (vect_slp_transform_bb): Change
18281 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
18282 * tree-vect-loop.c (vect_transform_loop): Ditto.
18284 2013-06-03 Jason Merrill <jason@redhat.com>
18287 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
18288 Use TARGET_EXPR for C++.
18290 2013-06-03 Jakub Jelinek <jakub@redhat.com>
18292 PR rtl-optimization/57268
18293 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
18294 if DEBUG_INSN_P (insn).
18297 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18299 PR rtl-optimization/57268
18300 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18301 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18303 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18305 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
18306 (ix86_avoid_lea_for_addr): Likewise.
18307 (exact_dependency_1): Likewise.
18308 (ix86_adjust_cost): Likewise.
18309 (swap_top_of_ready_list): Fix formatting and !reload_completed check
18311 (do_reorder_for_imul): Fix typo, formatting and
18312 !reload_completed check removed.
18313 (ix86_sched_reorder): Fix typo and formatting.
18314 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
18317 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
18319 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
18321 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
18323 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
18324 <STRING_CST>: Likewise.
18325 <VECTOR_CST>: Likewise.
18327 2013-06-01 Janus Weil <janus@gcc.gnu.org>
18328 Mikael Morin <mikael@gcc.gnu.org>
18330 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
18331 * config.in: Regenerated.
18332 * configure: Regenerated.
18334 2013-06-01 Jan Hubicka <jh@suse.cz>
18336 PR middle-end/57366
18337 * cgraphunit.c (compile): When weakref is not supported,
18338 set up transparent aliases before final output pass.
18339 * varasm.c (assemble_alias): Do not try to do it here.
18341 2013-06-01 Jan Hubicka <jh@suse.cz>
18343 PR middle-end/57467
18344 * passes.c (for_per_function): Skip unanalyzed functions.
18346 2013-06-01 Jan Hubicka <jh@suse.cz>
18348 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
18349 (lto_symtab_merge_symbols_1): ... this one.
18350 (lto_symtab_merge_cgraph_nodes): Rename to ...
18351 (lto_symtab_merge_symbols): ... this one; simplify.
18352 * cgraph.c (same_body_aliases_done): Rename to ...
18353 (cpp_implicit_aliases_done): ... this one.
18354 (cgraph_create_function_alias): Update.
18355 (cgraph_same_body_alias): Update.
18356 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
18357 (verify_edge_corresponds_to_fndecl): Simplify.
18358 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
18359 (cgraph_node): Remove same_body_alias.
18360 (varpool_node): Remove alias_of and extra_name_alias.
18361 (same_body_aliases_done): Rename to ..
18362 (cpp_implicit_aliases_done): ... this one.
18363 (symtab_alias_ultimate_target): Add default parameter.
18364 (symtab_resolve_alias): New function.
18365 (fixup_same_cpp_alias_visibility): Declare.
18366 (cgraph_function_node): Add default parameter.
18367 (cgraph_node_asm_name): Likewise.
18368 (cgraph_function_or_thunk_node): Add default parameter; do
18369 not ICE when it is NULL.
18370 (varpool_variable_node): Likewise.
18371 * tree-emutls.c (create_emultls_var): Update.
18372 (ipa_lower_emutls): Update.
18373 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
18374 (cgraph_reset_node): Reset alias info.
18375 (cgraph_finalize_function): Update.
18376 (fixup_same_cpp_alias_visibility): Move to symtab.c.
18377 (analyze_function): Simplify.
18378 (cgraph_process_same_body_aliases): Simplify.
18379 (analyze_functions): Fixup same body aliases.
18380 (handle_alias_pairs): Simplify.
18381 (assemble_thunk): Update.
18382 (assemble_thunks_and_aliases): Update.
18383 (output_weakrefs): Rewrite.
18384 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
18385 (lto_output_varpool_node): Likewise.
18386 (compute_ltrans_boundary): Remve assert.
18387 (get_alias_symbol): New functoin.
18388 (input_node): Rewrite alias handling.
18389 (input_varpool_node): Likewise.
18390 * ipa-pure-const.c (propagate_pure_const): Fix formating.
18391 * ipa.c (process_references): Handle weakrefs correctly.
18392 (symtab_remove_unreachable_nodes): Likewise.
18393 * trans-mem.c (get_cg_data): Update.
18394 (ipa_tm_create_version_alias): Update.
18395 (ipa_tm_execute): Update.
18396 * symtab.c (dump_symtab_base): Dump aliases.
18397 (verify_symtab_base): Verify aliases.
18398 (symtab_node_availability): New function.
18399 (symtab_alias_ultimate_target): Simplify.
18400 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
18401 handle all the fixup cases.
18402 (symtab_resolve_alias): New function.
18403 * passes.c (ipa_write_summaries): Handle weakrefs.
18404 * varpool.c (varpool_analyze_node): Simplify.
18405 (assemble_aliases): Update.
18406 (varpool_create_variable_alias): Simplify.
18407 (varpool_extra_name_alias): Simplify.
18408 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
18409 (lto_symtab_merge_symbols): ... this one.
18411 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
18414 PR rtl-optimization/57268
18415 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18416 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18418 2013-06-01 Tobias Burnus <burnus@net-b.de>
18420 Partially reverted:
18421 2013-05-31 Tobias Burnus <burnus@net-b.de>
18423 PR middle-end/57073
18424 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
18427 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18429 PR rtl-optimization/57268
18430 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18431 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18433 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18435 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
18436 unordered comparison operators when -fno-trapping-math is in effect
18438 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
18439 and implement unordered comparison operators properly on the e500.
18441 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18443 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
18444 for constant scalar integers.
18445 (simplify_relational_operation_1): Likewise.
18447 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
18449 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
18450 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
18453 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18454 Igor Zamyatin <igor.zamyatin@intel.com>
18456 Silvermont (SLM) architecture performance tuning.
18457 * config/i386/i386.h (enum ix86_tune_indices): Add
18458 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
18459 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
18461 * config/i386/i386.c (initial_ix86_tune_features)
18462 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
18463 (ix86_lea_outperforms): Handle Silvermont tuning.
18464 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
18466 (ix86_use_lea_for_mov): Likewise.
18467 (ix86_avoid_lea_for_addr): Likewise.
18468 (ix86_lea_for_add_ok): Likewise.
18469 (exact_dependency_1): New function.
18470 (exact_store_load_dependency): Likewise.
18471 (ix86_adjust_cost): Handle Silvermont tuning.
18472 (do_reoder_for_imul): Likewise.
18473 (swap_top_of_ready_list): New function.
18474 (ix86_sched_reorder): Changed to handle Silvermont tuning.
18476 * config/i386/i386.md (peepholes that split memory operand in fp
18479 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18481 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
18482 Remove un-necessary braces.
18484 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18486 * config/aarch64/aarch64.c (aarch64_classify_symbol):
18487 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
18489 2013-05-31 Tobias Burnus <burnus@net-b.de>
18491 PR middle-end/57073
18492 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
18494 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18497 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
18498 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
18499 * config/arm/neon.md (iordi3_neon): Remove.
18500 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
18501 * config/arm/predicates.md (imm_for_neon_logic_operand):
18502 Move to earlier in the file.
18503 (neon_logic_op2): Likewise.
18504 (arm_iordi_operand_neon): New predicate.
18506 2013-05-31 Richard Biener <rguenther@suse.de>
18508 PR tree-optimization/57478
18509 PR tree-optimization/57453
18510 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
18513 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18515 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
18516 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
18518 2013-05-30 Tobias Burnus <burnus@net-b.de>
18519 Thomas Koenig <tkoenig@gcc.gnu.org>
18521 PR middle-end/57073
18522 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
18523 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
18525 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
18527 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
18529 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
18531 * target.def (register_usage_leveling_p): New hook.
18532 * targhooks.c (default_register_usage_leveling_p): New.
18533 * targhooks.h (default_register_usage_leveling_p): New prototype.
18534 * lra-assigns.c (register_usage_leveling_p): Use the hook.
18535 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
18536 * doc/tm.texi: Update.
18537 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
18539 2013-05-30 Ian Bolton <ian.bolton@arm.com>
18541 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
18542 (*insv_reg<mode>): New define_insn.
18544 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
18546 PR rtl-optimization/57439
18547 * postreload.c (move2add_valid_value_p): Check that we have
18548 a zero subreg_regno_offset when accessing the register in
18549 the requested mode.
18551 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18552 Igor Zamyatin <igor.zamyatin@intel.com>
18554 Silvermont (SLM) architecture pipeline model, tuning and
18556 * config.gcc: Add slm config options and target.
18558 * config/i386/slm.md: New.
18560 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
18562 * config/i386/i386-c.c (ix86_target_macros_internal): New case
18564 (ix86_target_macros_internal): Likewise.
18566 * config/i386/i386.c (slm_cost): New cost.
18567 (m_SLM): New macro flag.
18568 (initial_ix86_tune_features): Set m_SLM.
18569 (x86_accumulate_outgoing_args): Likewise.
18570 (x86_arch_always_fancy_math_387): Likewise.
18571 (processor_target_table): Add slm cost.
18572 (cpu_names): Add slm cpu name.
18573 (x86_option_override_internal): Set SLM ISA.
18574 (ix86_issue_rate): New case PROCESSOR_SLM.
18575 (ia32_multipass_dfa_lookahead): Likewise.
18576 (fold_builtin_cpu): Add slm.
18578 * config/i386/i386.h (TARGET_SLM): New target macro.
18579 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
18580 (processor_type): Add PROCESSOR_SLM.
18582 * config/i386/i386.md (cpu): Add new value "slm".
18583 (slm.md): Include slm.md.
18585 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
18586 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18588 * config/arm/arm-protos.h: Add and update function protos.
18589 * config/arm/arm.c (use_simple_return_p): New added.
18590 (thumb2_expand_return): Check simple_return flag.
18591 * config/arm/arm.md: Add simple_return and conditional simple_return.
18592 * config/arm/iterators.md: Add iterator for return and simple_return.
18594 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18596 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
18597 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
18598 (arm_emit_vfp_multi_reg_pop): Likewise.
18599 (thumb2_emit_ldrd_pop): Likewise.
18600 (arm_expand_epilogue): Add misc REG_CFA notes.
18601 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
18603 2013-05-29 Lawrence Crowl <crowl@google.com>
18605 * config/arm/t-arm: Update for below.
18607 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
18608 Change type to hash_table. Update dependent calls and types.
18610 * config/i386/t-cygming: Update for below.
18612 * config/i386/t-interix: Update for below.
18614 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
18615 Change type to hash_table. Update dependent calls and types.
18616 (i386_find_on_wrapper_list::wrappers): Likewise.
18618 * config/ia64/t-ia64: Update for below.
18620 * config/ia64/ia64.c (bundle_state_table):
18621 Change type to hash_table. Update dependent calls and types.
18623 * config/mips/mips.c (mips_reorg_process_insns::htab):
18624 Change type to hash_table. Update dependent calls and types.
18626 * config/sol2.c (solaris_comdat_htab):
18627 Change type to hash_table. Update dependent calls and types.
18629 * config/t-sol2: Update for above.
18631 2013-05-29 Teresa Johnson <tejohnson@google.com>
18633 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
18634 functions are not yet marked as defined.
18636 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
18637 Pat Haugen <pthaugen@us.ibm.com>
18638 Peter Bergner <bergner@vnet.ibm.com>
18640 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
18644 (vrotl<mode>3): Likewise.
18645 (vashl<mode>3): Likewise.
18646 (vlshr<mode>3): Likewise.
18647 (vashr<mode>3): Likewise.
18649 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18650 support for power8 V2DI builtins.
18652 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
18653 power8 V2DI builtins.
18654 (vupkhsw): Likewise.
18655 (vupklsw): Likewise.
18656 (vaddudm): Likewise.
18657 (vminsd): Likewise.
18658 (vmaxsd): Likewise.
18659 (vminud): Likewise.
18660 (vmaxud): Likewise.
18661 (vpkudum): Likewise.
18662 (vpksdss): Likewise.
18663 (vpkudus): Likewise.
18664 (vpksdus): Likewise.
18669 (vsubudm): Likewise.
18670 (vcmpequd): Likewise.
18671 (vcmpgtsd): Likewise.
18672 (vcmpgtud): Likewise.
18673 (vcmpequd_p): Likewise.
18674 (vcmpgtsd_p): Likewise.
18675 (vcmpgtud_p): Likewise.
18676 (vupkhsw): Likewise.
18677 (vupklsw): Likewise.
18678 (vaddudm): Likewise.
18679 (vmaxsd): Likewise.
18680 (vmaxud): Likewise.
18681 (vminsd): Likewise.
18682 (vminud): Likewise.
18683 (vpksdss): Likewise.
18684 (vpksdus): Likewise.
18685 (vpkudum): Likewise.
18686 (vpkudus): Likewise.
18691 (vsubudm): Likewise.
18693 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
18694 support for power8 V2DI instructions.
18696 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
18697 power8 V2DI instructions. Combine pack and unpack insns to use an
18698 iterator for each mode. Check whether a particular mode supports
18699 Altivec instructions instead of just checking TARGET_ALTIVEC.
18700 (UNSPEC_VPKUWUM): Likewise.
18701 (UNSPEC_VPKSHSS): Likewise.
18702 (UNSPEC_VPKSWSS): Likewise.
18703 (UNSPEC_VPKUHUS): Likewise.
18704 (UNSPEC_VPKSHUS): Likewise.
18705 (UNSPEC_VPKUWUS): Likewise.
18706 (UNSPEC_VPKSWUS): Likewise.
18707 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
18708 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
18709 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
18710 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
18711 (UNSPEC_VUPKHSB): Likewise.
18712 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
18713 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
18714 (UNSPEC_VUPKHSH): Likewise.
18715 (UNSPEC_VUPKLSB): Likewise.
18716 (UNSPEC_VUPKLSH): Likewise.
18718 (VI_char): Likewise.
18719 (VI_scalar): Likewise.
18720 (VI_unit): Likewise.
18722 (VP_small): Likewise.
18723 (VP_small_lc): Likewise.
18724 (VU_char): Likewise.
18725 (add<mode>3): Likewise.
18726 (altivec_vaddcuw): Likewise.
18727 (altivec_vaddu<VI_char>s): Likewise.
18728 (altivec_vadds<VI_char>s): Likewise.
18729 (sub<mode>3): Likewise.
18730 (altivec_vsubcuw): Likewise.
18731 (altivec_vsubu<VI_char>s): Likewise.
18732 (altivec_vsubs<VI_char>s): Likewise.
18733 (altivec_vavgs<VI_char>): Likewise.
18734 (altivec_vcmpbfp): Likewise.
18735 (altivec_eq<mode>): Likewise.
18736 (altivec_gt<mode>): Likewise.
18737 (altivec_gtu<mode>): Likewise.
18738 (umax<mode>3): Likewise.
18739 (smax<mode>3): Likewise.
18740 (umin<mode>3): Likewise.
18741 (smin<mode>3): Likewise.
18742 (altivec_vpkuhum): Likewise.
18743 (altivec_vpkuwum): Likewise.
18744 (altivec_vpkshss): Likewise.
18745 (altivec_vpkswss): Likewise.
18746 (altivec_vpkuhus): Likewise.
18747 (altivec_vpkshus): Likewise.
18748 (altivec_vpkuwus): Likewise.
18749 (altivec_vpkswus): Likewise.
18750 (altivec_vpks<VI_char>ss): Likewise.
18751 (altivec_vpks<VI_char>us): Likewise.
18752 (altivec_vpku<VI_char>us): Likewise.
18753 (altivec_vpku<VI_char>um): Likewise.
18754 (altivec_vrl<VI_char>): Likewise.
18755 (altivec_vsl<VI_char>): Likewise.
18756 (altivec_vsr<VI_char>): Likewise.
18757 (altivec_vsra<VI_char>): Likewise.
18758 (altivec_vsldoi_<mode>): Likewise.
18759 (altivec_vupkhsb): Likewise.
18760 (altivec_vupkhs<VU_char>): Likewise.
18761 (altivec_vupkls<VU_char>): Likewise.
18762 (altivec_vupkhsh): Likewise.
18763 (altivec_vupklsb): Likewise.
18764 (altivec_vupklsh): Likewise.
18765 (altivec_vcmpequ<VI_char>_p): Likewise.
18766 (altivec_vcmpgts<VI_char>_p): Likewise.
18767 (altivec_vcmpgtu<VI_char>_p): Likewise.
18768 (abs<mode>2): Likewise.
18769 (vec_unpacks_hi_v16qi): Likewise.
18770 (vec_unpacks_hi_v8hi): Likewise.
18771 (vec_unpacks_lo_v16qi): Likewise.
18772 (vec_unpacks_hi_<VP_small_lc>): Likewise.
18773 (vec_unpacks_lo_v8hi): Likewise.
18774 (vec_unpacks_lo_<VP_small_lc>): Likewise.
18775 (vec_pack_trunc_v8h): Likewise.
18776 (vec_pack_trunc_v4si): Likewise.
18777 (vec_pack_trunc_<mode>): Likewise.
18779 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
18781 (vec_vmaxsd): Likewise.
18782 (vec_vmaxud): Likewise.
18783 (vec_vminsd): Likewise.
18784 (vec_vminud): Likewise.
18785 (vec_vpksdss): Likewise.
18786 (vec_vpksdus): Likewise.
18787 (vec_vpkudum): Likewise.
18788 (vec_vpkudus): Likewise.
18789 (vec_vrld): Likewise.
18790 (vec_vsld): Likewise.
18791 (vec_vsrad): Likewise.
18792 (vec_vsrd): Likewise.
18793 (vec_vsubudm): Likewise.
18794 (vec_vupkhsw): Likewise.
18795 (vec_vupklsw): Likewise.
18797 2013-05-29 Jan Hubicka <jh@suse.cz>
18799 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
18800 flags; reorder rest of fields in more consistent way.
18801 (varpool_node): Remove analyzed, finalized and alias.
18802 (cgraph_ndoe): Likewise.
18803 (symtab_alias_ultimate_target): New function.
18804 (cgraph_function_node): Move offline.
18805 (cgraph_reset_node): Declare.
18806 (cgraph_comdat_can_be_unshared_p): Remove.
18807 (varpool_remove_initializer): Declare.
18808 (varpool_first_defined_variable, varpool_next_defined_variable
18809 cgraph_first_defined_function, cgraph_next_defined_function): Update.
18810 (cgraph_function_with_gimple_body_p): Update.
18811 (varpool_all_refs_explicit_p): Update.
18812 (symtab_alias_target): New function.
18813 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
18814 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
18815 (cgraph_function_or_thunk_node): Simplify using
18816 symtab_alias_ultimate_target.
18817 (varpool_variable_node): Likewise.
18818 * cgraph.c (cgraph_create_function_alias): Update.
18819 (cgraph_add_thunk): Update.
18820 (cgraph_remove_node): Update.
18821 (dump_cgraph_node): Do not dump removed flags.
18822 (cgraph_function_body_availability): Update.
18823 (cgraph_propagate_frequency): Update.
18824 (verify_cgraph_node): Check sanity of local flag.
18825 (cgraph_function_node): Move here from cgraph.h; revamp for
18826 cgraph_function_or_thunk_node.
18827 * lto-symtab.c (lto_varpool_replace_node): Update.
18828 (lto_symtab_resolve_can_prevail_p): Update.
18829 (lto_symtab_merge_cgraph_nodes): Update.
18830 * ipa-cp.c (determine_versionability, initialize_node_lattices,
18831 propagate_constants_accross_call, devirtualization_time_bonus,
18832 ipcp_propagate_stage): Update.
18833 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
18834 * ipa-inline-transform.c (clone_inlined_nodes,
18835 preserve_function_body_p): Update.
18836 * ipa-reference.c (propagate): Update.
18837 (write_node_summary_p): Update.
18838 * toplev.c (wrapup_global_declaration_2): Update.
18839 * cgraphunit.c (cgraph_analyze_function): Rename to ...
18840 (analyze_function) ... this one.
18841 (cgraph_process_new_functions): Update.
18842 (cgraph_reset_node): Export.
18843 (cgraph_finalize_function): Update.
18844 (cgraph_add_new_function): Update.
18845 (process_function_and_variable_attributes): Update.
18846 (varpool_finalize_decl): Update.
18847 (symbol_finalized): Remove.
18848 (symbol_finalized_and_needed): Rename to ...
18849 (symbol_defined_and_needed): ... update.
18850 (cgraph_analyze_functions): Update.
18851 (handle_alias_pairs): Update.
18852 (mark_functions_to_output): Update.
18853 (assemble_thunk): Update.
18854 (output_in_order): Update.
18855 (output_weakrefs): Update.
18856 (finalize_compilation_unit): Update.
18857 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
18858 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
18859 input_node, input_varpool_node): Update.
18860 * dbxout.c (dbxout_expand_expr): Update.
18861 * cgraphclones.c (cgraph_clone_node): Update.
18862 (cgraph_copy_node_for_versioning): Update.
18863 (cgraph_materialize_clone): Update.
18864 (cgraph_materialize_all_clones): Update.
18865 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
18866 propagate_pure_const, propagate_nothrow): Update.
18867 * lto-streamer-out.c (lto_output, write_symbol): Update.
18868 * ipa-utils.c (ipa_reverse_postorder): Update.
18869 * ipa-inline.c (can_inline_edge_p): Update.
18870 (update_caller_keys, ipa_inline): Update.
18871 * dwarf2out.c (reference_to_unused,
18872 premark_types_used_by_global_vars_helper): Update.
18873 * tree-eh.c (tree_could_trap_p): Update.
18874 * ipa-split.c (consider_split, execute_split_functions): Update.
18875 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
18876 has_addr_references_p): Update; move ahead in file for better
18878 (process_references): Simplify.
18879 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
18880 bodies are removed.
18881 (cgraph_comdat_can_be_unshared_p): Make static.
18882 (cgraph_externally_visible_p): Update.
18883 (varpool_externally_visible_p): Update.
18884 (function_and_variable_visibility): Update.
18885 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
18886 ipa_tm_mark_force_output_node): Update.
18887 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
18888 estimate_edge_devirt_benefit, inline_generate_summary,
18889 inline_write_summary): Update.
18890 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
18891 * ipa-prop.c (ipa_compute_jump_functions): Update.
18892 (ipa_print_node_params, ipa_prop_read_section,
18893 ipa_update_after_lto_read, read_replacements_section): Update.
18894 * varasm.c (mark_decl_referenced): Update.
18895 (assemble_alias, dump_tm_clone_pairs): Update.
18896 * tree-inline.c (copy_bb): Update.
18897 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
18899 * symtab.c (dump_symtab_base): Print new flags.
18900 (verify_symtab_base): Verify new flags.
18901 (symtab_alias_ultimate_target): New function.
18902 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
18903 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
18905 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
18907 * i386.c (ix86_get_function_versions_dispatcher,
18908 ix86_generate_version_dispatcher_body): Update.
18909 (fold_builtin_cpu): Use varpool_add_new_variable.
18910 * varpool.c (varpool_remove_initializer): Break out from ...
18911 (varpool_remove_node): ... this one.
18912 (dump_varpool_node, varpool_node_for_asm,
18913 cgraph_variable_initializer_availability, varpool_analyze_node,
18914 varpool_assemble_decl, varpool_remove_unreferenced_decls,
18915 varpool_finalize_named_section_flags, varpool_create_variable_alias):
18918 2013-05-29 Jan Hubicka <jh@suse.cz>
18920 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
18922 2013-05-29 Easwaran Raman <eraman@google.com>
18924 PR tree-optimization/57442
18925 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
18926 when control exits the main loop.
18928 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
18930 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
18932 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
18933 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
18934 * rx/t-rx: Add rx100 under multi library matches option for nofpu
18937 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18939 PR tree-optimization/57441
18940 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
18941 Don't limit size of incr_vec to number of candidates.
18943 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
18945 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
18946 and mips16 directories.
18947 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
18948 (MULTILIB_DIRNAMES): Ditto.
18949 (MULTILIB_EXCEPTIONS): Add new exceptions.
18950 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
18951 (MULTILIB_DIRNAMES): Ditto.
18952 (MULTILIB_EXCEPTIONS): Add new exceptions.
18954 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18955 Marcus Shawcroft <marcus.shawcroft@arm.com>
18957 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
18958 SYMBOL_TINY_ABSOLUTE.
18959 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
18960 SYMBOL_TINY_ABSOLUTE.
18961 (aarch64_expand_mov_immediate): Likewise.
18962 (aarch64_classify_symbol): Likewise.
18963 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
18964 Permit SYMBOL_TINY_ABSOLUTE.
18965 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
18967 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18968 Marcus Shawcroft <marcus.shawcroft@arm.com>
18970 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
18971 Refactor if/switch. Replace gcc_assert with if.
18973 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
18975 * config/i386/i386.c (initial_ix86_tune_features): Enable
18976 FP Reassociation for AMD bdver1 and bdver2.
18978 2013-05-29 Martin Jambor <mjambor@suse.cz>
18980 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
18981 and IMAGPART_EXPR do not occur within other handled_components.
18983 2013-05-29 Richard Biener <rguenther@suse.de>
18985 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
18986 access on whether the use is in the BB we currently try to
18988 (vect_bb_vectorization_profitable_p): Pass the BB we currently
18989 vectorize to vect_bb_slp_scalar_cost.
18991 2013-05-29 Richard Biener <rguenther@suse.de>
18993 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
18994 computing scalar cost offsetted by stmts that are kept live
18996 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
18997 for computation of scalar cost.
18999 2013-05-28 Steve Ellcey <sellcey@mips.com>
19001 * config/mips/mips-cpus.def (mips32r2): Change processor type.
19003 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
19005 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
19006 array notation built-in reduction functions.
19007 * doc/passes.texi (Passes): Added documentation about changes done
19009 * doc/invoke.texi (C Dialect Options): Added documentation about
19010 the -fcilkplus flag.
19011 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
19012 (BUILTINS_DEF): Depend on cilkplus.def.
19013 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
19014 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
19015 * cilkplus.def: New file.
19017 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
19019 PR rtl-optimization/57439
19020 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
19022 2013-05-28 Easwaran Raman <eraman@google.com>
19024 PR tree-optimization/57337
19025 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
19026 (find_insert_point): Correctly identify the insertion point
19027 when two statements with the same UID is compared.
19029 2013-05-28 Richard Biener <rguenther@suse.de>
19031 PR tree-optimization/56787
19032 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
19033 from the list of data references.
19034 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
19036 (vect_analyze_loop_operations): Likewise.
19037 (vect_transform_loop): Remove clobbers.
19039 2013-05-28 Martin Jambor <mjambor@suse.cz>
19041 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
19042 and REALPART_EXPRs have scalar type.
19044 2013-05-28 Richard Biener <rguenther@suse.de>
19046 PR tree-optimization/57411
19047 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
19049 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
19050 virtual operand propagation.
19052 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19054 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
19055 destination register for bmasksi_vis.
19056 (vector_init_bshuffle): Likewise.
19057 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
19059 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19061 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
19062 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
19063 mode if the instruction isn't available in the original mode.
19064 * config/sparc/sparc.opt (mfix-ut699): New option.
19065 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
19066 (divdf3): Turn into expander.
19067 (divdf3_nofix): New insn.
19068 (divdf3_fix): Likewise.
19069 (divsf3): Disable if -mfix-ut699.
19070 (sqrtdf2): Turn into expander.
19071 (sqrtdf2_nofix): New insn.
19072 (sqrtdf2_fix): Likewise.
19073 (sqrtsf2): Disable if -mfix-ut699.
19075 2013-05-27 Richard Biener <rguenther@suse.de>
19077 PR middle-end/57412
19078 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
19079 block for the new loop.
19081 2013-05-27 Richard Biener <rguenther@suse.de>
19083 PR tree-optimization/57343
19084 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
19085 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
19086 (number_of_iterations_cond): Do not build the folded tree.
19088 2013-05-27 Richard Biener <rguenther@suse.de>
19091 PR middle-end/57381
19092 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19093 OEP_CONSTANT_ADDRESS_OF retained.
19095 PR tree-optimization/57417
19096 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
19097 for unchanged base.
19098 (set_ssa_val_to): Compare addresses using
19099 get_addr_base_and_unit_offset.
19101 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
19103 PR rtl-optimization/56833
19104 * postreload.c (move2add_record_mode): New function.
19105 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
19106 (move2add_use_add2_insn): Use move2add_record_sym_value.
19107 (move2add_use_add3_insn): Likewise.
19108 (reload_cse_move2add): Use move2add_valid_value_p and
19109 move2add_record_mode. Invalidate call-clobbered and REG_INC
19110 affected regs by setting reg_mode to VOIDmode.
19111 (move2add_note_store): Don't pretend the inside of a SUBREG is
19112 the actual destination. Invalidate single/leading registers by
19113 setting reg_mode to VOIDmode.
19114 Use move2add_record_sym_value, move2add_valid_value_p and
19115 move2add_record_mode.
19117 2013-05-27 Richard Biener <rguenther@suse.de>
19119 PR tree-optimization/57396
19120 * tree-affine.c (double_int_constant_multiple_p): Properly
19121 return false for val == 0 and div != 0.
19123 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19125 * config/mips/mips.h: Use #elif in preprocessor conditions.
19127 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19130 * config/mips/constraints.md (kl): New constraint.
19131 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
19132 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
19133 constraint for operand 0. Split after CSE for MIPS16. Emit a move
19134 from LO for MIPS16.
19135 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
19137 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19140 * config/mips/mips.c (mips_can_inline_p): New function.
19141 (TARGET_CAN_INLINE_P): Define.
19143 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
19145 * sched-int.h (ds_t, dw_t): Make unsigned int.
19146 Fix documentation that describes how all the ds_t bits are used.
19147 Reserve the last bit for delayed-branch scheduling.
19148 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
19149 (BITS_PER_DEP_WEAK): Fix definition and documentation.
19150 (gen_dep_weak_1): Remove prototype.
19151 * sched-deps.c (get_dep_weak_1): Make static.
19152 * target.def (speculate_insn, needs_block_p, gen_spec_check,
19153 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
19154 * doc/tm.texi: Regenerate.
19155 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
19157 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
19160 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
19162 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
19163 Sandra Loosemore <sandra@codesourcery.com>
19165 * config.gcc (powerpc-*): Allow native for with-cpu.
19167 2013-05-24 Jeff Law <law@redhat.com>
19169 PR tree-optimization/57124
19170 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
19171 conversion feeding a condition if the range has an overflow
19172 if -fstrict-overflow. Add warnings for when we do make the
19175 2013-05-24 Dehao Chen <dehao@google.com>
19177 * tree-cfg.c (locus_discrim_map): Fix the typo.
19178 (locus_discrim_hasher): Likewise.
19179 (locus_discrim_hasher::hash): Likewise.
19180 (locus_discrim_hasher::equal): Likewise.
19182 2013-05-24 Martin Jambor <mjambor@suse.cz>
19184 PR tree-optimization/57294
19185 * cgraph.h (ipa_record_stmt_references): Declare.
19186 * cgraphbuild.c (ipa_record_stmt_references): New function.
19187 (build_cgraph_edges): Use ipa_record_stmt_references.
19188 (rebuild_cgraph_edges): Likewise.
19189 (cgraph_rebuild_references): Likewise.
19190 * ipa-prop.c (ipa_modify_call_arguments): Discard references
19191 associated with the old statement and build references from the
19192 newly built statements.
19193 * ipa-ref.c (ipa_remove_stmt_references): New function.
19194 * ipa-ref.h (ipa_remove_stmt_references): Declare.
19196 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
19198 * lra-constraints.c (emit_spill_move): Use smaller mode for
19200 (check_and_process_move): Consider mem-reg moves for secondary
19202 (curr_insn_transform): Don't lose insns emitted before for
19203 secondary memory moves.
19204 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
19205 reg set up in the current insn.
19207 2013-05-24 Dehao Chen <dehao@google.com>
19209 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
19211 (locus_descrim_hasher::equal): Likewise.
19212 (build_gimple_cfg): New discriminator assignment algorithm.
19213 (make_edges): Likewise.
19214 (next_discriminator_for_locus): Likewise.
19215 (same_line_p): Likewise.
19216 (assign_discriminators): Likewise.
19217 (make_cond_expr_edges): Likewise.
19218 (make_gimple_switch_edges): Likewise.
19219 (make_goto_expr_edges): Likewise.
19220 (make_gimple_asm_edges): Likewise.
19222 2013-05-24 Ian Bolton <ian.bolton@arm.com>
19224 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
19225 X format specifier to only display bottom 16 bits.
19226 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
19227 immediate to match for operand 2, since it will be masked.
19229 2013-05-24 Richard Biener <rguenther@suse.de>
19231 PR tree-optimization/57287
19232 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
19233 all SSA names that occur in abnormal PHIs.
19235 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
19238 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
19239 that index is not negative.
19241 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
19243 PR rtl-optimization/55177
19244 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
19245 (simplify_byte_swapping_operation): New.
19246 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
19247 (simplify_relational_operation_1): Deal with BSWAP.
19249 2013-05-23 Richard Henderson <rth@redhat.com>
19252 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
19253 (ix86_reorg): Call it.
19255 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
19258 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
19259 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
19260 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
19262 2013-05-23 Christian Bruel <christian.bruel@st.com>
19265 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
19267 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19268 Marcus Shawcroft <marcus.shawcroft@arm.com>
19270 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
19271 * config/aarch64/constraints.md (Usa): Remove.
19272 * doc/md.texi (AArch64 Usa): Remove.
19274 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19275 Marcus Shawcroft <marcus.shawcroft@arm.com>
19277 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
19278 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
19279 * config/aarch64/predicates.md (aarch64_const_address): Remove.
19280 (aarch64_mov_operand): Use aarch64_mov_operand_p.
19282 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
19284 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
19285 instruction (AdvSIMD).
19286 * config/aarch64/aarch64-builtins.c
19287 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
19288 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
19290 2013-05-23 Martin Jambor <mjambor@suse.cz>
19292 PR middle-end/57347
19293 * tree.h (contains_bitfld_component_ref_p): Declare.
19294 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
19295 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
19297 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
19298 not access a bit-field. Assert all final offsets are byte-aligned.
19300 2013-05-23 Richard Biener <rguenther@suse.de>
19302 PR tree-optimization/57380
19303 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
19304 least one invariant or re-used load.
19305 * passes.c (init_optimization_passes): Move pass_phiprop before
19308 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
19310 * config/aarch64/aarch64-simd.md
19311 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
19313 2013-05-23 Richard Biener <rguenther@suse.de>
19315 PR middle-end/57381
19316 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19317 OEP_CONSTANT_ADDRESS_OF retained.
19319 2013-05-23 Jakub Jelinek <jakub@redhat.com>
19321 PR middle-end/57344
19322 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
19323 don't lower unit. Handle unit not being always BITS_PER_WORD.
19325 2013-05-23 Richard Biener <rguenther@suse.de>
19327 PR rtl-optimization/57341
19328 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
19329 instead of true_dependence.
19331 2013-05-22 David Malcolm <dmalcolm@redhat.com>
19333 * bb-reorder.c (branch_threshold): Make const.
19334 (exec_threshold): Ditto.
19336 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19337 Pat Haugen <pthaugen@us.ibm.com>
19338 Peter Bergner <bergner@vnet.ibm.com>
19340 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
19341 documentation for the power8 crypto builtins.
19343 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
19345 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
19346 macros for defining power8 builtin functions.
19347 (BU_P8V_AV_2): Likewise.
19348 (BU_P8V_AV_P): Likewise.
19349 (BU_P8V_VSX_1): Likewise.
19350 (BU_P8V_OVERLOAD_1): Likewise.
19351 (BU_P8V_OVERLOAD_2): Likewise.
19352 (BU_CRYPTO_1): Likewise.
19353 (BU_CRYPTO_2): Likewise.
19354 (BU_CRYPTO_3): Likewise.
19355 (BU_CRYPTO_OVERLOAD_1): Likewise.
19356 (BU_CRYPTO_OVERLOAD_2): Likewise.
19357 (XSCVSPDP): Fix typo, point to the correct instruction.
19358 (VCIPHER): Add power8 crypto builtins.
19359 (VCIPHERLAST): Likewise.
19360 (VNCIPHER): Likewise.
19361 (VNCIPHERLAST): Likewise.
19362 (VPMSUMB): Likewise.
19363 (VPMSUMH): Likewise.
19364 (VPMSUMW): Likewise.
19365 (VPERMXOR_V2DI): Likewise.
19366 (VPERMXOR_V4SI: Likewise.
19367 (VPERMXOR_V8HI: Likewise.
19368 (VPERMXOR_V16QI: Likewise.
19369 (VSHASIGMAW): Likewise.
19370 (VSHASIGMAD): Likewise.
19371 (VPMSUM): Likewise.
19372 (VPERMXOR): Likewise.
19373 (VSHASIGMA): Likewise.
19375 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
19376 __CRYPTO__ if the crypto instructions are available.
19377 (altivec_overloaded_builtins): Add support for overloaded power8
19380 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
19381 support for power8 crypto builtins.
19382 (builtin_function_type): Likewise.
19383 (altivec_init_builtins): Add support for builtins that take vector
19384 long long (V2DI) arguments.
19386 * config/rs6000/crypto.md: New file, define power8 crypto
19389 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19390 Pat Haugen <pthaugen@us.ibm.com>
19391 Peter Bergner <bergner@vnet.ibm.com>
19393 * doc/invoke.texi (Option Summary): Add power8 options.
19394 (RS/6000 and PowerPC Options): Likewise.
19396 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
19397 constraints.md instead of rs6000.h. Reorder w* constraints. Add
19398 wm, wn, wr documentation.
19400 * config/rs6000/constraints.md (wm): New constraint for VSX
19401 registers if direct move instructions are enabled.
19402 (wn): New constraint for no registers.
19403 (wq): New constraint for quad word even GPR registers.
19404 (wr): New constraint if 64-bit instructions are enabled.
19405 (wv): New constraint if power8 vector instructions are enabled.
19406 (wQ): New constraint for quad word memory locations.
19408 * config/rs6000/predicates.md (const_0_to_15_operand): New
19409 constraint for 0..15 for crypto instructions.
19410 (gpc_reg_operand): If VSX allow registers in VSX registers as well
19411 as GPR and floating point registers.
19412 (int_reg_operand): New predicate to match only GPR registers.
19413 (base_reg_operand): New predicate to match base registers.
19414 (quad_int_reg_operand): New predicate to match even GPR registers
19415 for quad memory operations.
19416 (vsx_reg_or_cint_operand): New predicate to allow vector logical
19417 operations in both GPR and VSX registers.
19418 (quad_memory_operand): New predicate for quad memory operations.
19419 (reg_or_indexed_operand): New predicate for direct move support.
19421 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
19422 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
19423 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
19424 (POWERPC_MASKS): Add power8 options.
19425 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
19428 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
19429 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
19431 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
19432 (-mpower8-fusion): New power8 options.
19433 (-mpower8-fusion-sign): Likewise.
19434 (-mpower8-vector): Likewise.
19435 (-mcrypto): Likewise.
19436 (-mdirect-move): Likewise.
19437 (-mquad-memory): Likewise.
19439 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
19440 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
19442 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
19443 (rs6000_debug_vector_unit): Add p8_vector.
19444 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
19445 definitions. Also print fusion state.
19446 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
19447 (rs6000_builtin_mask_calculate): Add power8 builtin support.
19448 (rs6000_option_override_internal): Add support for power8.
19449 (rs6000_common_init_builtins): Add debugging for skipped builtins
19450 if -mdebug=builtin.
19451 (rs6000_adjust_cost): Add power8 support.
19452 (rs6000_issue_rate): Likewise.
19453 (insn_must_be_first_in_group): Likewise.
19454 (insn_must_be_last_in_group): Likewise.
19455 (force_new_group): Likewise.
19456 (rs6000_register_move_cost): Likewise.
19457 (rs6000_opt_masks): Likewise.
19459 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
19460 power8 capable assembler, default to power7 options.
19461 (TARGET_DIRECT_MOVE): Likewise.
19462 (TARGET_CRYPTO): Likewise.
19463 (TARGET_P8_VECTOR): Likewise.
19464 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
19465 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
19466 (VECTOR_MEM_P8_VECTOR_P): Likewise.
19467 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
19468 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
19469 (TARGET_XSCVDPSPN): Likewise.
19470 (TARGET_XSCVSPDPN): Likewsie.
19471 (TARGET_SYNC_HI_QI): Likewise.
19472 (TARGET_SYNC_TI): Likewise.
19473 (MASK_CRYPTO): Likewise.
19474 (MASK_DIRECT_MOVE): Likewise.
19475 (MASK_P8_FUSION): Likewise.
19476 (MASK_P8_VECTOR): Likewise.
19477 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
19478 temporary used by some of the direct move instructions to get two FP
19479 temporary registers does not force creation of a stack frame.
19480 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
19481 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
19482 that any VSX registers are tieable, even if they are also an
19483 Altivec vector mode.
19484 (r6000_reg_class_enum): Add wm, wr, wv constraints.
19485 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
19486 (RS6000_BTM_CRYPTO): Likewise.
19487 (RS6000_BTM_COMMON): Likewise.
19489 * config/rs6000/rs6000.md (cpu attribute): Add power8.
19490 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
19491 (enum rs6000_vector): Add power8 vector support.
19493 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19497 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
19498 (any_sibcall_could_use_r3): this and handle indirect calls.
19499 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
19501 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19503 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
19505 2013-05-22 Richard Biener <rguenther@suse.de>
19507 PR middle-end/57349
19508 * profile.c (branch_prob): Do not split blocks that are
19509 abnormally receiving from ECF_RETURNS_TWICE functions.
19511 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19513 * recog.c (offsettable_address_addr_space_p): Fix calculation of
19514 address mode. Move pointer mode initialization to the same place.
19516 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
19518 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
19519 while it has any effect.
19521 2013-05-21 Easwaran Raman <eraman@google.com>
19523 PR tree-optimization/57322
19524 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
19525 UID of the statement added to the BB to be 1.
19527 2013-05-21 Jakub Jelinek <jakub@redhat.com>
19529 PR tree-optimization/57331
19530 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
19531 of conversion from pointer type to integral type with integer.
19533 2013-05-21 Martin Jambor <mjambor@suse.cz>
19536 * ipa-prop.c (ipa_read_node_info): Process param_used and
19537 controlled_uses in the same order as when writing.
19539 2013-05-21 Magnus Granberg <baldrick@free.fr>
19542 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
19544 2013-05-21 Richard Biener <rguenther@suse.de>
19546 PR tree-optimization/57318
19547 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
19548 estimate stmts with side-effects as likely eliminated.
19550 2013-05-21 Richard Biener <rguenther@suse.de>
19552 PR tree-optimization/57330
19553 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
19554 preserve the call stmts fntype.
19556 2013-05-21 Richard Biener <rguenther@suse.de>
19558 PR tree-optimization/57303
19559 * tree-ssa-sink.c (statement_sink_location): Improve killing
19560 stmt detection and properly handle self-assignments.
19562 2013-05-21 Christian Bruel <christian.bruel@st.com>
19564 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
19565 spanning registers. LEAF_REG_REMAP is supported only for contiguous
19566 registers. Set register size out of the PARALLEL loop.
19568 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
19571 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
19572 (*fmasf4, *fmasf4_media): New insns.
19574 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19576 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
19577 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
19578 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
19579 (mips_idiv_insns): Update the comments to say that the returned
19580 instruction counts are in units of BASE_INSN_LENGTH.
19581 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
19582 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
19583 using 2 rather than 4 as the length of indirect MIPS16 and
19584 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
19585 length of a NOP. Don't divide MIPS16 lengths by 2.
19586 (mips16_split_long_branches): Assume a branch is long if the
19587 length is greater than 4 rather than 8.
19588 * config/mips/mips.md (length): Give MIPS16 lengths directly,
19589 rather than multiplying them by 2. Multiply instruction counts
19590 by BASE_INSN_LENGTH rather than 4.
19591 (*jump_mips16, tls_get_tp_mips16_<mode>)
19592 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
19594 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19596 * config/mips/mips.md (extended_mips16): Remove branch case.
19597 (length): Remove duplicated extended_mips16 test.
19599 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19601 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
19603 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
19605 * recog.h (Recog_data): Rename to...
19606 (recog_data_d): ...this.
19607 (recog_data): Update accordingly.
19608 * recog.c (recog_data): Likewise.
19609 * reload.c (save_recog_data): Likewise.
19610 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
19611 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
19613 2013-05-17 Julian Brown <julian@codesourcery.com>
19615 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
19616 found in a REG_EQUAL note, invalidate it.
19618 2013-05-17 Easwaran Raman <eraman@google.com>
19620 * tree-ssa-reassoc.c (find_insert_point): New function.
19621 (insert_stmt_after): Likewise.
19622 (get_def_stmt): Likewise.
19623 (ensure_ops_are_available): Likewise.
19624 (not_dominated_by): Likewise.
19625 (rewrite_expr_tree): Do not move statements beyond what is
19626 necessary. Remove call to swap_ops_for_binary_stmt...
19627 (reassociate_bb): ... and move it here.
19628 (build_and_add_sum): Assign UIDs for new statements.
19629 (linearize_expr): Likewise.
19630 (do_reassoc): Renumber gimple statement UIDs.
19632 2013-05-17 Jan Hubicka <jh@suse.cz>
19634 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
19636 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
19637 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
19639 (output_weakrefs): Update.
19641 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
19642 Martin Jambor <mjambor@suse.cz>
19644 PR middle-end/57276
19645 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
19646 value that corresponds to the given aggval is found in values vector.
19648 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
19650 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
19651 sse, sse2, sse3, ssse3 and sse4a flags to options.
19653 2013-05-17 David Malcolm <dmalcolm@redhat.com>
19655 * gengtype-state.c: (s_expr_writer): New class, to handle
19656 prettifying of output layout of s-expressions.
19657 (state_writer): New class, to write out gtype.state.
19658 (state_written_type_count): Move this variable into member data of
19660 (s_expr_writer::s_expr_writer): New code: constructor for new class
19661 (state_writer::state_writer(): ditto
19662 (s_expr_writer::write_new_line): New function
19663 (s_expr_writer::write_any_indent): ditto
19664 (s_expr_writer::begin_s_expr): ditto
19665 (s_expr_writer::end_s_expr): ditto
19666 (write_state_fileloc): convert to method of state_writer...
19667 (state_writer:: write_state_fileloc): ...and use methods of
19668 s_expr_writer to write indentation into the gtype.state output file
19669 to visually represent the hierarchical structure of the list
19671 (write_state_fields): ditto, renaming to...
19672 (state_writer::write_state_fields)
19673 (write_state_a_string): ditto, renaming to...
19674 (state_writer::write_state_a_string)
19675 (write_state_string_option): ditto, renaming to...
19676 (state_writer::write_state_string_option)
19677 (write_state_type_option): ditto, renaming to...
19678 (state_writer::write_state_type_option)
19679 (write_state_nested_option): ditto, renaming to...
19680 (state_writer::write_state_nested_option)
19681 (write_state_option): ditto, renaming to...
19682 (state_writer::write_state_option)
19683 (write_state_options): ditto, renaming to...
19684 (state_writer::write_state_options)
19685 (write_state_lang_bitmap): ditto, renaming to...
19686 (state_writer::write_state_lang_bitmap)
19687 (write_state_version): ditto, renaming to...
19688 (state_writer::write_state_version)
19689 (write_state_scalar_type): ditto, renaming to...
19690 (state_writer::write_state_scalar_type)
19691 (write_state_string_type): ditto, renaming to...
19692 (state_writer::write_state_string_type)
19693 (write_state_undefined_type): ditto, renaming to...
19694 (state_writer::write_state_undefined_type)
19695 (write_state_struct_union_type): ditto, renaming to...
19696 (state_writer::write_state_struct_union_type)
19697 (write_state_struct_type): ditto, renaming to...
19698 (state_writer::write_state_struct_type)
19699 (write_state_user_struct_type): ditto, renaming to...
19700 (state_writer::write_state_user_struct_type)
19701 (write_state_lang_struct_type): ditto, renaming to...
19702 (state_writer::write_state_lang_struct_type)
19703 (write_state_param_struct_type): ditto, renaming to...
19704 (state_writer::write_state_param_struct_type)
19705 (write_state_pointer_type): ditto, renaming to...
19706 (state_writer::write_state_pointer_type)
19707 (write_state_array_type): ditto, renaming to...
19708 (state_writer::write_state_array_type)
19709 (write_state_gc_used): ditto, renaming to...
19710 (state_writer::write_state_gc_used)
19711 (write_state_common_type_content): ditto, renaming to...
19712 (state_writer::write_state_common_type_content)
19713 (write_state_type): ditto, renaming to...
19714 (state_writer::write_state_type)
19715 (write_state_pair_list): ditto, renaming to...
19716 (state_writer::write_state_pair_list)
19717 (write_state_pair): ditto, renaming to...
19718 (state_writer::write_state_pair)
19719 (write_state_typedefs): ditto, renaming to...
19720 (state_writer::write_state_typedefs)
19721 (write_state_structures): ditto, renaming to...
19722 (state_writer::write_state_structures)
19723 (write_state_param_structs): ditto, renaming to...
19724 (state_writer::write_state_param_structs)
19725 (write_state_variables): ditto, renaming to...
19726 (state_writer::write_state_variables)
19727 (write_state_srcdir): ditto, renaming to...
19728 (state_writer::write_state_srcdir)
19729 (write_state_files_list): ditto, renaming to...
19730 (state_writer::write_state_files_list)
19731 (write_state_languages): ditto, renaming to...
19732 (state_writer::write_state_languages)
19733 (write_state): create a state_writer instance and use it when
19734 writing out the state file
19736 2013-05-17 Mike Stump <mikestump@comcast.net>
19738 PR rtl-optimization/57304
19739 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
19740 accessing DF_REF_REAL_LOC.
19742 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19744 PR rtl-optimization/57281
19745 PR rtl-optimization/57300
19746 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
19747 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
19748 what the other splitter did if the registers are dead.
19750 2013-05-17 Richard Biener <rguenther@suse.de>
19752 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
19755 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19757 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
19760 2013-05-17 Marek Polacek <polacek@redhat.com>
19762 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
19763 length when doing non-zero store of storing '\0' to '\0'.
19765 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19767 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
19768 vect_external_def oprnd1 with loop_vinfo, try to emit
19769 optional cast, negation and and stmts on the loop preheader
19770 edge instead of into the pattern def seq.
19772 PR tree-optimization/57051
19773 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
19774 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
19776 2013-05-16 Nick Clifton <nickc@redhat.com>
19778 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
19779 (rl78_is_naked_func): New function.
19780 (rl78_expand_prologue): Skip prologue generation for naked functions.
19781 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
19782 * doc/extend.texi (naked): Add RL78 to the list of processors
19783 that supports this attribute.
19785 2013-05-16 Jeff Law <law@redhat.com>
19787 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
19789 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19791 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
19792 cache parameters using detect_caches_amd also for CYRIX,
19793 NSC and TM2 signatures.
19795 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19796 Dzianis Kahanovich <mahatma@eu.by>
19800 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
19801 VIA/Centaur processors and determine their cache parameters
19802 using detect_caches_amd.
19804 2013-05-16 Teresa Johnson <tejohnson@google.com>
19806 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
19807 (rtl_verify_edges): New function.
19808 (rtl_verify_bb_insns): Ditto.
19809 (rtl_verify_bb_pointers): Ditto.
19810 (rtl_verify_bb_insn_chain): Ditto.
19811 (rtl_verify_fallthru): Ditto.
19812 (rtl_verify_bb_layout): Ditto.
19813 (rtl_verify_flow_info_1): Outline checks into new functions.
19814 (rtl_verify_flow_info): Ditto.
19816 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
19818 * cfghooks.c (copy_bbs): Add update_dominance argument.
19819 * cfghooks.h (copy_bbs): Update prototype.
19820 * tree-cfg.c (gimple_duplicate_sese_region):
19821 Add update_dominance argument.
19822 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
19823 * tree-ssa-loop-ch.c (copy_loop_headers): Update
19824 gimple_duplicate_sese_region call.
19825 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
19826 Update copy_bbs call.
19827 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
19828 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
19830 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19832 * tree-vectorizer.h (NUM_PATTERNS): Increment.
19833 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
19834 vect_recog_rotate_pattern.
19835 (vect_recog_rotate_pattern): New function.
19837 2013-05-16 Jason Merrill <jason@redhat.com>
19839 * Makefile.in (LLINKER): New variable.
19840 (mostlyclean): Remove link mutex.
19841 * configure.ac: Handle --enable-link-mutex.
19842 * lock-and-run.sh: New script.
19844 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19847 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
19850 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19852 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
19854 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
19856 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
19857 * config/arm/arm.c (next_consecutive_mem): New function.
19858 (gen_movmem_ldrd_strd): Likewise.
19859 * config/arm/arm.md (movmemqi): Update condition and code.
19860 (unaligned_loaddi, unaligned_storedi): New patterns.
19862 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19864 * config.gcc: Obsolete *-*-solaris2.9*.
19865 * doc/install.texi (Specific, *-*-solaris2*): Document it.
19867 2013-05-16 Richard Biener <rguenther@suse.de>
19869 * passes.c (init_optimization_passes): Move pass_parallelize_loops
19870 earlier, after GRAPHITE transforms and IV canonicalization.
19872 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19874 * omp-low.c (extract_omp_for_data): For collapsed loops,
19875 if at least one of the loops is known at compile time to
19876 iterate zero times, set count to 0.
19877 (expand_omp_regimplify_p): New function.
19878 (expand_omp_for_generic): For collapsed loops, if at least
19879 one of the loops isn't known to iterate at least once,
19880 add runtime check with setting count to 0.
19881 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
19882 For unsigned types if it isn't known at compile time that
19883 the loop will iterate at least once, add runtime check to bypass
19884 the whole loop if initial condition isn't true.
19886 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
19888 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
19890 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
19892 PR middle-end/57286
19893 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
19894 transformations to avoid an infinite loop.
19896 2013-05-16 Marek Polacek <polacek@redhat.com>
19898 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
19900 2013-05-15 Leif Ekblad <leif@rdos.net>
19902 * config/i386/i386.c (ix86_decompose_address): Use
19903 DEFAULT_TLS_SEG_REG to access TLS segment register.
19904 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
19905 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
19906 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
19908 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
19911 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
19912 sibling calls to functions that would normally be lazily bound,
19913 unless $gp is call-clobbered.
19915 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
19917 * config/i386/i386.c (ix86_option_override_internal): Update
19918 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
19919 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
19920 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
19921 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
19923 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
19925 2013-05-15 Andreas Schwab <schwab@suse.de>
19927 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
19928 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
19931 2013-05-15 Teresa Johnson <tejohnson@google.com>
19933 * loop-unroll.c (report_unroll_peel): Check decision before
19934 emitting unroll/peel message.
19936 2013-05-15 Teresa Johnson <tejohnson@google.com>
19938 * function.h (has_bb_partition): New rtl_data flag.
19939 (bb_reorder_complete): Ditto.
19940 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
19941 instead of flag_reorder_blocks_and_partition.
19942 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
19943 with some enhancements.
19944 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
19945 * bb-reorder.c (connect_traces): Check for has_bb_partition
19946 instead of flag_reorder_blocks_and_partition.
19947 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
19948 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
19949 verify_hot_cold_block_grouping.
19950 (partition_hot_cold_basic_blocks): Set has_bb_partition.
19952 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19955 * config/arm/predicates.md (call_insn_operand): New predicate.
19956 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
19957 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
19958 if insn is not a tail call.
19959 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
19961 * config/arm/arm.h (enum reg_class): New caller save register class.
19962 (REG_CLASS_NAMES): Likewise.
19963 (REG_CLASS_CONTENTS): Likewise.
19964 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
19967 2013-05-15 Richard Biener <rguenther@suse.de>
19969 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
19970 of MSG_OPTIMIZED_LOCATIONS.
19971 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
19972 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
19974 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
19975 of MSG_OPTIMIZED_LOCATIONS.
19976 (execute_vect_slp): Likewise.
19977 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
19978 (vect_create_cond_for_alias_checks): Likewise.
19979 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
19980 (vect_recog_widen_mult_pattern): Likewise.
19981 (vect_recog_widen_sum_pattern): Likewise.
19982 (vect_recog_over_widening_pattern): Likewise.
19983 (vect_recog_widen_shift_pattern): Likewise.
19984 (vect_recog_vector_vector_shift_pattern): Likewise.
19985 (vect_recog_divmod_pattern): Likewise.
19986 (vect_recog_mixed_size_cond_pattern): Likewise.
19987 (vect_recog_bool_pattern): Likewise.
19988 (vect_pattern_recog_1): Likewise.
19990 2013-05-15 Martin Jambor <mjambor@suse.cz>
19992 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
19993 non-functions to builtin_unreachable.
19994 * ipa-inline-transform.c (inline_call): Do not assert estimates were
19995 correct when new direct edges were discovered.
19997 2013-05-15 Martin Jambor <mjambor@suse.cz>
19999 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
20000 header, print symbol order instead of node uid, print more information
20001 about indirect edge targets.
20002 (ipa_make_edge_direct_to_target): Print symbol order instead of node
20004 (ipa_make_edge_direct_to_target): Likewise.
20005 (remove_described_reference): Likewise.
20006 (propagate_controlled_uses): Likewise.
20007 (ipa_print_node_params): Also print symbol order.
20008 (ipcp_transform_function): Print symbol order instead of node uids.
20009 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
20010 (cgraph_get_create_real_symbol_node): Likewise.
20011 * ipa-cp.c (print_lattice): Likewise.
20012 (print_all_lattices): Likewise.
20013 (determine_versionability): Likewise.
20014 (initialize_node_lattices): Likewise.
20015 (estimate_local_effects): Likewise.
20016 (update_profiling_info): Likewise.
20017 (create_specialized_node): Likewise.
20018 (perhaps_add_new_callers): Likewise.
20019 (decide_about_value): Likewise.
20020 (decide_whether_version_node): Likewise.
20021 (identify_dead_nodes): Likewise.
20022 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
20023 (dump_inline_summary): Likewise.
20024 (estimate_node_size_and_time): Likewise.
20025 (inline_analyze_function): Likewise.
20026 * ipa-inline.c (report_inline_failed_reason): Likewise.
20027 (want_early_inline_function_p): Likewise.
20028 (edge_badness): Likewise.
20029 (update_edge_key): Likewise.
20030 (inline_small_functions): Likewise. Add dumping of order to two other
20032 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
20033 instead of node uids.
20034 (propagate_pure_const): Likewise.
20035 (propagate_pure_const): Likewise.
20036 * ipa-utils.c (dump_cgraph_node_set): Likewise.
20037 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
20038 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
20040 * tree-pretty-print.c (dump_function_header): Likewise.
20041 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
20042 Print symbol order instead of node uids.
20044 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20046 * config/s390/s390.c (s390_register_move_cost): Don't impose the
20047 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
20049 2013-05-15 Richard Biener <rguenther@suse.de>
20051 PR tree-optimization/57275
20052 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
20053 return value for fail to do runtime alias checks for gather loads.
20055 2013-05-15 Jan Hubicka <jh@suse.cz>
20059 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
20060 weakrefs are not external.
20061 (lto_symtab_merge_decls): Fix thinko when dealing with
20062 non-lto_symtab decls.
20063 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
20064 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
20065 * varpool.c (dump_varpool_node): Dump more flags.
20067 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20069 * config/i386/i386.c (processor_alias_table): Add instruction
20070 FSGSBASE for AMD bdver3 architecture.
20072 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20074 * tree.c (warn_deprecated_use): Print file:line using locus color.
20075 * diagnostic.c (diagnostic_report_current_module): Print file:line
20076 and file:line:column using locus color.
20078 2013-05-14 Mike Stump <mikestump@comcast.net>
20080 * gdbinit.in: Add __null.
20082 2013-05-14 Mike Stump <mikestump@comcast.net>
20084 * recog.h: Rename struct recog_data to Recog_data.
20085 * recog.c: Likewise.
20086 * reload.c (can_reload_into): Likewise.
20087 * config/picochip/picochip.c: Likewise.
20089 2013-05-14 Mike Stump <mikestump@comcast.net>
20091 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
20093 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
20095 * resource.h (struct resources): Remove unch_memory member.
20096 (CLEAR_RESOURCE): Don't clear unch_memory.
20097 * resource.c (mark_referenced_resources): Don't set it.
20098 (mark_set_resources): Likewise.
20099 (mark_target_live_regs): Don't clear it.
20100 (init_resource_info): Likewise.
20101 * reorg.c (resource_conflicts_p): Don't compare it.
20102 (redundant_insn): Don't set it.
20104 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
20106 * emit-rtl.c (next_label): Remove unused function.
20107 (skip_consecutive_labels, link_cc0_insns): Move to ...
20108 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
20109 only place where these functions are used, and make them static.
20111 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20113 * fold-const.c (fold_negate_expr): Handle vectors.
20114 (fold_truth_not_expr): Make it static.
20115 (fold_invert_truthvalue): New static function.
20116 (invert_truthvalue_loc): Handle vectors. Do not call
20117 fold_truth_not_expr directly.
20118 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
20119 <TRUTH_NOT_EXPR>: Do not cast to boolean.
20120 (fold_comparison): Handle vector constants.
20121 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
20122 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
20123 * tree.h (fold_truth_not_expr): Remove declaration.
20125 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
20127 * config/aarch64/aarch64-simd.md
20128 (aarch64_vcond_internal<mode>): Rename to...
20129 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
20130 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
20131 float modes. Clarify all iterator modes.
20132 (vcond<mode><mode>): Use new name for vcond expanders.
20133 (vcond<v_cmp_result><mode>): Likewise.
20134 (vcondu<mode><mode>: Likewise.
20135 * config/aarch64/iterators.md (VDQF_COND): New.
20137 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20140 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
20141 variable for the shift amount. Check that we shift by non-negative
20144 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
20147 * config/arm/arm.h (EPILOGUE_USES): Only return true
20148 for LR_REGNUM after epilogue_completed.
20150 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
20152 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
20153 is error_mark_node.
20155 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20158 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
20159 and Solaris 11+/x86 with gld.
20160 * configure: Regenerate.
20162 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20164 * expmed.c (expand_shift_1): Canonicalize rotates by
20165 constant bitsize / 2 to bitsize - 1.
20166 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
20167 case ROTATERT>: Likewise.
20170 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20172 * config/i386/i386.md (rotateinv): New code attr.
20173 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20174 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20175 roll $31, %eax, etc.
20177 2013-05-14 Richard Biener <rguenther@suse.de>
20179 PR middle-end/57235
20180 * tree-eh.c (sink_clobbers): Give up for successors with
20181 multiple predecessors and no virtual uses.
20183 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
20185 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
20186 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
20188 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20190 PR middle-end/57251
20191 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
20192 the case when both op0 and op1 have VOIDmode.
20194 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20196 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
20197 in multiply-accumulate mode.
20199 2013-05-13 Guozhi Wei <carrot@google.com>
20201 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
20203 2013-05-13 Kai Tietz <ktietz@redhat.com>
20206 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
20207 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
20208 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
20209 * config/i386/i386.c (ix86_option_override_internal): Likewise.
20210 (ix86_expand_prologue): Likewise.
20211 (ix86_expand_split_stack_prologue): Likewise.
20212 (legitimate_pic_address_disp_p): Likewise.
20213 (legitimize_pic_address): Likewise.
20214 (legitimize_tls_address): Likewise.
20215 (legitimize_pe_coff_symbol): Likewise.
20216 (output_pic_addr_const): Likewise.
20217 (construct_plt_address): Likewise.
20218 (ix86_expand_call): Likewise.
20219 (x86_output_mi_thunk): Likewise.
20220 (x86_function_profiler): Likewise.
20222 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
20224 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
20225 similar switch cases.
20226 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
20227 (aarch64_simd_mov_to_<mode>low): Delete.
20228 (aarch64_simd_mov_to_<mode>high): Delete.
20229 (move_lo_quad_<mode>): Add w<-r alternative.
20230 (aarch64_simd_move_hi_quad_<mode>): Likewise.
20231 (aarch64_simd_mov_from_*): Update type attribute.
20232 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
20235 2013-05-13 Jan Hubicka <jh@suse.cz>
20237 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
20238 * config/i386/i386.c (ix86_compute_frame_layout,
20239 ix86_expand_epilogue, emit_i387_cw_initialization,
20240 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
20241 ix86_local_alignment): Fix use of size/speed predicates.
20243 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20245 PR tree-optimization/45216
20246 PR tree-optimization/57157
20247 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
20248 the (-Y) & (B - 1) variant if OP is |.
20249 * expmed.c (expand_shift_1): For rotations by const0_rtx just
20250 return shifted. Use (-op1) & (prec - 1) as other_amount
20251 instead of prec - op1.
20253 2013-05-13 Martin Jambor <mjambor@suse.cz>
20255 PR middle-end/42371
20256 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
20257 (ipa_constant_data): New type.
20258 (ipa_jump_func): Use ipa_constant_data to hold information about
20259 constant jump functions.
20260 (ipa_get_jf_constant): Adjust to jump function type changes.
20261 (ipa_get_jf_constant_rdesc): New function.
20262 (ipa_param_descriptor): New field controlled_uses.
20263 (ipa_get_controlled_uses): New function.
20264 (ipa_set_controlled_uses): Likewise.
20265 * ipa-ref.h (ipa_find_reference): Declare.
20266 * ipa-prop.c (ipa_cst_ref_desc): New type.
20267 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
20269 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
20270 New parameter cs. Adjust all callers.
20271 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
20272 (remove_described_reference): New function.
20273 (jfunc_rdesc_usable): Likewise.
20274 (try_make_edge_direct_simple_call): Decrement controlled use count,
20275 attempt to remove reference if it hits zero.
20276 (combine_controlled_uses_counters): New function.
20277 (propagate_controlled_uses): Likewise.
20278 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
20279 (ipa_edge_duplication_hook): Duplicate reference descriptions.
20280 (ipa_print_node_params): Print described use counter.
20281 (ipa_write_jump_function): Adjust to jump function type changes.
20282 (ipa_read_jump_function): New parameter CS, pass it to
20283 ipa_set_jf_constant. Adjust caller.
20284 (ipa_write_node_info): Stream controlled use count
20285 (ipa_read_node_info): Likewise.
20286 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
20288 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
20289 count. Remove cloning-added reference if it reaches zero.
20290 * ipa-ref.c (ipa_find_reference): New function.
20292 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20294 * config/i386/i386.c (processor_target_table): Modified default
20295 alignment values for AMD BD and BT architectures.
20297 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
20299 * tree-vect-generic.c (uniform_vector_p): Move ...
20300 * tree.c (uniform_vector_p): ... here.
20301 * tree.h (uniform_vector_p): Declare it.
20302 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
20305 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20307 PR tree-optimization/57230
20308 * tree-ssa-strlen.c (handle_char_store): Record length for
20309 array store from STRING_CST.
20311 PR tree-optimization/57230
20312 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
20315 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
20317 * config/epiphany/epiphany.c (epiphany_init): Check size of
20318 NUM_MODES_FOR_MODE_SWITCHING.
20319 (epiphany_expand_prologue):
20320 Remove CONFIG_REGNUM initial value handling code.
20321 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
20322 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
20323 (emit_set_fp_mode, epiphany_mode_after): Likewise.
20324 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
20325 Don't return 1 for FP_MODE_NONE.
20326 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
20327 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
20328 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
20329 * config/epiphany/epiphany.md (save_config): New pattern.
20331 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
20333 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
20335 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
20337 * config/i386/i386.md (memory): Handle sseishft1.
20338 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
20339 (*vec_extractv2di_1): Ditto.
20341 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
20343 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
20346 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
20348 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
20349 Add mthumb/march=armv7-a multilib.
20350 Add mthumb/march=armv7-r multilib.
20351 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
20353 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
20355 * config/v850/t-rtems: Add more multilibs.
20357 2013-05-10 Richard Biener <rguenther@suse.de>
20359 PR tree-optimization/57214
20360 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
20361 not propagate from SSA names that occur in abnormal PHI nodes.
20363 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
20365 * stor-layout.c (element_precision): New function.
20366 * machmode.h (element_precision): Declare it.
20367 * tree.c (build_minus_one_cst): New function.
20368 (element_precision): Likewise.
20369 * tree.h (build_minus_one_cst): Declare new function.
20370 (element_precision): Likewise.
20371 * fold-const.c (operand_equal_p): Use element_precision.
20372 (fold_binary_loc): Handle vector types.
20373 * convert.c (convert_to_integer): Use element_precision.
20374 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
20377 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
20379 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
20380 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
20381 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
20382 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20383 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
20384 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
20385 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
20386 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20387 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
20388 (Uuw8): New constraints.
20389 (Usb4): Move into alphabetical order.
20390 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
20391 (sd8_operand, ub8_operand, uw8_operand): New predicates.
20392 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
20393 previously unnamed patterns.
20394 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
20395 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
20396 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
20397 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
20398 of set_attr_alternative/if_then_else. Use extended_mips16 instead
20399 of specific lengths.
20401 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20403 * config/i386/i386.md (rotateinv): New code attr.
20404 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20405 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20406 roll $31, %eax, etc.
20408 PR tree-optimization/45216
20409 PR tree-optimization/57157
20410 * tree-ssa-forwprop.c (simplify_rotate): New function.
20411 (ssa_forward_propagate_and_combine): Call it.
20413 2013-05-10 Richard Biener <rguenther@suse.de>
20415 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
20416 disable peeling when we version for aliasing.
20417 (vector_alignment_reachable_p): Honor explicit user alignment.
20418 (vect_supportable_dr_alignment): Likewise.
20419 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
20420 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
20421 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
20422 then peeling to arrange for the cost-model check to come first.
20424 2013-05-10 Alan Modra <amodra@gmail.com>
20426 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
20427 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
20428 * configure: Regenerate.
20430 2013-05-10 Alan Modra <amodra@gmail.com>
20433 * varasm.c (default_elf_select_section): Move !DECL_P check..
20434 (get_named_section): ..to here before calling get_section_name.
20436 (default_section_type_flags): Add DECL_P check.
20437 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
20438 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
20440 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
20442 * config/epiphany/epiphany.c (epiphany_expand_prologue):
20443 When using gen_stack_adjust_str with a register offset, add a
20444 REG_FRAME_RELATED_EXPR note.
20446 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
20448 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
20449 (*vec_extractv4si_zext_mem): Ditto.
20450 (*vec_extractv2di): Add 0->x and x->x alternatives.
20451 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
20452 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
20454 2013-05-09 Jason Merrill <jason@redhat.com>
20456 N3639 C++1y VLA support
20457 * gimplify.c (gimplify_vla_decl): Don't touch an existing
20460 * tree.c (build_constructor_va): New.
20461 * tree.h: Declare it.
20463 2013-05-09 Martin Jambor <mjambor@suse.cz>
20466 * gimple-fold.c (canonicalize_constructor_val): Call
20467 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
20469 2013-05-09 Jan Hubicka <jh@suse.cz>
20470 Richard Biener <rguenther@suse.de>
20473 * symtab.c (symtab_make_decl_local): Do not add private names.
20475 2013-05-09 Jan Hubicka <jh@suse.cz>
20478 * symtab.c (insert_to_assembler_name_hash): Handle clones.
20479 (unlink_from_assembler_name_hash): Likewise.
20480 (symtab_prevail_in_asm_name_hash, symtab_register_node,
20481 symtab_unregister_node, symtab_initialize_asm_name_hash,
20482 change_decl_assembler_name): Update.
20484 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
20486 * config/aarch64/aarch64.md: New movtf split.
20487 (*movtf_aarch64): Update.
20488 (aarch64_movdi_tilow): Handle TF modes and rename to
20489 aarch64_movdi_<mode>low.
20490 (aarch64_movdi_tihigh): Handle TF modes and rename to
20491 aarch64_movdi_<mode>high
20492 (aarch64_movtihigh_di): Handle TF modes and rename to
20493 aarch64_mov<mode>high_di
20494 (aarch64_movtilow_di): Handle TF modes and rename to
20495 aarch64_mov<mode>low_di
20496 (aarch64_movtilow_tilow): Remove spurious whitespace.
20497 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
20499 (aarch64_print_operand): Update.
20501 2013-05-09 Alan Modra <amodra@gmail.com>
20503 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
20505 * configure: Regenerate.
20507 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20509 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
20510 splitter preparation statements.
20511 * config/i386/sse.md (*vec_extract* splitters): Ditto.
20512 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
20515 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20517 * gimple-ssa-strength-reduction.c (count_candidates): Change
20518 return value to int.
20519 (analyze_candidates_and_replace): Change type of length to int.
20521 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20523 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
20524 (*vec_extract<mode>): Use VI12_128 mode iterator.
20525 (*vec_extract<mode>_mem): Ditto.
20526 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
20529 2013-05-08 Diego Novillo <dnovillo@google.com>
20534 2012-08-17 Diego Novillo <dnovillo@google.com>
20537 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
20538 * config.in: Regenerate.
20539 * configure: Regenerate.
20540 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
20542 2013-05-08 Jan Hubicka <jh@suse.cz>
20545 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
20546 * cgraph.h (symtab_node_base): Add unique_name.
20547 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20548 input_overwrite_node, input_varpool_node): Stream unique_name.
20549 * cgraphclones.c (cgraph_create_virtual_clone,
20550 cgraph_function_versioning): Set unique_name.
20551 * ipa.c (function_and_variable_visibility): Set unique_name.
20553 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20555 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
20556 (alloc_cand_and_find_basis): Restrict conditional candidate
20557 processing to CAND_MULTs.
20559 2013-05-08 Jan Hubicka <jh@suse.cz>
20562 lto-symtab.c (lto_symtab_symbol_p): New function.
20563 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
20564 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
20565 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
20566 Skip static symbols.
20568 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
20570 PR tree-optimization/57200
20571 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
20572 Only call inform if the preceding warning_at returns true.
20574 2013-05-07 Han Shen <shenhan@google.com>
20576 * cfgexpand.c (record_or_union_type_has_array_p): New function.
20577 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
20578 * common.opt (fstack-protector-strong): New option.
20579 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
20580 * doc/invoke.texi (Optimization Options): Document
20581 "-fstack-protector-strong".
20582 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
20584 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20586 * config/mips/mips.c (mips_machine_reorg2): Return 0.
20588 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
20590 * ira.c (update_equiv_regs): Add insn having equiv memory even if
20591 it is not lhs of the insn.
20592 (setup_reg_equiv): Remove insn having equiv memory which it is not
20594 * lra-constraints.c (process_address): Try to improve generation
20595 code for address base + disp.
20596 (lra_constraints): Make correct the code for checking insn setting
20597 up backward equivalence. Remove insn only if it is in the init
20599 * lra-eliminations.c (update_reg_eliminate): Change return value.
20600 (lra_eliminate): Use the result.
20602 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
20604 * config/i386/sse.md (ssescalarnummask): New mode attribute.
20605 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
20606 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
20607 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
20608 register target operands.
20609 (*vec_extractv8hi_sse2): New pattern.
20610 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
20611 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
20612 (*vec_extract<mode>_mem): New insn and split pattern.
20614 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
20616 * config/arm/arm.c (arm_asan_shadow_offset): New function.
20617 (TARGET_ASAN_SHADOW_OFFSET): Define.
20618 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
20619 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
20621 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20623 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
20624 (incr_vec_index): Return -1 if increment not found.
20625 (create_add_on_incoming_edge): Assert if increment not found.
20626 (record_increment): Limit number of increments recorded.
20627 (all_phi_incrs_profitable): Return false if an increment not found.
20628 (replace_profitable_candidates): Don't process increments that were
20630 (analyze_candidates_and_replace): Limit size of incr_vec.
20632 2013-05-07 Richard Biener <rguenther@suse.de>
20634 * calls.c (special_function_p): setjmp-like functions are leaf.
20635 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
20636 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
20638 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
20640 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
20641 (aarch64_simd_mov<mode>): New expander.
20642 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
20643 (aarch64_simd_mov_to_<mode>high): Likewise.
20644 (aarch64_simd_mov_from_<mode>low): Likewise.
20645 (aarch64_simd_mov_from_<mode>high): Likewise.
20646 (aarch64_dup_lane<mode>): Update.
20647 (aarch64_dup_lanedi): New instruction pattern.
20648 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
20649 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
20651 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20653 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
20654 (replace_mult_candidate): Remove unnecessary argument; remove
20655 unnecessary parameter from call to introduce_cast_before_cand.
20656 (replace_unconditional_candidate): Remove unnecessary parameter
20657 from call to replace_mult_candidate.
20658 (replace_conditional_candidate): Likewise.
20659 (insert_initializers): Use make_temp_ssa_name.
20660 (introduce_cast_before_cand): Remove unnecessary argument; use
20661 make_temp_ssa_name.
20662 (replace_one_candidate): Remove unnecessary argument; remove
20663 unnecessary parameter from calls to introduce_cast_before_cand.
20664 (replace_profitable_candidates): Remove unnecessary parameters
20665 from calls to replace_one_candidate.
20667 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20669 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
20670 phi def as possibly hiding a basis for a CAND_ADD whose operands
20671 have been commuted in the analysis.
20672 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
20674 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20676 * config/aarch64/aarch64.md
20677 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
20678 shift value between 0-4.
20680 2013-05-07 Richard Biener <rguenther@suse.de>
20682 * double-int.h (rshift): New overload.
20683 * double-int.c (rshift): New function.
20684 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
20685 (create_reference_ops_from_ref): Remove.
20686 (vn_reference_insert): Use shared ops for constructing the
20687 reference and copy it.
20689 2013-05-07 Richard Biener <rguenther@suse.de>
20691 PR middle-end/57190
20692 * tree-eh.c (sink_clobbers): Properly propagate
20693 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
20695 2013-05-07 Jakub Jelinek <jakub@redhat.com>
20697 PR tree-optimization/57149
20698 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
20699 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
20700 collect_phi_def_edges, execute_late_warn_uninitialized): Use
20701 uninit_undefined_value_p instead of ssa_undefined_value_p.
20704 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
20705 for modifier == EXPAND_INITIALIZER.
20707 2013-05-07 Anton Blanchard <anton@samba.org>
20709 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
20710 for powerpc64 little endian.
20711 * configure: Regenerate.
20713 2013-05-06 Graham Stott <grahams@btinternet.com>
20715 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
20716 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
20717 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
20718 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
20720 2013-05-06 Graham Stott <grahams@btinternet.com>
20722 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
20723 codes which allow non-lvalues.
20725 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
20727 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
20728 components are all 1s.
20729 (integer_minus_onep): New function.
20730 * tree.h (integer_minus_onep): Declare it.
20731 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
20732 integer_minus_onep instead of integer_all_onesp.
20734 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20737 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
20738 variations of these patterns.
20740 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20742 * config/i386/i386.md (isa): Add x64_sse4 member.
20743 (enabled): Handle x64_sse4.
20744 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
20745 instruction for 64bit SSE4_1 targets. Update insn attributes.
20746 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
20747 instruction for SSE4_1 targets. Update insn attributes.
20748 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20749 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
20750 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
20752 (*vec_extractv4si): Rename from *sse4_1_pextrd.
20753 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
20754 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
20756 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20759 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
20761 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
20763 * final.c (do_assembler_dialects): Don't handle curly braces and
20764 vertical bar escaped by % as dialect delimiters.
20765 (output_asm_insn): Print curly braces and vertical bar if escaped
20766 by % and ASSEMBLER_DIALECT defined.
20767 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
20768 * doc/tm.texi: Regenerated.
20770 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20772 * config/mips/mips.c: Include tree-pass.h.
20773 (mips_reorg): Split in pre- and post-dbr_schedule parts.
20774 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
20775 (pass_mips_machine_reorg2): New machine specific pass.
20776 (insert_pass_mips_machine_reorg2): New pass plugin definition.
20777 (mips_option_override): Register the new pass.
20778 * rtl.h (cleanup_barriers): Remove prototype.
20779 (dbr_schedule): Likewise.
20780 * jump.c (cleanup_barriers): Make static.
20781 * reorg.c (dbr_schedule): Likewise.
20783 2013-05-06 Richard Biener <rguenther@suse.de>
20785 PR tree-optimization/57185
20786 * tree-parloops.c (add_field_for_reduction): Handle anonymous
20787 SSA names properly.
20789 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20792 * config/i386/i386.c (add_parameter_dependencies): Add dependence
20793 between "first_arg" and "insn", not "last" and "insn".
20795 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20797 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
20798 (find_candidates_in_block): Re-enable slsr_process_phi.
20799 (create_phi_basis): Fix double counting of candidate adjustment.
20801 2013-05-06 Richard Biener <rguenther@suse.de>
20803 PR middle-end/57147
20804 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
20805 the edge is also fallthru, preserve it and just clear the
20807 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
20808 also complex, preserve that and just clear the fallthru flag.
20809 * tree-inline.c (update_ssa_across_abnormal_edges): Also
20810 update virtual operands.
20812 2013-05-06 Alan Modra <amodra@gmail.com>
20814 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
20815 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
20816 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
20817 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
20818 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20819 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
20821 2013-05-06 Alan Modra <amodra@gmail.com>
20823 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
20824 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
20825 (DEFAULT_ASM_ENDIAN): Define.
20826 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20827 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
20828 Update -K PIC clause from sysv4.h.
20829 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
20830 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
20832 2013-05-06 Alan Modra <amodra@gmail.com>
20834 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
20835 twice for little-endian.
20836 (ashrdi3_no_power, ashrdi3): Support little-endian.
20838 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20841 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
20842 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
20844 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
20845 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
20846 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
20849 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
20850 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20852 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
20853 * config/pa/pa.opt: Make mbig-switch a no-op.
20854 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
20855 (CASE_VECTOR_MODE): Always return SImode.
20856 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
20857 for the !TARGET_BIG_SWITCH case.
20858 * config/pa/pa-linux.h: Likewise.
20859 * config/pa/pa-openbsd.h: Likewise.
20860 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
20861 * config/pa/pa.md (short_jump): Remove define_insn.
20862 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
20863 (casesi0): Remove define_insn.
20864 (type): Remove btable_branch.
20865 (pa_combine_type): Likewise.
20866 (in_nullified_branch_delay): Likewise.
20867 (in_call_delay): Likewise.
20868 (define_delay): Likewise.
20869 (define_insn_reservation "Z3"): Likewise.
20870 (define_insn_reservation "Z4"): Likewise.
20871 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
20872 (pa_adjust_insn_length): Remove adjustment for btable branches.
20873 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
20876 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
20878 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20879 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
20880 Add m->r,x alternatives.
20881 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
20882 splitters using SWI48x mode iterator.
20883 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
20884 TARGET_64BIT. Add m->x alternative.
20885 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
20886 Add o->x alternative. Enable for TARGET_SSE.
20887 (sse_storeq): Remove expander.
20888 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
20889 with memory input operand.
20890 (*vec_extractv2di_1 splitter): New.
20891 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
20892 * config/i386/i386.md (ssevecmodelower): New mode attribute.
20894 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20896 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
20897 (INT_LOWPART): Delete.
20898 (extract_MB): Adjust.
20899 (extract_ME): Adjust.
20900 (print_operand): Adjust.
20902 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20904 * config/rs6000/predicates.md (reg_or_add_cint_operand,
20905 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20906 (reg_or_logical_cint_operand, easy_fp_constant,
20907 logical_const_operand): Delete "CONST_DOUBLE" case.
20908 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
20909 "HOST_BITS_PER_WIDE_INT == 64" test.
20910 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
20911 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
20912 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
20913 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
20915 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
20916 CONST_DOUBLE DImode/VOIDmode case.
20917 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
20918 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20919 CONST_DOUBLE VOIDmode case.
20920 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20921 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
20922 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
20923 Delete CONST_DOUBLE case.
20924 (splitters for mov FMOVE64 const_double): Delete
20925 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20926 "HOST_BITS_PER_WIDE_INT >= 64" test.
20927 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
20929 (mov DI const_double): Delete.
20931 2013-05-04 Jakub Jelinek <jakub@redhat.com>
20933 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
20934 on op shows all bits zero in mode of a lowpart subreg, return zero.
20936 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
20939 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
20940 to save TFmode registers and DImode to save TImode registers for
20941 caller save operations.
20942 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
20943 mark being partially clobbered since they only use the first
20946 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
20947 and TDmode only use the upper 64-bits of each VSX register.
20949 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20951 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
20952 (find_candidates_in_block): Disable slsr_process_phi.
20954 2013-05-03 Guozhi Wei <carrot@google.com>
20956 * coverage.c (coverage_obj_init): Move the construction of gcov
20958 (build_init_ctor): ... here.
20960 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20962 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
20963 (slsr_cand_d): Redefine def_phi.
20964 (stride_status, phi_adjust_status, count_phis_status): New enums.
20965 (find_phi_def): New.
20966 (find_basis_for_base_expr): New.
20967 (find_basis_for_candidate): Handle hidden bases.
20968 (alloc_cand_and_find_basis): Handle phi candidates.
20969 (slsr_process_phi): New.
20970 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
20971 (create_mul_imm_cand): Likewise.
20972 (create_add_ssa_cand): Exclude phi base candidates.
20973 (create_add_imm_cand): Likewise.
20974 (slsr_process_cast): Likewise.
20975 (slsr_process_copy): Likewise.
20976 (find_candidates_in_block): Handle phi candidates.
20977 (dump_candidate): Likewise.
20978 (unconditional_cands): Delete.
20979 (unconditional_cands_with_known_stride_p): Delete.
20980 (phi_dependent_cand_p): New.
20981 (cand_increment): Handle phi-dependent candidates.
20982 (replace_dependent): Delete.
20983 (replace_mult_candidate): New.
20984 (replace_unconditional_candidate): New.
20985 (incr_vec_index): Move to avoid forward reference.
20986 (create_add_on_incoming_edge): New.
20987 (create_phi_basis): New.
20988 (replace_dependents): Delete.
20989 (replace_conditional_candidate): New.
20990 (phi_add_costs): New.
20991 (replace_uncond_cands_and_profitable_phis): New.
20992 (record_increment): Handle phi adjustments.
20993 (record_phi_increments): New.
20994 (record_increments): Handle phi adjustments.
20995 (phi_incr_cost): New.
20996 (lowest_cost_path): Handle phis.
20997 (total_savings): Likewise.
20998 (analyze_increments): Likewise.
20999 (ncd_with_phi): New.
21000 (ncd_of_cand_and_phis): New.
21001 (nearest_common_dominator_for_cands): Handle phi increments.
21002 (all_phi_incrs_profitable): New.
21003 (replace_profitable_candidates): Handle phi-dependent candidates.
21004 (analyze_candidates_and_replace): Likewise.
21006 2013-05-03 Teresa Johnson <tejohnson@google.com>
21009 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
21010 do not exceed REG_BR_PROB_BASE.
21012 2013-05-03 Jeff Law <law@redhat.com>
21014 PR tree-optimization/57144
21015 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
21016 operand of the condition will bit into the new type when eliminating
21017 a cast feeding a condition.
21019 2013-05-03 Jakub Jelinek <jakub@redhat.com>
21021 PR rtl-optimization/57130
21022 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
21023 of COMPARE as in_code to the recursive call if needed.
21025 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
21027 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
21028 (enabled): Handle new members.
21029 * config/i386/sse.md (*vec_concatv2si): Merge from
21030 *vec_concatv2si_sse2 and vec_concatv2si_sse.
21031 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
21033 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
21035 PR tree-optimization/57027
21036 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
21037 for fnms opportunity, check we got the prerequisite kind
21038 of tree / gimple before using accessor functions.
21040 2013-05-03 Richard Biener <rguenther@suse.de>
21042 * double-int.h (lshift): New overload without precision
21043 and arith argument.
21044 (operator *=, operator +=, operator -=): Move ...
21045 * double-int.c (operator *=, operator +=, operator -=): ... here
21046 and implement more efficiently.
21047 (mul_double_with_sign): Remove.
21048 (lshift_double): Adjust to take unsinged shift argument, push
21049 dispatching code to callers.
21050 (mul_double_wide_with_sign): Add early out for callers that
21051 are not interested in high parts or overflow.
21052 (lshift): New function.
21053 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
21054 dispatch code here.
21055 (lrotate, rrotate): Use logical shifts.
21056 * expr.c (get_inner_reference): Use lshift.
21057 * fixed-value.c (do_fixed_divide): Likewise.
21058 * tree-dfa.c (get_ref_base_and_extent): Likewise.
21059 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
21060 (indirect_refs_may_alias_p): Likewise.
21061 (stmt_kills_ref_p_1): Likewise.
21063 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21065 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
21067 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21069 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
21070 scalar form of FABD instruction.
21072 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
21074 * lra-constraints.c (process_alt_operands): Add checking alt
21075 number to choose the best alternative.
21077 2013-05-02 Richard Biener <rguenther@suse.de>
21079 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
21080 bitmap and its handling.
21081 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
21083 2013-05-02 Richard Biener <rguenther@suse.de>
21085 PR middle-end/57140
21086 * tree-inline.c (copy_loops): Properly handle removed loops.
21087 (copy_cfg_body): Mark destination loops for fixup if source
21088 loops needed fixup.
21090 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
21093 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
21094 generating simple_return for naked functions.
21096 2013-05-02 Martin Jambor <mjambor@suse.cz>
21098 PR middle-end/56988
21099 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
21100 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
21102 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
21103 ipa_agg_replacement_value structures.
21104 (known_aggs_to_agg_replacement_list): Likewise.
21105 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
21106 (read_agg_replacement_chain): Likewise.
21107 (ipcp_transform_function): Also check that by_ref flags match.
21109 2013-05-02 Richard Biener <rguenther@suse.de>
21111 * graphds.h (struct graph): Add obstack member.
21112 * graphds.c (new_graph): Initialize obstack and allocate
21114 (add_edge): Allocate edge from the obstack.
21115 (free_graph): Free the obstack instead of all edges and vertices.
21117 2013-05-02 Teresa Johnson <tejohnson@google.com>
21119 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
21121 * cfg.c (update_bb_profile_for_threading): Ditto.
21122 * tree-inline.c (copy_bb): Ditto.
21123 (copy_edges_for_bb): Ditto.
21124 (initialize_cfun): Ditto.
21125 (copy_cfg_body): Ditto.
21126 (expand_call_inline): Ditto.
21127 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
21128 (estimate_node_size_and_time): Ditto.
21129 (inline_merge_summary): Ditto.
21130 * cgraphclones.c (cgraph_clone_edge): Ditto.
21131 (cgraph_clone_node): Ditto.
21132 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21133 (compute_trg_info): Ditto.
21135 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21137 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
21138 S reg when fp attribute set.
21139 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
21141 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21143 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
21145 (*and_one_cmplsi3_compare0_uxtw): Likewise.
21146 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
21147 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
21149 2013-05-02 Richard Biener <rguenther@suse.de>
21151 * tree-scalar-evolution.c (scev_info_hasher): Remove.
21152 (struct instantiate_cache_entry): New type.
21153 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
21154 (struct instantiate_cache_type): New type.
21155 (set_instantiated_value, get_instantiated_value): Remove.
21156 (get_instantiated_value_entry): New function.
21157 (instantiate_scev_name): Use the new cache and adjust.
21158 (instantiate_scev_poly): Adjust.
21159 (instantiate_scev_binary): Likewise.
21160 (instantiate_array_ref): Likewise.
21161 (instantiate_scev_convert): Likewise.
21162 (instantiate_scev_not): Likewise.
21163 (instantiate_scev_3): Likewise.
21164 (instantiate_scev_2): Likewise.
21165 (instantiate_scev_r): Likewise.
21166 (instantiate_scev): Likewise.
21167 (resolve_mixers): Likewise.
21169 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
21172 * lra-constraints.c (best_small_class_operands_num): Remove.
21173 (process_alt_operands): Remove small_class_operands_num. Take
21174 small classes operands into losers and only if the operand is not
21175 matched. Modify debugging output.
21176 (curr_insn_transform): Remove best_small_class_operands_num.
21179 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21181 * config/aarch64/aarch64-builtins.c
21182 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
21183 * config/aarch64/aarch64-simd-builtins.def
21184 (reduc_splus_): Add new modes.
21185 (reduc_uplus_): New.
21186 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
21187 (reduc_uplus_v4sf): Likewise.
21188 (reduc_splus_v4sf): Likewise.
21189 (aarch64_addv<mode>): Likewise.
21190 (reduc_uplus_<mode>): Likewise.
21191 (reduc_splus_<mode>): Likewise.
21192 (aarch64_addvv2di): Likewise.
21193 (reduc_uplus_v2di): Likewise.
21194 (reduc_splus_v2di): Likewise.
21195 (aarch64_addvv2si): Likewise.
21196 (reduc_uplus_v2si): Likewise.
21197 (reduc_splus_v2si): Likewise.
21198 (reduc_<sur>plus_<mode>): New.
21199 (reduc_<sur>plus_v2di): Likewise.
21200 (reduc_<sur>plus_v2si): Likewise.
21201 (reduc_<sur>plus_v4sf): Likewise.
21202 (aarch64_addpv4sf): Likewise.
21203 * config/aarch64/arm_neon.h
21204 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
21205 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
21206 add UNSPEC_SADDV, UNSPEC_UADDV.
21208 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
21210 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21212 * config/aarch64/arm_neon.h
21213 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
21215 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21217 * config/aarch64/aarch64-builtins
21218 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
21220 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21222 * config/aarch64/aarch64-simd-builtins.def
21223 (reduc_smax_): New.
21224 (reduc_smin_): Likewise.
21225 (reduc_umax_): Likewise.
21226 (reduc_umin_): Likewise.
21227 (reduc_smax_nan_): Likewise.
21228 (reduc_smin_nan_): Likewise.
21231 (smax): Update for V2SF, V4SF and V2DF modes.
21234 (smin_nan): Likewise.
21235 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
21236 (<su><maxmin><mode>3): ...This, refactor.
21237 (s<maxmin><mode>3): New.
21238 (<maxmin_uns><mode>3): Likewise.
21239 (reduc_<maxmin_uns>_<mode>): Refactor.
21240 (reduc_<maxmin_uns>_v4sf): Likewise.
21241 (reduc_<maxmin_uns>_v2si): Likewise.
21242 (aarch64_<fmaxmin><mode>: Remove.
21243 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
21245 (vmin<q>_f<32,64>): Likewise.
21246 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21248 (su): Add mappings for smax, smin, umax, umin.
21250 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21251 (FMAXMIN): Rename as...
21252 (FMAXMIN_UNS): ...This.
21254 (fmaxminv): Likewise.
21255 (fmaxmin): Likewise.
21257 (maxmin_uns_op): Likewise.
21259 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21261 * config/aarch64/arm_neon.h
21262 (vac<ge, gt><sd>_f<32, 64>): Rename to...
21263 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
21264 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
21266 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21268 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
21269 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
21271 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21273 * config/aarch64/aarch64-simd.md
21274 (vcond<mode>_internal): Handle special cases for constant masks.
21275 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
21276 (vcondu<mode><mode>): Likewise.
21277 (vcond<v_cmp_result><mode>): New.
21279 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21281 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
21282 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
21283 * config/aarch64/aarch64-simd-builtins.def
21284 (cmeq): Update to BUILTIN_VALLDI.
21289 * config/aarch64/arm_neon.h
21290 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
21291 to builtins or C as appropriate.
21293 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21295 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
21297 (cmhi): Rename to...
21299 * config/aarch64/aarch64-simd.md
21300 (simd_mode): Add SF.
21301 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
21302 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
21303 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
21304 (cstore<mode>_neg): ...This.
21305 * config/aarch64/iterators.md
21307 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
21308 (COMPARISONS): New.
21309 (UCOMPARISONS): Likewise.
21310 (optab): Add missing comparisons.
21316 (VCMP_S): Likewise.
21317 (VCMP_U): Likewise.
21318 (V_cmp_result): Add DF, SF modes.
21319 (v_cmp_result): Likewise.
21321 (vmtype): Likewise.
21322 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
21324 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
21326 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
21327 define_insn to define_insn_and_split.
21328 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
21329 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
21330 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
21331 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
21332 (thumb2_negscc): Likewise.
21334 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21336 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
21338 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21340 * config/arm/thumb2.md: Remove trailing whitespaces.
21342 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21344 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
21345 Use gen_int_mode rather than GEN_INT.
21347 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
21349 * value-prof.c (stream_in_histogram_value): Remove the strayed
21352 2013-04-30 Richard Biener <rguenther@suse.de>
21354 PR middle-end/57122
21355 * cfghooks.c (split_edge): Properly check for the loop latch edge.
21357 2013-04-30 Richard Biener <rguenther@suse.de>
21359 PR middle-end/57107
21360 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
21362 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
21364 PR rtl-optimization/56957
21365 PR rtl-optimization/57105
21366 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
21367 variable. Use just INSN_UID for determining whether an insn
21368 should be only disconnected from the insn stream.
21369 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
21371 2013-04-30 Jakub Jelinek <jakub@redhat.com>
21373 PR tree-optimization/57104
21374 * tsan.c (instrument_expr): Don't instrument accesses to
21375 DECL_HARD_REGISTER VAR_DECLs.
21377 2013-04-30 Richard Biener <rguenther@suse.de>
21379 * function.h (loops_for_fn): New inline function.
21380 (set_loops_for_fn): Likewise.
21381 * cfgloop.h (place_new_loop): Add struct function parameter.
21382 (get_loop): Likewise.
21383 (get_loops): Likewise.
21384 (number_of_loops): Likewise.
21385 (fel_next): Adjust.
21386 (fel_init): Likewise.
21387 * cfg.c (get_loop_copy): Adjust.
21388 * cfgloop.c (flow_loops_dump): Likewise.
21389 (record_loop_exits): Likewise.
21390 (verify_loop_structure): Likewise.
21391 * cfgloopanal.c (mark_irreducible_loops): Likewise.
21392 (estimate_reg_pressure_cost): Likewise.
21393 (mark_loop_exit_edges): Likewise.
21394 * cfgloopmanip.c (place_new_loop): Likewise.
21395 (add_loop): Likewise.
21396 (duplicate_loop): Likewise.
21397 * graph.c (draw_cfg_nodes): Likewise.
21398 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
21399 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
21400 (extract_affine_chrec): Likewise.
21401 (build_scop_iteration_domain): Likewise.
21402 * graphite.c (graphite_initialize): Likewise.
21403 * ira-build.c (create_loop_tree_nodes): Likewise.
21404 (more_one_region_p): Likewise.
21405 (rebuild_regno_allocno_maps): Likewise.
21406 (mark_loops_for_removal): Likewise.
21407 (mark_all_loops_for_removal): Likewise.
21408 (remove_unnecessary_regions): Likewise.
21409 (ira_build): Likewise.
21410 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21411 * loop-init.c (fix_loop_structure): Likewise.
21412 (gate_rtl_move_loop_invariants): Likewise.
21413 (gate_rtl_unswitch): Likewise.
21414 (gate_rtl_unroll_and_peel_loops): Likewise.
21415 (rtl_doloop): Likewise.
21416 * lto-streamer-in.c (input_cfg): Likewise.
21417 * lto-streamer-out.c (output_cfg): Likewise.
21418 * modulo-sched.c (sms_schedule): Likewise.
21419 * predict.c (tree_estimate_probability): Likewise.
21420 (tree_estimate_probability_driver): Likewise.
21421 (estimate_loops): Likewise.
21422 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
21423 (move_sese_region_to_fn): Likewise.
21424 (debug_loop_num): Likewise.
21425 * tree-chrec.c (chrec_evaluate): Likewise.
21426 (hide_evolution_in_other_loops_than_loop): Likewise.
21427 (chrec_component_in_loop_num): Likewise.
21428 (reset_evolution_in_loop): Likewise.
21429 (evolution_function_is_invariant_rec_p): Likewise.
21430 * tree-if-conv.c (main_tree_if_conversion): Likewise.
21431 * tree-inline.c (copy_loops): Likewise.
21432 (copy_cfg_body): Likewise.
21433 (tree_function_versioning): Likewise.
21434 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
21435 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
21437 (add_to_evolution_1): Likewise.
21438 (scev_const_prop): Likewise.
21439 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
21440 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
21441 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
21442 (tree_ssa_lim_initialize): Likewise.
21443 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
21444 (verify_loop_closed_ssa): Likewise.
21445 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
21446 (tree_ssa_loop_im): Likewise.
21447 (tree_ssa_loop_unswitch): Likewise.
21448 (tree_vectorize): Likewise.
21449 (check_data_deps): Likewise.
21450 (tree_ssa_loop_ivcanon): Likewise.
21451 (tree_ssa_loop_bounds): Likewise.
21452 (tree_complete_unroll): Likewise.
21453 (tree_complete_unroll_inner): Likewise.
21454 (tree_parallelize_loops): Likewise.
21455 (tree_ssa_loop_prefetch): Likewise.
21456 (tree_ssa_loop_ivopts): Likewise.
21457 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
21458 * tree-vectorizer.c (vectorize_loops): Likewise.
21460 2013-04-29 Mike Frysinger <vapier@gentoo.org>
21462 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
21463 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
21464 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
21465 with EABI_LINK_SPEC.
21467 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21470 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
21473 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
21476 * lra-constraints.c (process_alt_operands): Discourage a bit more
21477 using memory for pseudos. Print cost dump for alternatives.
21478 Modify cost values for conflicts with early clobbers.
21479 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
21481 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21484 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
21486 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21488 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
21489 from/to S register.
21490 (movdi_aarch64): Support LDR/STR from/to D register.
21492 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21494 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
21495 or higher by default.
21497 2013-04-29 Richard Biener <rguenther@suse.de>
21499 PR middle-end/57075
21500 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
21501 even if not adding abnormal edges for calls that can make
21504 2013-04-29 Richard Biener <rguenther@suse.de>
21506 PR middle-end/57103
21507 * tree-cfg.c (move_stmt_op): Fix condition under which to update
21509 (move_stmt_r): Remove redundant checking.
21511 2013-04-29 Teresa Johnson <tejohnson@google.com>
21514 * basic-block.h (apply_scale): New function.
21515 (apply_probability): Use apply_scale.
21516 * gimple-streamer-in.c (input_bb): Ditto.
21517 * lto-streamer-in.c (input_cfg): Ditto.
21518 * lto-cgraph.c (merge_profile_summaries): Ditto.
21519 * tree-optimize.c (execute_fixup_cfg): Ditto.
21520 * tree-inline.c (copy_bb): Update comment to use apply_scale.
21521 (copy_edges_for_bb): Ditto.
21522 (copy_cfg_body): Ditto.
21524 2013-04-29 Tom de Vries <tom@codesourcery.com>
21526 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
21527 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
21528 (tail_merge_optimize): Handle current_loops == NULL.
21530 2013-04-26 Jeff Law <law@redhat.com>
21532 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
21533 (simplify_cond_using_ranges): Generalize code to simplify
21534 COND_EXPRs where one argument is a constant and the other
21535 is an SSA_NAME created by an integral type conversion.
21537 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21539 * config/arm/arm.md (store_minmaxsi): Use only when
21540 optimize_insn_for_size_p.
21542 2013-04-29 Christian Bruel <christian.bruel@st.com>
21545 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
21547 2013-04-29 Richard Biener <rguenther@suse.de>
21549 PR middle-end/57089
21550 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
21551 loop tree make sure to schedule a fixup for the child as well.
21552 (expand_omp_for_generic): Properly add loops.
21553 (expand_omp_for_static_nochunk): Likewise.
21554 (expand_omp_for_static_chunk): Likewise.
21555 (expand_omp_for): For the degenerate case fixup loops.
21556 (expand_omp_sections): Fix default bb placement in loops.
21557 (expand_omp_atomic_pipeline): Properly add loops.
21559 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21561 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
21563 2013-04-29 Tom de Vries <tom@codesourcery.com>
21565 * tree-ssa-tail-merge.c: Update header comment.
21567 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21569 * config/aarch64/arm_neon.h
21570 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
21571 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
21572 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
21573 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
21574 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
21575 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
21576 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
21577 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
21579 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21581 * config/aarch64/aarch64-simd.md
21582 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
21583 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
21584 fix_trunc, fixuns_trunc.
21585 (ftrunc<VDQF:mode>2): New.
21586 * config/aarch64/iterators.md (optab): Add fix, fixuns.
21587 (fix_trunc_optab): New.
21589 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21591 * config/aarch64/aarch64-builtins.c
21592 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
21593 iceilf, lround, iroundf.
21595 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21598 * config/i386/i386.h (enum ix86_tune_indices)
21599 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
21600 New, split from X86_TUNE_INTER_UNIT_MOVES.
21601 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
21602 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
21603 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
21604 (TARGET_INTER_UNIT_MOVES): Remove.
21605 * config/i386/i386.c (initial_ix86_tune_features): Update.
21606 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
21607 (ix86_expand_convert_uns_didf_sse): Use
21608 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21609 (ix86_expand_vector_init_one_nonzero): Ditto.
21610 (ix86_expand_vector_init_interleave): Ditto.
21611 (inline_secondary_memory_needed): Return true for moves from SSE class
21612 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
21613 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
21614 * config/i386/constraints.md (Yi, Ym): Depend on
21615 TARGET_INTER_UNIT_MOVES_TO_VEC.
21616 (Yj, Yn): New constraints.
21617 * config/i386/i386.md (*movdi_internal): Change constraints of
21618 operand 1 from Yi to Yj and from Ym to Yn.
21619 (*movsi_internal): Ditto.
21620 (*movdf_internal): Ditto.
21621 (*movsf_internal): Ditto.
21622 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
21623 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21624 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
21625 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
21626 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
21627 * config/i386/sse.md (movdi_to_sse): Ditto.
21628 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
21629 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
21630 TARGET_INTER_UNIT_MOVES.
21631 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
21632 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
21633 instead of TARGET_INTER_UNIT_MOVES.
21634 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
21635 operand 1 from Yi to Yj and from Ym to Yn.
21637 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21639 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
21640 (float_truncate_hi_): Likewise.
21641 (float_extend_lo_): Likewise.
21642 (float_truncate_lo_): Likewise.
21643 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
21644 (aarch64_float_extend_lo_v2df): Likewise.
21645 (vec_unpacks_hi_v4sf): Likewise.
21646 (aarch64_float_truncate_lo_v2sf): Likewise.
21647 (aarch64_float_truncate_hi_v4sf): Likewise.
21648 (vec_pack_trunc_v2df): Likewise.
21649 (vec_pack_trunc_df): Likewise.
21651 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21653 * config/aarch64/aarch64-builtins.c
21654 (aarch64_fold_builtin): Fold float conversions.
21655 * config/aarch64/aarch64-simd-builtins.def
21656 (floatv2si, floatv4si, floatv2di): New.
21657 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
21658 * config/aarch64/aarch64-simd.md
21659 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
21660 * config/aarch64/iterators.md (FLOATUORS): New.
21661 (optab): Add float, floatuns.
21662 (su_optab): Likewise.
21664 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21666 * config/aarch64/aarch64-builtins.c
21667 (aarch64_builtin_vectorized_function): Use new names for
21669 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
21670 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
21671 (fcvtzu): Split as...
21672 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
21673 (fcvtas): Split as...
21674 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
21675 (fcvtau): Split as...
21676 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
21677 (fcvtps): Split as...
21678 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
21679 (fcvtpu): Split as...
21680 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
21681 (fcvtms): Split as...
21682 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
21683 (fcvtmu): Split as...
21684 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
21685 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
21686 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
21687 (lfrintnusf, lfrintnudf): Likewise.
21688 * config/aarch64/aarch64-simd.md
21689 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
21691 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
21692 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
21693 (fcvt_pattern): Likewise.
21695 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21697 * config/aarch64/aarch64-simd.md
21698 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
21699 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
21701 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21703 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
21704 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
21705 (vrnd<a,m,n,p>_f32): Implement using builtins.
21706 (vrnd<i,x><q>_f<32, 64>): New.
21708 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21710 * config/aarch64/aarch64-builtins.c
21711 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
21712 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
21713 (frintz): Rename to...
21715 (frintp): Rename to...
21717 (frintm): Rename to...
21719 (frinti): Rename to...
21720 (nearbyint): ...this.
21721 (frintx): Rename to...
21723 (frinta): Rename to...
21725 * config/aarch64/aarch64-simd.md
21726 (aarch64_frint<frint_suffix><mode>): Delete.
21727 (<frint_pattern><mode>2): Convert to insn.
21728 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
21729 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
21730 (frint_pattern): Likewise.
21731 (frint_suffix): Likewise.
21733 2013-04-29 Richard Biener <rguenther@suse.de>
21735 PR tree-optimization/57081
21736 * loop-init.c: Include tree-flow.h.
21737 (loop_optimizer_finalize): Free number of iteration estimates.
21738 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
21740 2013-04-29 Jakub Jelinek <jakub@redhat.com>
21742 PR tree-optimization/57083
21743 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
21744 non-singleton shift count range, zero extend low_bound for uns case.
21746 * config/i386/predicates.md (general_vector_operand): New predicate.
21747 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
21748 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
21749 if they aren't nonimmediate operands. If their original values
21750 satisfy const_vector_equal_evenodd_p, don't shift them.
21751 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
21752 predicates. For the SSE4.1 case force operands[{1,2}] into registers
21753 if not nonimmediate_operand.
21754 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
21755 instead of register_operand.
21756 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
21758 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
21760 * stor-layout.c (finalize_size_functions): Allocate a structure and
21761 reset cfun before dumping the functions.
21763 2013-04-27 Jakub Jelinek <jakub@redhat.com>
21765 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
21768 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
21769 use xop_pmacsdqh if uns_p.
21770 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
21771 the immediate rotate count.
21773 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
21775 * rtl.h (struct rtx_def): Add comment for field jump.
21776 (LRA_SUBREG_P): New macro.
21777 * recog.c (register_operand): Check LRA_SUBREG_P.
21778 * lra.c (lra): Add note at the end of RTL code. Align non-empty
21780 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
21781 (lra_final_code_change): Skip subreg change for operators.
21782 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
21783 if there are no operand changes.
21784 * lra-constraints.c (curr_insn_set): New.
21785 (match_reload): Set LRA_SUBREG_P.
21786 (emit_spill_move): Ditto.
21787 (check_and_process_move): Use curr_insn_set. Process only single
21788 set insns. Don't initialize sec_mem_p and change_p.
21789 (simplify_operand_subreg): Use LRA_SUBREG_P.
21790 (reg_in_class_p): New function.
21791 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
21792 of #ifdef. Add code to remove cycling.
21793 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
21794 non-null disp. Reload inner instead of disp when base and index
21795 are null. Try to put lo_sum into register.
21796 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
21797 (check_and_process_move): Move code for move cost check to
21798 simple_move_p. Remove equiv_substitution.
21799 (simple_move_p): New function.
21800 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
21801 curr_insn_set. Call check_and_process_move only for single set
21802 insns. Use the new function. Move call of check_and_process_move
21803 after operand equiv substitution and address process.
21805 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21808 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
21809 with nonlocal goto receivers or returns twice calls, ignore
21810 unininitialized values from abnormal edges to nl goto receiver
21811 or returns twice call.
21813 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21815 PR tree-optimization/57051
21816 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
21817 and VEC_RSHIFT_EXPR if shift count is a multiple of element
21820 2013-04-26 Richard Biener <rguenther@suse.de>
21822 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
21823 (expand_omp_taskreg): Likewise. Mark loops for fixup.
21824 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
21825 (fixup_loop_arrays_after_move): New function.
21826 (move_sese_region_to_fn): Properly outline the loop tree parts
21827 of the SESE region.
21829 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
21831 * config/i386/i386.md (type, unit): Fix long lines.
21833 2013-04-26 Richard Biener <rguenther@suse.de>
21835 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
21836 (lto-streamer-out.o): Likewise.
21837 * cfgloop.c (init_loops_structure): Export, add struct function
21838 argument and adjust.
21839 (flow_loops_find): Adjust.
21840 * cfgloop.h (enum loop_estimation): Add EST_LAST.
21841 (init_loops_structure): Declare.
21842 * lto-streamer-in.c: Include cfgloop.h.
21843 (input_cfg): Input the loop tree.
21844 * lto-streamer-out.c: Include cfgloop.h.
21845 (output_cfg): Output the loop tree.
21846 (output_struct_function_base): Do not drop PROP_loops.
21848 2013-03-26 Richard Biener <rguenther@suse.de>
21850 * tree-cfg.c (execute_build_cfg): Build the loop tree.
21851 (pass_build_cfg): Provide PROP_loops.
21852 (move_sese_region_to_fn): Remove loops that are outlined into fn
21854 * tree-inline.c: Include cfgloop.h.
21855 (initialize_cfun): Do not drop PROP_loops.
21856 (copy_loops): New function.
21857 (copy_cfg_body): Copy loop structure.
21858 (tree_function_versioning): Initialize destination loop tree.
21859 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
21860 (pass_parallelize_loops): Do IL verification.
21861 * loop-init.c (loop_optimizer_init): Fixup loops if required.
21862 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
21863 the CFG make sure we fixup loops as well.
21864 * tree-ssa-tail-merge.c: Include cfgloop.h.
21865 (replace_block_by): When merging loop latches mark loops for fixup.
21866 * lto-streamer-out.c (output_struct_function_base): Drop
21867 PROP_loops for now.
21868 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
21869 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
21870 * ipa-split.c: Include cfgloop.h.
21871 (split_function): Add the new return block to the loop tree root.
21872 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
21873 whether we have removed the forwarder block.
21874 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
21875 * cfgloop.h (place_new_loop): Declare.
21876 * cfgloopmanip.c (place_new_loop): Export.
21877 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
21878 (tree-switch-conversion.o): Likewise.
21879 (tree-complex.o): Likewise.
21880 (tree-inline.o): Likewise.
21881 (tree-ssa-tailmerge.o): Likewise.
21882 (ipa-split.o): Likewise.
21883 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
21884 (tree-ssa-copy.o): Likewise.
21885 * tree-switch-conversion.c: Include cfgloop.h
21886 (process_switch): If we emit a bit-test cascade, schedule loops
21888 * tree-complex.c: Include cfgloop.h.
21889 (expand_complex_div_wide): Properly add new basic-blocks to loops.
21890 * asan.c: Include cfgloop.h.
21891 (create_cond_insert_point): Properly add new basic-blocks to
21892 loops, schedule loop fixup.
21893 * cfgloop.c (verify_loop_structure): Check that looks are not
21895 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
21897 (expand_omp_for_generic): Likewise.
21898 (expand_omp_sections): Likewise.
21899 (expand_omp_atomic_pipeline): Schedule loops for fixup.
21900 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
21901 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
21902 is initialized, not when loops are present.
21903 * tree-parloops.c (parallelize_loops): Remove checking here.
21904 * passes.c (init_optimization_passes): Schedule a copy-propagation
21905 pass before complete unrolling of inner loops.
21907 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21909 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
21910 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
21911 (colorize_init): Add argument to _WIN32 version.
21912 * toplev.c: Include diagnostic-color.h.
21913 (process_options): Default to -fdiagnostics-color=auto if
21914 GCC_COLORS env var is in the environment.
21915 * common.opt (fdiagnostics-color=): Add Var and Init.
21916 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
21917 env var is in the environment, the default is auto rather than never.
21919 * diagnostic.h (file_name_as_prefix): Add context argument.
21920 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
21921 the string as locus.
21922 * langhooks.c (lhd_print_error_function): Adjust caller.
21924 2013-04-25 Lawrence Crowl <crowl@google.com>
21926 * var-tracking.c (shared_hash_def::htab):
21927 Change type to hash_table. Update dependent calls and types.
21929 2013-04-25 Lawrence Crowl <crowl@google.com>
21931 * Makefile.in: Update as needed below.
21933 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
21934 Move declaration to after the type's method definitons.
21936 * attribs.c (htab_t scoped_attributes::attribute_hash):
21937 Change type to hash_table. Update dependent calls and types.
21939 * bitmap.c (htab_t bitmap_desc_hash):
21940 Change type to hash_table. Update dependent calls and types.
21942 * cselib.c (htab_t cselib_hash_table):
21943 Change type to hash_table. Update dependent calls and types.
21945 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
21946 (hash_string_slot_node): Move implementation into lto-streamer.h
21947 struct string_slot_hasher.
21948 (eq_string_slot_node): Likewise.
21950 * data-streamer-out.c: Update output_block::string_hash_table
21951 dependent calls and types.
21953 * dwarf2cfi.c (htab_t trace_index):
21954 Change type to hash_table. Update dependent calls and types.
21956 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
21957 Change type to hash_table. Update dependent calls and types.
21958 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
21959 (htab_t optimize_external_refs::map): Likewise.
21960 (htab_t output_comp_unit::extern_map): Likewise.
21961 (htab_t output_comdat_type_unit::extern_map): Likewise.
21962 (htab_t output_macinfo::macinfo_htab): Likewise.
21963 (htab_t optimize_location_lists::htab): Likewise.
21964 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
21966 * except.c (htab_t ehspec_hash_type):
21967 Change type to hash_table. Update dependent calls and types.
21968 (assign_filter_values::ttypes): Likewise.
21969 (assign_filter_values::ehspec): Likewise.
21970 (sjlj_assign_call_site_values::ar_hash): Likewise.
21971 (convert_to_eh_region_ranges::ar_hash): Likewise.
21973 * gcse.c (htab_t pre_ldst_table):
21974 Change type to hash_table. Update dependent calls and types.
21976 * ggc-common.c (htab_t saving_htab):
21977 Change type to hash_table. Update dependent calls and types.
21978 (htab_t loc_hash): Likewise.
21979 (htab_t ptr_hash): Likewise.
21980 (call_count): Rename ggc_call_count.
21981 (call_alloc): Rename ggc_call_alloc.
21982 (loc_descriptor): Rename make_loc_descriptor.
21983 (add_statistics): Rename ggc_add_statistics.
21985 * ggc-common.c (saving_htab):
21986 Change type to hash_table. Update dependent calls and types.
21988 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
21989 (push_gimplify_context): Likewise.
21990 (pop_gimplify_context): Likewise.
21991 (struct gimple_temp_hash_elt): Added.
21992 (struct gimplify_hasher): Likewise.
21993 (struct gimplify_ctx.temp_htab):
21994 Change type to hash_table. Update dependent calls and types.
21996 * gimple-fold.c: Include gimplify-ctx.h.
21998 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
21999 Change type to hash_table. Update dependent calls and types.
22000 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
22001 avoid potential global name collision.
22003 * gimplify.c: Include gimplify-ctx.h.
22004 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
22005 (htab_t gimplify_ctx::temp_htab):
22006 Update dependent calls and types for new type hash_table.
22007 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
22008 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
22010 * gimplify-ctx.h: New.
22011 (struct gimple_temp_hash_elt): Move from gimplify.c.
22012 (class gimplify_hasher): New.
22013 (struct gimplify_ctx): Move from gimple.h.
22014 (htab_t gimplify_ctx::temp_htab):
22015 Change type to hash_table. Update dependent calls and types.
22017 * graphite-clast-to-gimple.c: Include graphite-htab.h.
22018 (htab_t ivs_params::newivs_index):
22019 Change type to hash_table. Update dependent calls and types.
22020 (htab_t ivs_params::params_index): Likewise.
22021 (htab_t print_generated_program::params_index): Likewise.
22022 (htab_t gloog::newivs_index): Likewise.
22023 (htab_t gloog::params_index): Likewise.
22025 * graphite.c: Include graphite-htab.h.
22026 4htab_t graphite_transform_loops::bb_pbb_mapping):
22027 Change type to hash_table. Update dependent calls and types.
22029 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
22030 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
22031 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
22033 * graphite-dependences.c: Include graphite-htab.h.
22034 (loop_is_parallel_p): Change hash table type of parameter.
22036 * graphite-htab.h: New.
22037 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
22038 (extern find_pbb_via_hash): Move from graphite-poly.h.
22039 (extern loop_is_parallel_p): Move from graphite-poly.h.
22040 (extern get_loop_body_pbbs): Move from graphite-poly.h.
22042 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
22043 (extern loop_is_parallel_p): Move to graphite-htab.h.
22044 (extern get_loop_body_pbbs): Move to graphite-htab.h.
22046 * haifa-sched.c (htab_t delay_htab):
22047 Change type to hash_table. Update dependent calls and types.
22048 (htab_t delay_htab_i2): Likewise.
22050 * ira-color.c (htab_t allocno_hard_regs_htab):
22051 Change type to hash_table. Update dependent calls and types.
22053 * ira-costs.c (htab_t cost_classes_htab):
22054 Change type to hash_table. Update dependent calls and types.
22056 * loop-invariant.c (htab_t merge_identical_invariants::eq):
22057 Change type to hash_table. Update dependent calls and types.
22059 * loop-iv.c (htab_t bivs):
22060 Change type to hash_table. Update dependent calls and types.
22062 * loop-unroll.c (htab_t opt_info::insns_to_split):
22063 Change type to hash_table. Update dependent calls and types.
22064 (htab_t opt_info::insns_with_var_to_expand): Likewise.
22066 * lto-streamer.h (struct string_slot): Move from data-streamer.h
22067 (struct string_slot_hasher): New.
22068 (htab_t output_block::string_hash_table):
22069 Change type to hash_table. Update dependent calls and types.
22071 * lto-streamer-in.c (freeing_string_slot_hasher): New.
22072 (htab_t file_name_hash_table):
22073 Change type to hash_table. Update dependent calls and types.
22075 * lto-streamer-out.c: Update output_block::string_hash_table dependent
22078 * lto-streamer.c (htab_t tree_htab):
22079 Change type to hash_table. Update dependent calls and types.
22081 * omp-low.c: Include gimplify-ctx.h.
22083 * passes.c (htab_t name_to_pass_map):
22084 Change type to hash_table. Update dependent calls and types.
22085 (pass_traverse): Rename to passes_pass_traverse.
22087 * plugin.c (htab_t event_tab):
22088 Change type to hash_table. Update dependent calls and types.
22090 * postreload-gcse.c (htab_t expr_table):
22091 Change type to hash_table. Update dependent calls and types.
22092 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
22094 * sese.c (debug_rename_map_1): Make extern.
22095 (htab_t copy_bb_and_scalar_dependences::rename_map):
22096 Change type to hash_table. Update dependent calls and types.
22098 * sese.h (extern debug_rename_map): Move to .c file.
22100 * store-motion.c (htab_t store_motion_mems_table):
22101 Change type to hash_table. Update dependent calls and types.
22103 * trans-mem.c (htab_t tm_new_mem_hash):
22104 Change type to hash_table. Update dependent calls and types.
22106 * tree-browser.c (htab_t TB_up_ht):
22107 Change type to hash_table. Update dependent calls and types.
22109 * tree-cfg.c (htab_t discriminator_per_locus):
22110 Change type to hash_table. Update dependent calls and types.
22112 * tree-complex.c: Include tree-hasher.h
22113 (htab_t complex_variable_components):
22114 Change type to hash_table. Update dependent calls and types.
22116 * tree-eh.c (htab_t finally_tree):
22117 Change type to hash_table. Update dependent calls and types.
22119 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
22120 struct int_tree_hasher.
22121 (extern int_tree_map_eq): Likewise.
22122 (uid_decl_map_hash): Removed.
22123 (extern decl_tree_map_eq): Likewise.
22125 * tree-hasher.h: New.
22126 (struct int_tree_hasher): New.
22127 (typedef int_tree_htab_type): New.
22129 * tree-inline.c: Include gimplify-ctx.h.
22131 * tree-mudflap.c: Include gimplify-ctx.h.
22133 * tree-parloops.c: Include tree-hasher.h.
22134 (htab_t eliminate_local_variables_stmt::decl_address):
22135 Change type to hash_table. Update dependent calls and types.
22136 (htab_t separate_decls_in_region::decl_copies): Likewise.
22138 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
22139 Change type to hash_table. Update dependent calls and types.
22141 * tree-sra.c (candidates):
22142 Change type to hash_table. Update dependent calls and types.
22144 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
22146 (int_tree_map_hash): Likewise.
22148 * tree-ssa-dom.c (htab_t avail_exprs):
22149 Change type to hash_table. Update dependent calls and types.
22151 * tree-ssa-live.c (var_map_base_init::tree_to_index):
22152 Change type to hash_table. Update dependent calls and types.
22154 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
22155 Change type to hash_table. Update dependent calls and types.
22157 * tree-ssa-phiopt.c (seen_ssa_names):
22158 Change type to hash_table. Update dependent calls and types.
22160 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
22161 Change type to hash_table. Update dependent calls and types.
22163 * tree-ssa-uncprop.c (equiv):
22164 Change type to hash_table. Update dependent calls and types.
22166 2013-04-25 Jakub Jelinek <jakub@redhat.com>
22168 PR rtl-optimization/57003
22169 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
22170 call note_stores with kill_clobbered_value callback again after
22171 killing regs_invalidated_by_call.
22173 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22175 * config/aarch64/aarch64-simd.md
22176 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
22177 (aarch64_simd_bsl<mode>): Likewise.
22178 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
22180 2013-04-25 Marek Polacek <polacek@redhat.com>
22182 PR tree-optimization/57066
22183 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
22185 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22187 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
22189 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22191 * config/aarch64/aarch64-builtins.c
22192 (aarch64_fold_builtin): New.
22193 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
22194 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
22195 * config/aarch64/aarch64-simd-builtins.def (abs): New.
22196 * config/aarch64/arm_neon.h
22197 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
22199 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22200 Tejas Belagod <tejas.belagod@arm.com>
22202 * config/aarch64/aarch64-builtins.c
22203 (aarch64_gimple_fold_builtin): New.
22204 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
22205 * config/aarch64/aarch64-simd-builtins.def (addv): New.
22206 * config/aarch64/aarch64-simd.md (addpv4sf): New.
22207 (addvv4sf): Update.
22208 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
22210 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22212 * config/aarch64/aarch64.md
22213 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
22215 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22217 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
22218 (*ngcsi_uxtw): New pattern.
22220 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22221 Julian Brown <julian@codesourcery.com>
22223 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
22224 (TB_DREG): Add T_V4HF.
22225 (v4hf_UP): New macro.
22226 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22227 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22228 Handle initialisation of V4HF. Adjust initialisation of reinterpret
22230 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22231 (arm_vector_mode_supported_p): Handle V4HF.
22232 (arm_mangle_map): Handle V4HFmode.
22233 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
22234 * config/arm/arm_neon_builtins.def: Add entries for
22235 vcvtv4hfv4sf, vcvtv4sfv4hf.
22236 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
22237 (neon_vcvtv4hfv4sf): Likewise.
22238 * config/arm/neon-gen.ml: Handle half-precision floating point
22240 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
22241 * config/arm/arm_neon.h: Regenerate.
22242 * config/arm/neon.ml (type elts): Add F16.
22243 (type vectype): Add T_float16x4, T_floatHF.
22244 (type vecmode): Add V4HF.
22245 (type features): Add Requires_FP_bit feature.
22246 (elt_width): Handle F16.
22247 (elt_class): Likewise.
22248 (elt_of_class_width): Likewise.
22249 (mode_of_elt): Refactor.
22250 (type_for_elt): Handle F16, fix error messages.
22251 (vectype_size): Handle T_float16x4.
22252 (vcvt_sh): New function.
22253 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
22254 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
22255 (string_of_mode): Handle V4HF.
22256 * doc/arm-neon-intrinsics.texi: Regenerate.
22258 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22260 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
22261 format specifier in 'X' case.
22263 2013-04-25 Alan Modra <amodra@gmail.com>
22266 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
22267 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
22268 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
22269 Repeat for many other rotate/shift and mask patterns using subregs.
22270 Name lshiftrt insns.
22271 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
22272 on WORDS_BIG_ENDIAN.
22274 2013-04-25 Alan Modra <amodra@gmail.com>
22276 * config.gcc: Support little-endian powerpc-linux targets.
22277 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
22278 (LINK_OS_LINUX_SPEC): Define.
22279 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
22280 Preserve MASK_LITTLE_ENDIAN.
22281 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
22282 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
22283 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
22284 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
22285 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
22286 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
22287 Correct fp word order for little-endian. Don't shift toc entries
22288 smaller than a word for little-endian.
22289 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
22290 (bswapdi2 splits): Correct low-part subreg for little-endian.
22291 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
22292 low/high where such is correct only for be.
22293 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
22294 little-endian for -mcall-aixdesc.
22296 2013-04-25 Alan Modra <amodra@gmail.com>
22298 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
22299 replace_equiv_address_nv.
22301 2013-04-25 Alan Modra <amodra@gmail.com>
22303 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
22305 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22308 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22309 * rtl.h (struct rtx_def): ...
22311 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22313 PR rtl-optimizations/57046
22314 * lra-constraints (split_reg): Set up lra_risky_transformations_p
22315 for multi-reg splits.
22317 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
22319 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
22321 2013-04-24 Sterling Augustine <saugustine@google.com>
22323 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
22324 (comp_dir_string, debug_str_dwo_section): New.
22325 (DEBUG_STR_DWO_SECTION): Rename to ...
22326 (DEBUG_DWO_STR_SECTION): ... this.
22327 (DEBUG_NORM_STR_SECTION): Delete.
22328 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
22329 (DEBUG_STR_DWO_SECTION_FLAGS): New.
22330 (find_AT_string): Move most logic to ...
22331 (find_AT_string_in_table): ... here. New.
22332 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
22333 add_skeleton_AT_string. Delete logic.
22334 (output_skeleton_debug_sections): Remove call to
22335 add_top_level_skeleton_die_attrs.
22336 (add_comp_dir_attribute): Move logic to comp_dir_string.
22337 (dwarf2out_init): Initialize debug_str_dwo_section.
22338 (output_indirect_string): Call find_string_form.
22339 (output_indirect_strings): Rewrite.
22340 (prune_unused_types): Empty skeleton_debug_str_hash.
22341 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
22342 (dwarf2out_finish): Call output_indirect_strings.
22344 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22346 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
22348 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22350 * rtl.h (struct rtx_def): Add comment for field jump.
22351 (LRA_SUBREG_P): New macro.
22352 * recog.c (register_operand): Check LRA_SUBREG_P.
22353 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22355 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22356 (lra_final_code_change): Skip subreg change for operators.
22357 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22358 if there are no operand changes.
22359 * lra-constraints.c (curr_insn_set): New.
22360 (match_reload): Set LRA_SUBREG_P.
22361 (emit_spill_move): Ditto.
22362 (check_and_process_move): Use curr_insn_set. Process only single
22363 set insns. Don't initialize sec_mem_p and change_p.
22364 (simplify_operand_subreg): Use LRA_SUBREG_P.
22365 (reg_in_class_p): New function.
22366 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22367 of #ifdef. Add code to remove cycling.
22368 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22369 non-null disp. Reload inner instead of disp when base and index
22370 are null. Try to put lo_sum into register.
22371 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22372 (check_and_process_move): Move code for move cost check to
22373 simple_move_p. Remove equiv_substitution.
22374 (simple_move_p): New function.
22375 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22376 curr_insn_set. Call check_and_process_move only for single set
22377 insns. Use the new function. Move call of check_and_process_move
22378 after operand equiv substitution and address process.
22380 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
22382 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
22383 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
22384 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
22386 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22388 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
22390 2013-04-24 Marek Polacek <polacek@redhat.com>
22392 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
22393 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
22394 (select_loops_exit_conditions): Likewise.
22395 (number_of_iterations_for_all_loops): Likewise.
22396 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
22397 (scev_analysis): Likewise.
22399 2013-04-02 Catherine Moore <clm@codesourcery.com>
22400 Chao-ying Fu <fu@mips.com>
22402 * config/mips/micromips.md (jraddiusp): New pattern.
22403 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
22404 instruction if possible.
22406 2013-04-24 Alan Modra <amodra@gmail.com>
22408 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
22410 2013-04-24 Julian Brown <julian@codesourcery.com>
22411 Chung-Lin Tang <cltang@codesourcery.com>
22413 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
22414 dependency behavior in enumeration type DIE generation. Add TODO note
22415 to comments about future DW_FORM_sdata/udata re-work of related code.
22417 2013-04-23 Lawrence Crowl <crowl@google.com>
22419 * Makefile.in: Update as needed below.
22421 * hash-table.h (class hash_table):
22422 Correct many methods with parameter types compare_type to the correct
22423 value_type. (Correct code was unlikely to notice the change.)
22424 (hash_table::elements_with_deleted) New.
22425 (class hashtable::iterator): New.
22426 (hashtable::begin()): New.
22427 (hashtable::end()): New.
22428 (FOR_EACH_HASH_TABLE_ELEMENT): New.
22430 * statistics.c (statistics_hashes):
22431 Change type to hash_table. Update dependent calls and types.
22433 * tree-into-ssa.c (var_infos):
22434 Change type to hash_table. Update dependent calls and types.
22436 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
22437 Change type to hash_table. Update dependent calls and types.
22439 * tree-ssa-loop-im.c (struct mem_ref.refs):
22440 Change type to hash_table. Update dependent calls and types.
22442 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
22443 Change type to hash_table. Update dependent calls and types.
22445 * tree-ssa-sccvn.c (vn_tables_s::nary):
22446 Change type to hash_table. Update dependent calls and types.
22447 (vn_tables_s::phis): Likewise.
22448 (vn_tables_s::references): Likewise.
22450 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
22451 (vn_reference_eq): Update parameter and return types.
22453 * tree-ssa-structalias.c (pointer_equiv_class_table):
22454 Change type to hash_table. Update dependent calls and types.
22455 (location_equiv_class_table): Likewise.
22457 * tree-vect-data-refs.c: Consequential changes for making
22458 peeling a hash_table.
22460 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
22461 (destroy_loop_vec_info): Dependent hash_table update.
22463 * tree-vectorizer.h (peeling_htab):
22464 Change type to hash_table. Update dependent calls and types.
22466 2013-04-23 Shiva Chen <shiva0217@gmail.com>
22468 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
22469 to check the register content is equal or not.
22470 * lra-constraints.c (match_reload): Use lra_assign_reg_val
22471 to assign register content record.
22472 * lra-eliminations.c (update_reg_eliminate): Use
22473 lra_update_reg_val_offset to update register content offset.
22474 * lra-int.h (struct lra_reg): Add offset member.
22475 (lra_reg_val_equal_p): New static inline function.
22476 (lra_update_reg_val_offset): New static inline function.
22477 (lra_assign_reg_val): New static inline function.
22478 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
22479 to assign register content record.
22480 (initialize_lra_reg_info_element): Initial offset to zero.
22482 2013-04-23 Catherine Moore <clm@codesourcery.com>
22484 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
22485 operands. Record compression.
22487 2013-04-23 Xinliang David Li <davidxl@google.com>
22489 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
22491 2013-04-23 Richard Biener <rguenther@suse.de>
22493 PR middle-end/57036
22494 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
22495 parameter, only add abnormal goto edges from the copied body
22496 if the call could perform abnormal gotos.
22497 (copy_cfg_body): Adjust.
22499 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
22501 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
22503 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
22505 * coretypes.h (gimple_stmt_iterator): Add struct to make
22508 2013-04-23 Richard Biener <rguenther@suse.de>
22510 PR tree-optimization/57026
22511 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
22512 from SSA names occuring in abnormal PHI nodes.
22514 2013-04-22 Andi Kleen <ak@linux.intel.com>
22516 * lto/lto.c (print_lto_report_1): Fix LTO report names.
22518 2013-04-22 Andi Kleen <ak@linux.intel.com>
22520 * lto/lto.c (print_lto_report_1): Declare early.
22521 (read_cgraph_and_symbols): Call print_lto_report_1 early.
22523 2013-04-22 Andi Kleen <ak@linux.intel.com>
22525 * common.opt (-flto-report-wpa): Add.
22526 * doc/invoke.texi (-flto-report-wpa): Add.
22527 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
22530 2013-04-22 Xinliang David Li <davidxl@google.com>
22532 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
22533 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
22534 * Makefile.in: New dependency
22536 David Daney <ddaney.cavm@gmail.com>
22538 * configure.ac (gcc_cv_as_micromips_support): Use the
22539 --fatal-warnings option.
22540 * configure: Regenerate.
22542 2013-04-22 Marek Polacek <polacek@redhat.com>
22545 * tsan.c (instrument_expr): Don't instrument expression
22546 in case its size is zero.
22548 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
22552 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22554 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22556 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22558 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
22559 (gimple_stmt_iterator): New typedef.
22560 * gimple.h (gimple_stmt_iterator): Rename to...
22561 (gimple_stmt_iterator_d): ... This.
22562 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
22563 trees be valid for GIMPLE and GENERIC.
22564 (TARGET_GIMPLE_FOLD_BUILTIN): New.
22565 * gimple-fold.c (gimple_fold_call): Call target hook
22566 gimple_fold_builtin.
22567 * hooks.c (hook_bool_gsiptr_false): New.
22568 * hooks.h (hook_bool_gsiptr_false): New.
22569 * target.def (fold_stmt): New.
22570 * doc/tm.texi: Regenerate.
22572 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
22575 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
22576 a set sp if no stack realignment.
22578 2013-04-22 Nick Clifton <nickc@redhat.com>
22580 * config.gcc (tilegx-linux): Extend extra_objs rather than
22582 (tilepro-linux): Likewise.
22584 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22586 * config/aarch64/aarch64-builtins.c
22588 (CF0, CF1, CF2, CF3, CF4, CF10): New.
22589 (VAR<1-12>): Add MAP parameter.
22590 (BUILTIN_*): Likewise.
22591 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
22592 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
22593 (aarch64_ushl_n<mode>): Likewise.
22594 (aarch64_sshr_n<mode>): Likewise.
22595 (aarch64_ushr_n<mode>): Likewise.
22596 (aarch64_<maxmin><mode>): Likewise.
22597 (aarch64_sqrt<mode>): Likewise.
22598 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
22599 (vshr<q>_n_*): Likewise.
22601 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22603 * config/aarch64/aarch64-builtins.c
22604 (aarch64_simd_builtin_type_mode): Handle SF types.
22606 (BUILTIN_GPF): Define.
22607 (aarch64_init_simd_builtins): Handle SF types.
22608 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
22609 (frecps): Likewise.
22610 (frecpx): Likewise.
22611 * config/aarch64/aarch64-simd.md
22612 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
22613 (aarch64_frecpe<mode>): New.
22614 (aarch64_frecps<mode>): Likewise.
22615 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
22616 (v8type): Add frecp<esx>.
22617 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
22618 (aarch64_frecps<mode>): Likewise.
22619 * config/aarch64/iterators.md (FRECP): New.
22620 (frecp_suffix): Likewise.
22621 * config/aarch64/arm_neon.h
22622 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
22624 2013-04-22 Christian Bruel <christian.bruel@st.com>
22627 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
22628 (REG_CLASS_NAMES): Idem.
22629 (REG_CLASS_CONTENTS): Idem.
22630 (REGCLASS_HAS_FP_REG): Idem.
22631 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
22632 (sh_conditional_register_usage): Idem.
22634 2013-04-21 Jeff Law <law@redhat.com>
22636 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
22637 (ssa_forward_propagate_and_combine): Use it.
22639 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22641 * lra.c: Update the flow chart diagram.
22643 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22645 PR rtl-optimization/56847
22646 * lra-constraints.c (process_alt_operands): Discourage alternative
22647 with non-matche doffsettable memory constraint fro memory with
22650 2013-04-19 Richard Biener <rguenther@suse.de>
22652 PR tree-optimization/56982
22653 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
22655 * gimplify.c (gimplify_call_expr): Notice special calls.
22656 (gimplify_modify_expr): Likewise.
22657 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
22658 abnormal control flow receivers.
22659 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
22660 in the same way as cfun->has_nonlocal_labels.
22661 (gimple_purge_dead_abnormal_call_edges): Likewise.
22662 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
22663 receivers start a basic-block.
22665 2013-04-19 Richard Biener <rguenther@suse.de>
22667 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
22669 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
22670 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
22671 (SLP_TREE_LOAD_PERMUTATION): Add.
22672 (vect_transform_slp_perm_load): Adjust prototype.
22673 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
22674 (vect_free_slp_instance): Likewise.
22675 (vect_create_new_slp_node): Likewise.
22676 (vect_supported_slp_permutation_p): Remove.
22677 (vect_slp_rearrange_stmts): Adjust.
22678 (vect_supported_load_permutation_p): Likewise. Inline
22679 vect_supported_slp_permutation_p here.
22680 (vect_analyze_slp_instance): Compute load permutations per
22681 slp node instead of per instance.
22682 (vect_get_slp_defs): Adjust.
22683 (vect_transform_slp_perm_load): Likewise.
22684 (vect_schedule_slp_instance): Remove redundant code.
22685 (vect_schedule_slp): Remove hack for PR56270, add it ...
22686 * tree-vect-stmts.c (vectorizable_load): ... here, do not
22687 CSE loads for SLP. Adjust.
22689 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22691 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
22692 spelling in two comments.
22694 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22697 * config/arm/arm.c (load_multiple_sequence): Require SP
22698 as base register for loads if SP is in the register list.
22700 2013-04-19 Martin Jambor <mjambor@suse.cz>
22702 PR tree-optimization/56718
22703 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
22704 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
22705 and made public. Adjusted all callers.
22706 (ipa_intraprocedural_devirtualization): New function.
22707 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
22708 (ipa_intraprocedural_devirtualization): Likewise.
22709 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
22711 2013-04-19 Richard Biener <rguenther@suse.de>
22713 PR tree-optimization/57000
22714 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
22716 2013-04-19 Terry Guo <terry.guo@arm.com>
22718 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
22720 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
22721 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
22722 (cortex_m4_fmacs): Use new reservations.
22723 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
22725 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
22727 PR rtl-optimization/56999
22728 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
22730 (lra_coalesce): Remove split_origin_bitmap and related code.
22731 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
22732 ranges if necessary.
22734 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
22736 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
22738 (ix86_expand_call): Remove clobbered_registers array and use
22739 x86_64_ms_sysv_extra_clobbered_registers instead.
22740 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
22742 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
22744 * config/i386/i386.md (*call_rex64_ms_sysv): Use
22745 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
22746 (*call_value_rex64_ms_sysv): Ditto.
22748 2013-04-18 Cary Coutant <ccoutant@google.com>
22750 * dwarf2out.c (output_pubnames): Check die_perennial_p of
22751 parent instead of die_mark.
22753 2013-04-18 Diego Novillo <dnovillo@google.com>
22755 * gimple.c (create_gimple_tmp): New.
22756 (get_expr_type): New.
22757 (build_assign): New.
22758 (build_type_cast): New.
22759 * gimple.h (enum ssa_mode): Define.
22760 (gimple_seq_set_location): New.
22761 * asan.c (build_check_stmt): Change some gimple_build_* calls
22762 to use build_assign and build_type_cast.
22764 2013-04-18 Richard Biener <rguenther@suse.de>
22766 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
22767 handle negative step. Remove redundant checks.
22768 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
22769 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
22770 for negative step and grouped loads fail to vectorize.
22772 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22774 * emit-rtl.c (reset_insn_used_flags): New function.
22775 (reset_all_used_flags): Use it.
22776 (verify_insn_sharing): New function.
22777 (verify_rtl_sharing): Fix verification for SEQUENCEs.
22779 2013-04-18 Jakub Jelinek <jakub@redhat.com>
22781 PR tree-optimization/56984
22782 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
22783 and (x >> M) >= N don't register any assertion if N << M is the
22786 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22788 * lower-subreg.c (resolve_simple_move): If called self-recursive,
22789 do not delete_insn insns that have not yet been emitted, only
22790 unlink them with remove_insn.
22791 * df-scan.c (df_insn_delete): Revert r197492.
22793 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
22795 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
22796 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
22798 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22800 * config/arm/arm.md (movsicc_insn): Convert define_insn into
22801 define_insn_and_split.
22802 (and_scc,ior_scc,negscc): Likewise.
22803 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
22805 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22807 * config/arm/arm.c (use_return_insn): Return 0 for targets that
22808 can benefit from using a sequence of LDRD instructions in epilogue
22809 instead of a single LDM instruction.
22811 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
22814 * doc/extend.texi: Fix typo.
22816 2013-04-17 Richard Biener <rguenther@suse.de>
22818 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
22819 (vect_build_slp_tree): ... here.
22820 (vect_build_slp_tree_1): Compute which stmts of the SLP group
22821 match. Remove special-casing of mismatched complex loads.
22822 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
22823 re-try the match with swapped commutative operands.
22824 (vect_supported_load_permutation_p): Remove special-casing of
22825 mismatched complex loads.
22826 (vect_analyze_slp_instance): Adjust.
22828 2013-04-17 Richard Biener <rguenther@suse.de>
22830 PR rtl-optimization/56921
22831 * cfgloop.h (struct loop): Add simple_loop_desc member.
22832 (struct niter_desc): Mark with GTY(()).
22833 (simple_loop_desc): Do not use aux field but simple_loop_desc.
22834 * loop-iv.c (get_simple_loop_desc): Likewise.
22835 (free_simple_loop_desc): Likewise.
22838 2013-04-16 Richard Biener <rguenther@suse.de>
22840 PR rtl-optimization/56921
22841 * loop-init.c (pass_rtl_move_loop_invariants): Add
22842 TODO_do_not_ggc_collect to todo_flags_finish.
22843 (pass_rtl_unswitch): Same.
22844 (pass_rtl_unroll_and_peel_loops): Same.
22845 (pass_rtl_doloop): Same.
22847 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
22849 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
22850 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
22851 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
22853 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
22854 * tree-streamer.c (record_common_node): Adjust reference in comment.
22856 2013-04-17 Terry Guo <terry.guo@arm.com>
22858 * config/arm/cortex-m4.md: Add a new bypass.
22860 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22862 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
22864 (*subs_<optab><mode>_multp2): New pattern.
22865 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22866 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22868 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22870 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
22871 (*subs_mul_imm_<mode>): New pattern.
22873 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
22876 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
22877 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
22878 (vsx_movti_32bit): Change j->wa to O->wa.
22880 2013-04-16 Richard Biener <rguenther@suse.de>
22882 PR rtl-optimization/56921
22883 * loop-init.c (pass_rtl_move_loop_invariants): Add
22884 TODO_do_not_ggc_collect to todo_flags_finish.
22885 (pass_rtl_unswitch): Same.
22886 (pass_rtl_unroll_and_peel_loops): Same.
22887 (pass_rtl_doloop): Same.
22889 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
22891 * config/arm/arm.c (emit_multi_reg_push): New declaration
22892 for an existing function.
22893 (arm_emit_strd_push): New function.
22894 (arm_expand_prologue): Used here.
22895 (arm_emit_ldrd_pop): New function.
22896 (arm_expand_epilogue): Used here.
22897 (arm_get_frame_offsets): Update condition.
22898 (arm_emit_multi_reg_pop): Add a special case for load of a single
22899 register with writeback.
22901 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
22903 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
22906 2013-04-16 Richard Biener <rguenther@suse.de>
22908 PR tree-optimization/56756
22909 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
22910 (first_mem_ref_loc): New.
22911 (execute_sm): Place the load temporarily before a previous
22912 access instead of in the latch edge to ensure its SSA dependencies
22913 are defined at points dominating the load.
22915 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
22917 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
22918 correct fix by moving header and footer insn to the footer of
22919 the merged basic block. Clear BB_END of the merged-away block.
22921 PR middle-end/43631
22922 * emit-rtl.c (make_note_raw): New function.
22923 (link_insn_into_chain): New static inline function.
22924 (add_insn): Use it.
22925 (add_insn_before, add_insn_after): Factor insn chain linking code...
22926 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
22927 using link_insn_into_chain.
22928 (note_outside_basic_block_p): New helper function for emit_note_after
22929 and emit_note_before.
22930 (emit_note_after): Use nobb variant of add_insn_after if the note
22931 should not be contained in a basic block.
22932 (emit_note_before): Use nobb variant of add_insn_before if the note
22933 should not be contained in a basic block.
22934 (emit_note_copy): Use make_note_raw.
22935 (emit_note): Likewise.
22936 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
22937 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
22938 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
22939 the moved barrier the tail of the basic block it follows.
22940 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
22942 2013-04-15 Jakub Jelinek <jakub@redhat.com>
22944 PR tree-optimization/56962
22945 * gimple-ssa-strength-reduction.c (record_increment): Only set
22946 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
22947 either rhs1 or rhs2 is equal to c->base_expr.
22949 2013-04-15 Richard Biener <rguenther@suse.de>
22951 PR tree-optimization/56933
22952 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
22954 (GROUP_READ_WRITE_DEPENDENCE): Remove.
22955 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
22956 * tree-vect-data-refs.c (vect_analyze_group_access): Move
22957 dependence check ...
22958 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
22960 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
22961 GROUP_READ_WRITE_DEPENDENCE.
22963 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22965 * emit-rtl.c (reset_all_used_flags): New function.
22966 (verify_rtl_sharing): Call reset_all_used_flags before and after
22967 performing the checks.
22969 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22971 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
22972 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
22973 * config/arm/constraints.md (De): New constraint.
22974 * config/arm/neon.md (anddi3_neon): Delete.
22975 (neon_vand<mode>): Expand to standard anddi3 pattern.
22976 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
22977 Move earlier in the file.
22978 (neon_inv_logic_op2): Likewise.
22979 (arm_anddi_operand_neon): New predicate.
22981 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22983 * configure.ac (gcc_cv_ld_as_needed): Set
22984 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
22985 Use -z ignore, -z record on *-*-solaris2*.
22986 (HAVE_LD_AS_NEEDED): Update comment.
22987 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
22988 * configure: Regenerate.
22989 * config.in: Regenerate.
22990 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
22991 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
22992 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
22993 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
22994 equivalents. Fix markup.
22995 * doc/tm.texi: Regenerate.
22997 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
22999 * config/i386/i386.opt: New option mstack-protector-guard=.
23000 * config/i386/i386-opts.h: Add enum stack_protector_guard.
23001 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
23002 TARGET_SSP_TLS_GUARD.
23003 * config/i386/i386.c (ix86_option_override_internal): Set
23004 ix86_stack_protector_guard.
23005 * config/i386/i386.md (stack_protect_set): Enable for
23006 TARGET_SSP_TLS_GUARD only.
23007 (stack_protect_set_<mode>): Ditto.
23008 (stack_protect_test): Ditto.
23009 (stack_protect_test_<mode>): Ditto.
23010 * doc/invoke.texi (i386 Option): Document.
23012 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
23015 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
23016 (S_MODES): Set H_MODE bit.
23017 (SF_MODES): Set only S_MODE and SF_MODE bits.
23018 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
23019 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
23020 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
23021 <MODE_FLOAT>: Likewise.
23023 2013-04-15 Joey Ye <joey.ye@arm.com>
23025 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
23027 2013-04-15 Joey Ye <joey.ye@arm.com>
23029 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
23031 (thumb_far_jump_used_p): Count instruction size and set
23034 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
23036 * reorg.c (fill_simple_delay_slots): Reindent block of code.
23037 * resource.c (mark_target_live_regs): Reformat conditional block.
23039 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23041 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
23042 notes, they are emitted only just before final.
23043 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
23045 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23047 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
23048 * cfgrtl.c (delete_insn): Call it here instead.
23049 * lra-spills.c (lra_final_code_change): Use delete_insn.
23050 * haifa-sched.c (sched_remove_insn): Likewise.
23051 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
23052 returning to the nop pool.
23053 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
23054 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
23056 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23058 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
23059 * doc/tm.texi: Regenerated.
23061 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
23063 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
23066 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23068 * df-core.c (df_find_def): Compare register numbers.
23069 (df_find_use): Likewise.
23071 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
23074 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
23075 lra_in_progress for return.
23077 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23079 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
23080 define_insn into define_insn_and_split and emit movsicc patterns.
23082 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23084 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
23086 2013-04-12 Richard Biener <rguenther@suse.de>
23088 * tree-pass.h (TODO_do_not_ggc_collect): New.
23089 * passes.c (execute_one_ipa_transform_pass): Honor
23090 TODO_do_not_ggc_collect.
23091 (execute_one_pass): Likewise.
23094 2013-04-10 Richard Biener <rguenther@suse.de>
23096 * passes.c (init_optimization_passes): Remove reload pass.
23097 * ira.c (do_reload): Merge into ...
23099 (rest_of_handle_reload): Remove.
23100 (pass_reload): Likewise.
23101 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23102 of reload for vzeroupper pass placement.
23104 2013-04-12 Jakub Jelinek <jakub@redhat.com>
23106 PR tree-optimization/56918
23107 PR tree-optimization/56920
23108 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
23109 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
23110 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
23111 use wide_mul_with_sign method.
23113 2013-04-12 Richard Biener <rguenther@suse.de>
23115 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
23116 not be considered a gimple constant.
23118 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
23120 * fold-const.c (const_binop): Handle vector shifts by a scalar.
23121 (fold_binary_loc): Call const_binop also for mixed vector-scalar
23124 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
23125 Jakub Jelinek <jakub@redhat.com>
23127 * opts.c: Include diagnostic-color.h.
23128 (common_handle_option): Handle OPT_fdiagnostics_color_.
23129 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
23130 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
23131 (diagnostic-color.o): New.
23132 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
23133 (diagnostic_color_rule): New enum.
23134 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
23135 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
23136 the location string.
23137 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
23138 either NULL, or color kind.
23139 * diagnostic-color.c: New file.
23140 * diagnostic-color.h: New file.
23141 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
23143 * doc/invoke.texi (-fdiagnostics-color): Document.
23144 * pretty-print.h (pp_show_color): Define.
23145 (struct pretty_print_info): Add show_color field.
23146 * diagnostic.c: Include diagnostic-color.h.
23147 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
23148 macros. Colorize error:, warning: etc. strings and also the location
23150 (diagnostic_show_locus): Colorize the caret line.
23151 * pretty-print.c: Include diagnostic-color.h.
23152 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
23153 inside of %< %> quotes or quoted through q format modifier.
23155 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23157 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
23159 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23161 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
23162 code in CC_NZ mode.
23163 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
23166 2013-04-11 Marek Polacek <polacek@redhat.com>
23168 PR tree-optimization/48184
23169 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
23171 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
23173 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
23174 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
23175 (skip_simple_arithmetic): Tidy up.
23176 * tree.h (skip_simple_constant_arithmetic): Declare.
23178 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23180 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
23182 2013-04-11 Richard Biener <rguenther@suse.de>
23184 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23185 generate vector constants.
23187 2013-04-11 Richard Biener <rguenther@suse.de>
23189 PR tree-optimization/56878
23190 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
23191 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
23193 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
23194 Prefer to align the DR with the most invariant base address.
23196 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23198 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
23201 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
23203 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
23204 floating-point vector comparisons against 0.
23206 2013-04-11 Jakub Jelinek <jakub@redhat.com>
23208 PR tree-optimization/56899
23209 * fold-const.c (extract_muldiv_1): Apply distributive law
23210 only if TYPE_OVERFLOW_WRAPS (ctype).
23212 2013-04-11 Bin Cheng <bin.cheng@arm.com>
23215 * ira-costs.c (scan_one_insn): Check whether the source rtx of
23216 loading has side effect.
23218 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
23220 * config/sparc/sparc.c: Include tree-pass.h.
23221 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
23222 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
23223 head of file. Change return type. Split off gate function.
23224 (sparc_gate_work_around_errata): New function.
23225 (pass_work_around_errata): New pass definition.
23226 (insert_pass_work_around_errata) New pass insert definition to
23227 insert pass_work_around_errata just after delayed-branch scheduling.
23228 (sparc_option_override): Insert the pass.
23229 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
23231 2013-04-10 David S. Miller <davem@davemloft.net>
23233 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
23234 or -mcpu=hypersparc.
23236 * target.def (cstore_mode): New hook.
23237 * target.h: Include insn-codes.h
23238 * targhooks.c: Likewise.
23239 (default_cstore_mode): New function.
23240 * targhooks.h: Declare it.
23241 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
23242 * doc/tm.texi: Rebuild.
23243 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
23244 target hook, rather than inspecting the insn_data.
23245 * config/sparc/sparc.c (sparc_cstore_mode): New function.
23246 (TARGET_CSTORE_MODE): Redefine.
23247 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
23249 * config/sparc/predicates.md (cstore_result_operand): New special
23251 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
23252 Use it for operand 0.
23253 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
23254 (*snesi_special): Likewise.
23255 (*snesi_zero): Likewise.
23256 (*seqsi_zero): Likewise.
23257 (*sltu_insn): Likewise.
23258 (*sgeu_insn): Likewise.
23259 (*seqdi_special): Make operand 0 and comparison operation be of
23261 (*snedi_special): Likewise.
23262 (*snedi_special_vis3): Likewise.
23263 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
23264 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
23265 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
23266 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
23267 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
23268 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
23269 (*sltu_extend_sp64): Likewise.
23270 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
23271 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
23272 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
23273 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
23274 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
23276 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
23278 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
23279 (aarch64_start_file): Use the new function.
23281 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23282 Jason Merrill <jason@redhat.com>
23284 * common.opt: Add -gdwarf.
23285 * opts.c (common_handle_option): Handle it.
23286 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
23288 2013-04-10 Richard Biener <rguenther@suse.de>
23290 * passes.c (execute_todo): Do not call ggc_collect conditional here.
23291 (execute_one_ipa_transform_pass): But unconditionally here.
23292 (execute_one_pass): And here.
23293 (init_optimization_passes): Remove reload pass.
23294 * tree-pass.h (TODO_ggc_collect): Remove.
23295 (pass_reload): Likewise.
23296 * ira.c (do_reload): Merge into ...
23298 (rest_of_handle_reload): Remove.
23299 (pass_reload): Likewise.
23300 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23301 of reload for vzeroupper pass placement.
23302 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
23303 and todo_flags_finish of all passes.
23305 2013-04-10 Richard Biener <rguenther@suse.de>
23307 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
23308 first_const_oprnd field, rename first_def_type to first_op_type.
23309 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
23310 (vect_get_and_check_slp_defs): Always use the type of the
23311 operand. Allow mixed vect_external_def, vect_constant_def types.
23312 (vect_get_constant_vectors): Handle mixed vect_external_def,
23313 vect_constant_def types.
23315 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
23317 PR tree-optimization/55524
23318 * tree-ssa-math-opts.c
23319 (convert_mult_to_fma): Don't use an fms construct
23320 when we don't have an fms operation, but fnma, and it looks
23321 likely that we'll be able to use the latter.
23323 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
23325 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
23327 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
23328 inline fail caused by overwritable functions.
23330 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
23332 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
23333 unnecessary bits in the constant power of two case.
23335 2013-04-10 Richard Biener <rguenther@suse.de>
23337 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
23338 broken code swapping operands.
23339 (vect_build_slp_tree): Do not compute load permutations here.
23340 (vect_analyze_slp_instance): Compute load permutations here,
23341 after building the SLP tree.
23343 2013-04-09 Christian Bruel <christian.bruel@st.com>
23345 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
23346 of next/prev_real_insn.
23348 2013-04-09 Jan Hubicka <jh@suse.cz>
23350 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
23351 Drop aliased parameter.
23352 (function_and_variable_visibility): Do not handle alias pairs.
23353 * cgraph.c (varpool_externally_visible_p): Update prototype.
23354 * varpool.c (varpool_add_new_variable): Update.
23356 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23358 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
23360 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
23362 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
23364 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
23366 2013-04-09 Marek Polacek <polacek@redhat.com>
23368 PR tree-optimization/48762
23369 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
23371 2013-04-09 Richard Biener <rguenther@suse.de>
23373 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
23375 (vect_build_slp_tree): Likewise.
23376 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
23377 calculating the cost of a SLP instance.
23378 (vect_analyze_slp_instance): Use it from here, after building
23381 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23383 PR middle-end/56883
23384 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
23385 expand_omp_for_static_chunk): Use simple_p = true in
23386 force_gimple_operand_gsi calls when assigning to addressable decls.
23388 2013-04-09 Jeff Law <law@redhat.com>
23390 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
23391 when the boolean was created by converting a wider object which
23392 had a boolean range.
23394 2013-04-09 Richard Biener <rguenther@suse.de>
23396 * tree-vectorizer.h (slp_void_p): Remove.
23397 (slp_tree): Typedef before _slp_tree declaration.
23398 (struct _slp_tree): Use a vector of slp_tree as children.
23399 (vect_get_place_in_interleaving_chain): Remove.
23400 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
23402 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
23404 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
23405 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
23406 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
23407 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
23408 Use slp_node instead of slp_void_p and adjust.
23410 2013-04-09 Richard Biener <rguenther@suse.de>
23412 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
23413 work that is not necessary.
23415 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23417 PR tree-optimization/56854
23418 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
23419 forward into clobber stmts if it would change MEM_REF lhs into
23422 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
23424 * tree.c (type_hash_lookup, type_hash_add): Make static.
23425 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
23427 2013-04-09 Richard Biener <rguenther@suse.de>
23429 * tree.h (unsave_expr_now): Remove.
23430 * tree-inline.c (mark_local_for_remap_r): Remove.
23431 (unsave_expr_1): Likewise.
23432 (unsave_r): Likewise.
23433 (unsave_expr_now): Likewise.
23434 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
23435 (propagate_tree_value): Likewise.
23437 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
23439 * doc/rtl.texi (sequence): Rewrite documentation to match the
23440 current use of SEQUENCE rtl objects.
23441 * rtl.def (SEQUENCE): Likewise.
23443 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
23444 Update documentation.
23445 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
23446 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
23448 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
23450 2013-04-08 Teresa Johnson <tejohnson@google.com>
23452 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
23453 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
23455 (estimate_edge_size_and_time): Add comment to suggest using rounding
23457 (estimate_node_size_and_time): Ditto.
23458 (remap_edge_change_prob): Use helper rounding divide methods.
23459 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
23460 (gimple_mod_pow2_value_transform): Ditto.
23461 (gimple_mod_subtract_transform): Ditto.
23462 (gimple_ic_transform): Ditto.
23463 (gimple_stringops_transform): Ditto.
23464 * stmt.c (conditional_probability): Ditto.
23465 (emit_case_dispatch_table): Ditto.
23466 * lto-cgraph.c (merge_profile_summaries): Ditto.
23467 * tree-optimize.c (execute_fixup_cfg): Ditto.
23468 * cfgcleanup.c (try_forward_edges): Ditto.
23469 * cfgloopmanip.c (scale_loop_profile): Ditto.
23471 (duplicate_loop_to_header_edge): Ditto.
23472 (lv_adjust_loop_entry_edge): Ditto.
23473 * tree-vect-loop.c (vect_transform_loop): Ditto.
23474 * profile.c (compute_branch_probabilities): Ditto.
23475 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
23476 * lto-streamer-in.c (input_cfg): Ditto.
23477 * gimple-streamer-in.c (input_bb): Ditto.
23478 * ipa-cp.c (update_profiling_info): Ditto.
23479 (update_specialized_profile): Ditto.
23480 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
23481 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
23483 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23484 (compute_trg_info): Ditto.
23485 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
23486 (purge_dead_edges): Ditto.
23487 * loop-unswitch.c (unswitch_loop): Ditto.
23488 * cgraphclones.c (cgraph_clone_edge): Ditto.
23489 (cgraph_clone_node): Ditto.
23490 * tree-inline.c (copy_bb): Ditto.
23491 (copy_edges_for_bb): Ditto.
23492 (initialize_cfun): Ditto.
23493 (copy_cfg_body): Ditto.
23494 (expand_call_inline): Ditto.
23496 2013-04-08 Kai Tietz <ktietz@redhat.com>
23498 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
23499 TARGET_CYGWIN64 by TARGET_64BIT.
23501 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23503 * config/epiphany/epiphany.md (GPR_1): New constant.
23504 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
23505 * config/epiphany/epiphany.c (gen_compare_reg):
23506 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
23507 is already in place.
23508 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
23509 Don't require being called during rtl expansion; If y operlaps r0,
23511 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
23512 (epiphany_expand_epilogue): Likewise.
23514 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
23515 Don't use CC_FPmode for ORDERED / UNORDERED.
23516 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
23518 * config/epiphany/constraints.md (CnL): New constraint.
23519 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
23520 * config/epiphany/predicates.md (add_operand): Allow 1024.
23522 * config/epiphany/epiphany.md (logical_op): New code iterator.
23523 (op_mnc): New code attribute.
23524 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
23525 (mov_f+1, mov_f+2): New peephole2 patterns.
23527 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
23528 (cstoresi4): Also allow re-use of zero result when doing a NE
23529 comparison to a non-zero operand.
23530 Use (clobber (scratch)) for first insn if the gpr output is not needed.
23532 * config/epiphany/epiphany.md (<insn_opname>v2si3):
23533 Use gen_addsi3_i / gen_subsi3_i.
23535 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23539 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
23540 contain anything but clobbers, at most one __builtin_stack_restore,
23541 optionally debug stmts and final resx, and if it has at least one
23542 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
23543 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
23544 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
23545 which isn't defaut definition, remove them.
23546 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
23547 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
23548 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
23549 with MEM_REF LHS with SSA_NAME address.
23551 2013-04-08 Jeff Law <law@redhat.com>
23553 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
23555 2013-04-08 Richard Biener <rguenther@suse.de>
23557 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
23559 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
23560 determined vector type.
23561 (vect_analyze_data_refs): Likewise.
23562 (vect_get_new_vect_var): Adjust.
23563 (vect_create_destination_var): Preserve SSA name versions.
23564 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
23565 not dump anything here.
23567 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23569 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
23570 Add member lr_slot_known.
23571 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
23573 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
23574 Remove code that sets lr_slot_offset according to what a previous
23575 version of epiphany_emit_save_restore used to do.
23576 (epiphany_emit_save_restore): When doing an lr save or restore,
23577 set/verify lr_slot_known and lr_slot_offset.
23579 2013-04-08 Xinyu Qi <xyqi@marvell.com>
23582 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
23585 2013-04-08 Richard Biener <rguenther@suse.de>
23587 * alias.c (find_base_term): Fix thinko in previous change.
23589 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23591 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
23592 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
23593 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
23594 if possible to compute val.
23595 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
23596 For QImode integers don't require anything about precision. Use
23597 const_with_all_bytes_same to find out if the constant doesn't have
23598 repeated bytes in it.
23600 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23602 * config/s390/s390.c (s390_expand_insv): Only accept insertions
23605 2013-04-08 Marek Polacek <polacek@redhat.com>
23607 PR rtl-optimization/48182
23608 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
23611 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23614 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
23615 nuses, make sure we have a label.
23617 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23620 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
23621 (rs6000_emit_swdiv_low_precision): Remove.
23622 (rs6000_emit_swdiv): Rewrite to handle between one and four
23623 iterations of Newton-Raphson generally; modify required number of
23624 iterations for some cases.
23625 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
23627 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
23629 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
23630 set-but-unused variable.
23632 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
23633 basic blocks of released function bodies garbage-collectable.
23635 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
23636 (struct rtl_opt_pass): Add TODO_df_finish.
23638 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
23640 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23642 * config/arm/constraints.md (q): New constraint.
23643 * config/arm/ldrdstrd.md: New file.
23644 * config/arm/arm.md (ldrdstrd.md) New include.
23645 (arm_movdi): Use "q" instead of "r" constraint
23646 for double-word memory access.
23647 (movdf_soft_insn): Likewise.
23648 * config/arm/vfp.md (movdi_vfp): Likewise.
23649 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
23650 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
23651 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
23652 (mem_ok_for_ldrd_strd): Likewise.
23653 (output_move_double): Update assertion.
23655 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23657 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
23659 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23661 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
23662 define_insn_and_split.
23663 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
23665 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23667 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
23668 define_insn_and_split.
23669 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
23670 (shiftsi3_compare): New pattern.
23671 (rrx): New pattern.
23672 * config/arm/unspecs.md (UNSPEC_RRX): New.
23674 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23676 * config/arm/arm.md (negdi_extendsidi): New pattern.
23677 (negdi_zero_extendsidi): Likewise.
23679 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23681 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
23682 define_insn_and_split.
23683 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
23684 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
23686 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23688 * config/arm/arm.md (arm_subdi3): Convert define_insn into
23689 define_insn_and_split.
23690 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
23691 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
23693 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23695 * config/arm/arm.md (subsi3_carryin): New pattern.
23696 (subsi3_carryin_const): Likewise.
23697 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
23698 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
23700 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23702 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
23704 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23706 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
23707 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
23709 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23711 * config/arm/arm.c (arm_expand_builtin): Change fcode
23712 type to unsigned int.
23714 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23716 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
23718 2013-04-04 Ian Lance Taylor <iant@google.com>
23720 * doc/standards.texi (Standards): The Go frontend supports the Go 1
23723 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
23725 PR middle-end/56729
23726 * df-scan.c (df_insn_delete): Disable failing assert.
23728 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23730 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
23731 New function prototype.
23732 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
23733 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
23734 (arm_builtin_vectorized_function): New function.
23736 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23738 * config/arm/arm_neon_builtins.def: New file.
23739 * config/arm/arm.c (neon_builtin_data): Move contents to
23740 arm_neon_builtins.def.
23741 (enum arm_builtins): Include neon builtin definitions.
23742 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
23743 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
23745 2013-04-04 Marek Polacek <polacek@redhat.com>
23747 PR tree-optimization/48186
23748 * predict.c (maybe_hot_frequency_p): Return false if
23749 HOT_BB_FREQUENCY_FRACTION is 0.
23750 (cgraph_maybe_hot_edge_p): Likewise.
23752 2013-04-04 Richard Biener <rguenther@suse.de>
23754 PR tree-optimization/56826
23755 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
23758 2013-04-04 Richard Biener <rguenther@suse.de>
23760 PR tree-optimization/56213
23761 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
23762 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
23764 2013-04-04 Richard Biener <rguenther@suse.de>
23766 PR tree-optimization/56837
23767 * tree-loop-distribution.c (classify_partition): For non-zero
23768 values require that the value has the same precision as its
23769 mode to be useful as memset value.
23771 2013-04-03 Nick Clifton <nickc@redhat.com>
23773 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
23774 (fmssf4): Use fmsf.s on E3V5 architectures.
23775 (fnmasf4): Use fnmaf.s on E3V5 architectures.
23776 (fnmssf4): Use fnmsf.s on E3V5 architectures.
23778 2013-04-03 Jeff Law <law@redhat.com>
23780 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
23781 (lra-eliminations.o): Likewise.
23783 2013-04-03 Teresa Johnson <tejohnson@google.com>
23785 * gcov-io.c (compute_working_sets): Moved most of body of old
23786 compute_working_sets here from profile.c.
23787 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
23788 (gcov_working_set_t): Moved typedef here from basic-block.h
23789 (compute_working_set): Declare.
23790 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
23791 (get_working_sets): Renamed from compute_working_set,
23792 replace most of body with call to new compute_working_sets.
23793 (get_exec_counts): Replace call to compute_working_sets
23794 to get_working_sets.
23795 * profile.h (get_working_sets): Renamed from compute_working_set.
23796 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
23797 to get_working_sets.
23798 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
23799 * gcov-dump.c (dump_working_sets): New function.
23801 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
23803 * hwint.c (sext_hwi, zext_hwi): New functions.
23804 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
23805 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
23806 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
23807 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
23808 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
23809 (sext_hwi, zext_hwi): New functions.
23811 2013-04-03 Jeff Law <law@redhat.com>
23813 PR tree-optimization/56799
23814 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
23815 back test for widening conversion erroneously dropped in prior change.
23817 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23820 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
23821 instead of next_real_insn.
23823 2013-04-03 Marek Polacek <polacek@redhat.com>
23826 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
23828 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23831 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
23833 (thumb1_output_casesi): Likewise.
23834 (thumb2_output_casesi): Likewise.
23836 2013-04-03 Richard Biener <rguenther@suse.de>
23838 PR tree-optimization/56817
23839 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
23841 (tree_unroll_loops_completely_1): ... new function to manually
23842 walk the loop tree, properly defering outer loops of unrolled
23843 loops to later iterations.
23845 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23847 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
23848 (vectorizable_load): Likewise.
23849 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
23850 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
23852 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23854 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
23857 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
23859 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
23861 2013-04-03 Bin Cheng <bin.cheng@arm.com>
23863 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
23865 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23867 PR tree-optimization/56790
23868 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
23871 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23873 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
23875 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
23876 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
23879 2013-04-03 Jakub Jelinek <jakub@redhat.com>
23882 * tree.h (force_folding_builtin_constant_p): New decl.
23883 * builtins.c (force_folding_builtin_constant_p): New variable.
23884 (fold_builtin_constant_p): Fold immediately also if
23885 force_folding_builtin_constant_p.
23887 2013-04-03 Richard Biener <rguenther@suse.de>
23889 PR tree-optimization/56812
23890 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
23891 DRs of the same interleaving chain are independent.
23893 2013-04-02 Jason Merrill <jason@redhat.com>
23895 * gdbinit.in (pbb): Use debug fn.
23897 2013-04-02 Lawrence Crowl <crowl@google.com>
23899 * sese.h (struct ivtype_map_elt_s): Remove unused.
23900 (extern debug_ivtype_map): Remove unused.
23901 (extern eq_ivtype_map_elts): Remove unused.
23902 * sese.c (debug_ivtype_map): Removed unused.
23903 (debug_ivtype_map_1): Removed unused.
23904 (debug_ivtype_elt): Remove unused.
23905 (eq_ivtype_map_elts): Remove unused.
23908 2013-04-02 Kai Tietz <ktietz@redhat.com>
23911 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
23912 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
23913 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
23915 (legitimize_pe_coff_symbol): Likewise.
23916 (is_imported_p): New helper-function.
23917 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
23918 for Windows x64 targets.
23919 (ix86_expand_prologue): Optimize for pe-coff targets.
23920 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
23921 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
23922 medium/large code-model.
23923 (legitimize_pic_address): Likewise.
23924 (legitimize_tls_address): Likewise.
23925 (ix86_expand_call): Likewise.
23926 (x86_output_mi_thunk): Likewise.
23927 (get_dllimport_decl): Add new beimport argument.
23928 (construct_plt_address): Don't assert for x64 pe-coff targets.
23929 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
23931 (SYMBOL_FLAG_STUBVAR): New macro.
23932 (SYMBOL_REF_STUBVAR_P): Likewise.
23933 * config/i386/winnt.c (stub_list): New structure.
23934 (stub_head): New local variable.
23935 (i386_pe_record_stub): New function.
23936 (i386_pe_file_end): Emit refptr-stubs.
23938 2013-04-02 Jakub Jelinek <jakub@redhat.com>
23940 PR rtl-optimization/56745
23941 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
23942 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
23945 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
23946 and both of them are MEM_REFs, just compare first argument for
23947 equality and attempt to deal even with differing offsets.
23950 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
23951 of gimple_clobber_p to be MEM_REF.
23952 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
23953 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
23954 after gimplification.
23955 * asan.c (get_mem_ref_of_assignment): Don't instrument
23956 gimple_clobber_p stmts.
23957 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
23958 gimple_clobber_p stmt if they have MEM_REF lhs and
23959 are dead because of another gimple_clobber_p stmt.
23960 * tree-ssa-live.c (clear_unused_block_pointer): Treat
23961 gimple_clobber_p stmts like debug stmts.
23962 (remove_unused_locals): Remove clobbers with MEM_REF lhs
23963 that refer to unused VAR_DECLs or uninitialized values.
23964 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
23965 gimple_clobber_p stmts if they refer to removed parameters.
23966 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
23969 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
23971 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
23972 using SWI48 mode attribute.
23974 2013-04-02 Wei Mi <wmi@google.com>
23976 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
23977 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
23978 *<rotate_insn><mode>3_mask in i386.md.
23980 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
23982 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
23984 2013-04-02 Richard Biener <rguenther@suse.de>
23986 PR tree-optimization/56778
23987 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
23988 Runtime alias tests are not supported for gather loads.
23989 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
23990 stmts referenced from SSA operands before updating SSA form.
23992 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
23993 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23995 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
23996 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
23997 * config/arm/cortex-a53.md: New file.
23998 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
23999 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
24000 * config/arm/arm.c (arm_issue_rate): Likewise.
24001 * config/arm/arm-tune.md: Regenerate
24002 * config/arm/arm-tables.opt: Regenerate.
24003 * config/arm/arm-cores.def: Add cortex-a53.
24005 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
24007 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
24010 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24012 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
24013 scalar load/store operations using B/H registers.
24014 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
24016 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24018 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
24020 * config/aarch64/aarch64.c
24021 (aarch64_simd_scalar_immediate_valid_for_move): New.
24022 * config/aarch64/aarch64-protos.h
24023 (aarch64_simd_scalar_immediate_valid_for_move): New.
24024 * config/aarch64/constraints.md (Dh, Dq): New.
24025 * config/aarch64/iterators.md (hq): New.
24027 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
24029 * reorg.c (get_branch_condition): Deal with conditional returns.
24030 (fill_simple_delay_slots): Remove dead code dealing with jumps.
24032 2013-04-01 Wei Mi <wmi@google.com>
24034 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
24035 Truncate operand 2 using %b asm operand modifier.
24036 (*<shift_insn><mode>3_mask): Ditto.
24037 (*<rotate_insn><mode>3_mask): Ditto.
24039 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
24041 PR middle-end/56798
24042 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
24044 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
24046 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
24048 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
24050 2013-03-30 Lawrence Crowl <crowl@google.com>
24052 * dse.c (clear_alias_sets): Remove never set.
24053 (disqualified_clear_alias_sets): Remove never set.
24054 (clear_alias_mode_pool): Remove never set.
24055 (dse_step0): Remove condition that is never true.
24056 (canon_address): Remove condition that is never true.
24057 (dse_step7): Remove condition that is never true.
24058 (rest_of_handle_dse): Remove condition that is never true.
24059 (rest_of_handle_dse::did_global): Remove never read from above.
24060 (dse_step2_spill): Remove never called from above.
24061 (dse_step5_spill): Remove never called from above.
24063 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
24065 * doc/md.texi (Standard Names) <casesi>: Update documentation for
24066 JUMP_TABLE_DATA changes.
24067 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
24068 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
24069 (Insns) <jump_table_data>: New entry.
24070 * doc/tm.texi: Regenerate.
24072 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
24074 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
24075 for table jump at the end of a basic block using tablejump_p.
24076 * targhooks.c (default_invalid_within_doloop): Likewise.
24077 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
24078 target hook implementation that is identical to the default hook.
24079 (rs6000_invalid_within_doloop): Remove.
24081 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
24082 unused variable from tablejump_p call.
24084 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
24085 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
24086 (INSN_DELETED_P): Likewise.
24087 (emit_jump_table_data): New prototype.
24088 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
24089 after 4th as unused.
24090 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
24091 * sched-vis.c (print_insn): Likewise.
24092 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
24093 insn for compatibility with back ends that use next_active_insn to
24094 identify jump table data.
24095 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
24096 (remove_insn): Likewise.
24097 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
24099 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
24100 (emit_jump_table_data): New function.
24102 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
24103 basic block, a JUMP_TABLE_DATA never is.
24104 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
24105 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
24106 off from code handling real insns.
24107 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
24108 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
24109 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
24110 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
24111 is not a NONDEBUG_INSN_P.
24112 * ira-costs.c (scan_one_insn): Likewise.
24113 * jump.c (mark_all_labels): Likewise.
24114 (mark_jump_label_1): Likewise.
24115 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24116 * lra.c (get_insn_freq): Expect all insns reaching here to be in
24118 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
24119 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
24120 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
24121 JUMP_TABLE_DATA_P insns.
24122 (calculate_elim_costs_all_insns): Likewise.
24123 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
24124 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
24125 (delete_output_reload): Code style fixups.
24126 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
24127 insn flags on this non-insn.
24128 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
24129 as scheduling barriers, for pre-change compatibility.
24130 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
24131 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
24133 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
24134 redundant JUMP_TABLE_DATA_P test.
24135 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
24136 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24137 (frv_for_each_packet): Likewise.
24138 * config/i386/i386.c (min_insn_size): Likewise.
24139 (ix86_avoid_jump_mispredicts): Likewise.
24140 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24141 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
24142 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24143 (mips16_insn_length): Robustify.
24144 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
24145 (mips16_split_long_branches): Likewise.
24146 * config/pa/pa.c (pa_combine_instructions): Likewise.
24147 * config/rs6000/rs6000.c (get_next_active_insn): Treat
24148 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
24149 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
24150 as contributing to pool range lengths.
24151 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
24152 Remove redundant JUMP_TABLE_DATA_P test.
24153 (sh_loop_align): Likewise.
24154 (split_branches): Likewise.
24155 (sh_insn_length_adjustment): Likewise.
24156 * config/spu/spu.c (get_branch_target): Likewise.
24158 2013-03-29 Jan Hubicka <jh@suse.cz>
24160 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
24161 gcov streaming; stream hot bb threshold to ltrans.
24162 * predict.c (get_hot_bb_threshold): Break out from ....
24163 (maybe_hot_count_p): ... here.
24164 (set_hot_bb_threshold): New function.
24165 * lto-section-in.c (lto_section_name): Add profile.
24166 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
24167 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
24168 and data-streamer.h
24169 (histogram_entry): New structure.
24170 (histogram, histogram_pool): New global vars.
24171 (histogram_hash): New structure.
24172 (histogram_hash::hash): New method.
24173 (histogram_hash::equal): Likewise.
24174 (account_time_size): New function.
24175 (cmp_counts): New function.
24176 (dump_histogram): New function.
24177 (ipa_profile_generate_summary): New function.
24178 (ipa_profile_write_summary): New function.
24179 (ipa_profile_read_summary): New function.
24180 (ipa_profile): Decide on threshold.
24181 (pass_ipa_profile): Add ipa_profile_write_summary and
24182 ipa_profile_read_summary.
24183 * Makefile.in (ipa.o): Update dependencies.
24184 * lto-streamer.h (LTO_section_ipa_profile): New section.
24186 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
24188 * tree.h (VAR_P): New.
24190 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
24193 * doc/invoke.texi ([-fwhole-program]): Fix typo.
24195 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
24197 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
24198 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24199 (control_flow_insn_p): Likewise.
24200 * cfgrtl.c (duplicate_insn_chain): Likewise.
24201 * final.c (get_attr_length_1): Likewise.
24202 (shorten_branches): Likewise.
24203 (final_scan_insn): Likewise.
24204 * function.c (instantiate_virtual_regs): Likewise.
24205 * gcse.c (insert_insn_end_basic_block): Likewise.
24206 * ira-costs.c (scan_one_insn): Likewise.
24207 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24208 * lra.c (check_rtl): Likewise.
24209 * reload1.c (elimination_costs_in_insn): Likewise.
24210 * reorg.c (follow_jumps): Likewise.
24212 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
24213 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24214 (thumb_far_jump_used_p): Likewise.
24215 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
24216 (workaround_speculation): Likewise.
24217 (add_sched_insns_for_speculation): Likewise.
24218 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
24219 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24220 (frv_for_each_packet): Likewise.
24221 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
24222 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
24223 (final_emit_insn_group_barriers): Likewise.
24224 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24225 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24226 (mips16_insn_length): Likewise.
24227 * config/pa/pa.c (pa_reorg): Likewise.
24228 (pa_combine_instructions): Likewise.
24229 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
24230 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
24231 (sh_reorg): Likewise.
24232 (split_branches): Likewise.
24233 * config/spu/spu.c (get_branch_target): Likewise.
24235 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
24238 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
24240 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
24241 Fix declaration name.
24243 2013-03-28 Lawrence Crowl <crowl@google.com>
24245 * graphds.h (struct graph.indicies): Remove unused.
24246 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
24247 (SCOP_ORIGINAL_PDDRS): Remove unused.
24248 * sese.h (extern insert_loop_close_phis): Removed unused.
24249 (extern insert_guard_phis): Removed unused.
24250 (extern ivtype_map_elt_info): Removed unused.
24251 (new_ivtype_map_elt): Removed unused.
24252 * sese.c (ivtype_map_elt_info): Removed unused.
24254 2013-03-28 Lawrence Crowl <crowl@google.com>
24256 * Makefile.in: Add several missing include dependences.
24258 (test-dump.o): New. This object is not added to any executable,
24259 but is present for ad-hoc testing.
24261 (debug (const bitmap_head_def &)): New.
24262 (debug (const bitmap_head_def *)): New.
24264 (extern debug (const bitmap_head_def &)): New.
24265 (extern debug (const bitmap_head_def *)): New.
24267 (debug (edge_def &)): New.
24268 (debug (edge_def *)): New.
24270 (debug (basic_block_def &)): New.
24271 (debug (basic_block_def *)): New.
24273 (dump_node (const_tree, int, FILE *)): Correct source file.
24275 (debug (die_struct &)): New.
24276 (debug (die_struct *)): New.
24278 (extern debug (die_struct &)): New.
24279 (extern debug (die_struct *)): New.
24280 * gimple-pretty-print.c
24281 (debug (gimple_statement_d &)): New.
24282 (debug (gimple_statement_d *)): New.
24283 * gimple-pretty-print.h
24284 (extern debug (gimple_statement_d &)): New.
24285 (extern debug (gimple_statement_d *)): New.
24287 (debug (ira_allocno_copy &)): New.
24288 (debug (ira_allocno_copy *)): New.
24289 (debug (ira_allocno &)): New.
24290 (debug (ira_allocno *)): New.
24292 (extern debug (ira_allocno_copy &)): New.
24293 (extern debug (ira_allocno_copy *)): New.
24294 (extern debug (ira_allocno &)): New.
24295 (extern debug (ira_allocno *)): New.
24297 (debug (live_range &)): New.
24298 (debug (live_range *)): New.
24300 (debug (lra_live_range &)): New.
24301 (debug (lra_live_range *)): New.
24303 (debug (lra_live_range &)): New.
24304 (debug (lra_live_range *)): New.
24306 (debug (omega_pb_d &)): New.
24307 (debug (omega_pb_d *)): New.
24309 (extern debug (omega_pb_d &)): New.
24310 (extern debug (omega_pb_d *)): New.
24312 (debug (const rtx_def &)): New.
24313 (debug (const rtx_def *)): New.
24315 (debug_tree (tree): Move within file.
24316 (debug_raw (const tree_node &)): New.
24317 (debug_raw (const tree_node *)): New.
24318 (dump_tree_via_hooks (const tree_node *, int)): New.
24319 (debug (const tree_node &)): New.
24320 (debug (const tree_node *)): New.
24321 (debug_verbose (const tree_node &)): New.
24322 (debug_verbose (const tree_node *)): New.
24323 (debug_head (const tree_node &)): New.
24324 (debug_head (const tree_node *)): New.
24325 (debug_body (const tree_node &)): New.
24326 (debug_body (const tree_node *)): New.
24327 (debug_vec_tree (tree): Move and reimplement in terms of dump.
24328 (debug (vec<tree, va_gc> &)): New.
24329 (debug (vec<tree, va_gc> *)): New.
24331 (extern debug (const rtx_def &)): New.
24332 (extern debug (const rtx_def *)): New.
24334 (debug_raw (simple_bitmap_def &)): New.
24335 (debug_raw (simple_bitmap_def *)): New.
24336 (debug (simple_bitmap_def &)): New.
24337 (debug (simple_bitmap_def *)): New.
24339 (extern debug (simple_bitmap_def &)): New.
24340 (extern debug (simple_bitmap_def *)): New.
24341 (extern debug_raw (simple_bitmap_def &)): New.
24342 (extern debug_raw (simple_bitmap_def *)): New.
24344 (debug (vinsn_def &)): New.
24345 (debug (vinsn_def *)): New.
24346 (debug_verbose (vinsn_def &)): New.
24347 (debug_verbose (vinsn_def *)): New.
24348 (debug (expr_def &)): New.
24349 (debug (expr_def *)): New.
24350 (debug_verbose (expr_def &)): New.
24351 (debug_verbose (expr_def *)): New.
24352 (debug (vec<rtx> &)): New.
24353 (debug (vec<rtx> *)): New.
24355 (extern debug (vinsn_def &)): New.
24356 (extern debug (vinsn_def *)): New.
24357 (extern debug_verbose (vinsn_def &)): New.
24358 (extern debug_verbose (vinsn_def *)): New.
24359 (extern debug (expr_def &)): New.
24360 (extern debug (expr_def *)): New.
24361 (extern debug_verbose (expr_def &)): New.
24362 (extern debug_verbose (expr_def *)): New.
24363 (extern debug (vec<rtx> &)): New.
24364 (extern debug (vec<rtx> *)): New.
24366 (_list_iter_cond_expr): Make inline instead of static.
24368 (debug (sreal &)): New.
24369 (debug (sreal *)): New.
24371 (extern debug (sreal &)): New.
24372 (extern debug (sreal *)): New.
24374 (extern debug_raw (const tree_node &)): New.
24375 (extern debug_raw (const tree_node *)): New.
24376 (extern debug (const tree_node &)): New.
24377 (extern debug (const tree_node *)): New.
24378 (extern debug_verbose (const tree_node &)): New.
24379 (extern debug_verbose (const tree_node *)): New.
24380 (extern debug_head (const tree_node &)): New.
24381 (extern debug_head (const tree_node *)): New.
24382 (extern debug_body (const tree_node &)): New.
24383 (extern debug_body (const tree_node *)): New.
24384 (extern debug (vec<tree, va_gc> &)): New.
24385 (extern debug (vec<tree, va_gc> *)): New.
24387 (debug (struct loop &)): New.
24388 (debug (struct loop *)): New.
24389 (debug_verbose (struct loop &)): New.
24390 (debug_verbose (struct loop *)): New.
24391 * tree-dump.c: Add header dependence.
24393 (extern debug (struct loop &)): New.
24394 (extern debug (struct loop *)): New.
24395 (extern debug_verbose (struct loop &)): New.
24396 (extern debug_verbose (struct loop *)): New.
24398 (debug (data_reference &)): New.
24399 (debug (data_reference *)): New.
24400 (debug (vec<data_reference_p> &)): New.
24401 (debug (vec<data_reference_p> *)): New.
24402 (debug (vec<ddr_p> &)): New.
24403 (debug (vec<ddr_p> *)): New.
24405 (extern debug (data_reference &)): New.
24406 (extern debug (data_reference *)): New.
24407 (extern debug (vec<data_reference_p> &)): New.
24408 (extern debug (vec<data_reference_p> *)): New.
24409 (extern debug (vec<ddr_p> &)): New.
24410 (extern debug (vec<ddr_p> *)): New.
24412 (debug (pt_solution &)): New.
24413 (debug (pt_solution *)): New.
24415 (extern debug (pt_solution &)): New.
24416 (extern debug (pt_solution *)): New.
24418 (debug (_var_map &)): New.
24419 (debug (_var_map *)): New.
24420 (debug (tree_live_info_d &)): New.
24421 (debug (tree_live_info_d *)): New.
24423 (extern debug (_var_map &)): New.
24424 (extern debug (_var_map *)): New.
24425 (extern debug (tree_live_info_d &)): New.
24426 (extern debug (tree_live_info_d *)): New.
24428 2013-03-28 Jan Hubicka <jh@suse.cz>
24430 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
24432 2013-03-28 Ian Bolton <ian.bolton@arm.com>
24434 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
24435 record only when desired or required.
24437 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
24439 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
24440 *vec_extractv2di_1_rex64. Use x64 isa attribute.
24442 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24444 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
24445 (*andsi3_compare0_uxtw): New pattern.
24446 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
24447 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
24449 2013-03-28 Jan Hubicka <jh@suse.cz>
24451 * data-streamer-in.c (streamer_read_gcov_count): New function.
24452 * gimple-streamer-out.c: Include value-prof.h.
24453 (output_gimple_stmt): Output histogram.
24454 (output_bb): Use streamer_write_gcov_count.
24455 * value-prof.c: Include data-streamer.h
24456 (dump_histogram_value): Add HIST_TYPE_MAX.
24457 (stream_out_histogram_value): New function.
24458 (stream_in_histogram_value): New function.
24459 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
24460 (stream_out_histogram_value, stream_in_histogram_value): Declare.
24461 * data-streamer-out.c (streamer_write_gcov_count): New function.
24462 (streamer_write_gcov_count_stream): New function.
24463 * lto-cgraph.c (lto_output_edge): Update counter streaming.
24464 (lto_output_node): Likewise.
24465 (input_node, input_edge): Likewise.
24466 * lto-streamer-out.c (output_cfg): Update streaming.
24467 * lto-streamer-in.c (input_cfg): Likewise.
24468 * data-streamer.h (streamer_write_gcov_count,
24469 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
24470 * gimple-streamer-in.c: Include value-prof.h
24471 (input_gimple_stmt): Input histograms.
24472 (input_bb): Update profile streaming.
24474 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24476 * genmodes.c (emit_max_int): New function.
24477 (emit_insn_modes_h): Added call to emit_max_function.
24478 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
24480 * machmode.def: Fixed comment.
24482 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24484 * combine.c (try_combine): Removed useless assert.
24485 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
24487 2013-03-28 Marek Polacek <polacek@redhat.com>
24488 Richard Biener <rguenther@suse.de>
24490 PR tree-optimization/56695
24491 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
24492 build signed result of a vector comparison.
24493 * tree-cfg.c (verify_gimple_comparison): Check that a result
24494 of a vector comparison has signed type.
24496 2013-03-28 Richard Biener <rguenther@suse.de>
24498 PR tree-optimization/37021
24499 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
24500 do not restrict gaps between groups.
24501 * tree-vect-stmts.c (vectorizable_load): Properly account for
24502 a gap between groups.
24504 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
24506 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
24507 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
24510 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
24512 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
24513 * doc/extend.texi (Named Address Spaces): Ditto.
24514 (Variable Attributes): Ditto.
24516 2013-03-27 Kai Tietz <ktietz@redhat.com>
24518 * config.build: Add support for cygwin x64 target.
24519 * config.gcc: Likewise.
24520 * config.host: Likewise.
24521 * configure.ac: Likewise
24522 * configure: Regenerated.
24524 2013-03-27 Kai Tietz <ktietz@redhat.com>
24526 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
24527 * config/i386/t-cygwin-w64: New file.
24528 * config/i386/cygwin-w64.h: New file.
24529 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
24530 and add support for x64-cygwin target.
24531 (CPP_SPEC): Likewise.
24532 (CXX_WRAP_SPEC_LIST): Undefine before define.
24533 (LIBGCJ_SONAME): Use 15 as version.
24535 2013-03-27 Richard Biener <rguenther@suse.de>
24537 PR tree-optimization/56716
24538 * tree-ssa-structalias.c (perform_var_substitution): Adjust
24539 dumping for ref nodes.
24541 2013-03-27 Martin Jambor <mjambor@suse.cz>
24543 PR tree-optimization/55334
24544 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
24545 restricted pointers to arrays.
24547 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
24549 * Makefile.in (.SUFFIXES): Add .cc.
24550 (.c.o): Apply same recipe for implicit rule .cc.o.
24552 2013-03-27 Richard Biener <rguenther@suse.de>
24554 PR tree-optimization/37021
24555 * tree-vect-data-refs.c (vect_check_strided_load): Allow
24556 REALPART/IMAGPART_EXPRs around the supported refs.
24557 * tree-ssa-structalias.c (find_func_aliases): Assume that
24558 floating-point values are not used to transfer pointers.
24560 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
24562 * target.def (TARGET_HAS_IFUNC_P): New target hook.
24563 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
24564 * doc/tm.texi: Regenerate.
24565 * targhooks.h (default_has_ifunc_p): New.
24566 * targhooks.c (default_has_ifunc_p): Ditto.
24567 * config/linux-protos.h: New file.
24568 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
24569 hook for linux which disables support of indirect functions in android.
24570 * config/linux-android.c: New file.
24571 * config/t-linux-android.c: Ditto.
24572 * config.gcc: Added new object file linux-android.o.
24573 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
24574 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
24575 * varasm.c (do_assemble_alias): Likewise.
24576 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
24577 doesn't support indirect functions.
24578 * configure: Regenerate.
24580 2013-03-27 Bin Cheng <bin.cheng@arm.com>
24583 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
24584 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
24587 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24589 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
24591 2013-03-27 Terry Guo <terry.guo@arm.com>
24593 * config/arm/arm-cores.def: Added core cortex-r7.
24594 * config/arm/arm-tune.md: Regenerated.
24595 * config/arm/arm-tables.opt: Regenerated.
24596 * doc/invoke.texi: Added entry for core cortex-r7.
24598 2013-03-27 Walter Lee <walt@tilera.com>
24600 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
24601 double-decrement of next_scratch_regno.
24603 2013-03-27 Walter Lee <walt@tilera.com>
24605 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
24607 (insn_v1mulus): Ditto.
24608 (insn_v2muls): Ditto.
24610 2013-03-27 Walter Lee <walt@tilera.com>
24612 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
24613 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
24615 2013-03-27 Walter Lee <walt@tilera.com>
24617 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
24618 (*sibcall_value): Ditto.
24620 2013-03-27 Walter Lee <walt@tilera.com>
24622 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
24623 (insn_mnz_v8qi): ... this ...
24624 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
24626 (insn_v<n>mnz): Replaced by ...
24627 (insn_v1mnz): ... this ...
24628 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
24630 (insn_mz_<mode>): Replaced by ...
24631 (insn_mz_v8qi): ... this ...
24632 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
24634 (insn_v<n>mz): Replaced by ...
24635 (insn_v1mz): ... this ...
24636 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
24639 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
24641 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
24643 2013-03-26 Roland McGrath <mcgrathr@google.com>
24645 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
24646 than fprintf with a non-constant, non-format string.
24648 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24650 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
24651 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
24652 operand 0 predicate.
24653 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
24654 attribute. Use general_x64nomem_operand as operand 1 predicate.
24655 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
24656 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24657 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
24658 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24659 (mov<mode>_insv_1): Remove expander. Merge insn with
24660 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
24661 Use general_x64nomem_operand as operand 1 predicate.
24662 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
24663 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
24664 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24665 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
24666 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
24667 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
24668 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24669 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
24670 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
24671 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
24672 (general_x64nomem_operand): Ditto.
24674 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
24676 * config/rtems.opt: Add -pthread option.
24678 2013-03-26 Richard Biener <rguenther@suse.de>
24680 * alias.c (find_base_term): Avoid redundant and not used recursion.
24681 (base_alias_check): Get the initial base term from the caller.
24682 (true_dependence_1): Compute and pass base terms to base_alias_check.
24683 (write_dependence_p): Likewise.
24684 (may_alias_p): Likewise.
24686 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
24688 * config/aarch64/aarch64.c (aarch64_classify_address): Support
24689 PC-relative load in SI modes and above only.
24691 2013-03-26 Xinyu Qi <xyqi@marvell.com>
24693 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
24694 * config/arm/iwmmxt.md (WCGR0): Update.
24695 (WCGR1, WCGR2, WCGR3): Likewise.
24697 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24699 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
24700 Use x64 and nox64 isa attributes.
24702 2013-03-26 Richard Biener <rguenther@suse.de>
24704 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
24705 alignment computations and rely on get_object_alignment_1
24706 for the !TYPE_P case.
24707 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
24709 2013-03-26 Walter Lee <walt@tilera.com>
24711 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
24712 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
24714 2013-03-25 Jeff Law <law@redhat.com>
24716 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
24717 check for INTEGRAL_TYPE_P that was missing due to checking in
24718 wrong version of prior patch.
24720 2013-03-25 Walter Lee <walt@tilera.com>
24722 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
24723 TILEGX_INSN_SHUFFLEBYTES1.
24724 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
24726 (tilegx_builtins): Ditto.
24727 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
24729 2013-03-25 Walter Lee <walt@tilera.com>
24731 * config/tilegx/tilegx.md (floatsisf2): New pattern.
24732 (floatunssisf2): New pattern.
24733 (floatsidf2): New pattern.
24734 (floatunssidf2): New pattern.
24736 2013-03-25 Walter Lee <walt@tilera.com>
24738 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
24739 tests for constraint J, K, N, P.
24741 2013-03-25 Walter Lee <walt@tilera.com>
24743 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
24744 Use indirect/pcrel encoding.
24745 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
24748 2013-03-25 Steve Ellcey <sellcey@mips.com>
24750 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
24751 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
24752 * config/mips/mips.c (mips_option_override): Set IMADD default.
24753 * config/mips/mips.h (PTF_AVOID_IMADD): New.
24754 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
24755 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
24756 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
24757 * doc/invoke.texi (-mimadd/-mno-imadd): New.
24759 2013-03-25 Jeff Law <law@redhat.com>
24761 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
24762 slightly to avoid creating and folding useless trees. Simplify
24763 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
24765 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
24767 * config/i386/i386.md (*zero_extendsidi2): Merge with
24768 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
24769 * config/i386/predicates.md (x86_64_zext_operand): Rename from
24770 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
24771 targets. Clarify comment.
24773 2013-03-25 Martin Jambor <mjambor@suse.cz>
24775 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
24776 pass-through jump functions differently.
24777 (ipa_read_jump_function): Likewise. Also use setter functions to set
24780 2013-03-25 Martin Jambor <mjambor@suse.cz>
24782 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
24783 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
24785 (ipa_get_indirect_edge_target): New function.
24786 (devirtualization_time_bonus): New parameter known_aggs, pass it to
24787 ipa_get_indirect_edge_target. Update all callers.
24788 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
24789 ipa_get_indirect_edge_target_1 instead of calling
24790 ipa_get_indirect_edge_target.
24791 (create_specialized_node): Pass aggvlas to
24792 ipcp_discover_new_direct_edges.
24794 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24796 * config/arm/arm.md (f_sels, f_seld): New types.
24797 (*cmov<mode>): New pattern.
24798 * config/arm/predicates.md (arm_vsel_comparison_operator): New
24801 2013-03-25 Kai Tietz <ktietz@redhat.com>
24803 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
24804 POSIX-printf for mingw-hosted builds.
24806 2013-03-25 Richard Biener <rguenther@suse.de>
24808 PR middle-end/56694
24809 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
24810 must-not-throw stmt location.
24812 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24814 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
24815 Emit load-acquire versions when acq is true.
24816 (arm_emit_store_exclusive): Add rel parameter.
24817 Emit store-release versions when rel is true.
24818 (arm_split_compare_and_swap): Use acquire-release instructions
24820 of barriers when appropriate.
24821 (arm_split_atomic_op): Likewise.
24822 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
24823 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
24824 (VUNSPEC_SLX): Likewise.
24825 (VUNSPEC_LDA): Likewise.
24826 (VUNSPEC_STL): Likewise.
24827 * config/arm/sync.md (atomic_load<mode>): New pattern.
24828 (atomic_store<mode>): Likewise.
24829 (arm_load_acquire_exclusive<mode>): Likewise.
24830 (arm_load_acquire_exclusivesi): Likewise.
24831 (arm_load_acquire_exclusivedi): Likewise.
24832 (arm_store_release_exclusive<mode>): Likewise.
24834 2013-03-25 Catherine Moore <clm@codesourcery.com>
24836 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
24837 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
24838 * config/mip/predicates.md (lwsp_swsp_operand,
24839 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
24840 sb16_operand, db4_operand, db7_operand, ib3_operand,
24841 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
24842 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
24843 andi16_operand): New predicates.
24844 * config/mips/mips.md (compression): New attribute.
24845 (enabled): New attribute.
24846 (length): Consider compression in computing length.
24847 (shift_compression): New code attribute.
24848 (*add<mode>3): New operands. Record compression.
24849 (sub<mode>3): Likewise.
24850 (one_cmpl<mode>2): Likewise.
24851 (*and<mode>3): Likewise.
24852 (*ior<mode>3): Likewise.
24853 (unnamed pattern for xor): Likewise.
24854 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
24855 (*<optab><mode>3): Likewise.
24856 (*mov<mode>_internal: Likewise.
24857 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
24858 (mips_unsigned_immediate_p): New.
24859 (umips_lwsp_swsp_address_p): New.
24860 (m16_based_address_p): New.
24861 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
24862 (mips_unsigned_immediate_p): New prototype.
24863 (lwsp_swsp_address_p): New prototype.
24864 (m16_based_address_p): New prototype.
24865 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
24866 (mips_signed_immediate_p): New function.
24867 (m16_based_address_p): New function.
24868 (lwsp_swsp_address_p): New function.
24869 (mips_print_operand_punctuation): Recognize short delay slot insns
24870 for microMIPS.add<mode>3"
24872 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24875 * config/arm/iterators.md (v_cmp_result): New mode attribute.
24876 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
24878 2013-03-25 Richard Biener <rguenther@suse.de>
24880 PR tree-optimization/56689
24881 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
24884 2013-03-25 Richard Biener <rguenther@suse.de>
24886 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
24888 (memory_references): Likewise.
24889 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
24890 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
24891 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
24892 (gather_mem_refs_in_loops): Fold into ...
24893 (analyze_memory_references): ... this. Move initialization
24894 to tree_ssa_lim_initialize.
24895 (fill_always_executed_in): Rename to ...
24896 (fill_always_executed_in_1): ... this.
24897 (fill_always_executed_in): Move contains_call computation to
24898 this new function from ...
24899 (tree_ssa_lim_initialize): ... here.
24900 (tree_ssa_lim): Call fill_always_executed_in.
24902 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24904 * postreload.c (reload_combine): Fix code detecting returns.
24906 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24908 * function.c (emit_use_return_register_into_block): On cc0 targets,
24909 do not emit the sequence between cc0 setter and user.
24911 2013-03-25 Kai Tietz <ktietz@redhat.com>
24913 * config/i386/predicates.md (local_symbolic_operand): Interpret
24914 dll-imported symbols as none-local.
24916 2013-03-25 Richard Biener <rguenther@suse.de>
24918 * tree-ssa-loop-im.c (struct depend): Remove.
24919 (struct lim_aux_data): Make depends a vec of gimples.
24920 (free_lim_aux_data): Adjust.
24921 (add_dependency): Likewise.
24922 (set_level): Likewise.
24924 2013-03-25 Richard Biener <rguenther@suse.de>
24926 PR middle-end/56434
24927 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
24928 the pointer returned by calls with ECF_MALLOC set.
24930 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24932 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
24934 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24936 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
24937 using MMXMODE mode iterator.
24938 (*move<mode>_internal): Merge with *movv2sf_internal and
24939 *movv2sf_internal_rex64 using MMXMODE mode iterator.
24941 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24943 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
24944 (record_last_mem_set_info): Likewise.
24946 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
24947 of XNEWVEC followed by memset.
24948 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
24950 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24952 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
24953 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
24954 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
24955 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
24956 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
24957 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
24958 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
24959 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
24960 BARRIER_P instead of GET_CODE.
24962 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24964 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
24965 inaccuracy in the probing code.
24967 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
24968 (ctrapdi4): Likewise.
24970 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24972 * calls.c (expand_call): Add missing guard to code handling return
24973 of non-BLKmode structures in MSB.
24974 * function.c (expand_function_end): Likewise.
24976 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24978 * combine.c (try_combine): Adjust comment. Do not add the set of
24979 insn #0 if the destination indirectly is set or dies in insn #2.
24980 Tidy up code to distribute a new note.
24982 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24984 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
24985 also for alternatives 16 and 17.
24987 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24989 * config/i386/sse.md (*mov<mode>_internal): Merge with
24990 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
24991 Emit insn template depending on type attribute. Use
24992 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24993 movd instead of movq mnemonic for interunit moves. Rewrite mode
24994 attribute calculation. Remove unit attribute calculation.
24995 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
24996 Set prefix_data16 attribute for DImode ssemov types.
24997 Use Ym instead of y for SSE-MMX conversion alternatives.
24998 Reorder operand constraints.
25000 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
25002 * df.h (df_insn_delete): Adjust prototype.
25003 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
25004 and let it decide whether mark the basic block dirty.
25005 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
25006 * df-scan.c (df_insn_info_delete): New helper function, split
25007 off from df_insn_delete.
25008 (df_scan_free_bb_info): Use it.
25009 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
25011 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
25012 that the insn is actually an insn and it has a non-NULL basic block.
25013 Do not mark basic block dirty if only deleting a DEBUG_INSN.
25015 2013-03-22 Richard Biener <rguenther@suse.de>
25017 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
25019 (mem_ref_alloc): Do not allocate them.
25020 (refs_independent_p): Do not query or maintain a cache.
25022 2013-03-22 Richard Biener <rguenther@suse.de>
25024 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
25025 (gather_mem_refs_in_loops): Do not compute it.
25026 (analyze_memory_references): Do not allocate it.
25027 (tree_ssa_lim_finalize): Do not free it.
25028 (for_all_locs_in_loop): Do not query all_refs_in_loop.
25030 2013-03-22 Richard Biener <rguenther@suse.de>
25032 * is-a.h (as_a): Use gcc_checking_assert.
25034 2013-03-22 Ian Bolton <ian.bolton@arm.com>
25036 * config/aarch64/aarch64.c (aarch64_print_operand): New
25037 format specifier for printing a constant in hex.
25038 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
25039 format specifier for printing second operand.
25041 2013-03-22 Richard Biener <rguenther@suse.de>
25043 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
25045 (gather_mem_refs_in_loops): Perform store accumulation here.
25046 (create_vop_ref_mapping_loop): Remove.
25047 (create_vop_ref_mapping): Likewise.
25048 (analyze_memory_references): Initialize refs_stored_in_loop.
25049 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
25050 (record_indep_loop): Remove.
25051 (record_dep_loop): New function.
25052 (ref_indep_loop_p_1): Adjust to only walk over references
25053 in the loop, not its subloops.
25054 (ref_indep_loop_p): Rename to ...
25055 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
25056 maintaining a more fine-grained cache.
25057 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
25058 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
25060 2013-03-22 Richard Biener <rguenther@suse.de>
25062 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
25063 (struct mem_ref): Make accesses_in_loop a vec of a vec of
25064 aggregate mem_ref_loc.
25065 (free_mem_ref_locs): Inline into ...
25066 (memref_free): ... this and adjust.
25067 (mem_ref_alloc): Adjust.
25068 (mem_ref_locs_alloc): Remove.
25069 (record_mem_ref_loc): Adjust.
25070 (get_all_locs_in_loop): Rewrite into ...
25071 (for_all_locs_in_loop): ... this iterator.
25072 (rewrite_mem_ref_loc): New functor.
25073 (rewrite_mem_refs): Use for_all_locs_in_loop.
25074 (sm_set_flag_if_changed): New functor.
25075 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
25076 (ref_always_accessed): New functor.
25077 (ref_always_accessed_p): Use for_all_locs_in_loop.
25079 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25081 * tree-pass.h (PROP_gimple_lvec): New.
25082 * passes.c (dump_properties): Handle PROP_gimple_lvec.
25083 (init_optimization_passes): Move pass_lower_vector.
25084 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
25086 (pass_lower_vector): Provide PROP_gimple_lvec.
25087 (pass_lower_vector_ssa): Likewise.
25088 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
25090 2013-03-21 Mark Wielaard <mjw@redhat.com>
25092 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
25094 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25096 * config/i386/i386.md (*movdi_internal): Disparage slightly
25097 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
25098 conversion alternatives.
25100 2013-03-21 Jakub Jelinek <jakub@redhat.com>
25102 PR middle-end/48087
25103 * diagnostic.def (DK_WERROR): New kind.
25104 * diagnostic.h (werrorcount): Define.
25105 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
25106 promoted to DK_ERROR, increment DK_WERROR counter instead of
25108 * toplev.c (toplev_main): Call print_ignored_options even if
25109 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
25110 even if just werrorcount is non-zero.
25113 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
25115 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
25116 (string_cst_pool_decl): New function.
25117 (optimize_one_addr_into_implicit_ptr): New function.
25118 (resolve_addr_in_expr): Optimize DWARF location expression
25119 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
25120 which doesn't live in memory, but has DW_AT_location or
25121 DW_AT_const_value, or refers to a string literal, into
25122 DW_OP_GNU_implicit_pointer.
25123 (optimize_location_into_implicit_ptr): New function.
25124 (resolve_addr): If removing DW_AT_location of a variable because
25125 it was DW_OP_addr of address of the variable, but the variable doesn't
25126 live in memory, try to emit const value attribute for the initializer.
25128 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25130 * tree.h (VECTOR_TYPE_P): New macro.
25131 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
25132 TYPE_MODE): Use it.
25133 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
25134 VEC_COND_EXPR cannot be lvalues.
25135 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
25137 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25139 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
25140 Restrict the transformation to equal modes.
25142 2013-03-21 Richard Biener <rguenther@suse.de>
25144 PR tree-optimization/39326
25145 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
25146 (MEM_ANALYZABLE): Adjust.
25147 (record_mem_ref_loc): Move bitmap ops ...
25148 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
25149 unanalyzable refs, do not record locations for it.
25150 (analyze_memory_references): Allocate ref zero as shared
25152 (refs_independent_p): Do not test for unanalyzed mems here.
25153 (ref_indep_loop_p_1): Special-case disambiguation against
25154 the unanalyzed ref.
25155 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
25157 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
25159 * config/arm/arm-protos.h (tune_params): Add
25160 prefer_neon_for_64bits field.
25161 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
25162 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
25163 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
25164 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
25165 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
25166 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
25167 (arm_option_override): Handle -mneon-for-64bits new option.
25168 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
25169 (prefer_neon_for_64bits): Declare new variable.
25170 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
25171 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
25172 (arch_enabled): Handle new arch types. Remove support for onlya8
25174 (one_cmpldi2): Use new arch names.
25175 (zero_extend<mode>di2, extend<mode>di2): Ditto.
25176 * config/arm/arm.opt (mneon-for-64bits): Add option.
25177 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
25178 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
25179 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
25181 * doc/invoke.texi (-mneon-for-64bits): Document.
25183 2013-03-21 Richard Biener <rguenther@suse.de>
25185 PR tree-optimization/39326
25186 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
25187 (sort_bbs_in_loop_postorder_cmp): New function.
25188 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
25190 2013-03-21 Richard Biener <rguenther@suse.de>
25192 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
25193 (vect_insert_into_interleaving_chain): Likewise.
25194 (vect_drs_dependent_in_basic_block): Inline ...
25195 (vect_slp_analyze_data_ref_dependence): ... here. New function,
25197 (vect_analyze_data_ref_dependence): ... here. Simplify.
25198 (vect_check_interleaving): Simplify.
25199 (vect_analyze_data_ref_dependences): Likewise. Split out ...
25200 (vect_slp_analyze_data_ref_dependences): ... this new function.
25201 (dr_group_sort_cmp): New function.
25202 (vect_analyze_data_ref_accesses): Compute data-reference groups
25203 here instead of in vect_analyze_data_ref_dependence. Use
25204 a more efficient algorithm.
25205 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
25206 vect_slp_analyze_data_ref_dependences. Call
25207 vect_analyze_data_ref_accesses earlier.
25208 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
25209 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
25210 (vect_slp_analyze_data_ref_dependences): New prototype.
25212 2013-03-21 Richard Biener <rguenther@suse.de>
25214 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
25215 ref is stored in the loop.
25216 (find_refs_for_sm): Walk only over all stores.
25217 (store_motion_loop): Allocate from lim_bitmap_obstack.
25218 (store_motion): Likewise.
25220 2013-03-21 Richard Biener <rguenther@suse.de>
25222 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
25223 Update virtual SSA form.
25225 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25227 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
25228 * configure: Regenerate.
25229 * config.in: Regenerate.
25230 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
25231 if !HAVE_LD_EH_FRAME_CIEV3.
25233 2013-03-21 Richard Biener <rguenther@suse.de>
25235 * tree-cfg.c (verify_expr_no_block): New function.
25236 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
25237 nor DECL_VALUE_EXPR have locations with associated blocks.
25238 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
25239 (clear_unused_block_pointer): Remove code dealing with
25240 blocks in DECL_DEBUG_EXPR locations.
25242 2013-03-21 Richard Biener <rguenther@suse.de>
25244 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
25245 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
25246 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
25247 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
25248 instead of DECL_DEBUG_EXPR_IS_FROM.
25249 * gimplify.c (gimplify_modify_expr): Likewise.
25250 * tree-cfg.c (verify_expr_location_1): Likewise.
25251 * tree-complex.c (create_one_component_var): Likewise.
25252 * tree-sra.c (create_access_replacement): Likewise.
25253 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
25254 (clear_unused_block_pointer): Likewise.
25255 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
25256 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
25257 * var-tracking.c (var_debug_decl): Likewise.
25258 (track_expr_p): Likewise.
25259 * tree-inline.c (add_local_variables): Likewise. Set
25260 DECL_HAS_DEBUG_EXPR_P after copying it.
25261 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
25262 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
25264 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25267 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
25268 * configure: Regenerate.
25269 * config.in: Regenerate.
25270 * config/i386/i386.md (*movdf_internal): Use
25271 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25272 movd instead of movq mnemonic for interunit moves.
25273 (*movdi_internal): Ditto.
25275 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25277 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
25278 (abd<mode>_3): New pattern.
25279 (aba<mode>_3): New pattern.
25280 (fabd<mode>_3): New pattern.
25282 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25284 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
25285 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
25286 occurrence of REGISTER_PREFIX as its empty string.
25288 2013-03-20 Jeff Law <law@redhat.com>
25290 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
25291 addititional equivalences for equality comparisons between an SSA_NAME
25292 and a constant where the SSA_NAME was set from a widening conversion.
25294 2013-03-20 Walter Lee <walt@tilera.com>
25296 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
25298 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25300 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
25301 depending on type attribute.
25302 (*movti_internal): Ditto.
25303 (*movtf_internal): Ditto.
25304 (*movxf_internal): Ditto.
25305 (*movdf_internal): Ditto.
25306 (*movsf_internal): Ditto.
25308 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25310 * config/i386/i386.md (*movti_internal): Set prefix attribute to
25311 maybe_vex for sselog1 and ssemov types.
25312 (*movdi_internal): Reorder operand constraints.
25313 (*movsi_internal): Ditto. Set prefix attribute to
25314 maybe_vex for sselog1 and ssemov types.
25315 (*movtf_internal): Set prefix attribute to maybe_vex
25316 for sselog1 and ssemov types.
25317 (*movdf_internal): Ditto. Set prefix_data16 attribute for
25318 DImode ssemov types. Reorder operand constraints.
25319 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
25320 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
25321 attribute for SImode ssemov types. Reorder operand constraints.
25323 2013-03-20 Martin Jambor <mjambor@suse.cz>
25325 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
25326 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
25328 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
25330 * config/rs6000/predicates.md (indexed_address, update_address_mem
25331 update_indexed_address_mem): New predicates.
25332 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
25333 attribute for load/store instructions.
25334 * config/rs6000/dfp.md (movsd_store): Likewise.
25335 (movsd_load): Likewise.
25336 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
25337 (unnamed HI->DI extend define_insn): Likewise.
25338 (unnamed SI->DI extend define_insn): Likewise.
25339 (unnamed QI->SI extend define_insn): Likewise.
25340 (unnamed QI->HI extend define_insn): Likewise.
25341 (unnamed HI->SI extend define_insn): Likewise.
25342 (unnamed HI->SI extend define_insn): Likewise.
25343 (extendsfdf2_fpr): Likewise.
25344 (movsi_internal1): Likewise.
25345 (movsi_internal1_single): Likewise.
25346 (movhi_internal): Likewise.
25347 (movqi_internal): Likewise.
25348 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
25349 attribute for load/store instructions.
25350 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
25352 (mov<mode>_softfloat): Likewise.
25353 (mov<mode>_hardfloat32): Likewise.
25354 (mov<mode>_hardfloat64): Likewise.
25355 (mov<mode>_softfloat64): Likewise.
25356 (movdi_internal32): Likewise.
25357 (movdi_internal64): Likewise.
25358 (probe_stack_<mode>): Likewise.
25360 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
25362 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
25363 floating point, and decimal floating point to reload iterator.
25365 * config/rs6000/constraints.md (wl constraint): New constraints to
25366 return FLOAT_REGS if certain options are used to reduce the number
25367 of separate patterns that exist in the file.
25368 (wx constraint): Likewise.
25369 (wz constraint): Likewise.
25371 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25372 -mdebug=reg, print wg, wl, wx, and wz constraints.
25373 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25374 Initialize the reload functions for 64-bit binary/decimal floating
25376 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25377 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25378 create the buffer on the stack to overcome not having a 32-bit
25380 (rs6000_emit_move): Likewise.
25381 (rs6000_secondary_memory_needed_rtx): Likewise.
25382 (rs6000_alloc_sdmode_stack_slot): Likewise.
25383 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25384 via xxlxor, just like DFmode 0.0.
25386 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
25387 define as 1 if we are running on a power7 or newer.
25388 (enum r6000_reg_class_enum): Add new constraints.
25390 * config/rs6000/dfp.md (movsd): Delete, combine with binary
25391 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25392 with other moves by using conditional constraits (wg). Use LFIWZX
25393 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
25394 (movsd splitter): Likewise.
25395 (movsd_hardfloat): Likewise.
25396 (movsd_softfloat): Likewise.
25398 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
25399 binary and decimal floating point moves.
25400 (fmove_ok): New attributes to combine binary and decimal floating
25401 point moves, and to combine power6x (mfpgpr) moves along normal
25403 (real_value_to_target): Likewise.
25404 (f32_lr): Likewise.
25405 (f32_lm): Likewise.
25406 (f32_li): Likewise.
25407 (f32_sr): Likewise.
25408 (f32_sm): Likewise.
25409 (f32_si): Likewise.
25410 (movsf): Combine binary and decimal floating point moves. Combine
25411 power6x (mfpgpr) moves with other moves by using conditional
25412 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
25413 (mov<mode> for SFmode/SDmode); Likewise.
25414 (SFmode/SDmode splitters): Likewise.
25415 (movsf_hardfloat): Likewise.
25416 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
25417 (movsf_softfloat): Likewise.
25418 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
25420 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
25421 wx and wz constraints.
25423 * config/rs6000/constraints.md (wg constraint): New constraint to
25424 return FLOAT_REGS if -mmfpgpr (power6x) was used.
25426 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
25429 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25430 -mdebug=reg, print wg, wl, wx, and wz constraints.
25431 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25432 Initialize the reload functions for 64-bit binary/decimal floating
25434 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25435 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25436 create the buffer on the stack to overcome not having a 32-bit
25438 (rs6000_emit_move): Likewise.
25439 (rs6000_secondary_memory_needed_rtx): Likewise.
25440 (rs6000_alloc_sdmode_stack_slot): Likewise.
25441 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25442 via xxlxor, just like DFmode 0.0.
25444 * config/rs6000/dfp.md (movdd): Delete, combine with binary
25445 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25446 with other moves by using conditional constraits (wg). Use LFIWZX
25447 and STFIWX for loading SDmode on power7.
25448 (movdd splitters): Likewise.
25449 (movdd_hardfloat32): Likewise.
25450 (movdd_softfloat32): Likewise.
25451 (movdd_hardfloat64_mfpgpr): Likewise.
25452 (movdd_hardfloat64): Likewise.
25453 (movdd_softfloat64): Likewise.
25455 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
25456 64-bit binary and decimal floating point moves.
25457 (FMOVE64X): Likewise.
25458 (movdf): Combine 64-bit binary and decimal floating point moves.
25459 Combine power6x (mfpgpr) moves with other moves by using
25460 conditional constraits (wg).
25461 (mov<mode> for DFmode/DDmode): Likewise.
25462 (DFmode/DDmode splitters): Likewise.
25463 (movdf_hardfloat32): Likewise.
25464 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
25465 (movdf_softfloat32): Likewise.
25466 (movdf_hardfloat64_mfpgpr): Likewise.
25467 (movdf_hardfloat64): Likewise.
25468 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
25469 (movdf_softfloat64): Likewise.
25470 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
25471 (reload_<mode>_load): Move to later in the file so they aren't in
25472 the middle of the floating point move insns.
25473 (reload_<mode>_store): Likewise.
25475 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
25478 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
25479 constraint if -mdebug=reg.
25480 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
25481 Enable using dd reload support if needed.
25483 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
25484 binary and decimal floating point moves in rs6000.md.
25485 (movtd_internal): Likewise.
25487 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
25488 decimal floating point moves.
25490 (movtf_internal): Likewise.
25491 (mov<mode>_internal, TDmode/TFmode): Likewise.
25492 (movtf_softfloat): Likewise.
25493 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
25495 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
25496 movdi_internal64, using wg constraint for move direct operations.
25497 (movdi_internal64): Likewise.
25499 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
25500 MODES_TIEABLE_P for selected modes. Print the numerical value of
25501 the various virtual registers. Use GPR/FPR first/last values,
25502 instead of hard coding the register numbers. Print which modes
25503 have reload functions registered.
25504 (rs6000_option_override_internal): If -mdebug=reg, trace the options
25505 settings before/after setting cpu, target and subtarget settings.
25506 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
25507 and for secondary reload failures in rs6000_secondary_reload_inner.
25508 (rs6000_secondary_reload_fail): Likewise.
25509 (rs6000_secondary_reload_inner): Likewise.
25511 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
25512 macros for first/last GPR and FPR registers.
25513 (LAST_GPR_REGNO): Likewise.
25514 (FIRST_FPR_REGNO): Likewise.
25515 (LAST_FPR_REGNO): Likewise.
25517 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
25518 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
25519 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
25520 (vcond<mode><mode>): Likewise.
25521 (vcondu<mode><mode>): Likewise.
25522 (vector_gtu<mode>): Likewise.
25523 (vector_gte<mode>): Likewise.
25524 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
25525 to prevent the compiler from converting DImode operations to TImode.
25526 (ior<mode>3): Likewise.
25527 (and<mode>3): Likewise.
25528 (one_cmpl<mode>2): Likewise.
25529 (nor<mode>3): Likewise.
25530 (andc<mode>3): Likewise.
25532 * config/rs6000/constraints.md (wt constraint): New constraint
25533 that returns VSX_REGS if TImode is allowed in VSX registers.
25535 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
25536 constant under VSX.
25538 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
25539 similar to TImode, but it is restricted to being in the GPRs.
25541 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
25542 TImode to occupy a single VSX register.
25544 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
25545 -mvsx-timode for power7/power8.
25546 (power7 cpu): Likewise.
25547 (power8 cpu): Likewise.
25549 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
25550 sure that TFmode/TDmode take up two registers if they are ever
25551 allowed in the upper VSX registers.
25552 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
25554 (rs6000_init_hard_regno_mode_ok): Likewise.
25555 (rs6000_debug_reg_global): Add debugging for PTImode and wt
25556 constraint. Print if LRA is turned on.
25557 (rs6000_option_override_internal): Give an error if -mvsx-timode
25558 and VSX is not enabled.
25559 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
25560 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
25561 to reg+offset addressing. Use PTImode when checking offset
25562 addresses for validity.
25563 (reg_offset_addressing_ok_p): Likewise.
25564 (rs6000_legitimate_offset_address_p): Likewise.
25565 (rs6000_legitimize_address): Likewise.
25566 (rs6000_legitimize_reload_address): Likewise.
25567 (rs6000_legitimate_address_p): Likewise.
25568 (rs6000_eliminate_indexed_memrefs): Likewise.
25569 (rs6000_emit_move): Likewise.
25570 (rs6000_secondary_reload): Likewise.
25571 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
25572 reloads to fpr registers to continue to use reg+offset addressing,
25573 but 64-bit reloads to altivec registers need reg+reg addressing.
25574 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
25575 it. Treat LO_SUM like a PLUS operation.
25576 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
25577 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
25578 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
25579 registers to share a register with a smaller sized type, since VSX
25580 puts scalars in the upper 64-bits.
25581 (print_operand): Add support for PTImode.
25582 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
25583 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
25584 registers, but don't have arithmetic support.
25585 (rs6000_memory_move_cost): Add test for VSX.
25586 (rs6000_opt_masks): Add -mvsx-timode.
25588 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
25591 (VSr): Use wt constraint for TImode.
25592 (VSv): Drop TImode support.
25593 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
25594 (vsx_movti_64bit): Likewise.
25595 (vsx_movti_32bit): Likewise.
25596 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
25597 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
25598 one '?' on the appropriate output constraint. Do not allow TImode
25599 logical operations on 32-bit systems.
25600 (vsx_ior<mode>3): Likewise.
25601 (vsx_xor<mode>3): Likewise.
25602 (vsx_one_cmpl<mode>2): Likewise.
25603 (vsx_nor<mode>3): Likewise.
25604 (vsx_andc<mode>3): Likewise.
25605 (vsx_concat_<mode>): Likewise.
25606 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
25608 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
25609 OPTION_MASK_VSX_TIMODE.
25610 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
25611 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
25613 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
25614 (TI2 iterator): New iterator for TImode, PTImode.
25615 (wd mode attribute): Add values for vector types.
25616 (movti_string): Replace TI move operations with operations for TImode
25617 and PTImode. Add support for TImode being allowed in VSX registers.
25618 (mov<mode>_string, TImode/PTImode): Likewise.
25619 (movti_ppc64): Likewise.
25620 (mov<mode>_ppc64, TImode/PTImode): Likewise.
25621 (TI mode splitters): Likewise.
25623 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
25626 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
25628 PR tree-optimization/56355
25629 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
25630 Also handle integers with undefined overflow.
25632 2013-03-20 Catherine Moore <clm@codesourcery.com>
25633 Maciej W. Rozycki <macro@codesourcery.com>
25634 Tom de Vries <tom@codesourcery.com>
25635 Nathan Sidwell <nathan@codesourcery.com>
25636 Iain Sandoe <iain@codesourcery.com>
25637 Nathan Froyd <froydnj@codesourcery.com>
25638 Chao-ying Fu <fu@mips.com>
25640 * doc/extend.texi: (micromips, nomicromips, nocompression):
25641 Document new function attributes.
25642 * doc/invoke.texi (minterlink-compressed, mmicromips,
25643 m14k, m14ke, m14kec): Document new options.
25644 (minterlink-mips16): Update documentation.
25645 * doc/md.texi (ZC, ZD): Document new constraints.
25646 * configure.ac (gcc_cv_as_micromips): Check if linker
25647 supports the .set micromips directive.
25648 * configure: Regenerate.
25649 * config.in: Regenerate.
25650 * config/mips/mips-tables.opt: Regenerate.
25651 * config/mips/micromips.md: New file.
25652 * constraints.md (ZC, ZD): New constraints.
25653 * config/mips/predicates.md (movep_src_register): New predicate.
25654 (movep_src_operand): New predicate.
25655 (non_volatile_mem_operand): New predicate.
25656 * config/mips/mips.md (multimem): New type.
25657 (length): Differentiate between 17-bit and 18-bit branch offsets.
25658 (MOVEP1, MOVEP2): New mode iterator.
25659 (mov_<load>l): Use ZC constraint.
25660 (mov_<load>r): Likewise.
25661 (mov_<store>l): Likewise.
25662 (mov_<store>r): Likewise.
25663 (*branch_equality<mode>_inverted): Add microMIPS support.
25664 (*branch_equality<mode>): Likewise.
25665 (*jump_absolute): Likewise.
25666 (indirect_jump_<mode>): Likewise.
25667 (tablejump_<mode>): Likewise.
25668 (<optab>_internal): Likewise.
25669 (sibcall_internal): Likewise.
25670 (sibcall_value_internal): Likewise.
25671 (prefetch): Use constraint ZD.
25672 * config/mips/mips.opt (minterlink-compressed): New option.
25673 (minterlink-mips16): Now an alias for minterlink-compressed.
25674 (mmicromips): New option.
25675 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
25676 (compare_and_swap_12): Likewise.
25677 (sync_add<mode>): Likewise.
25678 (sync_<optab>_12): Likewise.
25679 (sync_old_<optab>_12): Likewise.
25680 (sync_new_<optab>_12): Likewise.
25681 (sync_nand_12): Likewise.
25682 (sync_old_nand_12): Likewise.
25683 (sync_new_nand_12): Likewise.
25684 (sync_sub<mode>): Likewise.
25685 (sync_old_add<mode>): Likewise.
25686 (sync_old_sub<mode>): Likewise.
25687 (sync_new_add<mode>): Likewise.
25688 (sync_new_sub<mode>): Likewise.
25689 (sync_<optab><mode>): Likewise.
25690 (sync_old_<optab><mode>): Likewise.
25691 (sync_new_<optab><mode>): Likewise.
25692 (sync_nand<mode>): Likewise.
25693 (sync_old_nand<mode>): Likewise.
25694 (sync_new_nand<mode>): Likewise.
25695 (sync_lock_test_and_set<mode>): Likewise.
25696 (test_and_set_12): Likewise.
25697 (atomic_compare_and_swap<mode>): Likewise.
25698 (atomic_exchange<mode>_llsc): Likewise.
25699 (atomic_fetch_add<mode>_llsc): Likewise.
25700 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
25701 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
25702 (umips_save_restore_pattern_p): Likewise.
25703 (umips_load_store_pair_p): Likewise.
25704 (umips_output_load_store_pair): Likewise.
25705 (umips_movep_target_p): Likewise.
25706 (umips_12bit_offset_address_p): Likewise.
25707 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
25708 (mips_base_mips16): Rename this...
25709 (mips_base_compression_flags): ...to this. Update all uses.
25710 (mips_attribute_table): Add micromips, nomicromips and nocompression.
25711 (mips_mips16_decl_p): Delete.
25712 (mips_nomips16_decl_p): Delete.
25713 (mips_get_compress_on_flags): New function.
25714 (mips_get_compress_off_flags): New function.
25715 (mips_get_compress_mode): New function.
25716 (mips_get_compress_on_name): New function.
25717 (mips_get_compress_off_name): New function.
25718 (mips_insert_attributes): Support multiple compression types.
25719 (mips_merge_decl_attributes): Likewise.
25720 (umips_12bit_offset_address_p): New function.
25721 (mips_start_function_definition): Emit .set micromips directive.
25722 (mips_call_may_need_jalx_p): New function.
25723 (mips_function_ok_for_sibcall): Add microMIPS support.
25724 (mips_print_operand_punctuation): Support short delay slots and
25726 (umips_swm_mask, umips_swm_encoding): New.
25727 (umips_build_save_restore): New function.
25728 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
25729 (was_mips16_p): Remove.
25730 (old_compression_mode): New.
25731 (mips_set_compression_mode): New function.
25732 (mips_set_current_function): Add microMIPS support.
25733 (mips_option_override): Likewise.
25734 (umips_save_restore_pattern_p): New function.
25735 (umips_output_save_restore): New function.
25736 (umips_load_store_pair_p_1): New function.
25737 (umips_load_store_pair_p): New function.
25738 (umips_output_load_store_pair_1): New function.
25739 (umips_output_load_store_pair): New function.
25740 (umips_movep_target_p) New function.
25741 (mips_prepare_pch_save): Add microMIPS support.
25742 * config/mips/mips.h (TARGET_COMPRESSION): New.
25743 (TARGET_CPU_CPP_BUILTINS): Update macro
25744 to use new compression flags and to support microMIPS.
25745 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
25746 (MIPS_ARCH_FLOAT_SPEC): Likewise.
25747 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
25748 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
25749 (ASM_SPEC): Support mmicromips and mno-micromips.
25750 (M16STORE_REG_P): New macro.
25751 (MIPS_CALL): Support TARGET_MICROMIPS.
25752 (MICROMIPS_J): New macro.
25753 (mips_base_mips16): Rename this...
25754 (mips_base_compression_flags): ...to this.
25755 (UMIPS_12BIT_OFFSET_P): New macro.
25756 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
25757 (MULTILIB_DIRNAMES): Likewise.
25758 2013-03-20 Richard Biener <rguenther@suse.de>
25760 PR tree-optimization/56661
25761 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
25762 the result does not have to be distinct.
25764 2013-03-20 Richard Biener <rguenther@suse.de>
25766 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
25769 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
25770 Steven Bosscher <steven@gcc.gnu.org>
25772 PR rtl-optimization/56605
25773 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
25775 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25778 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
25779 that require movd instead of movq.
25781 2013-03-20 Richard Biener <rguenther@suse.de>
25783 * tree-ssa-structalias.c (struct variable_info): Add pointer
25784 to the first field of an aggregate with sub-vars. Make
25785 this and the pointer to the next subfield its ID.
25786 (vi_next): New function.
25787 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
25788 storedanything_id, integer_id): Increment by one.
25789 (new_var_info, get_call_vi, lookup_call_clobber_vi,
25790 get_call_clobber_vi): Adjust.
25791 (solution_set_expand): Simplify and speedup.
25792 (solution_set_add): Inline into ...
25793 (set_union_with_increment): ... this. Adjust accordingly.
25794 (do_sd_constraint): Likewise.
25795 (do_ds_constraint): Likewise.
25796 (do_complex_constraint): Simplify.
25797 (build_pred_graph): Adjust.
25798 (solve_graph): Likewise. Simplify and speedup.
25799 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
25800 get_constraint_for_component_ref, get_constraint_for_1,
25801 first_vi_for_offset, first_or_preceding_vi_for_offset,
25802 create_function_info_for, create_variable_info_for_1,
25803 create_variable_info_for, intra_create_variable_infos): Adjust.
25804 (init_base_vars): Push NULL for ID zero.
25805 (compute_points_to_sets): Adjust.
25807 2013-03-20 Richard Biener <rguenther@suse.de>
25809 * cfgloop.c (verify_loop_structure): Streamline and avoid
25810 ICEing on corrupt loop tree.
25811 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
25814 2013-03-20 Richard Biener <rguenther@suse.de>
25816 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
25817 check whether an SSA update is needed.
25819 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
25821 * config/mips/constraints.md (T): Rename to...
25825 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
25826 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
25828 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25830 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
25831 (*subsi3_carryin_uxtw): Likewise.
25833 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25835 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
25836 (*rorsi3_insn_uxtw): Likewise.
25838 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25840 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
25841 (*extrsi5_insn_uxtw): Likewise.
25843 2013-03-19 Richard Biener <rguenther@suse.de>
25845 PR tree-optimization/56273
25846 * passes.c (init_optimization_passes): Move second VRP after DOM.
25848 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
25850 * config/i386/i386.md (*movti_internal): Merge from
25851 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
25852 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
25853 nox64 isa attributes.
25855 2013-03-18 Richard Biener <rguenther@suse.de>
25857 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
25859 (merge_node_constraints): Likewise.
25860 (build_succ_graph): Likewise.
25861 (valid_graph_edge): Inline into single caller.
25862 (unify_nodes): Likewise. Use bitmap_set_bit return value
25864 (scc_visit): Fix formatting and variable use.
25865 (do_sd_constraint): Use gcc_checking_assert.
25866 (do_ds_constraint): Likewise.
25867 (do_complex_constraint): Likewise.
25868 (condense_visit): Likewise. Cleanup.
25869 (dump_pred_graph): New function.
25870 (perform_var_substitution): Dump the pred-graph before
25871 variable substitution.
25872 (find_equivalent_node): Use gcc_checking_assert.
25873 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
25875 2013-03-18 Richard Biener <rguenther@suse.de>
25877 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
25878 Remove cond_expr_stmt_list argument and do not gimplify the
25880 (vect_loop_versioning): Adjust.
25881 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
25882 Cleanup to use less temporaries.
25883 (vect_create_data_ref_ptr): Cleanup.
25885 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25887 PR tree-optimization/56635
25888 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
25889 require types_compatible_p types.
25891 2013-03-18 Nick Clifton <nickc@redhat.com>
25893 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
25894 spurious backslash.
25896 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
25897 Add missing line to comment describing function.
25899 2013-03-18 Richard Biener <rguenther@suse.de>
25901 PR tree-optimization/56210
25902 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
25903 Handle string / character search functions.
25904 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
25906 2013-03-18 Richard Biener <rguenther@suse.de>
25908 PR middle-end/56483
25909 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
25910 and implement properly.
25911 * gimple.h (gimple_cond_single_var_p): Remove.
25913 2013-03-18 Richard Biener <rguenther@suse.de>
25915 * tree-data-ref.h (find_data_references_in_loop): Declare.
25916 * tree-data-ref.c (get_references_in_stmt): Use a stack
25917 vector pre-allocated in the callers.
25918 (find_data_references_in_stmt): Adjust.
25919 (graphite_find_data_references_in_stmt): Likewise.
25920 (create_rdg_vertices): Likewise.
25921 (find_data_references_in_loop): Export.
25922 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
25923 Compute dependences here...
25924 (vect_analyze_data_refs): ...not here. When we encounter
25925 a non-vectorizable data reference in basic-block vectorization
25926 truncate the data reference vector. Do not bother to
25927 fixup data-dependence information for gather loads.
25928 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
25929 of data references, as reported.
25931 2013-03-18 Richard Biener <rguenther@suse.de>
25933 PR tree-optimization/3713
25934 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
25935 has_constants and expr.
25936 (stmt_has_constants): Properly valueize SSA names when deciding
25937 whether the stmt has constants.
25939 2013-03-18 Richard Biener <rguenther@suse.de>
25941 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
25942 whole function when there is nothing to do.
25943 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
25944 * tree-vectorizer.c (vectorize_loops): Update virtual and
25945 loop-closed SSA once.
25946 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
25948 2013-03-18 Richard Biener <rguenther@suse.de>
25950 PR middle-end/56113
25951 * domwalk.c (bb_postorder): New global static.
25952 (cmp_bb_postorder): New function.
25953 (walk_dominator_tree): Replace scheme imposing an order for
25954 visiting dominator sons by one sorting them at the time they
25955 are pushed on the stack.
25957 2013-03-18 Richard Biener <rguenther@suse.de>
25959 PR tree-optimization/39326
25960 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
25961 (struct mem_ref): Replace mem member with ao_ref typed member.
25962 (MEM_ANALYZABLE): Adjust.
25963 (memref_eq): Likewise.
25964 (mem_ref_alloc): Likewise.
25965 (gather_mem_refs_stmt): Likewise.
25966 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
25967 (execute_sm_if_changed_flag_set): Adjust.
25968 (execute_sm): Likewise.
25969 (ref_always_accessed_p): Likewise.
25970 (refs_independent_p): Likewise.
25971 (can_sm_ref_p): Likewise.
25973 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25976 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
25977 return 1 even for !unsignedp.
25979 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25981 * config/i386/i386.md (isa): Add x64 and nox64.
25982 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
25983 (*pushtf): Enable *roF alternative for x64 isa only.
25984 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
25985 mode attribute of integer alternatives to DImode for TARGET_64BIT.
25986 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
25987 (*movtf_internal): Merge from *movtf_internal_rex64 and
25988 *movtf_internal_sse. Use x64 and nox64 isa attributes.
25989 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
25990 nox64 isa attributes.
25991 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
25992 nox64 isa attributes.
25993 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
25995 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25997 * config/alpha/alpha.c (TARGET_LRA_P): New define.
25999 2013-03-17 Jakub Jelinek <jakub@redhat.com>
26002 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
26003 class names. Remove trailing comma after "ALL_REGS".
26005 2013-03-16 Jan Hubicka <jh@suse.cz>
26007 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
26008 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
26009 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
26010 of cgraph_get_create_node.
26011 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
26013 2013-03-16 Jason Merrill <jason@redhat.com>
26016 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
26017 with DW_AT_default_value.
26019 2013-03-16 Jakub Jelinek <jakub@redhat.com>
26021 * BASE-VER: Set to 4.9.0.
26023 2013-03-14 Andi Kleen <ak@linux.intel.com>
26026 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
26027 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
26028 Document _x* TSX intrinsics.
26030 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
26031 David Holsgrove <david.holsgrove@xilinx.com>
26033 * configure.ac: Add MicroBlaze TLS support detection.
26034 * configure: Regenerate.
26035 * config/microblaze/microblaze-protos.h
26036 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
26037 symbol_mentioned_p, label_mentioned_p): Add prototypes.
26038 * config/microblaze/microblaze.c (microblaze_address_type): Add
26039 ADDRESS_TLS and tls_reloc address types.
26040 (microblaze_address_info): Add tls_reloc.
26041 (TARGET_HAVE_TLS): Define.
26042 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
26043 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
26044 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
26045 load_tls_operand, microblaze_call_tls_get_addr,
26046 microblaze_legitimize_tls_address): New functions.
26047 (microblaze_classify_unspec): Handle UNSPEC_TLS.
26048 (get_base_reg): Use microblaze_tls_symbol_p.
26049 (microblaze_classify_address): Handle TLS.
26050 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
26051 label_mentioned_p and microblaze_tls_referenced_p.
26052 (microblaze_legitimize_address): Handle TLS.
26053 (microblaze_address_insns): Handle ADDRESS_TLS.
26054 (pic_address_needs_scratch): Handle TLS.
26055 (print_operand_address): Handle TLS.
26056 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
26057 (microblaze_expand_move): Handle TLS.
26058 (microblaze_legitimate_constant_p): Check
26059 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
26060 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
26061 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
26062 (PIC_OFFSET_TABLE_REGNUM): Set.
26063 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
26064 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
26065 (addsi3, movsi_internal2, movdf_internal): Update constraints
26066 * config/microblaze/predicates.md (arith_plus_operand): Define
26067 (move_operand): Redefine as move_src_operand,
26068 check microblaze_tls_referenced_p.
26070 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26072 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
26073 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
26075 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26077 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
26080 2013-03-14 Jakub Jelinek <jakub@redhat.com>
26082 PR tree-optimization/53265
26083 * common.opt (Waggressive-loop-optimizations): New option.
26084 * tree-ssa-loop-niter.c: Include tree-pass.h.
26085 (do_warn_aggressive_loop_optimizations): New function.
26086 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
26087 if number_of_latch_executions returned constant.
26088 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
26089 early. If number_of_latch_executions returned constant, set
26090 nb_iterations_upper_bound back to it.
26091 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
26093 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
26094 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
26096 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
26097 (MULTILIB_OSDIRNAMES): Set.
26098 * genmultilib: If defaultosdirname doesn't start with :: , set
26099 defaultosdirname2 instead, clear it and emit two . multilib_raw
26100 entries instead of just one.
26102 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
26104 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26105 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
26106 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26107 (SUBTARGET_OVERRIDE_OPTIONS): New.
26109 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
26112 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
26113 (musermode): Convert to Var(TARGET_USERMODE).
26114 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
26115 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
26116 * config/sh/sh.c (sh_option_override): Use
26117 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
26118 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
26120 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
26122 (udivsi3_i4_single, divsi3_i4_single): Use
26123 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
26125 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
26127 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
26130 2013-03-13 Richard Biener <rguenther@suse.de>
26132 PR tree-optimization/56608
26133 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
26134 calls when vectorizing basic-blocks.
26136 2013-03-13 Jakub Jelinek <jakub@redhat.com>
26139 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
26142 2013-03-12 Jakub Jelinek <jakub@redhat.com>
26144 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
26146 2013-03-11 Jan Hubicka <jh@suse.cz>
26149 * lto-streamer-out.c (output_symbol_p): Skip references from
26150 constructors of external variables.
26152 2013-03-11 Jan Hubicka <jh@suse.cz>
26154 PR middle-end/56571
26155 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
26157 * emit-rtl.c (verify_rtx_sharing): Likewise.
26158 (copy_insn_1): Likewise.
26159 * rtl.c (copy_rtx): Likewise.
26161 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
26164 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
26165 output_operand_lossage message.
26167 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
26170 * arm.c (shift_op): Validate RTL pattern on the fly.
26171 (arm_print_operand, case 'S'): Don't use shift_operator to validate
26174 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26177 * config/pa/pa.md (call_value): Check for calls to powf and direct to
26178 new call patterns that clobber %fr12.
26179 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
26180 split and postreload patterns.
26181 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
26182 registers %fr12 and %fr12R as call used.
26184 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
26186 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
26187 (canon_address, record_store, replace_read, check_mem_read_rtx,
26188 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
26189 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
26190 rest_of_handle_dse): Likewise.
26192 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
26194 PR middle-end/56524
26195 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
26197 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
26198 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
26199 (save_optabs_if_changed): Replace with...
26200 (init_tree_optimization_optabs): ...this.
26201 * optabs.c (save_optabs_if_changed): Rename to...
26202 (init_tree_optimization_optabs): ...this. Take the optimization node
26203 as argument. Do nothing if the base optabs are already correct.
26204 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
26205 to recompute optabs.
26206 * function.h (function): Remove optabs field.
26207 * function.c (invoke_set_current_function_hook): Call
26208 init_tree_optimization_optabs. Use the result to initialize
26211 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
26213 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
26214 if GTMA_HAS_NO_INSTRUMENTATION.
26215 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
26216 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
26217 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
26218 * gimple-pretty-print.c (dump_gimple_transaction): Handle
26219 GTMA_HAS_NO_INSTRUMENTATION.
26221 2013-03-08 Jakub Jelinek <jakub@redhat.com>
26223 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
26226 2013-03-08 Marek Polacek <polacek@redhat.com>
26227 Jakub Jelinek <jakub@redhat.com>
26229 PR tree-optimization/56478
26230 * predict.c (is_comparison_with_loop_invariant_p): Change the
26231 type of loop_step to tree.
26232 (predict_loops): Adjust.
26233 (predict_iv_comparison): Perform the computations on double_ints.
26235 2013-03-08 Richard Biener <rguenther@suse.de>
26237 PR tree-optimization/56570
26238 * tree-cfg.c (verify_expr_location_1): Verify locations for
26240 * tree-sra.c (create_access_replacement): Strip locations
26241 from DECL_DEBUG_EXPRs.
26243 2013-03-08 Richard Biener <rguenther@suse.de>
26245 * tree-inline.c (expand_call_inline): Do not associate
26246 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
26247 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
26249 2013-03-08 Richard Biener <rguenther@suse.de>
26251 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
26252 or block changes with -Og. Fix for location / block encoding
26253 changes and PHI arguments with locations.
26255 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
26257 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
26259 (struct output_info): Likewise.
26260 (register_overhead): Remove bad gcc_assert.
26261 (bitmap_find_bit): If there is only a single bitmap element, do not
26262 count a miss as a search.
26263 (print_statistics): Update for counter type changes.
26264 (dump_bitmap_statistics): Likewise. Print headers such that they
26265 are properly lined up with the printed counters.
26267 2013-03-07 Jakub Jelinek <jakub@redhat.com>
26269 PR tree-optimization/56559
26270 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
26271 check that it has only a single use.
26273 2013-03-07 Richard Biener <rguenther@suse.de>
26275 * doc/invoke.texi (fwhole-program): Discourage use in combination
26278 2013-03-06 Jakub Jelinek <jakub@redhat.com>
26280 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
26282 PR tree-optimization/56539
26283 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
26284 instead of GSI_CONTINUE_LINKING as last argument to
26285 force_gimple_operand_gsi. Adjust function comment.
26287 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
26290 PR middle-end/56548
26291 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
26292 promoted mode, convert the result back to the original mode.
26294 2013-03-06 Richard Biener <rguenther@suse.de>
26296 PR middle-end/56294
26297 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
26298 (insert_updated_phi_nodes_compare_uids): New function.
26299 (update_ssa): Sort symbols_to_rename after UID before
26300 traversing it to insert PHI nodes.
26302 2013-03-06 Richard Biener <rguenther@suse.de>
26304 PR middle-end/50494
26305 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
26306 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
26309 2013-02-13 Richard Biener <rguenther@suse.de>
26312 * varasm.c (output_constant_def_1): Get the decl representing
26313 the constant as argument.
26314 (output_constant_def): Wrap output_constant_def_1.
26315 (make_decl_rtl): Use output_constant_def_1 with the decl
26316 representing the constant.
26317 (build_constant_desc): Optionally re-use a decl already
26318 representing the constant.
26319 (tree_output_constant_def): Adjust.
26321 2013-03-06 Joey Ye <joey.ye@arm.com>
26324 * gcc.c (convert_white_space): New function.
26325 (main): Handles white space in function name.
26327 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
26330 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
26331 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
26332 to SH_DIV_CALL_TABLE for TARGET_SH2.
26333 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
26335 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
26336 call-table options.
26338 2013-03-05 Sterling Augustine <saugustine@google.com>
26339 Cary Coutant <ccoutant@google.com>
26342 * dwarf2out.c (resolve_addr): Don't call
26343 remove_loc_list_addr_table_entries a second time for the same
26346 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26349 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
26350 (avoid_complex_debug_insns): New function.
26351 (expand_debug_locations): Call it.
26353 PR rtl-optimization/56484
26354 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
26355 lifetimes of hard registers on small register class machines.
26357 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26359 * config/microblaze/microblaze-protos.h: Rename
26360 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
26361 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
26363 (microblaze_fast_interrupt_function_p): New function.
26364 (microblaze_is_interrupt_handler): Rename to
26365 microblaze_is_interrupt_variant and add fast_interrupt check.
26366 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
26367 (save_restore_insns): Likewise.
26368 (compute_frame_size): Likewise.
26369 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
26370 (microblaze_globalize_label): Likewise.
26371 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
26372 * config/microblaze/microblaze.md: Use wrapper
26373 microblaze_is_interrupt_variant.
26375 2013-03-05 Kai Tietz <ktietz@redhat.com>
26377 * sdbout.c (sdbout_one_type): Switch to current function's section
26378 supporting cold/hot.
26380 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26382 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
26385 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26387 PR middle-end/56461
26388 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
26389 if VALGRIND_GET_VBITS is defined, temporarily make object
26390 memory all defined, and restore previous valgrind addressability
26391 and definability afterwards. Free this_object at the end.
26393 PR middle-end/56461
26394 * lra.c (lra): Call lra_clear_live_ranges if live_p,
26395 right before calling lra_create_live_ranges, also call it
26396 when clearing live_p. Only call lra_clear_live_ranges
26397 at the end if live_p.
26399 PR middle-end/56461
26400 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
26402 2013-03-05 Richard Biener <rguenther@suse.de>
26404 PR tree-optimization/56521
26405 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
26408 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
26411 * except.h (remove_unreachable_eh_regions): New prototype.
26412 * except.c (remove_eh_handler_splicer): New function, split out
26413 of remove_eh_handler.
26414 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
26415 warning about running it on many EH regions one at a time.
26416 (remove_unreachable_eh_regions_worker): New function, walk the
26417 EH tree in depth-first order and remove non-marked regions.
26418 (remove_unreachable_eh_regions): New function.
26419 * tree-eh.c (mark_reachable_handlers): New function, split out
26420 from remove_unreachable_handlers.
26421 (remove_unreachable_handlers): Use mark_reachable_handlers and
26422 remove_unreachable_eh_regions.
26423 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
26424 and remove_unreachable_eh_regions.
26426 2013-03-05 Richard Biener <rguenther@suse.de>
26428 PR middle-end/56525
26429 * loop-init.c (fix_loop_structure): Remove loops in two stages,
26430 not freeing them until the end.
26432 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26434 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
26436 2013-03-05 Richard Biener <rguenther@suse.de>
26438 PR tree-optimization/56270
26439 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
26440 of loads after scheduling an SLP instance.
26442 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26444 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
26446 (check_gcc_parallelize): Run guality.exp as a separate job from
26447 vect.exp with unsorted.exp and $(dg_target_exps) separately from
26448 struct-layout-1.exp with stackalign.exp.
26450 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
26452 PR middle-end/56461
26453 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
26454 load_index sbitmap even if some bit in it isn't set.
26456 PR middle-end/56461
26457 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
26458 (discover_iteration_bound_by_body_walk): Change queues to
26459 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
26460 spelling in comment. Call safe_push on queues[bound_index] directly.
26461 Release queues[queue_index] in every iteration unconditionally.
26462 Release bounds vector.
26464 PR middle-end/56461
26465 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
26466 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
26467 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
26470 2013-03-05 Richard Biener <rguenther@suse.de>
26473 * tree-inline.c (remap_blocks_to_null): New function.
26474 (expand_call_inline): When expanding a call stmt without
26475 an associated block inline remap all callee blocks to NULL.
26477 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26479 PR rtl-optimization/56494
26480 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
26481 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
26482 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
26484 PR middle-end/56461
26485 * sel-sched-ir.c (free_sched_pools): Release
26486 succs_info_pool.stack[succs_info_pool.max_top] vectors too
26487 if succs_info_pool.max_top isn't -1.
26490 * opts.c (opts_obstack, opts_concat): Moved to...
26491 * opts-common.c (opts_obstack, opts_concat): ... here.
26493 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26495 PR middle-end/56461
26496 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
26498 2013-03-04 Martin Jambor <mjambor@suse.cz>
26500 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
26501 all appropriate places.
26503 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
26505 PR tree-optimization/56424
26506 * ipa-split.c (split_function): Do not set the RSO flag if result is
26507 not by reference and its type is a register type.
26509 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
26511 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
26512 (microblaze_legitimate_pic_operand): Likewise
26513 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
26514 new function microblaze_legitimate_pic_operand
26515 * config/microblaze/microblaze-protos.h
26516 (microblaze_legitimate_pic_operand): Declare.
26518 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26520 * config/microblaze/predicates.md (call_insn_simple_operand):
26521 New predicate for supported rtx code types.
26522 * config/microblaze/microblaze.md (call_internal1): Use
26523 call_insn_simple_operand predicate.
26525 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26527 PR middle-end/56461
26528 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
26529 partitions.ordered_remove.
26531 PR middle-end/56461
26532 * tree-vect-stmts.c (vectorizable_conversion): Don't call
26533 vec_oprnds0.create (1) for modifier == NONE.
26535 PR middle-end/56461
26536 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
26537 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
26538 vec_oprnds1 right before pushing anything to it for
26541 PR middle-end/56461
26542 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
26543 set nbbs to 0 instead of having separate code path.
26544 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
26545 instead of false as last argument if returning NULL.
26547 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
26549 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
26550 the attribute is now called "target" instead of "option".
26551 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
26552 * doc/tm.texi.in (Target Attributes): Likewise document the correct
26553 attribute/pragma name for TARGET_OPTION_VALID_P and
26554 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
26555 * doc/tm.texi: Regenerated.
26557 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
26559 * config/microblaze/microblaze.c:
26560 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
26561 * config/microblaze/microblaze.h: Add -mxl-reorder to
26563 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
26564 instructions emitted if TARGET_REORDER.
26565 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
26566 or 0 for -m/-mno case, but initialises as 2 to detect default use case
26569 2013-03-01 Xinliang David Li <davidxl@google.com>
26571 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
26574 2013-03-01 Jakub Jelinek <jakub@redhat.com>
26576 PR middle-end/56461
26577 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
26578 vector even when returning true. Fix up function comment formatting.
26580 PR middle-end/56461
26581 * ira-build.c (ira_loop_nodes_count): New variable.
26582 (create_loop_tree_nodes): Initialize it.
26583 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
26585 PR middle-end/56461
26586 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
26587 method on dr_chain and result_chain.
26588 * tree-vect-stmts.c (vectorizable_store): Only call
26589 result_chain.create if j == 0.
26591 PR middle-end/56461
26592 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
26593 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
26594 before overwriting it.
26596 2013-03-01 Tobias Burnus <burnus@net-b.de>
26598 * doc/extended.texi (C Extensions): Change order in @menu
26600 (Other MIPS Built-in Functions): Move last MIPS entry before
26601 "picoChip Built-in Functions".
26602 (SH Built-in Functions): Move after RX Built-in Functions.
26603 * doc/gcc.texi (Introduction): Change order in @menu
26605 * doc/md.texi (Constraints): Ditto.
26606 * gty.texi (Type Information): Ditto.
26607 (User-provided marking routines for template types): Make
26609 * doc/invoke.texi (AArch64 Options): Move before
26610 "Adapteva Epiphany Options".
26612 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
26613 Jakub Jelinek <jakub@redhat.com>
26616 * asan.c (gate_asan): Lookup no_sanitize_address instead of
26617 no_address_safety_analysis attribute.
26618 * doc/extend.texi (no_address_safety_attribute): Rename to
26619 no_sanitize_address attribute, mention no_address_safety_analysis
26620 attribute as deprecated alias.
26622 2013-02-28 Jakub Jelinek <jakub@redhat.com>
26624 PR middle-end/56461
26625 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
26626 type to vec<vec<tree> > *.
26627 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
26628 to be vec<tree> instead of vec<tree> *, set vec_defs
26629 to vNULL and call vec_defs.create (number_of_vects), adjust other
26631 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
26632 vectorizable_condition): Adjust vect_get_slp_defs callers.
26634 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26636 * config/aarch64/aarch64.c
26637 (aarch64_float_const_representable): Remove unused variable.
26639 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26641 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
26643 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26645 * config/aarch64/aarch64-builtins.c
26646 (aarch64_init_simd_builtins): Make static.
26648 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26650 * config/aarch64/aarch64.c
26651 (aarch64_simd_make_constant): Make static.
26653 2013-02-28 Martin Jambor <mjambor@suse.cz>
26655 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
26656 with no initialization to the RHS of debug statements.
26658 2013-02-28 Martin Jambor <mjambor@suse.cz>
26660 PR tree-optimization/56294
26661 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
26663 (get_access_replacement): Do not call create_access_replacement.
26664 Assert a replacement exists.
26665 (get_repl_default_def_ssa_name): Create the replacement declaration
26668 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26670 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
26671 final_end_function.
26673 2013-02-28 Marek Polacek <polacek@redhat.com>
26675 PR rtl-optimization/56466
26676 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
26677 if we're changing a loop.
26678 (peel_loops_completely): Likewise.
26680 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
26683 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
26685 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26688 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
26689 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
26690 INTX_FTYPE_FX, FX_FTYPE_INTX.
26691 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
26693 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26695 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
26696 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
26697 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
26698 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
26699 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
26700 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
26701 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
26702 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
26703 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
26704 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
26705 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
26706 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
26707 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
26708 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
26709 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
26710 (avrxmega6): Increase max flash segments from 5 to 6.
26711 * config/avr/t-multilib: Regenerate.
26712 * config/avr/avr-tables.opt: Regenerate.
26713 * doc/avr-mmcu.texi: Regenerate.
26715 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26717 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
26718 (avr_device_to_arch): Rename to avr_device_to_ld.
26719 (avr_device_to_as): New prototype.
26720 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
26721 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
26722 * config/avr/driver-avr.c (avr_device_to_as): New.
26723 (avr_device_to_arch): Rename to avr_device_to_ld.
26725 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26727 PR middle-end/56461
26728 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
26729 method on dr_chain and result_chain.
26731 PR middle-end/56461
26732 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
26733 pointer_set_destroy on not_executed_last_iteration.
26735 PR middle-end/56461
26736 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
26738 PR middle-end/56461
26739 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
26740 FOR_EACH_DEFINED_FUNCTION when freeing state.
26742 PR middle-end/56461
26743 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
26745 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
26748 PR middle-end/56461
26749 * ipa-cp.c (decide_whether_version_node): Call vec_free on
26750 known_aggs[i].items and release known_aggs vector.
26752 PR middle-end/56461
26753 * ipa-reference.c (propagate): Free node_info even for alias nodes.
26755 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26757 * config/microblaze/microblaze.c (microblaze_emit_compare):
26758 Use xor for EQ/NE comparisions.
26759 * config/microblaze/microblaze.md (cstoresf4): Add constraints
26760 (cbranchsf4): Adjust operator to comparison_operator.
26762 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26764 PR middle-end/56461
26765 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
26767 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
26768 vec_safe_push, always update *slot.
26769 (redirect_edge_var_map_clear): Use vec_free.
26770 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
26771 (free_var_map_entry): Use vec_free.
26772 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
26773 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
26775 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
26777 PR middle-end/45472
26778 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
26779 when the may_trap_p bit of the exprs being merged differs.
26780 Reorder tests for speculativeness in the logical and operator.
26782 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26784 * incpath.c (add_standard_paths): Use reconcat instead of concat
26785 where appropriate and avoid leaking memory.
26787 * opts.h: Include obstack.h.
26788 (opts_concat): New prototype.
26789 (opts_obstack): New declaration.
26790 * opts.c (opts_concat): New function.
26791 (opts_obstack): New variable.
26792 (init_options_struct): Call gcc_init_obstack on opts_obstack.
26793 (finish_options): Use opts_concat instead of concat
26794 and XOBNEWVEC instead of XNEWVEC.
26795 * opts-common.c (generate_canonical_option, decode_cmdline_option,
26796 generate_option): Likewise.
26797 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
26798 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
26801 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
26802 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
26804 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26806 PR middle-end/56461
26807 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
26809 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26811 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
26812 (arm_block_move_unaligned_straight): Likewise.
26813 (arm_adjust_block_mem): Likewise.
26815 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26818 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
26819 temp, cond and label.
26820 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
26823 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
26824 * config/c6x/c6x.h (dbx_register_map): Update declaration.
26827 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
26828 of prologue/epilogue functions.
26831 * config/tilegx/tilegx.c (tilegx_expand_prologue):
26832 Remove unused variable cfa_offset.
26833 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
26836 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
26837 type promotion to unsigned.
26840 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
26841 for HOST_WIDE_INT of 32 bit / same size as int.
26842 (arm_block_move_unaligned_straight): Likewise.
26843 (arm_adjust_block_mem): Likewise.
26846 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
26849 2013-02-26 Marek Polacek <polacek@redhat.com>
26851 PR tree-optimization/56426
26852 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
26854 2013-02-26 Richard Biener <rguenther@suse.de>
26857 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
26858 unused variable loops.
26860 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26862 PR tree-optimization/56448
26863 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
26864 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
26865 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
26866 later operands of the references, or even first operand for
26867 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
26869 PR tree-optimization/56443
26870 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
26871 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
26872 to type_for_mode langhook.
26874 2013-02-25 Matt Turner <mattst88@gmail.com>
26876 * doc/invoke.texi: Document r4700.
26878 2013-02-25 Richard Biener <rguenther@suse.de>
26880 PR tree-optimization/56175
26881 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
26883 (simplify_bitwise_binary): ... here. Also guard the conversion
26884 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
26886 2013-02-25 Catherine Moore <clm@codesourcery.com>
26889 2013-02-24 Catherine Moore <clm@codesourcery.com>
26890 Maciej W. Rozycki <macro@codesourcery.com>
26891 Tom de Vries <tom@codesourcery.com>
26892 Nathan Sidwell <nathan@codesourcery.com>
26893 Iain Sandoe <iain@codesourcery.com>
26894 Nathan Froyd <froydnj@codesourcery.com>
26895 Chao-ying Fu <fu@mips.com>
26897 * doc/extend.texi: (micromips, nomicromips, nocompression):
26898 Document new function attributes.
26899 * doc/invoke.texi (minterlink-compressed, mmicromips,
26900 m14k, m14ke, m14kec): Document new options.
26901 (minterlink-mips16): Update documentation.
26902 * doc/md.texi (ZC, ZD): Document new constraints.
26903 * configure.ac (gcc_cv_as_micromips): Check if linker
26904 supports the .set micromips directive.
26905 * configure: Regenerate.
26906 * config.in: Regenerate.
26907 * config/mips/mips-tables.opt: Regenerate.
26908 * config/mips/micromips.md: New file.
26909 * constraints.md (ZC, AD): New constraints.
26910 * config/mips/predicates.md (movep_src_register): New predicate.
26911 (movep_src_operand): New predicate.
26912 (non_volatile_mem_operand): New predicate.
26913 * config/mips/mips.md (multimem): New type.
26914 (length): Differentiate between 17-bit and 18-bit branch offsets.
26915 (MOVEP1, MOVEP2): New mode iterator.
26916 (mov_<load>l): Use ZC constraint.
26917 (mov_<load>r): Likewise.
26918 (mov_<store>l): Likewise.
26919 (mov_<store>r): Likewise.
26920 (*branch_equality<mode>_inverted): Add microMIPS support.
26921 (*branch_equality<mode>): Likewise.
26922 (*jump_absolute): Likewise.
26923 (indirect_jump_<mode>): Likewise.
26924 (tablejump_<mode>): Likewise.
26925 (<optab>_internal): Likewise.
26926 (sibcall_internal): Likewise.
26927 (sibcall_value_internal): Likewise.
26928 (prefetch): Use constraint ZD.
26929 * config/mips/mips.opt (minterlink-compressed): New option.
26930 (minterlink-mips16): Now an alias for minterlink-compressed.
26931 (mmicromips): New option.
26932 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26933 (compare_and_swap_12): Likewise.
26934 (sync_add<mode>): Likewise.
26935 (sync_<optab>_12): Likewise.
26936 (sync_old_<optab>_12): Likewise.
26937 (sync_new_<optab>_12): Likewise.
26938 (sync_nand_12): Likewise.
26939 (sync_old_nand_12): Likewise.
26940 (sync_new_nand_12): Likewise.
26941 (sync_sub<mode>): Likewise.
26942 (sync_old_add<mode>): Likewise.
26943 (sync_old_sub<mode>): Likewise.
26944 (sync_new_add<mode>): Likewise.
26945 (sync_new_sub<mode>): Likewise.
26946 (sync_<optab><mode>): Likewise.
26947 (sync_old_<optab><mode>): Likewise.
26948 (sync_new_<optab><mode>): Likewise.
26949 (sync_nand<mode>): Likewise.
26950 (sync_old_nand<mode>): Likewise.
26951 (sync_new_nand<mode>): Likewise.
26952 (sync_lock_test_and_set<mode>): Likewise.
26953 (test_and_set_12): Likewise.
26954 (atomic_compare_and_swap<mode>): Likewise.
26955 (atomic_exchange<mode>_llsc): Likewise.
26956 (atomic_fetch_add<mode>_llsc): Likewise.
26957 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26958 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26959 (umips_save_restore_pattern_p): Likewise.
26960 (umips_load_store_pair_p): Likewise.
26961 (umips_output_load_store_pair): Likewise.
26962 (umips_movep_target_p): Likewise.
26963 (umips_12bit_offset_address_p): Likewise.
26964 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26965 (mips_base_mips16): Rename this...
26966 (mips_base_compression_flags): ...to this. Update all uses.
26967 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26968 (mips_mips16_decl_p): Delete.
26969 (mips_nomips16_decl_p): Delete.
26970 (mips_get_compress_on_flags): New function.
26971 (mips_get_compress_off_flags): New function.
26972 (mips_get_compress_mode): New function.
26973 (mips_get_compress_on_name): New function.
26974 (mips_get_compress_off_name): New function.
26975 (mips_insert_attributes): Support multiple compression types.
26976 (mips_merge_decl_attributes): Likewise.
26977 (umips_12bit_offset_address_p): New function.
26978 (mips_start_function_definition): Emit .set micromips directive.
26979 (mips_call_may_need_jalx_p): New function.
26980 (mips_function_ok_for_sibcall): Add microMIPS support.
26981 (mips_print_operand_punctuation): Support short delay slots and
26983 (umips_swm_mask, umips_swm_encoding): New.
26984 (umips_build_save_restore): New function.
26985 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26986 (was_mips16_p): Remove.
26987 (old_compression_mode): New.
26988 (mips_set_compression_mode): New function.
26989 (mips_set_current_function): Add microMIPS support.
26990 (mips_option_override): Likewise.
26991 (umips_save_restore_pattern_p): New function.
26992 (umips_output_save_restore): New function.
26993 (umips_load_store_pair_p_1): New function.
26994 (umips_load_store_pair_p): New function.
26995 (umips_output_load_store_pair_1): New function.
26996 (umips_output_load_store_pair): New function.
26997 (umips_movep_target_p) New function.
26998 (mips_prepare_pch_save): Add microMIPS support.
26999 * config/mips/mips.h (TARGET_COMPRESSION): New.
27000 (TARGET_CPU_CPP_BUILTINS): Update macro
27001 to use new compression flags and to support microMIPS.
27002 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27003 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27004 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27005 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27006 (ASM_SPEC): Support mmicromips and mno-micromips.
27007 (M16STORE_REG_P): New macro.
27008 (MIPS_CALL): Support TARGET_MICROMIPS.
27009 (MICROMIPS_J): New macro.
27010 (mips_base_mips16): Rename this...
27011 (mips_base_compression_flags): ...to this.
27012 (UMIPS_12BIT_OFFSET_P): New macro.
27013 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27014 (MULTILIB_DIRNAMES): Likewise.
27016 2013-02-25 Tom de Vries <tom@codesourcery.com>
27018 PR rtl-optimization/56131
27019 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
27020 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
27021 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
27023 2013-02-25 Tobias Burnus <burnus@net-b.de>
27025 * doc/invoke.texi (-fsanitize=): Move from optimization
27026 to debugging options.
27028 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27030 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
27032 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27033 Alexander Monakov <amonakov@ispras.ru>
27035 PR middle-end/56077
27036 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
27037 flush pending lists also on non-jumps. Adjust comment.
27039 2013-02-24 Catherine Moore <clm@codesourcery.com>
27040 Maciej W. Rozycki <macro@codesourcery.com>
27041 Tom de Vries <tom@codesourcery.com>
27042 Nathan Sidwell <nathan@codesourcery.com>
27043 Iain Sandoe <iain@codesourcery.com>
27044 Nathan Froyd <froydnj@codesourcery.com>
27045 Chao-ying Fu <fu@mips.com>
27047 * doc/extend.texi: (micromips, nomicromips, nocompression):
27048 Document new function attributes.
27049 * doc/invoke.texi (minterlink-compressed, mmicromips,
27050 m14k, m14ke, m14kec): Document new options.
27051 (minterlink-mips16): Update documentation.
27052 * doc/md.texi (ZC, ZD): Document new constraints.
27053 * configure.ac (gcc_cv_as_micromips): Check if linker
27054 supports the .set micromips directive.
27055 * configure: Regenerate.
27056 * config.in: Regenerate.
27057 * config/mips/mips-tables.opt: Regenerate.
27058 * config/mips/micromips.md: New file.
27059 * constraints.md (ZC, AD): New constraints.
27060 * config/mips/predicates.md (movep_src_register): New predicate.
27061 (movep_src_operand): New predicate.
27062 (non_volatile_mem_operand): New predicate.
27063 * config/mips/mips.md (multimem): New type.
27064 (length): Differentiate between 17-bit and 18-bit branch offsets.
27065 (MOVEP1, MOVEP2): New mode iterator.
27066 (mov_<load>l): Use ZC constraint.
27067 (mov_<load>r): Likewise.
27068 (mov_<store>l): Likewise.
27069 (mov_<store>r): Likewise.
27070 (*branch_equality<mode>_inverted): Add microMIPS support.
27071 (*branch_equality<mode>): Likewise.
27072 (*jump_absolute): Likewise.
27073 (indirect_jump_<mode>): Likewise.
27074 (tablejump_<mode>): Likewise.
27075 (<optab>_internal): Likewise.
27076 (sibcall_internal): Likewise.
27077 (sibcall_value_internal): Likewise.
27078 (prefetch): Use constraint ZD.
27079 * config/mips/mips.opt (minterlink-compressed): New option.
27080 (minterlink-mips16): Now an alias for minterlink-compressed.
27081 (mmicromips): New option.
27082 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27083 (compare_and_swap_12): Likewise.
27084 (sync_add<mode>): Likewise.
27085 (sync_<optab>_12): Likewise.
27086 (sync_old_<optab>_12): Likewise.
27087 (sync_new_<optab>_12): Likewise.
27088 (sync_nand_12): Likewise.
27089 (sync_old_nand_12): Likewise.
27090 (sync_new_nand_12): Likewise.
27091 (sync_sub<mode>): Likewise.
27092 (sync_old_add<mode>): Likewise.
27093 (sync_old_sub<mode>): Likewise.
27094 (sync_new_add<mode>): Likewise.
27095 (sync_new_sub<mode>): Likewise.
27096 (sync_<optab><mode>): Likewise.
27097 (sync_old_<optab><mode>): Likewise.
27098 (sync_new_<optab><mode>): Likewise.
27099 (sync_nand<mode>): Likewise.
27100 (sync_old_nand<mode>): Likewise.
27101 (sync_new_nand<mode>): Likewise.
27102 (sync_lock_test_and_set<mode>): Likewise.
27103 (test_and_set_12): Likewise.
27104 (atomic_compare_and_swap<mode>): Likewise.
27105 (atomic_exchange<mode>_llsc): Likewise.
27106 (atomic_fetch_add<mode>_llsc): Likewise.
27107 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27108 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27109 (umips_save_restore_pattern_p): Likewise.
27110 (umips_load_store_pair_p): Likewise.
27111 (umips_output_load_store_pair): Likewise.
27112 (umips_movep_target_p): Likewise.
27113 (umips_12bit_offset_address_p): Likewise.
27114 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27115 (mips_base_mips16): Rename this...
27116 (mips_base_compression_flags): ...to this. Update all uses.
27117 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27118 (mips_mips16_decl_p): Delete.
27119 (mips_nomips16_decl_p): Delete.
27120 (mips_get_compress_on_flags): New function.
27121 (mips_get_compress_off_flags): New function.
27122 (mips_get_compress_mode): New function.
27123 (mips_get_compress_on_name): New function.
27124 (mips_get_compress_off_name): New function.
27125 (mips_insert_attributes): Support multiple compression types.
27126 (mips_merge_decl_attributes): Likewise.
27127 (umips_12bit_offset_address_p): New function.
27128 (mips_start_function_definition): Emit .set micromips directive.
27129 (mips_call_may_need_jalx_p): New function.
27130 (mips_function_ok_for_sibcall): Add microMIPS support.
27131 (mips_print_operand_punctuation): Support short delay slots and
27133 (umips_swm_mask, umips_swm_encoding): New.
27134 (umips_build_save_restore): New function.
27135 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27136 (was_mips16_p): Remove.
27137 (old_compression_mode): New.
27138 (mips_set_compression_mode): New function.
27139 (mips_set_current_function): Add microMIPS support.
27140 (mips_option_override): Likewise.
27141 (umips_save_restore_pattern_p): New function.
27142 (umips_output_save_restore): New function.
27143 (umips_load_store_pair_p_1): New function.
27144 (umips_load_store_pair_p): New function.
27145 (umips_output_load_store_pair_1): New function.
27146 (umips_output_load_store_pair): New function.
27147 (umips_movep_target_p) New function.
27148 (mips_prepare_pch_save): Add microMIPS support.
27149 * config/mips/mips.h (TARGET_COMPRESSION): New.
27150 (TARGET_CPU_CPP_BUILTINS): Update macro
27151 to use new compression flags and to support microMIPS.
27152 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27153 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27154 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27155 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27156 (ASM_SPEC): Support mmicromips and mno-micromips.
27157 (M16STORE_REG_P): New macro.
27158 (MIPS_CALL): Support TARGET_MICROMIPS.
27159 (MICROMIPS_J): New macro.
27160 (mips_base_mips16): Rename this...
27161 (mips_base_compression_flags): ...to this.
27162 (UMIPS_12BIT_OFFSET_P): New macro.
27163 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27164 (MULTILIB_DIRNAMES): Likewise.
27166 2013-02-24 Jakub Jelinek <jakub@redhat.com>
27169 * target-globals.c (save_target_globals): For init_reg_sets and
27170 target_reinit remporarily set this_fn_optabs to this_target_optabs.
27172 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
27174 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
27175 * config/aarch64/t-aarch64
27176 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
27178 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
27180 PR inline-asm/56148
27181 * lra-constraints.c (process_alt_operands): Reload operand
27182 conflicting with earlier clobber only if no more other conflicting
27185 2013-02-22 Jakub Jelinek <jakub@redhat.com>
27188 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
27189 if not linking a shared library.
27191 2013-02-22 Seth LaForge <sethml@google.com>
27193 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
27195 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
27197 * config/arm/arm.md (split for extendsidi): Update condition.
27198 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
27199 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
27200 (qhs_zextenddi_cstr): Likewise.
27202 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27204 PR middle-end/56420
27205 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
27206 avoid signed wrapping.
27207 (expand_mult): Handle properly multiplication by
27208 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
27209 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
27210 in the compiler if coeff is HOST_WIDE_INT_MIN.
27211 (expand_divmod): Don't make ext_op1 static, change it's type to
27212 uhwi. Avoid undefined behavior in -INTVAL (op1).
27214 PR rtl-optimization/50339
27215 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
27217 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
27218 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
27219 into splitting_ashiftrt field.
27220 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
27222 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
27225 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
27227 PR middle-end/56108
27228 * trans-mem.c (execute_tm_mark): Do not expand transactions that
27229 are sure to go irrevocable.
27231 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
27233 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
27234 scalars are valid operands.
27236 2013-02-21 Martin Jambor <mjambor@suse.cz>
27238 PR tree-optimization/56310
27239 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
27240 only matching indices and non-negative final offsets.
27241 (intersect_aggregates_with_edge): Pass src_idx to
27242 agg_replacements_to_vector. Pass src_idx insstead of index to
27243 intersect_with_agg_replacements.
27245 2013-02-21 Martin Jambor <mjambor@suse.cz>
27247 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
27248 instead of hard-wired defaults.
27250 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
27252 * doc/invoke.texi (MIPS Options): Update documentation of the
27253 floating-point multiply-accumulate instruction restrictions.
27255 2013-02-21 Kostya Serebryany <kcc@google.com>
27257 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
27258 asan_shadow_offset on x86_64 linux.
27260 2013-02-21 Richard Biener <rguenther@suse.de>
27262 PR tree-optimization/56415
27264 2013-02-11 Richard Biener <rguenther@suse.de>
27266 PR tree-optimization/56273
27267 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27270 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27273 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
27276 PR inline-asm/56405
27277 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
27278 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
27280 2013-02-20 Jan Hubicka <jh@suse.cz>
27282 PR tree-optimization/56265
27283 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
27284 when target is referenced for first time.
27286 2013-02-20 Richard Biener <rguenther@suse.de>
27288 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
27289 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
27290 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
27291 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
27292 not return anything.
27293 (rename_ssa_copies): Do not remove unused locals.
27294 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
27295 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
27296 * passes.c (execute_function_todo): Do not schedule unused locals
27297 removal if cleanup_tree_cfg did something.
27298 * tree-ssa-live.c (remove_unused_locals): Dump statistics
27299 about the number of removed locals.
27301 2013-02-20 Richard Biener <rguenther@suse.de>
27303 PR tree-optimization/56398
27304 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
27306 2013-02-20 Martin Jambor <mjambor@suse.cz>
27308 PR tree-optimization/55334
27309 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
27310 restricted pointers to arrays.
27312 2013-02-20 Richard Biener <rguenther@suse.de>
27313 Jakub Jelinek <jakub@redhat.com>
27315 PR tree-optimization/56396
27316 * tree-ssa-ccp.c (n_const_val): New static variable.
27317 (get_value): Return NULL for SSA names we don't have a lattice
27319 (ccp_initialize): Initialize n_const_val.
27320 * tree-ssa-copy.c (n_copy_of): New static variable.
27321 (init_copy_prop): Initialize n_copy_of.
27322 (get_value): Return NULL_TREE for SSA names we don't have a
27325 2013-02-20 Martin Jambor <mjambor@suse.cz>
27327 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
27329 2013-02-20 Richard Biener <rguenther@suse.de>
27331 * genpreds.c (write_lookup_constraint): Do not compare first
27332 letter of the constraint again.
27334 2013-02-20 Richard Biener <rguenther@suse.de>
27336 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
27338 (get_use_iv_cost): Terminate hashtable walk when coming across
27341 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
27343 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
27344 reassociation for avx2 targets.
27346 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27348 * config/microblaze/microblaze.c: microblaze_has_clz = 0
27349 Add version check for v8.10.a to enable microblaze_has_clz
27350 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
27351 version and TARGET_PATTERN_COMPARE check
27352 * config/microblaze/microblaze.md: New clzsi2 instruction
27354 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27356 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
27357 function before branching.
27359 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27361 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
27363 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
27365 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27367 PR middle-end/55889
27368 * sel-sched.c: Include ira.h.
27369 (implicit_clobber_conflict_p): New function.
27370 (moveup_expr): Use it.
27371 * Makefile.in (sel-sched.o): Depend on ira.h.
27373 2013-02-19 Richard Biener <rguenther@suse.de>
27375 PR tree-optimization/56384
27376 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
27377 (vn_hash_type): Split out from ...
27378 (vn_hash_constant_with_type): ... here.
27379 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
27380 (vn_phi_eq): Compare types from vn_phi_s structure.
27381 (vn_phi_lookup): Populate vn_phi_s type.
27382 (vn_phi_insert): Likewise.
27384 2013-02-19 Jakub Jelinek <jakub@redhat.com>
27386 PR tree-optimization/56350
27387 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
27388 if haven't found reduction or nested cycle operand, rather than
27389 asserting we must find it.
27391 PR tree-optimization/56381
27392 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
27395 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
27396 Jakub Jelinek <jakub@redhat.com>
27399 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
27400 (swap_optab_enable): Same.
27401 (init_all_optabs): Use argument instead of global.
27402 * tree.h (struct tree_optimization_option): New field target_optabs.
27403 * expr.h (init_all_optabs): Add argument to prototype.
27404 (TREE_OPTIMIZATION_OPTABS): New.
27405 (save_optabs_if_changed): Protoize.
27406 * optabs.h: Declare this_fn_optabs.
27407 * optabs.c (save_optabs_if_changed): New.
27408 Declare this_fn_optabs.
27409 (init_optabs): Add argument to init_all_optabs() call.
27410 * function.c (invoke_set_current_function_hook): Handle per
27412 * function.h (struct function): New field optabs.
27413 * config/mips/mips.c (mips_set_mips16_mode): Handle when
27414 optimization_current_node has changed.
27415 * target-globals.h (save_target_globals_default_opts): Protoize.
27416 * target-globals.c (save_target_globals_default_opts): New.
27418 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27421 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
27422 registers %fr12 and %fr12R as call used.
27425 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
27426 and HImode, require all displacements to be an integer multiple of
27428 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
27429 only allow QImode and HImode when reload is in progress and strict is
27430 true. Likewise for symbolic addresses. Use base14_operand to check
27431 displacements in REG+BASE addresses.
27433 2013-02-18 Richard Biener <rguenther@suse.de>
27435 PR tree-optimization/56366
27436 * tree-vect-loop.c (get_initial_def_for_induction): Properly
27437 handle sign-conversion of outer-loop initial induction value.
27439 2013-02-18 Richard Biener <rguenther@suse.de>
27441 PR middle-end/56349
27442 * cfghooks.c (merge_blocks): If we merge a latch into another
27443 block adjust references to it.
27444 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
27445 (verify_loop_structure): Verify that a recorded latch is in fact
27448 2013-02-18 Richard Biener <rguenther@suse.de>
27450 PR tree-optimization/56321
27451 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
27452 order SSA name release and virtual operand unlinking.
27454 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27456 * config/microblaze/microblaze.md (save_stack_block): Define.
27457 (restore_stack_block): Likewise.
27459 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27461 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
27462 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
27463 * config/microblaze/microblaze.c (microblaze_option_override):
27464 Bail out early for PIC modes when target does not support PIC.
27466 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27468 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
27469 Replace with a microblaze version.
27470 (microblaze_trampoline_init): Adapt for microblaze.
27471 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
27474 2013-02-16 Jakub Jelinek <jakub@redhat.com>
27475 Dodji Seketeli <dodji@redhat.com>
27478 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
27479 (instrument_mem_region_access): Do not forget to always put
27480 instrumentation of the of 'base' and 'base + len' in a "if (len !=
27481 0) statement, even for cases where either 'base' or 'base + len'
27482 are not instrumented -- because they have been previously
27483 instrumented. Simplify the logic by putting all the statements
27484 instrument 'base + len' inside a sequence, and then insert that
27485 sequence right before the current insertion point. Then, to
27486 instrument 'base + len', just get an iterator on that statement.
27487 And do not forget to update the pointer to iterator the function
27488 received as argument.
27490 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
27492 PR rtl-optimization/56348
27493 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
27495 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
27497 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
27498 (clean_graph_dump_file): Pass base to start_graph_dump.
27500 2013-02-14 Richard Henderson <rth@redhat.com>
27503 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
27505 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
27507 * collect2-aix.h: Define F_LOADONLY.
27509 2013-02-14 Richard Biener <rguenther@suse.de>
27512 * varasm.c (output_constant_def_1): Get the decl representing
27513 the constant as argument.
27514 (output_constant_def): Wrap output_constant_def_1.
27515 (make_decl_rtl): Use output_constant_def_1 with the decl
27516 representing the constant.
27517 (build_constant_desc): Optionally re-use a decl already
27518 representing the constant.
27519 (tree_output_constant_def): Adjust.
27521 2013-02-14 Dodji Seketeli <dodji@redhat.com>
27524 * asan.c (instrument_builtin_call): Really put the length of the
27525 second source argument into src1_len.
27527 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27529 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
27530 argument. If it is false, don't create edge from then_bb to
27532 (insert_if_then_before_iter): Pass true to it.
27533 (build_check_stmt): Pass false to it.
27534 (transform_statements): Flush hash table only on extended basic
27535 block boundaries, rather than at the beginning of every bb.
27536 Don't flush hash table on nonfreeing_call_p calls.
27537 * tree-flow.h (nonfreeing_call_p): New prototype.
27538 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
27540 2013-02-13 David S. Miller <davem@davemloft.net>
27542 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
27544 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
27547 * ira.c (max_regno_before_ira): Move from ...
27549 (fix_reg_equiv_init): Use max_regno_before_ira instead of
27552 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27554 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
27556 2013-02-13 Richard Biener <rguenther@suse.de>
27559 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
27560 globals in MEM_REFs.
27562 2013-02-13 Richard Biener <rguenther@suse.de>
27564 * loop-init.c (loop_optimizer_init): Clear loop state when
27565 re-initializing preserved loops.
27566 * loop-unswitch.c (unswitch_single_loop): Return whether
27567 we unswitched the loop. Do not verify loop state here.
27568 (unswitch_loops): When we unswitched a loop discover new loops.
27570 2013-02-13 Kostya Serebryany <kcc@google.com>
27572 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
27574 * sanitizer.def: Rename __asan_init to __asan_init_v1.
27576 2013-02-12 Dodji Seketeli <dodji@redhat.com>
27578 Avoid instrumenting duplicated memory access in the same basic block
27579 * Makefile.in (asan.o): Add new dependency on hash-table.h
27580 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
27581 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
27582 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
27583 (free_mem_ref_resources, has_mem_ref_been_instrumented)
27584 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
27585 (get_mem_ref_of_assignment): New functions.
27586 (get_mem_refs_of_builtin_call): Extract from
27587 instrument_builtin_call and tweak a little bit to make it fit with
27589 (instrument_builtin_call): Use the new
27590 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
27591 of is_gimple_builtin_call.
27592 (instrument_derefs, instrument_mem_region_access): Insert the
27593 instrumented memory reference into the hash table.
27594 (maybe_instrument_assignment): Renamed instrument_assignment into
27595 this, and change it to advance the iterator when instrumentation
27596 actually happened and return true in that case. This makes it
27597 homogeneous with maybe_instrument_assignment, and thus give a
27598 chance to callers to be more 'regular'.
27599 (transform_statements): Clear the memory reference hash table
27600 whenever we enter a new BB, when we cross a function call, or when
27601 we are done transforming statements. Use
27602 maybe_instrument_assignment instead of instrumentation. No more
27603 need to special case maybe_instrument_assignment and advance the
27604 iterator after calling it; it's now handled just like
27605 maybe_instrument_call. Update comment.
27607 2013-02-13 Richard Biener <rguenther@suse.de>
27609 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
27610 Fix loop discovery code.
27612 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
27614 PR inline-asm/56148
27615 * lra-constraints.c (process_alt_operands): Match early clobber
27616 operand with itself. Check conflicts with earlyclobber only if
27617 the operand is not reloaded. Prefer to reload conflicting operand
27618 if earlyclobber and matching operands are the same.
27620 2013-02-12 Richard Biener <rguenther@suse.de>
27623 * lto-streamer-out.c (write_symbol): Do not output symbols
27624 for hard register variables.
27626 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
27629 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
27630 (umulsidi3_insn, mulsidi3_insn): New insns.
27632 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
27634 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
27635 (struct tune_params): Add vec_costs field.
27636 * config/arm/arm.c (arm_builtin_vectorization_cost)
27637 (arm_add_stmt_cost): New functions.
27638 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
27639 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
27640 (arm_default_vec_cost): New struct of type cpu_vec_costs.
27641 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
27642 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
27643 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
27644 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
27646 2013-02-12 Richard Biener <rguenther@suse.de>
27649 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
27650 decls again if possible.
27652 2013-02-12 Richard Biener <rguenther@suse.de>
27654 PR middle-end/56288
27655 * tree-ssa.c (verify_ssa_name): Fix check, move
27656 SSA_NAME_IN_FREE_LIST check up.
27658 2013-02-12 Jakub Jelinek <jakub@redhat.com>
27659 Steven Bosscher <steven@gcc.gnu.org>
27661 PR rtl-optimization/56151
27662 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
27663 equal to op0 or op1, and last_insn pattern is CODE operation
27664 with MEM dest and one of the operands matches that MEM.
27666 2013-02-11 Sriraman Tallam <tmsriram@google.com>
27668 * doc/extend.texi: Document Function Multiversioning and "default"
27669 parameter string to target attribute.
27670 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
27671 target attribute parameter is "default".
27672 (ix86_compare_version_priority): Remove checks for target attribute.
27673 (ix86_mangle_function_version_assembler_name): Change error to sorry.
27674 Remove check for target attribute equal to NULL. Add assert.
27675 (ix86_generate_version_dispatcher_body): Change error to sorry.
27677 2013-02-11 Iain Sandoe <iain@codesourcery.com>
27678 Jack Howarth <howarth@bromo.med.uc.edu>
27679 Patrick Marlier <patrick.marlier@gmail.com>
27682 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
27683 define ENDFILE_SPEC as TM_DESTRUCTOR.
27684 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
27686 2013-02-11 Alexander Potapenko <glider@google.com>
27687 Jack Howarth <howarth@bromo.med.uc.edu>
27688 Jakub Jelinek <jakub@redhat.com>
27691 * config/darwin.c (cdtor_record): Rename ctor_record.
27692 (sort_cdtor_records): Rename sort_ctor_records.
27693 (finalize_dtors): New routine to sort destructors by
27694 priority before use in assemble_integer.
27695 (machopic_asm_out_destructor): Use finalize_dtors if needed.
27697 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
27699 PR rtl-optimization/56275
27700 * simplify-rtx.c (avoid_constant_pool_reference): Check that
27701 offset is non-negative and less than cmode size before
27702 calling simplify_subreg.
27704 2013-02-11 Richard Biener <rguenther@suse.de>
27706 PR tree-optimization/56264
27707 * cfgloop.h (fix_loop_structure): Adjust prototype.
27708 * loop-init.c (fix_loop_structure): Return the number of
27709 newly discovered loops.
27710 * tree-cfgcleanup.c (repair_loop_structures): When new loops
27711 are discovered, do a full loop-closed SSA rewrite.
27713 2013-02-11 Richard Biener <rguenther@suse.de>
27715 PR tree-optimization/56273
27716 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27718 (check_array_ref): Fix missing newline in dumps.
27719 (search_for_addr_array): Likewise.
27721 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
27723 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
27725 2013-02-09 Jakub Jelinek <jakub@redhat.com>
27728 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
27730 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
27732 PR rtl-optimization/56246
27733 * lra-constraints.c (simplify_operand_subreg): Try to reuse
27735 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
27736 constraints are satisfied.
27738 2013-02-08 Jeff Law <law@redhat.com>
27741 * emit-rtl.c (reg_is_parm_p): New function.
27742 * regs.h (reg_is_parm_p): New prototype.
27743 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
27744 callee-clobbered registers.
27746 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
27749 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
27750 If there is no implicit builtin declaration, just return NULL.
27752 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27754 * config/i386/sse.md (FMAMODEM): New mode iterator.
27755 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
27756 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
27758 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27760 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
27761 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
27762 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
27764 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27766 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
27767 (microblaze*-*-elf): Likewise.
27768 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
27770 * config/microblaze/microblaze-c.c: Add builtin defines for
27771 _LITTLE_ENDIAN and _BIG_ENDIAN.
27772 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
27773 add to TARGET_DEFAULT flags.
27774 Expand ASM_SPEC and LINK_SPEC.
27775 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
27776 * config/microblaze/microblaze.md: Update extendsidi2 and
27777 movdi_internal instructions to use low-order / high-order reg
27779 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
27780 options and inversemask / mask of LITTLE_ENDIAN.
27781 * config/microblaze/t-microblaze: Expand multilib options to
27782 include mlittle-endian (le) and update exceptions patterns.
27784 2013-02-08 Jakub Jelinek <jakub@redhat.com>
27786 PR rtl-optimization/56195
27787 * lra-constraints.c (get_reload_reg): Don't reuse regs
27788 if they have smaller mode than requested, if they have
27789 wider mode than requested, try to return a SUBREG.
27791 PR tree-optimization/56250
27792 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
27793 if type is unsigned and code isn't MULT_EXPR.
27795 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27797 PR tree-optimization/56064
27798 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
27799 bits according to mode.
27800 * fixed-value.h (fixed_from_double_int)
27801 (const_fixed_from_double_int): Adjust comments.
27803 2013-02-08 Richard Biener <rguenther@suse.de>
27806 * lto-streamer.h (struct data_in): Remove current_file, current_line
27807 and current_col members.
27808 * lto-streamer-out.c (lto_output_location): Stream changed bits
27809 en-block for efficiency.
27810 * lto-streamer-in.c (clear_line_info): Remove.
27811 (lto_input_location): Cache current file, line and column
27812 globally via local statics. Read changed bits en-block.
27813 (input_function): Do not call clear_line_info.
27814 (lto_read_body): Likewise.
27815 (lto_input_toplevel_asms): Likewise.
27817 2013-02-08 Michael Matz <matz@suse.de>
27819 PR tree-optimization/52448
27820 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
27821 (nt_call_phase): New static.
27822 (add_or_mark_expr): Only mark accesses with newer phase than any
27824 (nonfreeing_call_p): New.
27825 (nt_init_block): Update nt_call_phase, mark blocks as visited.
27826 (nt_fini_block): Keep blocks marked as visited.
27827 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
27829 2013-02-08 Richard Biener <rguenther@suse.de>
27831 * ira.c (ira): Free broken dominator information.
27833 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27835 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
27837 2013-02-08 Marek Polacek <polacek@redhat.com>
27839 * cfgloop.c (verify_loop_structure): Add more checking of headers.
27841 2013-02-08 Richard Biener <rguenther@suse.de>
27843 PR middle-end/56181
27844 * cfgloop.h (flow_loops_find): Adjust.
27845 (bb_loop_header_p): Declare.
27846 * cfgloop.c (bb_loop_header_p): New function split out from ...
27847 (flow_loops_find): ... here. Adjust function signature,
27848 support incremental loop structure update.
27849 (verify_loop_structure): Cleanup. Verify a loop is a loop.
27850 * cfgloopmanip.c (fix_loop_structure): Move ...
27851 * loop-init.c (fix_loop_structure): ... here.
27852 (apply_loop_flags): Split out from ...
27853 (loop_optimizer_init): ... here.
27854 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
27855 in incremental mode, only remove dead loops here.
27857 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27860 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
27861 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
27862 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
27863 (*round<mode>3.libgcc): New insns for fixed-modes.
27864 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
27865 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
27866 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
27867 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
27868 implementations. Define to __builtin_avr_absFX,
27869 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
27870 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
27871 __builtin_avr_countlsFX, respectively.
27872 * config/avr/avr-c.c (target.h): Include it.
27873 (enum avr_builtin_id): New enum.
27874 (avr_resolve_overloaded_builtin): New static function.
27875 (avr_register_target_pragmas): Use it to set
27876 targetm.resolve_overloaded_builtin.
27877 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
27878 tree nodes used by DEF_BUILTIN.
27879 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
27880 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
27881 <AVR_BUILTIN_xxBITS>: Same.
27883 2013-02-08 Richard Biener <rguenther@suse.de>
27885 * cfgloop.c (verify_loop_structure): Properly handle
27886 a loop exiting to another loop header.
27887 * ira-int.h (ira_loops): Remove.
27888 * ira.c (ira_loops): Remove.
27889 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
27890 (do_reload): Use loop_optimizer_finalize.
27891 * ira-build.c (create_loop_tree_nodes): Use get_loops and
27892 number_of_loops to access the loop tree.
27893 (more_one_region_p): Likewise.
27894 (finish_loop_tree_nodes): Likewise.
27895 (rebuild_regno_allocno_maps): Likewise.
27896 (mark_loops_for_removal): Likewise.
27897 (mark_all_loops_for_removal): Likewise.
27898 (remove_unnecessary_regions): Likewise.
27899 (ira_build): Likewise.
27900 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
27902 2013-02-08 Richard Biener <rguenther@suse.de>
27904 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
27905 * ipa-pure-const.c (analyze_function): Avoid calling
27906 mark_irreducible_loops twice.
27907 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
27909 2013-02-07 David S. Miller <davem@davemloft.net>
27911 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
27913 * var-tracking.c (vt_add_function_parameter): Test the presence of
27914 HAVE_window_save properly and do not remap argument registers when
27915 we have a leaf function.
27917 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
27920 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
27922 * config/i386/i386.c (ix86_print_operand): Ditto.
27924 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27926 * lra-constraints.c (process_alt_operands): Fix recently added comment.
27928 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27930 PR rtl-optimization/56225
27931 * lra-constraints.c (process_alt_operands): Check that reload hard
27932 reg can hold value for strict_low_part.
27934 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27937 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
27938 dwarf2out_end_function.
27939 (in_first_function_p, maybe_at_text_label_p,
27940 first_loclabel_num_not_at_text_label): New variables.
27941 (dwarf2out_var_location): In the first function find out
27942 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
27943 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
27946 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
27948 PR rtl-optimization/56178
27949 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
27950 SUBREG of a register. Tidy up related block of code.
27951 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
27952 note if the source is a register or a SUBREG of a register.
27954 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27957 * config/rs6000/rs6000.md (ptrm): New mode attr.
27958 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
27959 call_value_indirect_aix<pttrsize>,
27960 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
27963 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
27965 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
27966 if -bnortl. Convert to strcmp and strncmp.
27968 2013-02-07 Alan Modra <amodra@gmail.com>
27971 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
27972 addresses won't wrap when offsetting.
27973 (rs6000_secondary_reload): Provide secondary reloads needed for
27974 wrapping LO_SUM addresses.
27976 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
27978 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
27979 MACH, just __MACH__.
27981 2013-02-06 Richard Biener <rguenther@suse.de>
27983 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
27984 instead of calling fix_loop_structure.
27986 2013-02-06 Jakub Jelinek <jakub@redhat.com>
27988 PR middle-end/56217
27989 * omp-low.c (use_pointer_for_field): Return false if
27990 lower_send_shared_vars doesn't generate any copy-out code.
27992 2013-02-06 Tom de Vries <tom@codesourcery.com>
27994 PR rtl-optimization/56131
27995 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
27996 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
27997 of the label is NULL. Add comment.
27999 2013-02-05 Jakub Jelinek <jakub@redhat.com>
28001 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
28004 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
28005 (STATIC_LIBTSAN_LIBS): Likewise.
28006 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
28007 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
28008 is defined, don't add anything else beyond that.
28009 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
28010 (LINK_COMMAND_SPEC): Use them.
28012 PR tree-optimization/56205
28013 * tree-stdarg.c (check_all_va_list_escapes): Return true if
28014 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
28015 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
28017 2013-02-05 Richard Biener <rguenther@suse.de>
28019 PR tree-optimization/53342
28020 PR tree-optimization/53185
28021 * tree-vectorizer.h (vect_check_strided_load): Remove.
28022 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
28023 not disallow peeling for vectorized strided loads.
28024 (vect_check_strided_load): Make static and simplify.
28025 (vect_analyze_data_refs): Adjust.
28026 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
28027 correctly when vectorizing strided loads.
28029 2013-02-05 Richard Biener <rguenther@suse.de>
28031 * doc/install.texi: Refer to ISL, not PPL.
28033 2013-02-05 Jan Hubicka <jh@suse.cz>
28035 PR tree-optimization/55789
28036 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
28038 2013-02-05 Jan Hubicka <jh@suse.cz>
28040 PR tree-optimization/55789
28041 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
28042 the dead call anyway.
28044 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
28047 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
28049 2013-02-04 Alexander Potapenko <glider@google.com>
28050 Jack Howarth <howarth@bromo.med.uc.edu>
28051 Jakub Jelinek <jakub@redhat.com>
28054 * config/darwin.c (sort_ctor_records): Stabilized qsort
28055 on constructor priority by using original position.
28056 (finalize_ctors): New routine to sort constructors by
28057 priority before use in assemble_integer.
28058 (machopic_asm_out_constructor): Use finalize_ctors if needed.
28060 2013-02-04 Jakub Jelinek <jakub@redhat.com>
28063 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
28064 about visibility on artificial decls.
28065 * config/sol2.c (solaris_assemble_visibility): Likewise.
28067 2013-02-04 Kai Tietz <ktietz@redhat.com>
28070 * config/i386/i386.c (function_value_ms_64): Add additional valtype
28071 argument and improve checking of return-argument types for 16-byte
28073 (ix86_function_value_1): Add additional valtype argument on call
28074 of function_value_64.
28075 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
28076 handling infunction_value_64 function.
28078 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
28080 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
28082 2013-02-04 Richard Biener <rguenther@suse.de>
28084 PR tree-optimization/56188
28085 * tree-ssa-structalias.c (label_visit): Consider case with
28086 initially non-empty points-to set.
28087 (perform_var_substitution): Dump node mapping and clean up.
28089 2013-02-04 Richard Guenther <rguenther@suse.de>
28092 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
28093 node prevail as last resort.
28094 (lto_symtab_merge_decls): Remove guard on LTRANS here.
28095 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
28097 2013-02-04 Richard Biener <rguenther@suse.de>
28099 PR tree-optimization/56113
28100 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
28102 (equiv_class_lookup_or_add): ... this.
28103 (label_visit): Adjust and fix error in previous patch.
28104 (perform_var_substitution): Adjust.
28106 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
28108 * config/sh/divtab.c: Fix formatting and comments throughout the file.
28109 * config/sh/sh4-300.md: Likewise.
28110 * config/sh/sh4a.md: Likewise.
28111 * config/sh/constraints.md: Likewise.
28112 * config/sh/sh.md: Likewise.
28113 * config/sh/netbsd-elf.h: Likewise.
28114 * config/sh/predicates.md: Likewise.
28115 * config/sh/sh-protos.h: Likewise.
28116 * config/sh/ushmedia.h: Likewise.
28117 * config/sh/linux.h: Likewise.
28118 * config/sh/sh.c: Likewise.
28119 * config/sh/superh.h: Likewise.
28120 * config/sh/elf.h: Likewise.
28121 * config/sh/sh4.md: Likewise.
28122 * config/sh/sh.h: Likewise.
28124 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28126 * config/pa/constraints.md: Adjust unused letters. Change "T"
28127 constraint to match_test floating_point_store_memory_operand().
28128 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
28129 (base14_operand): New.
28130 (floating_point_store_memory_operand): New.
28131 (integer_store_memory_operand): Revise to use base14_operand and
28132 reg_plus_base_memory_operand.
28133 (move_dest_operand): Allow symbolic_memory_operands.
28134 (symbolic_memory_operand): Check for LO_SOM.
28135 (symbolic_operand): Change default case to break.
28136 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
28137 CONST_DOUBLE values to be reloaded by putting them into memory when
28138 the destination is a floating point register.
28139 (movdf): Remove code to handle CONST_DOUBLE.
28141 (reload_indf_r1): New.
28142 (reload_insf_r1): New.
28143 Consistently use "Q" and "T" constraints with integer and floating
28144 point move instructions, respectively.
28145 (movdi): Remove FAIL.
28146 Change predicate for source operand unamed DImode move from
28147 general_operand to move_src_operand.
28148 (umulsidi3): Change predicate for destination operand to
28150 Likewise for similar unamed patterns.
28151 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
28152 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
28153 (hppa_legitimize_address): Simplify mask calculation.
28154 (pa_emit_move_sequence): Revised handling of secondary reloads from
28155 REG+D addresses for floating point loads and stores. Directly handle
28156 loading CONST0_RTX (mode) to a floating point register.
28157 (pa_secondary_reload): Handle reloading DF and SFmode constant values
28158 to floating point registers. Don't restrict secondary reloads to
28159 floating point registers to integer modes. Revise some comments and
28161 (TARGET_LEGITIMATE_ADDRESS_P): Define.
28162 (pa_legitimate_address_p): New.
28163 (pa_legitimize_reload_address): New.
28164 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
28165 (STRICT_REG_OK_FOR_BASE_P): New.
28166 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
28167 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
28169 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
28170 Andrew Dixie <andrewd@gentrack.com>
28172 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
28175 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
28177 * expmed.c (extract_bit_field_1): Pass the full width of the
28178 structure to get_best_reg_extraction_insn.
28180 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28183 * configure.ac (use_cxa_atexit): Add AIX.
28184 * configure: Regenerate.
28186 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
28188 2013-02-01 Jakub Jelinek <jakub@redhat.com>
28191 * final.c (need_profile_function): New variable.
28192 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
28193 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
28194 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
28195 notes, targetm.asm_out.function_prologue doesn't emit anything,
28196 HAVE_prologue and profiler should be emitted before prologue,
28197 set need_profile_function instead of emitting it.
28198 (final_scan_insn): If need_profile_function, emit
28199 profile_function on the first NOTE_INSN_BASIC_BLOCK or
28200 NOTE_INSN_FUNCTION_BEG note.
28202 2013-02-01 Richard Henderson <rth@redhat.com>
28204 * config/rs6000/rs6000.md (smulditi3): New.
28207 * config/alpha/alpha.md (umulditi3): New.
28209 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28211 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
28212 (ASM_OUTPUT_ALIGNED_LOCAL): New.
28214 2013-02-01 Richard Biener <rguenther@suse.de>
28216 PR tree-optimization/56113
28217 * tree-ssa-structalias.c (label_visit): Reduce work for
28218 single-predecessor nodes.
28220 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
28222 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
28223 range isn't testing for zero.
28225 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
28227 PR middle-end/56113
28228 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
28230 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
28231 Nick Clifton <nickc@redhat.com>
28233 * config/v850/constraints.md (Q): Define as a memory constraint.
28234 * config/v850/predicates.md (label_ref_operand): New predicate.
28235 (e3v5_shift_operand): New predicate.
28236 (ior_operator): New predicate.
28237 * config/v850/t-v850: Add e3v5 multilib.
28238 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
28239 (v850_gen_movdi): Prototype.
28240 * config/v850/v850.c: Add support for e3v5 architecture.
28241 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
28243 (construct_save_jarl): Add e3v5 long JARL support.
28244 (v850_adjust_insn_length): New function. Adjust length of call
28245 insns when using e3v5 instructions.
28246 (v850_gen_movdi): New function: Generate instructions to move a
28248 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
28249 (CPP_SPEC): Define __v850e3v5__ as appropriate.
28250 (TARGET_USE_FPU): Enable for e3v5.
28251 (CONST_OK_FOR_W): New macro.
28252 (ADJUST_INSN_LENGTH): Define.
28253 * config/v850/v850.md (UNSPEC_LOOP): Define.
28254 (attr cpu): Add v850e3v5.
28255 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
28256 (movdi): New pattern.
28257 (movdi_internal): New pattern.
28258 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
28259 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
28260 (cstoresf4): Likewise.
28261 (cstoredf4): Likewise.
28262 (insv): New pattern.
28263 (rotlso3_a): New pattern.
28264 (rotlsi3_b): New pattern
28265 (rotlsi3_v850e3v5): New pattern.
28266 (doloop_begin): New pattern.
28267 (fix_loop_counter): New pattern.
28268 (doloop_end): New pattern.
28269 (branch_normal): Add e3v5 long branch support.
28270 (branch_invert): Likewise.
28271 (branch_z_normal): Likewise.
28272 (branch_z_invert): Likewise.
28273 (branch_nz_normal): Likewise.
28274 (branch_nz_invert): Likewise.
28275 (call_internal_short): Add e3v5 register-indirect JARL support.
28276 (call_internal_long): Likewise.
28277 (call_value_internal_short): Likewise.
28278 (call_value_internal_long): Likewise.
28279 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
28280 (mloop): New option.
28281 * config.gcc: Add support for configuring v840e3v5 target.
28282 * doc/invoke.texi: Document new v850 specific command line options.
28284 2013-01-31 Paul Koning <ni1d@arrl.net>
28288 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
28289 children if parent is a class.
28290 (prune_unused_types_prune): Don't add DW_AT_declaration.
28292 2013-01-31 Richard Biener <rguenther@suse.de>
28294 PR tree-optimization/56157
28295 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
28296 match up operand with SLP child.
28298 2013-01-31 Jason Merrill <jason@redhat.com>
28301 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
28302 parameters the first time.
28303 (gen_scheduled_generic_parms_dies): Check completeness here.
28305 2013-01-31 Richard Biener <rguenther@suse.de>
28307 PR middle-end/53073
28308 * common.opt (faggressive-loop-optimizations): New flag,
28309 enabled by default.
28310 * doc/invoke.texi (faggressive-loop-optimizations): Document.
28311 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
28312 infer_loop_bounds_from_undefined by it.
28314 2013-01-31 Richard Biener <rguenther@suse.de>
28316 PR tree-optimization/56150
28317 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
28318 visit virtual operands.
28319 (find_uses_to_rename_bb): Likewise.
28321 2013-01-31 Richard Biener <rguenther@suse.de>
28323 PR tree-optimization/56150
28324 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
28325 mixed store non-store stmts.
28327 2013-01-30 Jakub Jelinek <jakub@redhat.com>
28330 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
28331 LIBASAN_EARLY_SPEC is defined.
28332 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
28333 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
28335 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
28338 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
28339 invalid args instead of ICEing on it.
28340 (ix86_valid_target_attribute_tree): Return error_mark_node if
28341 ix86_valid_target_attribute_inner_p failed.
28342 (ix86_valid_target_attribute_p): Return false only if
28343 ix86_valid_target_attribute_tree returned error_mark_node. Allow
28344 target("default") attribute.
28345 (sorted_attr_string): Change argument from const char * to tree,
28346 merge in all target attribute arguments rather than just one.
28347 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
28348 instead of free. Avoid using strcat.
28349 (ix86_mangle_function_version_assembler_name): Mangle
28350 target("default") as if no target attribute is present. Adjust
28351 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
28352 instead of xmalloc and XDELETEVEC instead of free.
28353 (ix86_function_versions): Don't return true if one of the decls
28354 doesn't have target attribute. If they don't and one of the decls
28355 is DECL_FUNCTION_VERSIONED, report an error. Adjust
28356 sorted_attr_string caller. Use XDELETEVEC instead of free.
28357 (ix86_supports_function_versions): Remove.
28358 (make_name): Fix up formatting.
28359 (make_dispatcher_decl): Remove resolver_name and its initialization.
28360 Avoid leaking memory.
28361 (is_function_default_version): Return true if there is
28362 target("default") attribute rather than no target attribute at all.
28363 (make_resolver_func): Avoid leaking memory.
28364 (ix86_generate_version_dispatcher_body): Likewise.
28365 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
28366 * target.def (supports_function_versions): Remove.
28367 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
28368 * doc/tm.texi: Regenerated.
28370 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
28372 PR rtl-optimization/56144
28373 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
28374 for values with side effects.
28376 2013-01-30 Richard Biener <rguenther@suse.de>
28378 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
28379 (sparseset_pop): Likewise.
28380 * cfganal.c (compute_idf): Likewise. Increase work-stack size
28381 to be able to use quick_push in the worker loop.
28383 2013-01-30 Marek Polacek <polacek@redhat.com>
28385 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
28387 2013-01-30 Richard Biener <rguenther@suse.de>
28390 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
28392 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
28394 PR tree-optimization/56064
28395 * fixed-value.c (fixed_from_double_int): New function.
28396 * fixed-value.h (fixed_from_double_int): New prototype.
28397 (const_fixed_from_double_int): New static inline function.
28398 * fold-const.c (native_interpret_fixed): New static function.
28399 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
28400 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
28401 (native_encode_fixed): New static function.
28402 (native_encode_expr) <FIXED_CST>: Use it.
28403 (native_interpret_int): Move double_int worker code to...
28404 * double-int.c (double_int::from_buffer): ...this new static method.
28405 * double-int.h (double_int::from_buffer): Prototype it.
28407 2013-01-30 Richard Biener <rguenther@suse.de>
28409 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
28410 New pointer-map and obstack.
28411 (init_alias_vars): Allocate pointer-map and obstack.
28412 (delete_points_to_sets): Free them.
28413 (find_what_var_points_to): Cache result.
28414 (find_what_p_points_to): Adjust for changed interface of
28415 find_what_var_points_to.
28416 (compute_points_to_sets): Likewise.
28417 (ipa_pta_execute): Likewise.
28419 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
28421 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
28422 * configure: Regenerate.
28423 * config.in: Regenerate.
28424 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
28425 #nobits/#progbits if supported.
28427 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
28430 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
28431 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
28432 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
28434 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28436 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
28437 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
28439 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28441 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
28442 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
28444 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28446 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
28448 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
28449 * config/arm/cortex-a7.md: New bypasses using
28450 arm_mac_accumulator_is_result.
28452 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28454 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
28455 (cortex_a7_neon_mla): Likewise.
28456 (cortex_a7_fpfmad): New reservation.
28457 (cortex_a7_fpmacs): Use ffmas and update required units.
28458 (cortex_a7_fpmuld): Update required units and latency.
28459 (cortex_a7_fpmacd): Likewise.
28460 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
28461 (cortex_a7_neon). Likewise.
28462 (bypass) Update participating units.
28464 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28466 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
28467 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
28469 * config/arm/vfp11.md (vfp_farith): Use ffmas.
28470 (vfp_fmul): Use ffmad.
28471 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
28472 (cortex_r4_fmacd): Use ffmad.
28473 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
28474 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
28475 (cortex_a9_fmacd): Use ffmad.
28476 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
28477 (cortex_a8_vfp_macd): Use ffmad.
28478 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
28479 (cortex_a5_fpmacd): Use ffmad.
28480 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
28481 (cortex_a15_vfp_macd): Use ffmad.
28482 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
28484 2013-01-29 Jason Merrill <jason@redhat.com>
28487 * varasm.c (default_assemble_visibility): Don't warn about
28488 visibility on artificial decls.
28490 2013-01-29 Richard Biener <rguenther@suse.de>
28492 PR tree-optimization/56113
28493 * tree-ssa-structalias.c (equiv_class_lookup): Also return
28495 (label_visit): Free duplicate bitmaps and record the leader instead.
28496 (perform_var_substitution): Adjust.
28498 2013-01-29 Richard Biener <rguenther@suse.de>
28500 PR tree-optimization/55270
28501 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
28502 the CFG, schedule loops for fixup.
28504 2013-01-29 Nick Clifton <nickc@redhat.com>
28506 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
28509 2013-01-28 Leif Ekblad <leif@rdos.net>
28511 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
28512 * config/i386/i386.h (TARGET_RDOS): New macro.
28513 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
28514 * config/i386/i386.c (ix86_option_override_internal): For 64bit
28515 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
28516 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
28517 DEFAULT_LARGE_SECTION_THRESHOLD.
28518 * config/i386/i386.md (R14_REG, R15_REG): New constants.
28519 * config/i386/rdos.h: New file.
28520 * config/i386/rdos64.h: New file.
28522 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
28525 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
28528 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28530 PR rtl-optimization/56117
28531 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
28532 call cselib_lookup_from_insn on the MEM before calling
28533 add_insn_mem_dependence.
28535 2013-01-28 Richard Biener <rguenther@suse.de>
28537 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
28538 to a stmt that didn't have one.
28539 (copy_phis_for_bb): Likewise for PHI arguments.
28540 (copy_debug_stmt): Likewise for debug stmts.
28542 2013-01-28 Richard Biener <rguenther@suse.de>
28544 PR tree-optimization/56034
28545 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
28546 (partition_builtin_p): Adjust.
28547 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
28548 it is the last partition.
28549 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
28550 up the vertex for the definition.
28551 (classify_partition): Classify whether a partition is a
28552 PKIND_REDUCTION, thus has uses outside of the loop.
28553 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
28554 Merge all PKIND_REDUCTION partitions into the last partition.
28555 (tree_loop_distribution): Seed partitions from reductions as well.
28557 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28559 PR tree-optimization/56125
28560 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
28561 pow(x,c) into sqrt(x) * powi(x, n/2) or
28562 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
28563 optimizing for size.
28564 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
28565 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
28568 PR tree-optimization/56094
28569 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
28570 to UNKNOWN_LOCATION while gimplifying expr.
28572 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
28575 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
28576 operand 0 in movabs insn template for -masm=intel asm alternative.
28577 (*movabs<mode>_2): Ditto for operand 1.
28579 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
28582 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
28585 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28587 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
28590 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
28592 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
28593 Undef to avoid warning.
28595 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28597 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
28598 * configure: Regenerate.
28600 2013-01-25 Jakub Jelinek <jakub@redhat.com>
28602 PR tree-optimization/56098
28603 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
28604 for stmts with volatile ops.
28605 (cond_store_replacement): Don't optimize if assign has volatile ops.
28606 (cond_if_else_store_replacement_1): Don't optimize if either
28607 then_assign or else_assign have volatile ops.
28608 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
28611 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28613 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
28615 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28617 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
28618 missing ':' in asm example.
28620 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
28622 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
28623 entries into lane and laneq entries.
28624 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
28625 Remove AdvSIMD scalar modes.
28626 (aarch64_sq<r>dmulh_laneq<mode>): New.
28627 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
28629 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
28630 builtin implementations to relfect changes in RTL in aarch64-simd.md.
28631 * config/aarch64/iterators.md (VCOND): New.
28634 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28637 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
28638 Add NULL LIBNAME argument to existing definitions.
28639 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
28640 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
28641 * config/avr/avr.c (DEF_BUILTIN): Same.
28642 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
28643 (avr_expand_builtin): Expand to a vanilla call if a libgcc
28644 implementation is available (DECL_ASSEMBLER_NAME is set).
28645 (avr_fold_absfx): New static function.
28646 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
28647 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
28648 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
28649 AVR_BUILTIN_ABSLLK.
28650 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
28651 (abshk, absk, abslk, absllk): Provide as static inline functions.
28653 2013-01-25 Marek Polacek <polacek@redhat.com>
28655 PR tree-optimization/56035
28656 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
28658 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
28660 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
28661 (*movtf_internal_rex64): Add (!o,C) alternative
28662 (*movxf_internal_rex64): Ditto.
28663 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
28665 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
28667 * doc/invoke.texi: fix typo.
28668 * doc/objc.texi: fix typo.
28670 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
28672 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
28673 for the first two alternatives.
28675 2013-01-24 Diego Novillo <dnovillo@google.com>
28677 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
28678 (ggc-zone.o): Remove.
28679 * configure.ac: Remove option --with-gc.
28680 * configure: Re-generate.
28681 * doc/install.texi: Remove documentation for --with-gc.
28682 * gengtype.c (write_enum_defn): Remove. Update all users.
28683 (write_Types_process_field): Remove generation of gt_e_* argument.
28684 (output_type_enum): Remove. Update all users.
28685 (write_enum_defn): Remove. Update all users.
28686 (enum alloc_zone): Remove. Update all users.
28687 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
28688 * ggc-common.c (ggc_splay_alloc): Remove first argument.
28689 Update all callers.
28690 (struct ptr_data): Remove field TYPE. Update all users.
28691 (gt_pch_note_object): Remove argument TYPE. Update all users.
28692 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
28694 * ggc-none.c (ggc_alloc_typed_stat): Remove.
28695 (struct alloc_zone): Remove.
28696 (ggc_internal_alloc_zone_stat): Remove.
28697 (ggc_internal_cleared_alloc_zone_stat): Remove.
28698 * ggc-page.c (ggc_alloc_typed_stat): Remove.
28699 (ggc_pch_count_object): Remove last argument. Update all users.
28700 (ggc_pch_alloc_object): Remove last argument. Update all users.
28701 (struct alloc_zone): Remove.
28702 * ggc-zone.c: Remove.
28703 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
28704 (struct alloc_zone): Remove.
28705 (ggc_alloc_typed_stat): Remove.
28706 (ggc_alloc_typed): Remove.
28707 (ggc_splay_alloc): Remove first argument.
28708 (rtl_zone): Remove. Update all users.
28709 (tree_zone): Remove. Update all users.
28710 (tree_id_zone): Remove. Update all users.
28711 (ggc_internal_zone_alloc_stat): Remove. Update all users.
28712 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
28713 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
28714 * tree-ssanames.c: Remove references to zone allocator in comments.
28716 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
28718 * config/avr/avr.c (avr_out_fract): Make register numbers that
28719 might be outside of source operand signed.
28721 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
28723 * config/i386/constraints.md (Yf): New constraint.
28724 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
28725 of f constraint to conditionaly disable x87 register preferences.
28726 (*movdf_internal): Ditto.
28727 (*movsf_internal): Ditto.
28729 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
28731 PR inline-asm/55934
28732 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
28733 that have operands with impossible constraints.
28734 Add a FIXME for a speed-up opportunity.
28735 * lra-constraints.c (process_alt_operands): Verify that a class
28736 selected from constraints on asms is valid for the operand mode.
28737 (curr_insn_transform): Remove incorrect comment.
28739 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
28741 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
28742 TOC operand is a valid symbol ref in the constant pool.
28744 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28746 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
28748 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
28751 * config/avr/stdfix.h: New file.
28752 * t-avr (stdfix-gcc.h): New rule to build it.
28753 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
28755 2013-01-23 Kostya Serebryany <kcc@google.com>
28757 * config/darwin.h: remove dependency on
28758 CoreFoundation (asan on Mac OS).
28760 2013-01-23 Jakub Jelinek <jakub@redhat.com>
28763 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
28764 instead of cmpdi_operand for first comparison operand.
28765 Don't assert that comparison operands aren't both constants.
28767 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28769 * doc/install.texi (Downloading the Source): Update references to
28770 downloading separate components.
28772 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28774 * doc/extend.texi (__int128): Improve grammar.
28776 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
28779 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
28780 alternative to (o,r).
28781 (*movdi_internal_rex64): Remove (!o,n) alternative.
28782 (DImode immediate->memory splitter): Remove.
28783 (DImode immediate->memory peephole2): Remove.
28784 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
28785 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
28786 alternative to (!o,*r).
28787 (*movtf_internal_sse): New pattern.
28788 (*movxf_internal_rex64): New pattern.
28789 (*movxf_internal): Disable for TARGET_64BIT.
28790 (*movdf_internal_rex64): Remove (!o,F) alternative.
28792 2013-01-22 Jakub Jelinek <jakub@redhat.com>
28794 PR middle-end/56074
28795 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
28796 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
28797 * tree-vect-loop-manip.c (find_loop_location): Also ignore
28798 stmt locations where LOCATION_LOCUS of the stmt location is
28799 UNKNOWN_LOCATION or BUILTINS_LOCATION.
28802 * config/i386/i386.md (UNSPEC_STOS): New.
28803 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
28804 *strsetqi_1): Add UNSPEC_STOS.
28806 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
28809 * doc/invoke.texi: Remove left over -Wsynth example.
28811 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28813 PR tree-optimization/56051
28814 * fold-const.c (fold_binary_loc): Don't fold
28815 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
28816 a narrowing conversion, or widening conversion from signed
28819 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28821 PR rtl-optimization/56023
28822 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
28823 dependent on debug instruction.
28825 2013-01-21 Martin Jambor <mjambor@suse.cz>
28827 PR middle-end/56022
28828 * function.c (allocate_struct_function): Call
28829 invoke_set_current_function_hook earlier.
28831 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28833 * reload1.c (init_reload): Only initialize reload_obstack
28834 during the first call.
28836 2013-01-21 Marek Polacek <polacek@redhat.com>
28838 * cfgloop.c (verify_loop_structure): Fix up grammar.
28840 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
28842 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
28843 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
28845 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28848 * config/arm/marvell-pj4.md: Update copyright year.
28849 Fix up use of alu to alu_reg and simple_alu_imm.
28851 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28853 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
28855 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
28858 * lra-constraints.c (curr_insn_transform): Don't reuse original
28859 insn for secondary memory move when memory mode should be different.
28861 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28863 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
28864 atomic_storedi_1): New patterns.
28866 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
28868 btver2 pipeline descriptions.
28869 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
28871 * config/i386/i386.md (btver2_decode): New type attributes.
28872 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
28874 * config/i386/btver2.md: New file describing btver2 pipelines.
28876 2013-01-19 Andrew Pinski <apinski@cavium.com>
28878 PR tree-optimization/52631
28879 * tree-ssa-sccvn (visit_use): Before looking up the original
28880 statement, try looking up the simplified expression.
28882 2013-01-19 Anthony Green <green@moxielogic.com>
28884 * config/moxie/moxie.c (moxie_expand_prologue): Set
28885 current_function_static_stack_size.
28887 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28889 PR tree-optimization/56029
28890 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
28891 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
28893 2013-01-18 Sharad Singhai <singhai@google.com>
28895 PR tree-optimization/55995
28896 * dumpfile.c (dump_loc): Print location only if available.
28897 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
28899 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
28902 * lra-constraints.c (curr_insn_transform): Reuse original insn for
28903 secondary memory move.
28904 (inherit_reload_reg): Use rclass instead of cl for
28905 check_secondary_memory_needed_p.
28907 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28909 PR middle-end/56015
28910 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
28911 the case where writing real complex part of target modifies op1.
28913 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
28915 * config/aarch64/aarch64-simd.md
28916 (aarch64_vcond_internal<mode>): Handle unordered cases.
28917 * config/aarch64/iterators.md (v_cmp_result): New.
28919 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
28920 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28922 * config/arm/marvell-pj4.md: New file.
28923 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
28924 * config/arm/arm.md (generic_sched): Add marvell_pj4.
28925 (generic_vfp): Likewise.
28926 * config/arm/arm-cores.def: Add marvell-pj4.
28927 * config/arm/arm-tune.md: Regenerate.
28928 * config/arm/arm-tables.opt: Regenerate.
28929 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
28930 * doc/invoke.texi: Document marvell-pj4.
28932 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
28934 * config/aarch64/arm_neon.h: Map scalar types to standard types.
28936 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28941 * var-tracking.c (negative_power_of_two_p): New.
28942 (global_get_addr_cache, local_get_addr_cache): New.
28943 (get_addr_from_global_cache, get_addr_from_local_cache): New.
28944 (vt_canonicalize_addr): Rewrite using the above. Adjust the
28946 (vt_stack_offset_p): Remove.
28947 (vt_canon_true_dep): Always canonicalize loc's address.
28948 (clobber_overlapping_mems): Make sure we have a MEM.
28949 (local_get_addr_clear_given_value): New.
28950 (val_reset): Clear local cached entries.
28951 (compute_bb_dataflow): Create and release the local cache.
28952 Disable duplicate MEMs clobbering.
28953 (emit_notes_in_bb): Clobber MEMs likewise.
28954 (vt_emit_notes): Create and release the local cache.
28955 (vt_initialize, vt_finalize): Create and release the global
28956 cache, respectively.
28957 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
28959 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28961 PR libmudflap/53359
28962 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
28963 not found in the symtab.
28965 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28968 PR rtl-optimization/55547
28969 PR rtl-optimization/53827
28972 * alias.c (offset_overlap_p): New, factored out of...
28973 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
28974 the conservative special case for symbolic constants. Don't
28975 adjust zero sizes on alignment.
28977 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
28979 PR rtl-optimization/52573
28980 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
28981 REG_UNUSED for the same register.
28983 2013-01-17 Richard Biener <rguenther@suse.de>
28984 Marek Polacek <polacek@redhat.com>
28986 PR rtl-optimization/55833
28987 * loop-unswitch.c (unswitch_loops): Move loop verification...
28988 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
28989 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
28990 Set it to true when we're removing a loop from hierarchy tree in
28991 an irreducible region.
28992 (fix_bb_placements): Adjust caller.
28993 (fix_loop_placements): Likewise.
28995 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
28997 * config/avr/builtins.def (DEF_BUILTIN): Factor out
28998 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
28999 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
29000 Remove ID. Adjust comments.
29001 * config/avr/avr-c.c (avr_builtin_name): Remove.
29002 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
29003 * config/avr/avr.c (avr_tolower): New static function.
29004 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
29005 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
29006 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
29009 2013-01-17 Jan Hubicka <jh@suse.cz>
29011 PR tree-optimization/55273
29012 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
29014 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
29017 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
29018 store through atomic_store<mode>_1.
29019 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
29021 2013-01-17 Martin Jambor <mjambor@suse.cz>
29023 PR tree-optimizations/55264
29024 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
29025 for virtual methods.
29026 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
29027 virtual methods before inlining is over.
29028 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
29030 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
29033 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
29035 PR rtl-optimization/56005
29036 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
29037 pending reads for prefetch.
29039 2013-01-16 Ian Bolton <ian.bolton@arm.com>
29041 * config/aarch64/aarch64.md
29042 (*cstoresi_neg_uxtw): New pattern.
29043 (*cmovsi_insn_uxtw): New pattern.
29044 (*<optab>si3_uxtw): New pattern.
29045 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
29046 (*<optab>si3_insn_uxtw): New pattern.
29047 (*bswapsi2_uxtw): New pattern.
29049 2013-01-16 Richard Biener <rguenther@suse.de>
29051 * tree-inline.c (tree_function_versioning): Remove set but
29052 never used variable.
29054 2013-01-16 Richard Biener <rguenther@suse.de>
29056 PR tree-optimization/55964
29057 * tree-flow.h (rename_variables_in_loop): Remove.
29058 (rename_variables_in_bb): Likewise.
29059 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
29060 (copy_loop_before): Adjust and delete update-ssa status.
29061 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
29062 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
29063 (rename_variables_in_loop): Remove.
29064 (slpeel_update_phis_for_duplicate_loop): Likewise.
29065 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
29066 use available cfg machinery instead of duplicating it.
29067 Update PHI nodes and perform poor-mans SSA update here.
29068 (slpeel_tree_peel_loop_to_edge): Adjust.
29070 2013-01-16 Richard Biener <rguenther@suse.de>
29072 PR tree-optimization/54767
29073 PR tree-optimization/53465
29074 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
29075 (vrp_visit_phi_node): For PHI arguments coming via backedges
29076 drop all symbolical range information.
29077 (execute_vrp): Compute backedges.
29079 2013-01-16 Richard Biener <rguenther@suse.de>
29081 * doc/install.texi: Update CLooG and ISL requirements to
29084 2013-01-16 Christian Bruel <christian.bruel@st.com>
29087 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
29088 (broken_move): Handle UNSPECV_SP_SWITCH_B.
29089 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
29091 2013-01-16 DJ Delorie <dj@redhat.com>
29093 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
29094 (UNSPECV_SP_SWITCH_E): New.
29095 (sp_switch_1): Change to an unspec.
29096 (sp_switch_2): Change to an unspec. Don't use post-inc when we
29099 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
29101 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
29102 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
29103 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
29104 (expand_mem_thread_fence): Ditto.
29105 (expand_mem_signal_fence): Ditto.
29106 (expand_atomic_load): Ditto.
29107 (expand_atomic_store): Ditto.
29109 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
29111 PR rtl-optimization/55547
29112 PR rtl-optimization/53827
29115 * alias.c (memrefs_conflict_p): Set sizes to negative after
29118 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29121 * function.c (thread_prologue_and_epilogue_insns): Always
29122 add crtl->drap_reg to set_up_by_prologue.set, even if
29123 stack_realign_drap is false.
29125 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
29127 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
29128 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
29129 *call): Fix indention.
29131 2013-01-15 Tom de Vries <tom@codesourcery.com>
29134 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
29137 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
29139 PR rtl-optimization/55153
29140 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
29142 2013-01-15 Martin Jambor <mjambor@suse.cz>
29144 PR tree-optimization/55920
29145 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
29146 accesses as grp_to_be_debug_replaced.
29148 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29150 PR tree-optimization/55920
29151 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
29152 there is non-useless type conversion needed from debug rhs to lhs,
29153 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
29155 2013-01-15 Joseph Myers <joseph@codesourcery.com>
29156 Mikael Pettersson <mikpe@it.uu.se>
29159 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
29161 (ASM_OUTPUT_CASE_LABEL): Remove.
29162 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
29163 * final.c (shorten_branches): Update alignment of labels before
29164 jump tables if CASE_VECTOR_SHORTEN_MODE.
29166 2013-01-15 Richard Biener <rguenther@suse.de>
29169 * system.h: Do not include gmp.h for building host tools.
29171 2013-01-15 Richard Biener <rguenther@suse.de>
29173 PR middle-end/55882
29174 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
29175 account for bitpos when computing alignment.
29177 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
29179 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
29180 (ix86_target_macros_internal): Likewise.
29182 * config/i386/i386.c (m_CORE2I7): Removed.
29183 (m_CORE_HASWELL): New macro.
29184 (m_CORE_ALL): Likewise.
29185 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
29186 (initial_ix86_arch_features): Likewise.
29187 (processor_target_table): Initializations for Core avx2.
29188 (cpu_names): New names "core-avx2".
29189 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
29190 PROCESSOR_CORE_HASWELL.
29191 (ix86_issue_rate): New case.
29192 (ia32_multipass_dfa_lookahead): Likewise.
29193 (ix86_sched_init_global): Likewise.
29195 * config/i386/i386.h (TARGET_HASWELL): New macro.
29196 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
29197 (processor_type): New PROCESSOR_HASWELL.
29199 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29201 PR tree-optimization/55955
29202 * tree-vect-loop.c (vectorizable_reduction): Give up early on
29203 *SHIFT_EXPR and *ROTATE_EXPR codes.
29205 PR tree-optimization/48766
29206 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
29207 -ftrapv disable -fwrapv.
29209 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29212 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
29213 etc. to 1 and not to __flash.
29214 Use LL suffix for __INT24_MAX__ with -mint8.
29215 Use ULL suffix for __UINT24_MAX__ with -mint8.
29217 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29219 * config/avr/avr-arch.h
29220 (struct base_arch_s): Use typedef avr_arch_t instead.
29221 (struct arch_info_s): Use typedef avr_arch_info_t instead.
29222 (struct mcu_type_s): Use typedef avr_mcu_t instead.
29223 * config/avr/avr.c: Same.
29224 * config/avr/avr-devices.c: Same.
29225 * config/avr/driver-avr.c: Same.
29226 * config/avr/gen-avr-mmcu-texi.c: Same.
29227 * config/avr/avr-mcus.def: Adjust comment.
29229 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
29231 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
29232 * config/aarch64/iterators.md (VALLDI): New.
29234 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
29235 Andi Kleen <ak@linux.intel.com>
29238 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
29239 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
29242 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29244 * config/avr/avr-stdint.h: Remove trailing blanks.
29245 * config/avr/avr-log.h: Same.
29246 * config/avr/avr-arch.h: Same.
29247 * config/avr/avr-devices.c: Same.
29248 * config/avr/avr-dimode.md: Same.
29249 * config/avr/predicates.md: Same.
29250 * config/avr/avr-c.c: Same. And fix typo.
29252 * config/avr/avr-protos.h: Same. And:
29253 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
29254 (init_cumulative_args): Rename to avr_init_cumulative_args.
29255 (expand_prologue): Rename to avr_expand_prologue.
29256 (expand_epilogue): Rename to avr_expand_epilogue.
29257 (adjust_insn_length): Rename to avr_adjust_insn_length.
29258 (notice_update_cc): Rename to avr_notice_update_cc.
29259 (final_prescan_insn): Rename to avr_final_prescan_insn.
29260 * config/avr/avr.c: Same.
29261 * config/avr/avr.h: Same.
29262 * config/avr/avr.md: Remove trailing blanks.
29263 (prologue): Use avr_expand_prologue.
29264 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
29266 2013-01-14 Richard Biener <rguenther@suse.de>
29268 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
29269 verify_location, collect_subblocks): New functions.
29270 (verify_gimple_in_cfg): Verify that locations only reference
29271 BLOCKs in the functions BLOCK tree.
29273 2013-01-14 Richard Biener <rguenther@suse.de>
29275 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
29277 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
29279 (insert_out_of_ssa_copy_on_edge): Likewise.
29280 (rewrite_close_phi_out_of_ssa): Likewise.
29281 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
29283 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
29284 propagated constants.
29285 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
29288 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29290 * config/avr/avr-modes.def: Add GPL copyright notice.
29292 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
29294 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
29295 MEMMODEL_MASK to determine memory model.
29296 (atomic_store<mode>): Ditto from operands[2].
29297 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
29299 2013-01-13 Jakub Jelinek <jakub@redhat.com>
29302 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
29303 (fold_gimple_assign): Don't call unshare_expr here.
29304 (fold_ctor_reference): Call unshare_expr.
29306 2013-01-13 Terry Guo <terry.guo@arm.com>
29308 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
29309 * doc/fragments.texi: Document MULTILIB_REUSE.
29310 * gcc.c (multilib_reuse): New internal spec.
29311 (set_multilib_dir): Also search multilib from multilib_reuse.
29312 * genmultilib (tmpmultilib3): Refactor code.
29313 (tmpmultilib4): Ditto.
29314 (multilib_reuse): New multilib argument.
29316 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
29318 * Makefile.in: Update copyright.
29320 2013-01-12 Tom de Vries <tom@codesourcery.com>
29322 PR middle-end/55890
29323 * calls.c (expand_call): Check if arg_nr is valid.
29325 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
29327 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
29328 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
29329 documentation. Add missing '__' in front of
29330 __builtin_ia32_packssdw256.
29332 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29335 * config/s390/s390.c (s390_preferred_reload_class): Do not return
29336 NO_REGS for larl operands.
29337 (s390_reload_larl_operand): Use s390_load_address instead of
29340 2013-01-11 Richard Biener <rguenther@suse.de>
29342 * tree-cfg.c (verify_node_sharing_1): Split out from ...
29343 (verify_node_sharing): ... here.
29344 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
29346 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
29348 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
29349 Substitute TREECHECKING.
29350 * configure: Regenerate.
29351 * Makefile.in (TREECHECKING): New.
29353 2013-01-11 Richard Guenther <rguenther@suse.de>
29355 PR tree-optimization/44061
29356 * tree-vrp.c (extract_range_basic): Compute zero as
29357 value-range for __builtin_constant_p of function parameters.
29359 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
29361 Update copyright years.
29363 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
29365 PR rtl-optimization/55672
29366 * lra-eliminations.c (mark_not_eliminable): Permit addition with
29367 const to be eliminable.
29369 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
29371 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
29372 * configure: Regenerate.
29374 2013-01-10 Richard Biener <rguenther@suse.de>
29376 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
29378 2013-01-10 Richard Biener <rguenther@suse.de>
29381 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
29382 locations for virtual PHI arguments.
29383 (rewrite_update_phi_arguments): Likewise.
29385 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
29387 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
29390 2013-01-10 Jakub Jelinek <jakub@redhat.com>
29392 PR tree-optimization/55921
29393 * tree-complex.c (expand_complex_asm): New function.
29394 (expand_complex_operations_1): Call it for GIMPLE_ASM.
29396 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29399 * config/s390/s390.c (s390_symref_operand_p)
29400 (s390_loadrelative_operand_p): Merge the two functions.
29401 (s390_check_qrst_address, print_operand_address): Add parameters
29402 to s390_loadrelative_operand_p invokation.
29403 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
29404 (s390_reload_larl_operand, s390_secondary_reload): Use
29405 s390_loadrelative_operand_p instead of s390_symref_operand_p.
29406 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
29408 2013-01-09 Mike Stump <mikestump@comcast.net>
29410 * dse.c (record_store): Remove unnecessary assert.
29412 2013-01-09 Jan Hubicka <jh@suse.cz>
29414 PR tree-optimization/55569
29415 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
29416 * cfgloop.h (scale_loop_profile): Likewise.
29418 2013-01-09 Jan Hubicka <jh@suse.cz>
29421 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
29423 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
29425 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
29427 PR middle-end/55114
29428 * expr.h (maybe_emit_group_store): Declare.
29429 * expr.c (maybe_emit_group_store): New function.
29430 * builtins.c (expand_builtin_int_roundingfn): Call it.
29431 (expand_builtin_int_roundingfn_2): Likewise.
29433 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
29435 PR rtl-optimization/55829
29436 * lra-constraints.c (match_reload): Add code for absent output.
29437 (curr_insn_transform): Add code for reloads of matched inputs
29440 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
29442 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
29443 attribute of movddup insn to DF.
29444 (*vec_interleave_lowv2df): Ditto.
29445 (vec_dupv2df): Ditto.
29447 2013-01-09 Jan Hubicka <jh@suse.cz>
29449 PR tree-optimiation/55875
29450 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
29451 EVERY_ITERATION parameter.
29452 (number_of_iterations_exit): Check if exit is executed every iteration.
29453 (idx_infer_loop_bounds): Similarly here.
29454 (n_of_executions_at_most): Simplify
29455 to only test for cases where statement is dominated by the
29456 particular bound; handle correctly the "postdominance" test.
29457 (scev_probably_wraps_p): Use max loop iterations info
29458 as a global bound first.
29460 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
29461 Nick Clifton <nickc@redhat.com>
29463 * config/v850/v850.md (cbranchsf4): New pattern.
29464 (cstoresf4): New pattern.
29465 (cbranchdf4): New pattern.
29466 (cstoredf4): New pattern.
29467 (movsicc): Disallow floating point comparisons.
29468 (cmpsf_le_insn): Fix order of operators.
29469 (cmpsf_lt_insn): Likewise.
29470 (cmpsf_eq_insn): Likewise.
29471 (cmpdf_le_insn): Likewise.
29472 (cmpdf_lt_insn): Likewise.
29473 (cmpdf_eq_insn): Likewise.
29474 (cmpsf_ge_insn): Use LE comparison.
29475 (cmpdf_ge_insn): Likewise.
29476 (cmpsf_gt_insn): Use LT comparison.
29477 (cmpdf_gt_insn): Likewise.
29478 (cmpsf_ne_insn): Delete pattern.
29479 (cmpdf_ne_insn): Delete pattern.
29480 * config/v850/v850.c (v850_gen_float_compare): Use
29481 gen_cmpdf_eq_insn for NE comparison.
29482 (v850_float_z_comparison_operator)
29483 (v850_float_nz_comparison_operator): Move from here ...
29484 * config/v850/predicates.md: ... to here. Move GT and GE
29485 comparisons into v850_float_z_comparison_operator.
29486 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
29488 (v850_float_nz_comparison_operator): Likewise.
29490 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29492 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
29493 with calls to gen_insvsi/gen_insvdi.
29495 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29497 * config/i386/i386.c (initial_ix86_tune_features): Set up
29498 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
29500 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
29501 Jakub Jelinek <jakub@redhat.com>
29503 PR tree-optimization/48189
29504 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
29505 If nitercst is 0, don't predict the exit edge.
29507 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29509 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
29510 in asm_fprintf with reg_names.
29511 (aarch64_print_operand_address): Likewise.
29512 (aarch64_return_addr): Likewise.
29513 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
29515 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29517 * config/pa/pa.h (VAL_U6_BITS_P): Define.
29518 (INT_U6_BITS): Likewise.
29519 * config/pa/predicates.md (uint6_operand): New predicate.
29520 (shift5_operand, shift6_operand): Likewise.
29521 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
29523 (lshrdi3): Use shift6_operand.
29524 (shrpsi4, shrpdi4): New insn patterns.
29525 (extzv): Delete expander.
29526 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
29527 predicates in unamed zero extract patterns. Tighten common constraint.
29528 (extv): Delete expander.
29529 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
29530 predicates in unamed sign extract patterns. Tighten common constraint.
29531 (insv): Delete expander.
29532 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
29533 predicates in unamed insert patterns. Tighten common constraint.
29534 Change uint32_operand predicate to uint6_operand predicate in unamed
29535 DImode pattern to insert constant values of type 1...1xxxx.
29537 2013-01-04 Jan Hubicka <jh@suse.cz>
29539 PR tree-optimization/55823
29540 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
29543 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29544 Uros Bizjak <ubizjak@gmail.com>
29546 PR rtl-optimization/55845
29547 * df-problems.c (can_move_insns_across): Stop scanning at
29548 volatile_insn_p source instruction or give up if
29549 across_from .. across_to range contains any volatile_insn_p
29552 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29554 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
29555 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
29557 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
29558 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
29560 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29563 * asan.c (asan_clear_shadow): New function.
29564 (asan_emit_stack_protection): Use it.
29566 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29568 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
29569 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
29570 with tab instead of space.
29572 2013-01-08 Nick Clifton <nickc@redhat.com>
29574 * config/rl78/rl78.c (rl78_expand_prologue): Always select
29575 register bank 0 at the start of an interrupt handler.
29576 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
29579 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29581 * config/aarch64/aarch64-simd.md
29582 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
29583 (aarch64_simd_bsl): Likewise.
29584 (aarch64_vcond_internal<mode>): Likewise.
29585 (vcond<mode><mode>): Likewise.
29586 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
29587 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
29589 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29591 * config/aarch64/aarch64-builtins.c
29592 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
29594 2013-01-08 Martin Jambor <mjambor@suse.cz>
29597 * tree-sra.c (analyze_access_subtree): Return true also after
29598 potentially creating a debug-only replacement.
29600 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29602 PR middle-end/55890
29603 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
29605 PR tree-optimization/54120
29606 * tree-vrp.c (range_fits_type_p): Don't allow
29607 src_precision < precision from signed vr to unsigned_p
29608 if vr->min or vr->max is negative.
29609 (simplify_float_conversion_using_ranges): Test can_float_p
29610 against CODE_FOR_nothing.
29612 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29613 Richard Biener <rguenther@suse.de>
29615 PR middle-end/55851
29616 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
29617 types instead of just INTEGER_TYPE types.
29619 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
29621 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
29624 2013-01-07 Steve Ellcey <sellcey@mips.com>
29627 * config/mips/mips.opt: Change mad to mmad to match documentation.
29629 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29632 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
29633 .progmemx.data now.
29635 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29638 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
29639 (avr_addrspace_t): Add .section_name field.
29640 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
29642 (avr_addrspace): Same. Initialize .section_name. Remove last
29643 NULL entry. Put __memx into .progmemx.data.
29644 (progmem_section_prefix): Remove.
29645 (avr_asm_init_sections): No need to initialize progmem_section.
29646 (avr_asm_named_section): Use avr_addrspace[].section_name to get
29647 section name prefix.
29648 (avr_asm_select_section): Ditto. And use get_unnamed_section to
29649 retrieve the progmem section.
29650 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
29651 boundary to run over avr_addrspace[].
29652 (avr_register_target_pragmas): Ditto.
29654 2013-01-06 Jakub Jelinek <jakub@redhat.com>
29656 * varasm.c (output_constant_def_contents): For asan_protect_global
29657 protected strings, adjust DECL_ALIGN if needed, before testing for
29659 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
29660 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
29662 (output_object_block): For asan protected decls, emit asan padding
29663 after their contents.
29664 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
29665 (asan_finish_file): Test it here instead.
29667 2013-01-07 Nick Clifton <nickc@redhat.com>
29668 Matthias Klose <doko@debian.org>
29669 Doug Kwan <dougkwan@google.com>
29670 H.J. Lu <hongjiu.lu@intel.com>
29673 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
29675 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
29677 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
29679 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
29681 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
29683 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29686 * doc/install.texi (Cross-Compiler-Specific Options): Document
29689 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
29691 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
29692 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
29693 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
29694 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
29695 vqmovun_high_s64): Fix source operand number and update copyright.
29697 2013-01-07 Richard Biener <rguenther@suse.de>
29699 PR middle-end/55890
29700 * gimple.h (gimple_call_builtin_p): New overload.
29701 * gimple.c (validate_call): New function.
29702 (gimple_call_builtin_p): Likewise.
29703 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
29704 Use gimple_call_builtin_p.
29705 (find_func_clobbers): Likewise.
29706 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
29707 (strlen_optimize_stmt): Likewise.
29709 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29711 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
29712 (vld1q_dup_*): Likewise.
29713 (vld1_*): Likewise.
29714 (vld1q_*): Likewise.
29715 (vld1_lane_*): Likewise.
29716 (vld1q_lane_*): Likewise.
29718 2013-01-07 Richard Biener <rguenther@suse.de>
29720 * lto-streamer.h (LTO_minor_version): Bump to 2.
29722 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29724 * config/aarch64/aarch64-protos.h
29725 (aarch64_const_double_zero_rtx_p): Rename to...
29726 (aarch64_float_const_zero_rtx_p): ...this.
29727 (aarch64_float_const_representable_p): New.
29728 (aarch64_output_simd_mov_immediate): Likewise.
29729 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
29730 move immediate case.
29731 * config/aarch64/aarch64.c
29732 (aarch64_const_double_zero_rtx_p): Rename to...
29733 (aarch64_float_const_zero_rtx_p): ...this.
29734 (aarch64_print_operand): Allow printing of new constants.
29735 (aarch64_valid_floating_const): New.
29736 (aarch64_legitimate_constant_p): Check for valid floating-point
29738 (aarch64_simd_valid_immediate): Likewise.
29739 (aarch64_vect_float_const_representable_p): New.
29740 (aarch64_float_const_representable_p): Likewise.
29741 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
29742 (aarch64_output_simd_mov_immediate): New.
29743 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
29744 (*movdf_aarch64): Likewise.
29745 * config/aarch64/constraints.md (Ufc): New.
29746 (Y): call aarch64_float_const_zero_rtx.
29747 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
29749 2013-01-07 Richard Biener <rguenther@suse.de>
29751 PR tree-optimization/55888
29752 PR tree-optimization/55862
29753 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
29754 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
29755 not if it is contained therein.
29757 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29759 * config/avr/t-avr: Typo.
29761 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29764 * config/avr/t-avr: Don't automatically rebuild
29765 $(srcdir)/config/avr/t-multilib
29766 $(srcdir)/config/avr/avr-tables.opt
29767 $(srcdir)/doc/avr-mmcu.texi
29768 (avr-mcus): New phony target to build them on request.
29769 (s-avr-mlib, s-avr-mmcu-texi): Remove.
29770 * avr/avr-mcus.def: Adjust comments.
29772 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
29774 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
29776 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29778 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
29780 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29782 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
29784 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
29786 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
29787 to generate profiling.
29788 * config/rs6000/aix64.h (LIB_SPEC): Same.
29790 2013-01-04 Andrew Pinski <apinski@cavium.com>
29792 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
29794 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
29796 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
29798 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
29800 (ix86_expand_move): Ditto.
29801 (ix86_zero_extend_to_Pmode): Ditto.
29802 (ix86_expand_call): Ditto.
29803 (ix86_expand_special_args_builtin): Ditto.
29804 (ix86_expand_builtin): Ditto.
29806 2013-01-04 Richard Biener <rguenther@suse.de>
29808 PR tree-optimization/55862
29809 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
29810 translating them through PHI nodes.
29812 2013-01-04 Martin Jambor <mjambor@suse.cz>
29814 PR tree-optimization/55755
29815 * tree-sra.c (sra_modify_assign): Do not check that an access has no
29816 children when trying to avoid producing a VIEW_CONVERT_EXPR.
29818 2013-01-04 Marek Polacek <polacek@redhat.com>
29820 PR middle-end/55859
29821 * opts.c (default_options_optimization): Clarify error message.
29823 2013-01-04 Richard Biener <rguenther@suse.de>
29825 PR middle-end/55863
29826 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
29829 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29832 * config/pa/pa.md (movsi): Revert previous change.
29833 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
29836 2013-01-03 Richard Henderson <rth@redhat.com>
29838 * config/i386/i386.c (ix86_expand_move): Always assign to op1
29839 after eliminating TLS symbols.
29841 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
29844 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
29845 * graphite-poly.c (debug_gmp_value): Likewise.
29847 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
29850 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
29851 selected code model, define __code_mode_small__, __code_model_medium__,
29852 __code_model_large__, __code_model_32__ or __code_model_kernel__.
29853 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
29854 xchg temporary register with %k. Declare temporary register as
29856 [__x86_64__]: For medium and large code models, preserve %rbx register.
29858 2013-01-03 Richard Biener <rguenther@suse.de>
29860 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
29861 (dump_subscript): Adjust.
29862 (finalize_ddr_dependent): Do not dump redundant info.
29863 (analyze_siv_subscript): Adjust.
29864 (subscript_dependence_tester): Likewise.
29865 (compute_affine_dependence): Likewise.
29867 2013-01-03 Richard Biener <rguenther@suse.de>
29870 2013-01-03 Richard Biener <rguenther@suse.de>
29872 PR tree-optimization/55857
29873 * tree-vect-stmts.c (vectorizable_load): Do not setup
29874 re-alignment for invariant loads.
29876 2013-01-02 Richard Biener <rguenther@suse.de>
29878 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29879 invariant load do not generate a vector load from the scalar location.
29881 2013-01-03 Richard Biener <rguenther@suse.de>
29883 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
29884 for not vectorizing.
29885 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
29886 not build INDIRECT_REFs, call get_name once only.
29887 (vect_create_data_ref_ptr): Likewise. Dump base object kind
29888 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
29890 2013-01-03 Richard Biener <rguenther@suse.de>
29892 PR tree-optimization/55857
29893 * tree-vect-stmts.c (vectorizable_load): Do not setup
29894 re-alignment for invariant loads.
29896 2013-01-03 Richard Biener <rguenther@suse.de>
29899 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
29900 prefer a built-in decl.
29902 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29904 * gcc.c (process_command): Update copyright notice dates.
29905 * gcov.c (print_version): Likewise.
29906 * gcov-dump.c (print_version): Likewise.
29908 PR rtl-optimization/55838
29909 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
29910 iv0.step, iv1.step and step.
29912 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29913 Marc Glisse <marc.glisse@inria.fr>
29915 PR tree-optimization/55832
29916 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
29917 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
29918 integer_{one,zero}_node.
29920 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29923 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
29924 * var-tracking.c (reverse_op): Don't add reverse ops to
29925 VALUEs that have already
29926 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
29928 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
29930 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
29932 2013-01-02 Teresa Johnson <tejohnson@google.com>
29934 * dumpfile.c (dump_loc): Print filename with location.
29935 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
29936 new location_t parameter to emit complete unroll message with
29937 new dump framework.
29938 (canonicalize_loop_induction_variables): Compute loops location
29939 and pass to try_unroll_loop_completely.
29940 * loop-unroll.c (report_unroll_peel): New function.
29941 (peel_loops_completely): Use new dump format with location
29942 for main dumpfile message, and invoke report_unroll_peel on success.
29943 (decide_unrolling_and_peeling): Ditto.
29944 (decide_peel_once_rolling): Remove old dumpfile message subsumed
29945 by report_unroll_peel.
29946 (decide_peel_completely): Ditto.
29947 (decide_unroll_constant_iterations): Ditto.
29948 (decide_unroll_runtime_iterations): Ditto.
29949 (decide_peel_simple): Ditto.
29950 (decide_unroll_stupid): Ditto.
29951 * cfgloop.c (get_loop_location): New function.
29952 * cfgloop.h (get_loop_location): Declare.
29954 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29956 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
29959 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29961 PR middle-end/55198
29962 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
29963 BLKmode objects when EXPAND_MEMORY is specified.
29965 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29967 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
29969 (fold_builtin_cpu): Do not share cpu model decls across statements.
29971 2013-01-02 Jason Merrill <jason@redhat.com>
29974 * tree.c (build_array_type_1): Revert earlier change.
29976 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
29978 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
29980 * config/aarch64/aarch64-tune.md: Re-generate.
29982 2013-01-02 Richard Biener <rguenther@suse.de>
29984 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29985 invariant load do not generate a vector load from the scalar location.
29987 2013-01-02 Richard Biener <rguenther@suse.de>
29990 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
29991 * configure: Regenerate.
29993 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29995 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
29996 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
29997 (expand_builtin_int_roundingfn_2): Keep the original target around
29998 for the fallback case.
30000 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30002 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
30003 to be clear for sign changes.
30005 2013-01-01 Jan Hubicka <jh@suse.cz>
30007 * ipa-inline-analysis.c: Fix formatting.
30009 2013-01-01 Jakub Jelinek <jakub@redhat.com>
30011 PR tree-optimization/55831
30012 * tree-vect-loop.c (get_initial_def_for_induction): Use
30013 gsi_after_labels instead of gsi_start_bb.
30015 Copyright (C) 2013 Free Software Foundation, Inc.
30017 Copying and distribution of this file, with or without modification,
30018 are permitted in any medium without royalty provided the copyright
30019 notice and this notice are preserved.