453176f8822c157a0b82ecbb26552d2c288f0c8a
[gcc.git] / gcc / ChangeLog
1 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
2
3 * config/i386/bdver3.md : Added two additional decoder units
4 to support issue rate of 4 and remodeled vector unit.
5 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
6 architectures is set to 4.
7 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
8 lookahead is set to 4 for BD architectures.
9
10 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11
12 * config/rs6000/rs6000.c (rs6000_option_override_internal):
13 Remove restriction against use of VSX instructions when generating
14 code for little endian mode.
15
16 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17
18 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
19 for both big and little endian.
20 (mulv8hi3): Swap input operands for merge high and merge low
21 instructions for little endian.
22
23 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24
25 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
26 define_insn to define_expand that uses even patterns for big
27 endian and odd patterns for little endian.
28 (vec_widen_smult_even_v16qi): Likewise.
29 (vec_widen_umult_even_v8hi): Likewise.
30 (vec_widen_smult_even_v8hi): Likewise.
31 (vec_widen_umult_odd_v16qi): Likewise.
32 (vec_widen_smult_odd_v16qi): Likewise.
33 (vec_widen_umult_odd_v8hi): Likewise.
34 (vec_widen_smult_odd_v8hi): Likewise.
35 (altivec_vmuleub): New define_insn.
36 (altivec_vmuloub): Likewise.
37 (altivec_vmulesb): Likewise.
38 (altivec_vmulosb): Likewise.
39 (altivec_vmuleuh): Likewise.
40 (altivec_vmulouh): Likewise.
41 (altivec_vmulesh): Likewise.
42 (altivec_vmulosh): Likewise.
43
44 2013-11-05 Mike Stump <mikestump@comcast.net>
45
46 * Makefile.in (mostlyclean): Remove c-family objects.
47
48 2013-11-05 Ian Lance Taylor <iant@google.com>
49
50 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
51 If possible, add .cfi directives to record change to bx.
52 * config/i386/i386.c (ix86_emit_cfi): New function.
53 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
54
55 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
56
57
58 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
59 immediately after a label for a tablejump pattern.
60
61 * config/arm/arm.c (is_jump_table): Remove.
62 (create_fix_barrier): Use tablejump_p instead.
63 (arm_reorg): Likewise.
64 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
65 (thumb2_output_casesi): Likewise.
66 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
67 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
68 casesi_shift_media, casesi_load_media): Likewise.
69 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
70 * config/microblaze/microblaze.md: Likewise.
71
72 2013-11-05 Tobias Burnus <burnus@net-b.de>
73
74 * doc/invoke.texi (-Wdate-time): Document.
75
76 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
77
78 * double-int.c (lshift_double, rshift_double): Remove
79 SHIFT_COUNT_TRUNCATED handling.
80
81 2013-11-05 Jeff Law <law@redhat.com>
82
83 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
84 * common.opt (-fisolate-erroneous-paths): Add option and
85 documentation.
86 * gimple-ssa-isolate-paths.c: New file.
87 * gimple.c (check_loadstore): New function.
88 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
89 Verify OP is in the argument list and the argument corresponding
90 to OP is a pointer type. Use operand_equal_p rather than
91 pointer equality when testing if OP is on the nonnull list.
92 Use check_loadstore rather than count_ptr_derefs. Handle
93 GIMPLE_RETURN statements.
94 * tree-vrp.c (infer_nonnull_range): Remove.
95 * gimple.h (infer_nonnull_range): Declare.
96 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
97 * passes.def: Add pass_isolate_erroneous_paths.
98 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
99 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
100 * tree-ssa.c (struct count_ptr_d): Remove.
101 (count_ptr_derefs, count_uses_and_derefs): Remove.
102 * tree-ssa.h (count_uses_and_derefs): Remove.
103
104 2013-11-05 Jakub Jelinek <jakub@redhat.com>
105
106 PR rtl-optimization/58997
107 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
108 get_iv_value to be in iv->mode rather than iv->extend_mode.
109 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
110 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
111 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
112 mode.
113
114 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
115
116 * gimple.h: Move some prototypes to gimple-expr.h and add to include
117 list.
118 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
119 Move to gimple-expr.h.
120 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
121 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
122 is_gimple_id, virtual_operand_p, is_gimple_addressable,
123 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
124 Relocate here.
125 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
126 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
127 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
128 is_gimple_address, is_gimple_invariant_address,
129 is_gimple_ip_invariant_address, is_gimple_min_invariant,
130 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
131 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
132 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
133 gimple_decl_printable_name, useless_type_conversion_p,
134 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
135 gimple-expr.[ch].
136 * gimple-expr.c: New File.
137 (useless_type_conversion_p, gimple_set_body, gimple_body,
138 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
139 gimple_can_coalesce_p, extract_ops_from_tree_1,
140 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
141 is_gimple_address, is_gimple_invariant_address,
142 is_gimple_ip_invariant_address, is_gimple_min_invariant,
143 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
144 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
145 is_gimple_mem_ref_addr): Relocate here.
146 * Makefile.in (OBJS): Add gimple-expr.o.
147
148 2013-11-05 David Malcolm <dmalcolm@redhat.com>
149
150 * gengtype-parse.c (struct_field_seq): Support empty structs.
151
152 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
153
154 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
155
156 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
157
158 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
159 for !TARGET_80387.
160 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
161 _SOFT_FLOAT here.
162 (LONG_DOUBLE_TYPE_SIZE): New define.
163 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
164
165 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
166
167 PR c++/58724
168 * doc/extend.texi [visibility ("visibility_type")]: Add example
169 about visibility attribute on namespace declaration.
170
171 2013-11-05 Richard Biener <rguenther@suse.de>
172
173 PR ipa/58492
174 * passes.def (all_passes): Start with pass_fixup_cfg again.
175
176 2013-11-05 Richard Biener <rguenther@suse.de>
177
178 PR tree-optimization/58955
179 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
180 edge direction.
181
182 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
183
184 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
185 little endian.
186 (vec_pack_ufix_trunc_v2df): Likewise.
187
188 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
189
190 PR middle-end/58981
191 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
192 pattern, instead of word_mode.
193
194 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
195 Pmode for size.
196 (set_storage_via_setmem): Likewise.
197
198 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
199
200 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
201 gimple.h and the rest of the condition in eliminate_build.
202 (eliminate_build): Call new routine.
203 * gimple.h (phi_ssa_name_p): Delete.
204
205 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
206
207 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
208 case of no prefix and reserving zero slots, because when that's the
209 case we'll never get here.
210 * vec.h (va_heap::reserve): Don't try and handle
211 vec_prefix::calculate_allocation returning zero because that should
212 never happen.
213
214 2013-11-05 Richard Biener <rguenther@suse.de>
215
216 PR middle-end/58941
217 * tree-dfa.c (get_ref_base_and_extent): Merge common code
218 in MEM_REF and TARGET_MEM_REF handling. Make sure to
219 process trailing array detection before diving into the
220 view-converted object (and possibly apply some extra offset).
221
222 2013-11-05 Joseph Myers <joseph@codesourcery.com>
223
224 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
225 New function.
226 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
227
228 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
229
230 PR tree-optimization/58958
231 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
232 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
233
234 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
235
236 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
237 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
238
239 2013-11-05 Jakub Jelinek <jakub@redhat.com>
240
241 PR tree-optimization/58984
242 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
243 set *SIZE_P if non-NULL on success.
244 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
245 callers.
246 (ipcp_transform_function): Likewise. Punt if size of access
247 is different from TYPE_SIZE on v->value's type.
248
249 2013-11-05 Tobias Burnus <burnus@net-b.de>
250
251 * doc/invoke.texi (-fopenmp-simd): Document new option.
252 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
253 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
254 * tree.c (attribute_value_equal): Ditto.
255
256 2013-11-04 Wei Mi <wmi@google.com>
257
258 * sched-rgn.c (add_branch_dependences): Keep insns in
259 a SCHED_GROUP at the end of BB to remain their location.
260
261 2013-11-04 Wei Mi <wmi@google.com>
262
263 * gcc/config/i386/i386.c (memory_address_length): Extract a part
264 of code to rip_relative_addr_p.
265 (rip_relative_addr_p): New Function.
266 (ix86_macro_fusion_p): Ditto.
267 (ix86_macro_fusion_pair_p): Ditto.
268 * gcc/config/i386/i386.h: Add new tune features about macro-fusion.
269 * gcc/config/i386/x86-tune.def (DEF_TUNE): Ditto.
270 * gcc/doc/tm.texi: Generated.
271 * gcc/doc/tm.texi.in: Ditto.
272 * gcc/haifa-sched.c (try_group_insn): New Function.
273 (group_insns_for_macro_fusion): Ditto.
274 (sched_init): Call group_insns_for_macro_fusion.
275 * gcc/target.def: Add two hooks: macro_fusion_p and
276 macro_fusion_pair_p.
277
278 2013-11-04 Kostya Serebryany <kcc@google.com>
279
280 Update to match the changed asan API.
281 * asan.c (asan_function_start): New function.
282 (asan_emit_stack_protection): Update the string stored in the
283 stack red zone to match new API. Store the PC of the current
284 function in the red zone.
285 (asan_global_struct): Update the __asan_global definition to match
286 the new API.
287 (asan_add_global): Ditto.
288 * asan.h (asan_function_start): New prototype.
289 * final.c (final_start_function): Call asan_function_start.
290 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
291 to __asan_init_v3.
292
293 2013-11-04 Wei Mi <wmi@google.com>
294
295 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): Separate
296 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
297 * gcc/config/i386/i386.c (ix86_option_override_internal): Ditto.
298 (ix86_issue_rate): Ditto.
299 (ix86_adjust_cost): Ditto.
300 (ia32_multipass_dfa_lookahead): Ditto.
301 (ix86_sched_init_global): Ditto.
302 (get_builtin_code_for_version): Ditto.
303 * gcc/config/i386/i386.h (enum target_cpu_default): Ditto.
304 (enum processor_type): Ditto.
305 * gcc/config/i386/x86-tune.def (DEF_TUNE): Ditto.
306
307 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
308
309 PR rtl-optimization/58967
310 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
311 !lra_in_progress for mode sizes bigger word.
312
313 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
314
315 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
316 arguments to merge instruction for little endian.
317 (vec_widen_umult_lo_v16qi): Likewise.
318 (vec_widen_smult_hi_v16qi): Likewise.
319 (vec_widen_smult_lo_v16qi): Likewise.
320 (vec_widen_umult_hi_v8hi): Likewise.
321 (vec_widen_umult_lo_v8hi): Likewise.
322 (vec_widen_smult_hi_v8hi): Likewise.
323 (vec_widen_smult_lo_v8hi): Likewise.
324
325 2013-11-04 Ian Lance Taylor <iant@google.com>
326
327 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
328 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
329 builtins that take pointers.
330 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
331 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
332 (run_gcc): Pass -fnon-call-exceptions.
333
334 2013-11-04 Jakub Jelinek <jakub@redhat.com>
335
336 * optabs.c (expand_vec_perm): Revert one incorrect line from
337 2013-10-31 change.
338
339 PR tree-optimization/58978
340 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
341 use_stmt by single_imm_use directly. Only call single_imm_use
342 on SSA_NAMEs.
343
344 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
345
346 PR rtl-optimization/58968
347 * lra-spills.c (return_regno_p): New function.
348 (lra_final_code_change): Use it.
349
350 2013-11-04 Joseph Myers <joseph@codesourcery.com>
351
352 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
353 * target.def (float_exceptions_rounding_supported_p): New hook.
354 * targhooks.c (default_float_exceptions_rounding_supported_p): New
355 function.
356 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
357 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
358 New @hook.
359 * doc/tm.texi: Regenerate.
360 * config.gcc (powerpc*-*-linux*): Set extra_objs.
361 * config/rs6000/rs6000-linux.c: New file.
362 * config/rs6000/rs6000-protos.h
363 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
364 * config/rs6000/linux.h
365 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
366 * config/rs6000/linux64.h
367 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
368 * config/rs6000/t-linux (rs6000-linux.o): New rule.
369 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
370
371 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
372
373 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
374 Replace the define_insn_and_split with a define_insn and two
375 define_splits, with the split after reload re-permuting the source
376 register to its original value.
377 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
378 (*vsx_le_perm_store_v8hi): Likewise.
379 (*vsx_le_perm_store_v16qi): Likewise.
380
381 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
382
383 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
384 little endian.
385
386 2013-11-04 Jakub Jelinek <jakub@redhat.com>
387
388 PR tree-optimization/58946
389 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
390 bbs with bbinfo[idx].op != NULL before all blocks with
391 bbinfo[idx].op == NULL.
392
393 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
394
395 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
396 (avr_dump_double_int_hex): Likewise.
397 (avr_log_vadump): Remove %D and %X handling.
398 * config/avr/avr.c (avr_double_int_push_digit): Delete.
399 (avr_map_op_t): Change map from double_int to unsigned int.
400 (avr_map_op): Update accordingly.
401 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
402 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
403 unsigned ints rather than double_ints.
404
405 2013-11-03 Marek Polacek <polacek@redhat.com>
406
407 Implement -fsanitize=vla-bound.
408 * opts.c (common_handle_option): Handle vla-bound.
409 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
410 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
411 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
412
413 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
414
415 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
416 little endian.
417
418 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
419
420 * config/i386/constraints.md (Ts, Tv): New address constrains.
421 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
422 constraint for address_no_seg_operand.
423 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
424 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
425 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
426 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
427 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
428 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
429 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
430 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
431 constraint for vsib_address_operand.
432
433 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
434
435 * gcse.c (pre_delete): Remove references to regmove from comments.
436 * recog.c: (validate_replace_rtx_1): Likewise.
437 * config/rl78/rl78.c: Likewise.
438 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
439 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
440 * common/config/mmix/mmix-common.c: Likewise.
441
442 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
443
444 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
445 * gimplify.c (gimplify_compound_lval): Likewise.
446 * graphite-clast-to-gimple.c (gloog): Likewise.
447 * graphite-dependences.c (loop_is_parallel_p): Likewise.
448 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
449 (limit_scop); Likewise.
450 (build_scops): Likewise.
451 (dot_scop): Likewise.
452 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
453 (build_scop_drs): Likewise.
454 (insert_stmts): Likewise.
455 (insert_out_of_ssa_copy): Likewise.
456 (remove_phi): Likewise.
457 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
458 * hw-doloop.c (discover_loop): Likewise.
459 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
460 * tree-dfa.c (dump_enumerated_decls): Likewise.
461 * tree-if-conv.c (if_convertable_loop_p): Likewise.
462 * tree-inline.c (tree_function_versioning): Likewise.
463 * tree-loop-distribution.c (build_rdg): Likewise.
464 (rdg_flag_vertex_and_dependent): Likewise.
465 (distribute_loop): Likewise.
466 * tree-parloops.c (loop_parallel_p): Likewise.
467 (eliminate_local_variables): Likewise.
468 (separate_decls_in_region): Likewise.
469 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
470 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
471 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
472 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
473 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
474 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
475 (vectorizable_condition): Likewise.
476
477 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
478
479 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
480 * configure: Regenerate.
481 * config/i386/i386.md (*movdi_internal): Change
482 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
483 (*movdf_internal): Ditto.
484 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
485 * config/i386/sse.md (vec_concatv2di): Output interunit movq
486 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
487
488 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
489
490 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
491 after comment.
492
493 2013-10-31 David Malcolm <dmalcolm@redhat.com>
494
495 Automated part of renaming of symtab_node_base to symtab_node.
496
497 Patch autogenerated by rename_symtab.py from
498 https://github.com/davidmalcolm/gcc-refactoring-scripts
499 revision 58bb219cc090b2f4516a9297d868c245495ee622
500 with ChangeLog entry fixed up by hand.
501
502 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
503 symtab_node.
504 (cgraph_node_for_asm): Likewise.
505 * cgraph.h (symtab_node_base): Likewise.
506 (cgraph_node): Likewise.
507 (varpool_node): Likewise.
508 (is_a_helper <cgraph_node>::test): Likewise.
509 (is_a_helper <varpool_node>::test): Likewise.
510 (symtab_nodes): Likewise.
511 (symtab_register_node): Likewise.
512 (symtab_unregister_node): Likewise.
513 (symtab_remove_node): Likewise.
514 (symtab_get_node): Likewise.
515 (symtab_node_for_asm): Likewise.
516 (symtab_node_asm_name): Likewise.
517 (symtab_node_name): Likewise.
518 (symtab_insert_node_to_hashtable): Likewise.
519 (symtab_add_to_same_comdat_group): Likewise.
520 (symtab_dissolve_same_comdat_group_list): Likewise.
521 (dump_symtab_node): Likewise.
522 (debug_symtab_node): Likewise.
523 (dump_symtab_base): Likewise.
524 (verify_symtab_node): Likewise.
525 (verify_symtab_base): Likewise.
526 (symtab_used_from_object_file_p): Likewise.
527 (symtab_alias_ultimate_target): Likewise.
528 (symtab_resolve_alias): Likewise.
529 (fixup_same_cpp_alias_visibility): Likewise.
530 (symtab_for_node_and_aliases): Likewise.
531 (symtab_nonoverwritable_alias): Likewise.
532 (availability symtab_node_availability): Likewise.
533 (symtab_semantically_equivalent_p): Likewise.
534 (fixup_same_cpp_alias_visibility): Likewise.
535 (symtab_prevail_in_asm_name_hash): Likewise.
536 (cgraph): Likewise.
537 (varpool): Likewise.
538 (varpool_first_variable): Likewise.
539 (varpool_next_variable): Likewise.
540 (varpool_first_static_initializer): Likewise.
541 (varpool_next_static_initializer): Likewise.
542 (varpool_first_defined_variable): Likewise.
543 (varpool_next_defined_variable): Likewise.
544 (cgraph_first_defined_function): Likewise.
545 (cgraph_next_defined_function): Likewise.
546 (cgraph_first_function): Likewise.
547 (cgraph_next_function): Likewise.
548 (cgraph_first_function_with_gimple_body): Likewise.
549 (cgraph_next_function_with_gimple_body): Likewise.
550 (symtab_alias_target): Likewise.
551 (symtab_real_symbol_p): Likewise.
552 (symtab_can_be_discarded): Likewise.
553 * cgraphbuild.c (mark_address): Likewise.
554 (mark_load): Likewise.
555 (mark_store): Likewise.
556 * cgraphunit.c (decide_is_symbol_needed): Likewise.
557 (first): Likewise.
558 (enqueue_node): Likewise.
559 (referred_to_p): Likewise.
560 (cgraph_process_same_body_aliases): Likewise.
561 (analyze_functions): Likewise.
562 (handle_alias_pairs): Likewise.
563 (output_weakrefs): Likewise.
564 (compile): Likewise.
565 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
566 * ipa-inline-analysis.c (inline_write_summary): Likewise.
567 * ipa-prop.c (remove_described_reference): Likewise.
568 (try_decrement_rdesc_refcount): Likewise.
569 (ipa_edge_duplication_hook): Likewise.
570 * ipa-ref.c (ipa_record_reference): Likewise.
571 (ipa_maybe_record_reference): Likewise.
572 (ipa_clone_ref): Likewise.
573 (ipa_clone_references): Likewise.
574 (ipa_clone_referring): Likewise.
575 (ipa_find_reference): Likewise.
576 (ipa_remove_stmt_references): Likewise.
577 (ipa_clear_stmts_in_references): Likewise.
578 * ipa-ref.h (symtab_node_base): Likewise.
579 (ipa_ref): Likewise.
580 (ipa_record_reference): Likewise.
581 (ipa_maybe_record_reference): Likewise.
582 (ipa_clone_references): Likewise.
583 (ipa_clone_referring): Likewise.
584 (ipa_clone_ref): Likewise.
585 (ipa_find_reference): Likewise.
586 (ipa_remove_stmt_references): Likewise.
587 (ipa_clear_stmts_in_references): Likewise.
588 * ipa-reference.c (ipa_reference_write_optimization_summary):
589 Likewise.
590 * ipa.c (enqueue_node): Likewise.
591 (process_references): Likewise.
592 (walk_polymorphic_call_targets): Likewise.
593 (symtab_remove_unreachable_nodes): Likewise.
594 (address_taken_from_non_vtable_p): Likewise.
595 (comdat_can_be_unshared_p_1): Likewise.
596 (comdat_can_be_unshared_p): Likewise.
597 (can_replace_by_local_alias): Likewise.
598 (function_and_variable_visibility): Likewise.
599 * is-a.h: Likewise (within example in comment).
600 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
601 (lto_symtab_encoder_encode): Likewise.
602 (lto_symtab_encoder_delete_node): Likewise.
603 (lto_symtab_encoder_in_partition_p): Likewise.
604 (lto_set_symtab_encoder_in_partition): Likewise.
605 (output_refs): Likewise.
606 (compute_ltrans_boundary): Likewise.
607 (output_symtab): Likewise.
608 (input_node): Likewise.
609 (input_ref): Likewise.
610 (input_edge): Likewise.
611 (input_cgraph_1): Likewise.
612 (input_refs): Likewise.
613 (output_cgraph_opt_summary): Likewise.
614 (input_node_opt_summary): Likewise.
615 (input_cgraph_opt_section): Likewise.
616 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
617 Likewise.
618 * lto-streamer-out.c (lto_output): Likewise.
619 (output_symbol_p): Likewise.
620 (produce_symtab): Likewise.
621 * lto-streamer.h (lto_encoder_entry): Likewise.
622 (lto_free_function_in_decl_state_for_node): Likewise.
623 (lto_symtab_encoder_encode): Likewise.
624 (lto_symtab_encoder_delete_node): Likewise.
625 (lto_symtab_encoder_in_partition_p): Likewise.
626 (lto_set_symtab_encoder_in_partition): Likewise.
627 (lto_symtab_encoder_lookup): Likewise.
628 (lsei_node): Likewise.
629 (lto_symtab_encoder_deref): Likewise.
630 * symtab.c (symtab_hash): Likewise.
631 (assembler_name_hash): Likewise.
632 (symtab_nodes): Likewise.
633 (hash_node): Likewise.
634 (eq_node): Likewise.
635 (hash_node_by_assembler_name): Likewise.
636 (eq_assembler_name): Likewise.
637 (insert_to_assembler_name_hash): Likewise.
638 (unlink_from_assembler_name_hash): Likewise.
639 (symtab_prevail_in_asm_name_hash): Likewise.
640 (symtab_register_node): Likewise.
641 (symtab_insert_node_to_hashtable): Likewise.
642 (symtab_unregister_node): Likewise.
643 (symtab_get_node): Likewise.
644 (symtab_remove_node): Likewise.
645 (symtab_initialize_asm_name_hash): Likewise.
646 (symtab_node_for_asm): Likewise.
647 (symtab_add_to_same_comdat_group): Likewise.
648 (symtab_dissolve_same_comdat_group_list): Likewise.
649 (symtab_node_asm_name): Likewise.
650 (symtab_node_name): Likewise.
651 (dump_symtab_base): Likewise.
652 (dump_symtab_node): Likewise.
653 (dump_symtab): Likewise.
654 (debug_symtab_node): Likewise.
655 (verify_symtab_base): Likewise.
656 (verify_symtab_node): Likewise.
657 (verify_symtab): Likewise.
658 (symtab_used_from_object_file_p): Likewise.
659 (symtab_node_availability): Likewise.
660 (symtab_alias_ultimate_target): Likewise.
661 (fixup_same_cpp_alias_visibility): Likewise.
662 (symtab_resolve_alias): Likewise.
663 (symtab_for_node_and_aliases): Likewise.
664 (symtab_for_node_and_aliases): Likewise.
665 (symtab_nonoverwritable_alias_1): Likewise.
666 (symtab_nonoverwritable_alias): Likewise.
667 (symtab_semantically_equivalent_p): Likewise.
668 * value-prof.c (init_node_map): Likewise.
669 * varasm.c (find_decl): Likewise.
670 * varpool.c (varpool_node_for_asm): Likewise.
671 (varpool_remove_unreferenced_decls): Likewise.
672
673 2013-10-31 David Malcolm <dmalcolm@redhat.com>
674
675 Manual part of renaming of symtab_node_base to symtab_node.
676
677 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
678 clashes with the preferred name for the base class.
679 (const_symtab_node): Remove redundant typedef.
680
681 2013-10-31 Jakub Jelinek <jakub@redhat.com>
682
683 * optabs.c (expand_vec_perm): Avoid vector mode punning
684 SUBREGs in SET_DEST.
685 * expmed.c (store_bit_field_1): Likewise.
686 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
687 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
688 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
689 * config/i386/i386.c (ix86_expand_vector_move_misalign,
690 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
691 ix86_expand_sse_unpack, ix86_expand_args_builtin,
692 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
693 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
694 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
695 expand_vec_perm_vpshufb2_vpermq,
696 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
697 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
698 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
699 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
700 than *d itself.
701
702 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
703
704 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
705 Use gcc_unreachable for unhandled modes. Do not check results of
706 expand_simple_binop. If not expanded to target, move the result.
707
708 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
709 Shiva Chen <shiva0217@gmail.com>
710
711 * config.gcc (nds32*-*-*): Add nds32 target.
712 * config/nds32/nds32.c: New file.
713 * config/nds32/nds32.h: New file.
714 * config/nds32/nds32.md: New file.
715 * config/nds32/constants.md: New file.
716 * config/nds32/constraints.md: New file.
717 * config/nds32/iterators.md: New file.
718 * config/nds32/nds32-doubleword.md: New file.
719 * config/nds32/nds32-intrinsic.md: New file.
720 * config/nds32/nds32_intrinsic.h: New file.
721 * config/nds32/nds32-modes.def: New file.
722 * config/nds32/nds32-multiple.md: New file.
723 * config/nds32/nds32.opt: New file.
724 * config/nds32/nds32-opts.h: New file.
725 * config/nds32/nds32-protos.h: New file.
726 * config/nds32/nds32-peephole2.md: New file.
727 * config/nds32/pipelines.md: New file.
728 * config/nds32/predicates.md: New file.
729 * config/nds32/t-mlibs: New file.
730 * common/config/nds32: New directory and files.
731
732 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
733 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
734 * doc/install.texi (Cross-Compiler-Specific Options): Document
735 --with-nds32-lib for nds32 target.
736 * doc/extend.texi (Function Attributes, Target Builtins): Document
737 nds32 specific attributes.
738
739 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
740
741 * lra-constraints (process_alt_operands): Use the result
742 elimination register for operand when matching constraints.
743
744 2013-10-31 Jakub Jelinek <jakub@redhat.com>
745
746 * tree-vrp.c (maybe_set_nonzero_bits): New function.
747 (remove_range_assertions): Call it.
748
749 * tree.c (tree_ctz): New function.
750 * tree.h (tree_ctz): New prototype.
751 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
752 first argument from tree to const_tree.
753 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
754 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
755 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
756 static.
757 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
758 * tree-vect-loop.c (vect_analyze_loop_operations,
759 vect_transform_loop): Don't force scalar loop for bound just because
760 number of iterations is unknown, only do it if it is not known to be
761 a multiple of vectorization_factor.
762 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
763
764 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
765 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
766 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
767 (set_nonzero_bits, get_nonzero_bits): New prototypes.
768 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
769 a default def isn't partially constant.
770 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
771 is known to be partially zero.
772 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
773 to see if a default def isn't partially constant.
774 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
775 creation of a range, if VR_RANGE, try to improve nonzero_bits from
776 the range.
777 (set_nonzero_bits, get_nonzero_bits): New functions.
778
779 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
780 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
781 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
782 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
783 uses but in the condition and ASSERT_EXPR and the other successor of
784 the predecessor bb is __builtin_unreachable (), set_range_info of the
785 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
786
787 2013-10-31 Martin Jambor <mjambor@suse.cz>
788
789 PR rtl-optimization/58934
790 Revert:
791 2013-10-30 Martin Jambor <mjambor@suse.cz>
792 PR rtl-optimization/10474
793 * ira.c (find_moveable_pseudos): Do not calculate dominance info
794 nor df analysis.
795 (interesting_dest_for_shprep): New function.
796 (split_live_ranges_for_shrink_wrap): Likewise.
797 (ira): Calculate dominance info and df analysis. Call
798 split_live_ranges_for_shrink_wrap.
799
800 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
801 Yury Gribov <y.gribov@samsung.com>
802
803 PR sanitizer/58543
804 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
805 shadow pointer to avoid clobbering the main one.
806
807 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
808
809 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
810
811 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
812
813 PR bootstrap/58933
814 * ira-color.c (update_costs_from_copies): Add new parameter. Use
815 it for calling update_costs_from_allocno.
816 (assign_hard_reg): Call restore_costs_from_copies only for
817 !retry_p. Pass new argument to update_costs_from_copies.
818 (color_pass): Pass new argument to update_costs_from_copies.
819 (ira_mark_allocation_change): Ditto.
820
821 2013-10-30 Sharad Singhai <singhai@google.com>
822
823 PR middle-end/58134
824 * opts.c (common_handle_option): Remove deprecated option
825 -ftree-vectorizer-verbose.
826 * doc/invoke.texi (Debugging Options): Ditto.
827 * opts-global.c (handle_common_deferred_options): Ditto.
828 (dump_remap_tree_vectorizer_verbose): Delete.
829 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
830
831 2013-10-30 DJ Delorie <dj@redhat.com>
832
833 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
834 that take no arguments.
835
836 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
837
838 PR other/58545
839 * reload1.c (update_eliminables_and_spill): New function, broken
840 out of reload.
841 (reload): Use it. Check for frame size change after frame size
842 alignment, and call update_eliminables_and_spill first if continue-ing.
843
844 2013-10-30 Cong Hou <congh@google.com>
845
846 PR target/58762
847 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
848 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
849 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
850
851 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
852
853 PR rtl-optimization/58369
854 * reload1.c (compute_reload_subreg_offset): New function.
855 (choose_reload_regs): Use it to pass endian-correct
856 offset to subreg_regno_offset.
857
858 2013-10-30 Tobias Burnus <burnus@net-b.de>
859
860 PR other/33426
861 * gcc/tree-cfg.c (replace_loop_annotate): Replace warning by
862 warning_at.
863
864 2013-10-30 Jason Merrill <jason@redhat.com>
865
866 * configure.ac (loose_warn): Add -Wno-format if
867 --disable-build-format-warnings.
868
869 2013-10-30 David Malcolm <dmalcolm@redhat.com>
870
871 * cgraphunit.c (analyze_functions): Split symtab_node declarations
872 onto multiple lines to make things easier for rename_symtab.py.
873
874 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
875 (symtab_semantically_equivalent_p): Likewise.
876
877 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
878
879 PR target/58784
880 * lra.c (check_rtl): Remove address check before LRA work.
881
882 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
883
884 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
885 POINTER_PLUS_EXPR in the defining statement.
886
887 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
888
889 * regmove.c: Remove.
890 * tree-pass.h (make_pass_regmove): Remove.
891 * timevar.def (TV_REGMOVE): Remove.
892 * passes.def (pass_regmove): Remove.
893 * opts.c (default_options_table): Remove entry for regmove.
894 * doc/passes.texi: Remove regmove pass description.
895 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
896 options.
897 (-fdump-rtl-regmove): Ditto.
898 * common.opt (foptimize-register-move, fregmove): Ignore.
899 * Makefile.in (OBJS): Remove regmove.o.
900 * regmove.c: Remove.
901 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
902 and type.
903 (struct ira_allocno) New member allocno_prefs.
904 (ALLOCNO_PREFS): New macro.
905 (ira_prefs, ira_prefs_num): New external vars.
906 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
907 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
908 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
909 Ditto.
910 (ira_add_allocno_copy_to_list): Remove prototype.
911 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
912 (ira_pref_iterator): New type.
913 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
914 (FOR_EACH_PREF): New macro.
915 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
916 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
917 code.
918 (ira_setup_alts): New function.
919 (decrease_live_ranges_number): New function.
920 (ira): Call the above function.
921 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
922 (ira_create_allocno): Initialize allocno prefs.
923 (pref_pool, pref_vec): New static vars.
924 (initiate_prefs, find_allocno_pref, ira_create_pref): New
925 functions.
926 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
927 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
928 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
929 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
930 (ira_add_allocno_copy_to_list): Make static. Rename to
931 add_allocno_copy_to_list.
932 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
933 swap_allocno_copy_ends_if_necessary.
934 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
935 ira_remove_allocno_prefs.
936 (ira_flattening): Ditto.
937 (ira_build): Call initiate_prefs, print_prefs.
938 (ira_destroy): Call finish_prefs.
939 * ira-color.c (struct update_cost_record): New.
940 (struct allocno_color_data): Add new member update_cost_records.
941 (update_cost_record_pool): New static var.
942 (init_update_cost_records, get_update_cost_record): New functions.
943 (free_update_cost_record_list, finish_update_cost_records): Ditto.
944 (struct update_cost_queue_elem): Add member from.
945 (initiate_cost_update): Call init_update_cost_records.
946 (finish_cost_update): Call finish_update_cost_records.
947 (queue_update_cost, get_next_update_cost): Add new param from.
948 (Update_allocno_cost, update_costs_from_allocno): New functions.
949 (update_costs_from_prefs): Ditto.
950 (update_copy_costs): Rename to update_costs_from_copies.
951 (restore_costs_from_copies): New function.
952 (update_conflict_hard_regno_costs): Don't go back.
953 (assign_hard_reg): Call restore_costs_from_copies. Add printing
954 more debug info.
955 (pop_allocnos): Add priniting more debug info.
956 (color_allocnos): Remove prefs for conflicting hard regs.
957 Call update_costs_from_prefs.
958 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
959 (get_dup_num): Rename, modify, and move to ira.c
960 (process_regs_for_copy): Add prefs.
961 (add_insn_allocno_copies): Put src as first arg of
962 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
963 * ira-costs.c (record_reg_classes): Modify and move code into
964 record_operands_costs.
965 (find_costs_and_classes): Create prefs for the hard reg of small
966 reg class.
967 (process_bb_node_for_hard_reg_moves): Add prefs.
968
969 2013-10-30 Richard Biener <rguenther@suse.de>
970
971 PR middle-end/57100
972 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
973 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
974 as worker for ...
975 (pre_and_rev_post_order_compute): ... which now wraps it.
976 * graph.c (draw_cfg_nodes_no_loops): Use
977 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
978 on cfun.
979
980 2013-10-30 Christian Bruel <christian.bruel@st.com>
981
982 * gcc/config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
983 (sh_expand_cmpstr): Handle known align and schedule improvements.
984 * gcc/config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
985 * gcc/config/sh/sh.md (cmpstrnsi): New pattern.
986
987 2013-10-30 Martin Jambor <mjambor@suse.cz>
988
989 PR rtl-optimization/10474
990 * ira.c (find_moveable_pseudos): Do not calculate dominance info
991 nor df analysis.
992 (interesting_dest_for_shprep): New function.
993 (split_live_ranges_for_shrink_wrap): Likewise.
994 (ira): Calculate dominance info and df analysis. Call
995 split_live_ranges_for_shrink_wrap.
996
997 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
998
999 PR target/58854
1000 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
1001
1002 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1003
1004 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
1005 * tree.h (POINTER_BOUNDS_P): New.
1006 (BOUNDED_TYPE_P): New.
1007 (BOUNDED_P): New.
1008 (pointer_bounds_type_node): New.
1009 * tree.c (build_common_tree_nodes): Initialize
1010 pointer_bounds_type_node.
1011 * gimple.h (gimple_call_get_nobnd_arg_index): New.
1012 (gimple_call_num_nobnd_args): New.
1013 (gimple_call_nobnd_arg): New.
1014 (gimple_return_retbnd): New.
1015 (gimple_return_set_retbnd): New
1016 * gimple.c (gimple_build_return): Increase number of ops
1017 for return statement.
1018 (gimple_call_get_nobnd_arg_index): New.
1019 * gimple-pretty-print.c (dump_gimple_return): Print second op.
1020
1021 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1022
1023 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
1024 with "chkp ctor" and "bnd_legacy" attributes.
1025 * gimplify.c (gimplify_init_constructor): Avoid infinite
1026 loop during gimplification of bounds initializer.
1027
1028 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1029
1030 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
1031 (handle_bnd_legacy): New.
1032 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
1033 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
1034 attributes.
1035
1036 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
1037
1038 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
1039 (BT_FN_PTR_CONST_PTR): New.
1040 (BT_FN_CONST_PTR_CONST_PTR): New.
1041 (BT_FN_PTR_CONST_PTR_SIZE): New.
1042 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
1043 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
1044 (BT_FN_VOID_CONST_PTR_SIZE): New.
1045 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
1046 * chkp-builtins.def: New.
1047 * builtins.def: include chkp-builtins.def.
1048 (DEF_CHKP_BUILTIN): New.
1049 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
1050 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
1051 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
1052 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
1053 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
1054 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
1055 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
1056 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
1057 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
1058 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
1059 * common.opt (fcheck-pointer-bounds): New.
1060 * toplev.c (process_options): Check Pointer Bounds Checker is
1061 supported.
1062 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
1063
1064 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1065
1066 * target.def (builtin_chkp_function): New.
1067 (chkp_bound_type): New.
1068 (chkp_bound_mode): New.
1069 (fn_abi_va_list_bounds_size): New.
1070 (load_bounds_for_arg): New.
1071 (store_bounds_for_arg): New.
1072 * targhooks.h (default_load_bounds_for_arg): New.
1073 (default_store_bounds_for_arg): New.
1074 (default_fn_abi_va_list_bounds_size): New.
1075 (default_chkp_bound_type): New.
1076 (default_chkp_bound_mode): New.
1077 (default_builtin_chkp_function): New.
1078 * targhooks.c (default_load_bounds_for_arg): New.
1079 (default_store_bounds_for_arg): New.
1080 (default_fn_abi_va_list_bounds_size): New.
1081 (default_chkp_bound_type): New.
1082 (default_chkp_bound_mode); New.
1083 (default_builtin_chkp_function): New.
1084 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
1085 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
1086 (TARGET_STORE_BOUNDS_FOR_ARG): New.
1087 (TARGET_BUILTIN_CHKP_FUNCTION): New.
1088 (TARGET_CHKP_BOUND_TYPE): New.
1089 (TARGET_CHKP_BOUND_MODE): New.
1090 * doc/tm.texi: Regenerated.
1091 * langhooks.h (lang_hooks): Add chkp_supported field.
1092 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
1093 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
1094
1095 2013-10-29 Andrew Pinski <apinski@cavium.com>
1096
1097 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
1098 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
1099 fails, combining the branches anyways.
1100 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
1101 increases the number of combinings.
1102 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
1103
1104 2013-10-29 Mike Stump <mikestump@comcast.net>
1105
1106 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
1107 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
1108 (make_vector_mode): Increase namebuf to 16.
1109 (emit_insn_modes_h): When processing BImode, don't
1110 also match partial int modes.
1111 (emit_class_narrowest_mode): Likewise.
1112
1113 * config/bfin/bfin-modes.def: Add precision to PDI.
1114 * config/m32c/m32c-modes.def: Add precision to PSI.
1115 * config/msp430/msp430-modes.def: Add precision to PSI.
1116 * config/rs6000/rs6000-modes.def: Add precision to PTI.
1117 * config/sh/sh-modes.def: Add precision to PSI and PDI.
1118
1119 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
1120
1121 PR target/54236
1122 * config/sh/sh.md (*addc): Rename existing variations to ...
1123 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
1124 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
1125 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
1126 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
1127
1128 2013-10-29 Teresa Johnson <tejohnson@google.com>
1129
1130 PR ipa/58862
1131 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
1132 insanities when updating probabilities.
1133
1134 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1135
1136 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
1137 prettyprinter for cgraph_node to reflect the conversion of the
1138 symtable types to a C++ class hierarchy: it now *is* a
1139 symtab_node_base, rather than having one (named "symbol").
1140
1141 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
1142
1143 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
1144 __cilkrts_pop_frame. If matched, then return true for built-in
1145 function name.
1146 (expand_builtin): Added BUILT_IN_CILK_DETACH and
1147 BUILT_IN_CILK_POP_FRAME case.
1148 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
1149 (lhs_cilk_detect_spawn): Likewise.
1150 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
1151 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
1152 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
1153 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
1154 (LANG_HOOKS_CILKPLUS): Likewise.
1155 * tree.h (CILK_SPAWN_FN): Likewise.
1156 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
1157 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
1158 (OBJS): Added cilk-common.o.
1159 (BUILTINS_DEF): Added cilk-builtins.def.
1160 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
1161 (lhd_cilk_detect_spawn): Likewise.
1162 * langhooks.h (lang_hooks_for_cilkplus): New struct.
1163 (struct lang_hooks): Added new field called "cilkplus."
1164 * cilk-common.c: New file.
1165 * cilk.h: Likewise.
1166 * cilk-builtins.def: Likewise.
1167 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
1168 "__cilk" macro and set it to 200.
1169 * function.h (struct function::cilk_frame_decl): New field.
1170 (struct function::is_cilk_function): Likewise.
1171 (struct function::calls_cilk_spawn): Likewise.
1172 * gimplify.c (gimplify_call_expr): Added a check if the function call
1173 being gimplified is a spawn detach point. If so, then add pop_frame
1174 and detach function calls.
1175 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
1176 for gimplifying _Cilk_spawn and _Cilk_sync statements.
1177 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
1178 function. If so, added a _Cilk_sync and gimplified it.
1179 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
1180 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
1181 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
1182 spawner function.
1183 (can_inline_edge_p): Prevent inling of spawnee function.
1184 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
1185 for functions that use Cilk keywords.
1186 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
1187 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
1188 CILK_SYNC_STMT cases.
1189 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
1190 trees.
1191 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
1192 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
1193 * passes.texi (Cilk Keywords): New section that describes the compiler
1194 code changes for handling Cilk Keywords.
1195
1196 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1197
1198 Patch autogenerated by refactor_symtab.py from
1199 https://github.com/davidmalcolm/gcc-refactoring-scripts
1200 revision 58bb219cc090b2f4516a9297d868c245495ee622
1201
1202 * asan.c (asan_finish_file): Update for conversion of symtab types to
1203 a true class hierarchy.
1204 * cfgexpand.c (estimated_stack_frame_size): Likewise.
1205 * cgraph.c (cgraph_get_body): Likewise.
1206 (cgraph_get_create_real_symbol_node): Likewise.
1207 (verify_cgraph_node): Likewise.
1208 (verify_edge_corresponds_to_fndecl): Likewise.
1209 (verify_edge_count_and_frequency): Likewise.
1210 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
1211 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1212 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
1213 (cgraph_node_cannot_return): Likewise.
1214 (cgraph_set_pure_flag_1): Likewise.
1215 (cgraph_set_const_flag_1): Likewise.
1216 (cgraph_set_nothrow_flag_1): Likewise.
1217 (cgraph_make_node_local_1): Likewise.
1218 (cgraph_for_node_and_aliases): Likewise.
1219 (cgraph_for_node_thunks_and_aliases): Likewise.
1220 (cgraph_node_can_be_local_p): Likewise.
1221 (cgraph_node_cannot_be_local_p_1): Likewise.
1222 (cgraph_function_body_availability): Likewise.
1223 (dump_cgraph_node): Likewise.
1224 (cgraph_rtl_info): Likewise.
1225 (cgraph_mark_address_taken_node): Likewise.
1226 (cgraph_remove_node): Likewise.
1227 (cgraph_release_function_body): Likewise.
1228 (cgraph_update_edges_for_call_stmt_node): Likewise.
1229 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
1230 (cgraph_make_edge_direct): Likewise.
1231 (cgraph_resolve_speculation): Likewise.
1232 (cgraph_speculative_call_info): Likewise.
1233 (cgraph_turn_edge_to_speculative): Likewise.
1234 (cgraph_create_edge_1): Likewise.
1235 (cgraph_set_call_stmt): Likewise.
1236 (cgraph_node_for_asm): Likewise.
1237 (cgraph_add_thunk): Likewise.
1238 (cgraph_same_body_alias): Likewise.
1239 (cgraph_create_function_alias): Likewise.
1240 (cgraph_create_node): Likewise.
1241 (cgraph_create_empty_node): Likewise.
1242 (record_function_versions): Likewise.
1243 (used_from_object_file_p): Likewise.
1244 * cgraph.h (symtab_can_be_discarded): Likewise.
1245 (symtab_real_symbol_p): Likewise.
1246 (cgraph_mark_force_output_node): Likewise.
1247 (cgraph_edge_recursive_p): Likewise.
1248 (symtab_alias_target): Likewise.
1249 (varpool_all_refs_explicit_p): Likewise.
1250 (varpool_can_remove_if_no_refs): Likewise.
1251 (cgraph_only_called_directly_or_aliased_p): Likewise.
1252 (cgraph_next_function_with_gimple_body): Likewise.
1253 (cgraph_first_function_with_gimple_body): Likewise.
1254 (cgraph_function_with_gimple_body_p): Likewise.
1255 (cgraph_next_function): Likewise.
1256 (cgraph_first_function): Likewise.
1257 (cgraph_next_defined_function): Likewise.
1258 (cgraph_first_defined_function): Likewise.
1259 (varpool_next_defined_variable): Likewise.
1260 (varpool_first_defined_variable): Likewise.
1261 (varpool_next_static_initializer): Likewise.
1262 (varpool_first_static_initializer): Likewise.
1263 (varpool_next_variable): Likewise.
1264 (varpool_first_variable): Likewise.
1265 (varpool_node_name): Likewise.
1266 (varpool): Likewise.
1267 (cgraph): Likewise.
1268 (is_a_helper <varpool_node>::test): Likewise.
1269 (is_a_helper <cgraph_node>::test): Likewise.
1270 (varpool_variable_node): Likewise.
1271 (cgraph_function_or_thunk_node): Likewise.
1272 (varpool_alias_target): Likewise.
1273 (cgraph_alias_target): Likewise.
1274 (cgraph_node_name): Likewise.
1275 (varpool_node_asm_name): Likewise.
1276 (cgraph_node_asm_name): Likewise.
1277 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
1278 (cgraph_rebuild_references): Likewise.
1279 (rebuild_cgraph_edges): Likewise.
1280 (record_eh_tables): Likewise.
1281 (build_cgraph_edges): Likewise.
1282 (mark_store): Likewise.
1283 (mark_load): Likewise.
1284 (mark_address): Likewise.
1285 (record_type_list): Likewise.
1286 (record_reference): Likewise.
1287 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
1288 (cgraph_materialize_clone): Likewise.
1289 (cgraph_function_versioning): Likewise.
1290 (cgraph_copy_node_for_versioning): Likewise.
1291 (update_call_expr): Likewise.
1292 (cgraph_find_replacement_node): Likewise.
1293 (cgraph_create_virtual_clone): Likewise.
1294 (cgraph_clone_node): Likewise.
1295 * cgraphunit.c (compile): Likewise.
1296 (output_weakrefs): Likewise.
1297 (output_in_order): Likewise.
1298 (expand_function): Likewise.
1299 (assemble_thunks_and_aliases): Likewise.
1300 (expand_thunk): Likewise.
1301 (mark_functions_to_output): Likewise.
1302 (handle_alias_pairs): Likewise.
1303 (analyze_functions): Likewise.
1304 (walk_polymorphic_call_targets): Likewise.
1305 (varpool_finalize_decl): Likewise.
1306 (process_function_and_variable_attributes): Likewise.
1307 (cgraph_process_same_body_aliases): Likewise.
1308 (analyze_function): Likewise.
1309 (cgraph_add_new_function): Likewise.
1310 (cgraph_finalize_function): Likewise.
1311 (referred_to_p): Likewise.
1312 (cgraph_reset_node): Likewise.
1313 (cgraph_process_new_functions): Likewise.
1314 (enqueue_node): Likewise.
1315 (decide_is_symbol_needed): Likewise.
1316 * coverage.c (coverage_compute_profile_id): Likewise.
1317 * dbxout.c (dbxout_expand_expr): Likewise.
1318 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
1319 (reference_to_unused): Likewise.
1320 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1321 * gimplify.c (unvisit_body): Likewise.
1322 (unshare_body): Likewise.
1323 * ipa-cp.c (ipcp_generate_summary): Likewise.
1324 (ipcp_decision_stage): Likewise.
1325 (identify_dead_nodes): Likewise.
1326 (decide_whether_version_node): Likewise.
1327 (decide_about_value): Likewise.
1328 (perhaps_add_new_callers): Likewise.
1329 (create_specialized_node): Likewise.
1330 (update_profiling_info): Likewise.
1331 (ipcp_propagate_stage): Likewise.
1332 (estimate_local_effects): Likewise.
1333 (good_cloning_opportunity_p): Likewise.
1334 (devirtualization_time_bonus): Likewise.
1335 (propagate_constants_accross_call): Likewise.
1336 (initialize_node_lattices): Likewise.
1337 (ipcp_cloning_candidate_p): Likewise.
1338 (determine_versionability): Likewise.
1339 (print_all_lattices): Likewise.
1340 (print_lattice): Likewise.
1341 (ipcp_discover_new_direct_edges): Likewise.
1342 * ipa-devirt.c (ipa_devirt): Likewise.
1343 (likely_target_p): Likewise.
1344 (update_type_inheritance_graph): Likewise.
1345 (possible_polymorphic_call_target_p): Likewise.
1346 (dump_possible_polymorphic_call_targets): Likewise.
1347 (devirt_variable_node_removal_hook): Likewise.
1348 (record_binfo): Likewise.
1349 (maybe_record_node): Likewise.
1350 (build_type_inheritance_graph): Likewise.
1351 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1352 (inline_generate_summary): Likewise.
1353 (inline_analyze_function): Likewise.
1354 (do_estimate_growth): Likewise.
1355 (simple_edge_hints): Likewise.
1356 (estimate_node_size_and_time): Likewise.
1357 (estimate_edge_devirt_benefit): Likewise.
1358 (compute_inline_parameters): Likewise.
1359 (estimate_function_body_sizes): Likewise.
1360 (compute_bb_predicates): Likewise.
1361 (initialize_inline_failed): Likewise.
1362 (dump_inline_summary): Likewise.
1363 (dump_inline_edge_summary): Likewise.
1364 * ipa-inline-transform.c (inline_transform): Likewise.
1365 (preserve_function_body_p): Likewise.
1366 (save_inline_function_body): Likewise.
1367 (inline_call): Likewise.
1368 (clone_inlined_nodes): Likewise.
1369 (can_remove_node_now_p): Likewise.
1370 (can_remove_node_now_p_1): Likewise.
1371 * ipa-inline.c (early_inliner): Likewise.
1372 (early_inline_small_functions): Likewise.
1373 (inline_always_inline_functions): Likewise.
1374 (ipa_inline): Likewise.
1375 (flatten_function): Likewise.
1376 (inline_small_functions): Likewise.
1377 (speculation_useful_p): Likewise.
1378 (recursive_inlining): Likewise.
1379 (update_caller_keys): Likewise.
1380 (reset_edge_caches): Likewise.
1381 (update_edge_key): Likewise.
1382 (edge_badness): Likewise.
1383 (relative_time_benefit): Likewise.
1384 (want_inline_self_recursive_call_p): Likewise.
1385 (want_inline_small_function_p): Likewise.
1386 (want_early_inline_function_p): Likewise.
1387 (num_calls): Likewise.
1388 (can_early_inline_edge_p): Likewise.
1389 (can_inline_edge_p): Likewise.
1390 (report_inline_failed_reason): Likewise.
1391 * ipa-profile.c (ipa_profile): Likewise.
1392 (ipa_propagate_frequency): Likewise.
1393 (ipa_propagate_frequency_1): Likewise.
1394 (ipa_profile_generate_summary): Likewise.
1395 * ipa-prop.c (ipcp_transform_function): Likewise.
1396 (read_replacements_section): Likewise.
1397 (ipa_prop_read_section): Likewise.
1398 (ipa_modify_call_arguments): Likewise.
1399 (ipa_print_node_params): Likewise.
1400 (propagate_controlled_uses): Likewise.
1401 (update_indirect_edges_after_inlining): Likewise.
1402 (remove_described_reference): Likewise.
1403 (ipa_make_edge_direct_to_target): Likewise.
1404 (ipa_analyze_node): Likewise.
1405 (ipa_analyze_params_uses): Likewise.
1406 (ipa_compute_jump_functions): Likewise.
1407 (ipa_get_callee_param_type): Likewise.
1408 (ipa_print_node_jump_functions): Likewise.
1409 (ipa_initialize_node_params): Likewise.
1410 (ipa_populate_param_decls): Likewise.
1411 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
1412 (write_agg_replacement_chain): Likewise.
1413 (ipa_write_node_info): Likewise.
1414 (ipa_edge_duplication_hook): Likewise.
1415 (try_decrement_rdesc_refcount): Likewise.
1416 * ipa-pure-const.c (propagate_nothrow): Likewise.
1417 (propagate_pure_const): Likewise.
1418 (pure_const_read_summary): Likewise.
1419 (pure_const_write_summary): Likewise.
1420 (analyze_function): Likewise.
1421 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
1422 (ipa_ref_referring_ref_list): Likewise.
1423 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
1424 (ipa_remove_stmt_references): Likewise.
1425 (ipa_find_reference): Likewise.
1426 (ipa_dump_referring): Likewise.
1427 (ipa_dump_references): Likewise.
1428 (ipa_record_reference): Likewise.
1429 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
1430 (ipa_reference_write_optimization_summary): Likewise.
1431 (write_node_summary_p): Likewise.
1432 (propagate): Likewise.
1433 (read_write_all_from_decl): Likewise.
1434 (generate_summary): Likewise.
1435 (analyze_function): Likewise.
1436 (propagate_bits): Likewise.
1437 (ipa_reference_get_not_written_global): Likewise.
1438 (ipa_reference_get_not_read_global): Likewise.
1439 * ipa-split.c (execute_split_functions): Likewise.
1440 (split_function): Likewise.
1441 * ipa-utils.c (ipa_merge_profiles): Likewise.
1442 (dump_cgraph_node_set): Likewise.
1443 (ipa_reverse_postorder): Likewise.
1444 (ipa_edge_within_scc): Likewise.
1445 (ipa_get_nodes_in_cycle): Likewise.
1446 (ipa_free_postorder_info): Likewise.
1447 (ipa_reduced_postorder): Likewise.
1448 (searchc): Likewise.
1449 (recursive_call_p): Likewise.
1450 * ipa.c (ipa_cdtor_merge): Likewise.
1451 (record_cdtor_fn): Likewise.
1452 (function_and_variable_visibility): Likewise.
1453 (varpool_externally_visible_p): Likewise.
1454 (cgraph_externally_visible_p): Likewise.
1455 (comdat_can_be_unshared_p): Likewise.
1456 (comdat_can_be_unshared_p_1): Likewise.
1457 (address_taken_from_non_vtable_p): Likewise.
1458 (ipa_discover_readonly_nonaddressable_vars): Likewise.
1459 (symtab_remove_unreachable_nodes): Likewise.
1460 (walk_polymorphic_call_targets): Likewise.
1461 (process_references): Likewise.
1462 (enqueue_node): Likewise.
1463 (has_addr_references_p): Likewise.
1464 (cgraph_non_local_node_p_1): Likewise.
1465 * is-a.h (varpool_analyze_node): Likewise.
1466 * lto-cgraph.c (input_symtab): Likewise.
1467 (merge_profile_summaries): Likewise.
1468 (input_cgraph_1): Likewise.
1469 (input_edge): Likewise.
1470 (input_varpool_node): Likewise.
1471 (input_node): Likewise.
1472 (input_overwrite_node): Likewise.
1473 (compute_ltrans_boundary): Likewise.
1474 (output_refs): Likewise.
1475 (lto_output_varpool_node): Likewise.
1476 (lto_output_node): Likewise.
1477 (reachable_from_other_partition_p): Likewise.
1478 (referenced_from_other_partition_p): Likewise.
1479 (lto_output_edge): Likewise.
1480 (output_node_opt_summary): Likewise.
1481 (add_node_to): Likewise.
1482 (reachable_from_this_partition_p): Likewise.
1483 (lto_set_symtab_encoder_in_partition): Likewise.
1484 (lto_symtab_encoder_in_partition_p): Likewise.
1485 (lto_set_symtab_encoder_encode_initializer): Likewise.
1486 (lto_symtab_encoder_encode_initializer_p): Likewise.
1487 (lto_set_symtab_encoder_encode_body): Likewise.
1488 (lto_symtab_encoder_encode_body_p): Likewise.
1489 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1490 Likewise.
1491 * lto-streamer-in.c (lto_read_body): Likewise.
1492 (fixup_call_stmt_edges): Likewise.
1493 (fixup_call_stmt_edges_1): Likewise.
1494 * lto-streamer-out.c (produce_symtab): Likewise.
1495 (output_symbol_p): Likewise.
1496 (write_symbol): Likewise.
1497 (lto_output): Likewise.
1498 (copy_function): Likewise.
1499 (output_function): Likewise.
1500 * passes.c (function_called_by_processed_nodes_p): Likewise.
1501 (ipa_write_optimization_summaries): Likewise.
1502 (ipa_write_summaries): Likewise.
1503 (do_per_function_toporder): Likewise.
1504 (do_per_function): Likewise.
1505 (dump_passes): Likewise.
1506 * symtab.c (symtab_semantically_equivalent_p): Likewise.
1507 (symtab_nonoverwritable_alias): Likewise.
1508 (symtab_nonoverwritable_alias_1): Likewise.
1509 (symtab_for_node_and_aliases): Likewise.
1510 (symtab_resolve_alias): Likewise.
1511 (fixup_same_cpp_alias_visibility): Likewise.
1512 (symtab_alias_ultimate_target): Likewise.
1513 (symtab_used_from_object_file_p): Likewise.
1514 (verify_symtab_base): Likewise.
1515 (dump_symtab_base): Likewise.
1516 (symtab_node_name): Likewise.
1517 (symtab_node_asm_name): Likewise.
1518 (symtab_dissolve_same_comdat_group_list): Likewise.
1519 (symtab_add_to_same_comdat_group): Likewise.
1520 (symtab_unregister_node): Likewise.
1521 (symtab_insert_node_to_hashtable): Likewise.
1522 (symtab_register_node): Likewise.
1523 (unlink_from_assembler_name_hash): Likewise.
1524 (insert_to_assembler_name_hash): Likewise.
1525 (eq_assembler_name): Likewise.
1526 (hash_node_by_assembler_name): Likewise.
1527 (eq_node): Likewise.
1528 (hash_node): Likewise.
1529 * toplev.c (wrapup_global_declaration_2): Likewise.
1530 * trans-mem.c (ipa_tm_execute): Likewise.
1531 (ipa_tm_transform_clone): Likewise.
1532 (ipa_tm_transform_transaction): Likewise.
1533 (ipa_tm_transform_calls_redirect): Likewise.
1534 (ipa_tm_insert_gettmclone_call): Likewise.
1535 (ipa_tm_insert_irr_call): Likewise.
1536 (ipa_tm_create_version): Likewise.
1537 (ipa_tm_create_version_alias): Likewise.
1538 (ipa_tm_mark_forced_by_abi_node): Likewise.
1539 (ipa_tm_mark_force_output_node): Likewise.
1540 (ipa_tm_diagnose_tm_safe): Likewise.
1541 (ipa_tm_mayenterirr_function): Likewise.
1542 (ipa_tm_scan_irr_function): Likewise.
1543 (ipa_tm_note_irrevocable): Likewise.
1544 (ipa_tm_scan_calls_clone): Likewise.
1545 (get_cg_data): Likewise.
1546 * tree-eh.c (tree_could_trap_p): Likewise.
1547 * tree-emutls.c (ipa_lower_emutls): Likewise.
1548 (create_emultls_var): Likewise.
1549 (lower_emutls_function_body): Likewise.
1550 (gen_emutls_addr): Likewise.
1551 (emutls_decl): Likewise.
1552 (new_emutls_decl): Likewise.
1553 * tree-inline.c (tree_function_versioning): Likewise.
1554 (optimize_inline_calls): Likewise.
1555 (expand_call_inline): Likewise.
1556 (estimate_num_insns): Likewise.
1557 (copy_bb): Likewise.
1558 (delete_unreachable_blocks_update_callgraph): Likewise.
1559 * tree-nested.c (gimplify_all_functions): Likewise.
1560 (create_nesting_tree): Likewise.
1561 (check_for_nested_with_variably_modified): Likewise.
1562 * tree-pretty-print.c (dump_function_header): Likewise.
1563 * tree-profile.c (tree_profiling): Likewise.
1564 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
1565 (modify_function): Likewise.
1566 (convert_callers): Likewise.
1567 (convert_callers_for_node): Likewise.
1568 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
1569 (associate_varinfo_to_alias): Likewise.
1570 (create_variable_info_for): Likewise.
1571 (get_constraint_for_ssa_var): Likewise.
1572 * tree-vectorizer.c (increase_alignment): Likewise.
1573 * tree.c (find_decls_types_in_var): Likewise.
1574 (find_decls_types_in_node): Likewise.
1575 (free_lang_data_in_decl): Likewise.
1576 * value-prof.c (gimple_ic_transform): Likewise.
1577 (gimple_ic): Likewise.
1578 (check_ic_target): Likewise.
1579 (init_node_map): Likewise.
1580 * varasm.c (decl_binds_to_current_def_p): Likewise.
1581 (default_binds_local_p_1): Likewise.
1582 (dump_tm_clone_pairs): Likewise.
1583 (assemble_alias): Likewise.
1584 (find_decl): Likewise.
1585 (mark_decl_referenced): Likewise.
1586 * varpool.c (varpool_for_node_and_aliases): Likewise.
1587 (varpool_extra_name_alias): Likewise.
1588 (varpool_create_variable_alias): Likewise.
1589 (add_new_static_var): Likewise.
1590 (varpool_finalize_named_section_flags): Likewise.
1591 (varpool_remove_unreferenced_decls): Likewise.
1592 (enqueue_node): Likewise.
1593 (varpool_assemble_decl): Likewise.
1594 (assemble_aliases): Likewise.
1595 (varpool_analyze_node): Likewise.
1596 (cgraph_variable_initializer_availability): Likewise.
1597 (varpool_add_new_variable): Likewise.
1598 (ctor_for_folding): Likewise.
1599 (dump_varpool_node): Likewise.
1600 (varpool_remove_initializer): Likewise.
1601 (varpool_remove_node): Likewise.
1602 (varpool_node_for_decl): Likewise.
1603 (varpool_create_empty_node): Likewise.
1604 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
1605 (ix86_get_function_versions_dispatcher): Likewise.
1606
1607 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1608
1609 * cgraph.h (symtab_node_base): Convert to a class;
1610 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
1611 chain_next/prev from symtab_node_def.
1612 (cgraph_node): Inherit from symtab_node; add GTY option
1613 tag ("SYMTAB_FUNCTION").
1614 (varpool_node): Inherit from symtab_node; add GTY option
1615 tag ("SYMTAB_VARIABLE").
1616 (symtab_node_def): Remove.
1617 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
1618 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
1619 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
1620 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
1621
1622 * ipa-ref.h (symtab_node_def): Drop.
1623 (symtab_node): Change underlying type from symtab_node_def to
1624 symtab_node_base.
1625 (const_symtab_node): Likwise.
1626
1627 * is-a.h: Update examples in comment.
1628
1629 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
1630 (assembler_name_hash): Likewise.
1631
1632 2013-10-29 Martin Liska <marxin.liska@gmail.com>
1633
1634 * doc/tree-ssa.texi (gimple_phi_result): Document.
1635 (gimple_phi_num_args, gimple_phi_arg): Likewise.
1636 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
1637 (PHI_RESULT, PHI_NUM_ARGS): Remove.
1638 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
1639
1640 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1641
1642 * expr.h: Revert change and include tree-core.h.
1643 * rtl.h: Revert change and don't include tree-core.h.
1644
1645 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1646
1647 * config/darwin.c: Include gimple.h.
1648 * config/i386/winnt.c: Likewise.
1649
1650 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
1651
1652 PR tree-optimization/19831
1653 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
1654
1655 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1656
1657 * tree-outof-ssa.h: Remove include files.
1658 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
1659 * expr.c: Likewise.
1660 * tree-ssa-coalesce.c: Likewise.
1661 * cfgexpand.c: Likewise.
1662 * tree-ssa-ter.c: Likewise.
1663 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
1664 * lto-streamer.h: Likewise.
1665 * cgraphbuild.c: Add gimple.h to include list.
1666 * data-streamer-in.c: Likewise.
1667 * ipa-cp.c: Likewise.
1668 * tree-streamer.c: Likewise.
1669 * lto-compress.c: Likewise.
1670 * ipa-reference.c: Likewise.
1671 * data-streamer-out.c: Likewise.
1672 * lto-cgraph.c: Likewise.
1673 * cgraphclones.c: Likewise.
1674 * ipa-utils.c: Likewise.
1675 * data-streamer.c: Likewise.
1676 * ipa-split.c: Likewise.
1677 * lto-section-in.c: Likewise.
1678 * tree-streamer-out.c: Likewise.
1679 * ipa-prop.c: Likewise.
1680 * tree-streamer-in.c: Likewise.
1681 * symtab.c: Likewise.
1682 * opts-global.c: Likewise.
1683 * lto-opts.c: Likewise.
1684 * lto-section-out.c: Likewise.
1685 * lto-streamer.c: Likewise.
1686 * rtl.h: Add tree-core.h to include list.
1687 * expr.h: Remove tree-core.h from include list.
1688 * gimple.h: Likewise.
1689 * ipa-utils.h: Likewise.
1690 * streamer-hooks.h: Likewise.
1691 * streamer-hooks.c: Include input.h.
1692
1693 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1694
1695 * config/arm/arm.c (cortexa7_extra_costs): New table.
1696 (arm_cortex_a7_tune): New.
1697 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
1698
1699 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
1700
1701 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
1702
1703 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1704
1705 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
1706 to use autogenerated markers for a class-hierarchy, every class
1707 must have a GTY marker.
1708 * gengtype.h (struct type): Add linked list of subclasses to
1709 the "s" member of the union.
1710 (add_subclass): New decl.
1711 * gengtype-state.c (read_state_struct_type): Set up subclass
1712 linked list.
1713 * gengtype.c (get_ultimate_base_class): New.
1714 (add_subclass): New.
1715 (new_structure): Set up subclass linked list.
1716 (set_gc_used_type): Propagate usage information to subclasses.
1717 (output_mangled_typename): Use get_ultimate_base_class.
1718 (walk_subclasses): Use the subclass linked list, avoiding an
1719 O(N^2) when writing out all types.
1720 (walk_type): Issue an error if the base class is missing a tag,
1721 rather than generating bogus C code. Add a gcc_unreachable
1722 default case, in case people omit tags from concrete subclasses,
1723 or get the values wrong.
1724 (write_func_for_structure): Issue an error for subclasses for
1725 which the base doesn't have a "desc", since otherwise the
1726 autogenerated routines for the base would silently fail to visit
1727 any subclass fields.
1728 (write_root): Use get_ultimate_base_class, tweaking constness of
1729 tp to match that function's signature.
1730
1731 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1732
1733 * doc/gty.texi (GTY Options): Add note about inheritance to
1734 description of desc and tag.
1735 (Inheritance and GTY): New.
1736
1737 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1738
1739 * gengtype-parse.c (opts_have): Drop "static" so that
1740 we can use this from gengtype.c.
1741 * gengtype.c (set_gc_used_type): Mark any base class as used;
1742 update field traversal to visit inherited fields.
1743 (output_mangled_typename): Convert references to classes within
1744 an inheritance hierarchy to reference the ultimate base class,
1745 since only it will have gt_ functions.
1746 (get_string_option): New.
1747 (walk_subclasses): New.
1748 (walk_type): Treat GTY structs that have a "desc" as being the
1749 root of an inheritance hierarchy. Generate a switch on it
1750 within the marking function which walks all subclasses, adding
1751 cases for them via walk_subclasses. For subclasses, visit all
1752 fields of the type (including inherited ones).
1753 (write_func_for_structure): Don't write fns for subclasses, only
1754 for the ultimate base class within an inheritance hierarchy.
1755 Subclasses-marking will be handled by the base class marking functions.
1756 (write_types): Likewise.
1757 (write_local_func_for_structure): Likewise.
1758 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
1759 a "tag" option (and are thus concrete subclasses).
1760 (write_root): Use the marker function for the ultimate base class.
1761 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
1762 (opts_have): Add declaration.
1763
1764 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
1765
1766 * lra-spills.c (lra_final_code_change): Remove useless move insns
1767 originated from moves of pseudos.
1768
1769 2013-10-28 Jeff Law <law@redhat.com>
1770
1771 * lower-subreg.c (resolve_simple_move): Fix comment typo.
1772
1773 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
1774
1775 * df-scan.c (df_collection_rec): Adjust.
1776 (copy_defs): New constant.
1777 (copy_uses): Likewise.
1778 (copy_eq_uses): Likewise.
1779 (copy_mw): Likewise.
1780 (copy_all): Likewise.
1781 (df_insn_rescan): Adjust.
1782 (df_notes_rescan): Likewise.
1783 (df_swap_refs): Likewise.
1784 (df_sort_and_compress_refs): Likewise.
1785 (df_sort_and_compress_mws): Likewise.
1786 (df_install_refs): Likewise.
1787 (df_install_mws): Likewise.
1788 (df_refs_add_to_chains): Add flags parameter controlling which vectors
1789 are coppied.
1790 (df_bb_refs_record): Adjust.
1791 (df_record_entry_block_defs): Likewise.
1792 (df_record_exit_block_defs): Likewise.
1793 (df_refs_verify): Likewise.
1794 (df_mws_verify): Likewise.
1795 (df_insn_refs_verify): Likewise.
1796 (df_bb_verify): Likewise.
1797 * ipa-pure-const.c (finish_state): Remove.
1798 (propagate): Adjust.
1799 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
1800 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
1801 var-tracking.c: Adjust.
1802 * vec.c (stack_vecs): Remove.
1803 (register_stack_vec): Likewise.
1804 (stack_vec_register_index): Likewise.
1805 (unregister_stack_vec): Likewise.
1806 * vec.h (struct va_stack): Remove.
1807 (struct vec<T, A, vl_ptr>): Specialize as
1808 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
1809 allocation strategy compatable with the vl_ptr layout.
1810 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
1811 specialization anyway.
1812 (class stack_vec): New class.
1813 (vec_stack_alloc): Remove.
1814 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
1815
1816 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1817 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1818 Sergey Lega <sergey.s.lega@intel.com>
1819 Anna Tikhonova <anna.tikhonova@intel.com>
1820 Ilya Tocar <ilya.tocar@intel.com>
1821 Andrey Turetskiy <andrey.turetskiy@intel.com>
1822 Ilya Verbin <ilya.verbin@intel.com>
1823 Kirill Yukhin <kirill.yukhin@intel.com>
1824 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1825
1826 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
1827 (*prefetch_avx512pf_<mode>): New.
1828 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
1829 (avx512f_maskcmp<mode>3): Ditto.
1830 (vashrv16si3): Ditto.
1831
1832 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1833 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1834 Sergey Lega <sergey.s.lega@intel.com>
1835 Anna Tikhonova <anna.tikhonova@intel.com>
1836 Ilya Tocar <ilya.tocar@intel.com>
1837 Andrey Turetskiy <andrey.turetskiy@intel.com>
1838 Ilya Verbin <ilya.verbin@intel.com>
1839 Kirill Yukhin <kirill.yukhin@intel.com>
1840 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1841
1842 * config/i386/i386.md (any_truncate): New.
1843 (trunsuffix): Ditto.
1844 * config/i386/predicates.md (const_8_to_9_operand): New.
1845 (const_10_to_11_operand): Ditto.
1846 (const_12_to_13_operand): Ditto.
1847 (const_14_to_15_operand): Ditto.
1848 (const_16_to_19_operand): Ditto.
1849 (const_20_to_23_operand): Ditto.
1850 (const_24_to_27_operand): Ditto.
1851 (const_28_to_31_operand): Ditto.
1852 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
1853 (cvtusi2<ssescalarmodesuffix>32): New.
1854 (cvtusi2<ssescalarmodesuffix>64): Ditto.
1855 (ufloatv16siv16sf2): Ditto.
1856 (avx512f_fix_notruncv16sfv16si): Ditto.
1857 (avx512f_ufix_notruncv16sfv16si): Ditto.
1858 (avx512f_vcvtss2usi): Ditto.
1859 (avx512f_vcvtss2usiq): Ditto.
1860 (avx512f_vcvttss2usi): Ditto.
1861 (avx512f_vcvttss2usiq): Ditto.
1862 (avx512f_vcvtsd2usi): Ditto.
1863 (avx512f_vcvtsd2usiq): Ditto.
1864 (avx512f_vcvttsd2usi): Ditto.
1865 (avx512f_vcvttsd2usiq): Ditto.
1866 (ufloatv8siv8df): Ditto.
1867 (avx512f_cvtdq2pd512_2): Ditto.
1868 (avx512f_cvtpd2dq512): Ditto.
1869 (avx512f_ufix_notruncv8dfv8si): Ditto.
1870 (avx512f_cvtpd2ps512): Ditto.
1871 (vec_unpacks_lo_v16sf): Ditto.
1872 (vec_unpacks_hi_v16sf): Ditto.
1873 (vec_unpacks_float_hi_v16si): Ditto.
1874 (vec_unpacks_float_lo_v16si): Ditto.
1875 (avx512f_unpckhps512): Ditto.
1876 (avx512f_unpcklps512): Ditto.
1877 (avx512f_movshdup512): Ditto.
1878 (avx512f_movsldup512): Ditto.
1879 (vec_extract_lo_v32hi): Ditto.
1880 (vec_extract_hi_v32hi): Ditto.
1881 (vec_extract_lo_v64qi): Ditto.
1882 (vec_extract_hi_v64qi): Ditto.
1883 (avx512f_unpckhpd512): Ditto.
1884 (avx512f_movddup512): Ditto.
1885 (avx512f_unpcklpd512): Ditto.
1886 (*avx512f_unpcklpd512): Ditto.
1887 (avx512f_shufps512_1): Ditto.
1888 (avx512f_shufpd512_1): Ditto.
1889 (avx512f_interleave_highv8di): Ditto.
1890 (avx512f_interleave_lowv8di): Ditto.
1891 (PMOV_DST_MODE): Ditto.
1892 (pmov_src_mode): Ditto.
1893 (pmov_src_lower): Ditto.
1894 (pmov_suff): Ditto.
1895 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
1896 (*avx512f_<code>v8div16qi2): Ditto.
1897 (*avx512f_<code>v8div16qi2_store): Ditto.
1898 (vec_widen_umult_even_v16si): Ditto.
1899 (*vec_widen_umult_even_v16si): Ditto.
1900 (vec_widen_smult_even_v16si): Ditto.
1901 (*vec_widen_smult_even_v16si): Ditto.
1902 (avx512f_interleave_highv16si): Ditto.
1903 (avx512f_interleave_lowv16si): Ditto.
1904 (avx512f_<code>v16qiv16si2): Ditto.
1905 (avx512f_<code>v16hiv16si2): Ditto.
1906 (avx512f_<code>v8qiv8di2): Ditto.
1907 (avx512f_<code>v8hiv8di2): Ditto.
1908 (avx512f_<code>v8siv8di2): Ditto.
1909 (avx512cd_maskb_vec_dupv8di): Ditto.
1910 (avx512cd_maskw_vec_dupv16si): Ditto.
1911 (avx512f_vcvtph2ps512): Ditto.
1912 (avx512f_vcvtps2ph512): Ditto.
1913 (VEC_EXTRACT_MODE): Extened with wider modes.
1914 (VEC_PERM_AVX2): Ditto.
1915 (VEC_PERM_CONST): Ditto.
1916
1917 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
1918
1919 * config/arc/arc.c (arc_ccfsm_post_advance):
1920 Add comment about TYPE_RETURN.
1921
1922 2013-10-28 Bin Cheng <bin.cheng@arm.com>
1923
1924 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
1925 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
1926 (strip_offset): Convert offset to unsigned number.
1927
1928 2013-10-27 Tom de Vries <tom@codesourcery.com>
1929
1930 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
1931 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
1932
1933 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
1934
1935 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
1936 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
1937 (TARGET_BIG_ENDIAN): New macro.
1938 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
1939 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
1940 * config/sh/sh.c: Likewise.
1941 * config/sh/sh.h: Likewise.
1942
1943 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
1944
1945 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
1946 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
1947
1948 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1949
1950 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
1951 Fix formatting.
1952 (cmpstr_t, cmpstrsi): Fix formatting.
1953
1954 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1955
1956 PR target/52483
1957 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
1958 addressing, do not use general_operand for memory operands.
1959
1960 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
1961
1962 Revert:
1963 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1964 * lra-spills.c (lra_final_code_change): Remove useless move insns.
1965
1966 2013-10-26 Jeff Law <law@redhat.com>
1967
1968 * predict.c (PRED_MUDFLAP): Remove.
1969 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
1970
1971 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
1972 (OBJS): Remove tree-nomudflap.o
1973 (GTFILES): Remove tree-mudflap.c
1974 * builtins.c (expand_builtin_alloc): Remove mudflap support.
1975 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
1976 (mfwrap_spec, mflib_spec): Likewise.
1977 (cpp_unique_options, cc1_options, static_specs): Likewise.
1978 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
1979 * passes.def: Likewise.
1980 * toplev.c (compile_file, process_options): Likewise.
1981 * tree-inline.c (copy_tree_r): Likewise.
1982 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
1983 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
1984 (build_constant_desc, output_constant_def_contents): Likewise.
1985 (categorize_decl_for_section): Likewise.
1986 * tree-mudflap.c: Removed.
1987 * tree-mudflap.h: Removed.
1988 * tree-nomudflap.c: Removed.
1989 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
1990 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
1991 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
1992 * config/sol2.h (MFLIB_SPEC): Likewise.
1993 * doc/install.texi: Remove mudflap references.
1994 * doc/passes.texi: Similarly.
1995 * doc/sourcebuild.texi: Similarly.
1996 * doc/invoke.texi: Remove mudlfap related options.
1997
1998 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1999
2000 PR rtl-optimization/58759
2001 * lra-constraints.c (lra_constraints): Remove wrong condition to
2002 remove insn setting up an equivalent pseudo.
2003
2004 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2005
2006 * config/rs6000/rs6000-protos.h
2007 (rs6000_secondary_memory_needed_mode): New prototype.
2008 * config/rs6000/rs6000.c: Include ira.h.
2009 (TARGET_LRA_P): Redefine.
2010 (rs6000_legitimate_offset_address_p): Call
2011 legitimate_constant_pool_address_p in strict mode for LRA.
2012 (rs6000_legitimate_address_p): Ditto.
2013 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
2014 (rs6000_emit_move): Add LRA version of code to generate load/store
2015 of SDmode values.
2016 (rs6000_secondary_memory_needed_mode): New.
2017 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
2018 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
2019 constants, memory, and FP registers.
2020 (rs6000_lra_p): New.
2021 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
2022 * config/rs6000/rs6000.opt (mlra): New option.
2023 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2024
2025 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
2026
2027 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
2028 has_single_use () and not do the conversion if has_single_use ()
2029 returns false for the multiplication result.
2030
2031 2013-10-25 David Malcolm <dmalcolm@redhat.com>
2032
2033 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
2034 to "NODE", since this works on a "tree", not an
2035 "enum tree_code".
2036 (CONSTANT_CLASS_P): Likewise.
2037 (TYPE_P): Likewise.
2038 (DECL_P): Likewise.
2039 (INDIRECT_REF_P): Likewise.
2040 (REFERENCE_CLASS_P): Likewise.
2041 (COMPARISON_CLASS_P): Likewise.
2042 (UNARY_CLASS_P): Likewise.
2043 (BINARY_CLASS_P): Likewise.
2044 (STATEMENT_CLASS_P): Likewise.
2045 (VL_EXP_CLASS_P): Likewise.
2046 (EXPRESSION_CLASS_P): Likewise.
2047 (IS_TYPE_OR_DECL_P): Likewise.
2048
2049 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
2050
2051 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
2052 ADDR_EXPR in the defining statement.
2053
2054 2013-10-25 Richard Biener <rguenther@suse.de>
2055
2056 PR tree-optimization/58626
2057 * tree-loop-distribution.c (enum rdg_dep_type): Remove
2058 anti_dd, output_dd and input_dd.
2059 (struct rdg_edge): Remove level and relation members.
2060 (RDGE_LEVEL, RDGE_RELATION): Remove.
2061 (dot_rdg_1): Adjust.
2062 (create_rdg_edge_for_ddr): Remove.
2063 (create_rdg_edges_for_scalar): Adjust.
2064 (create_edge_for_control_dependence): Likewise.
2065 (create_rdg_edges): Split into ...
2066 (create_rdg_flow_edges): ... this
2067 (create_rdg_cd_edges): ... and this.
2068 (free_rdg): Adjust.
2069 (build_rdg): Likewise, do not compute data dependences or
2070 add edges for them.
2071 (pg_add_dependence_edges): New function.
2072 (pgcmp): Likewise.
2073 (distribute_loop): First apply all non-dependence based
2074 partition mergings. Then compute dependences between partitions
2075 and merge and order partitions according to them.
2076
2077 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2078
2079 PR rtl-optimization/58831
2080 * alias.c (init_alias_analysis): At the beginning of each iteration,
2081 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
2082
2083 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2084
2085 * recog.c (search_ofs): New static variable moved from...
2086 (peep2_find_free_register): ...here.
2087 (peephole2_optimize): Initialize it.
2088
2089 2013-10-25 Tobias Burnus <burnus@net-b.de>
2090
2091 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
2092
2093 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
2094
2095 * config/i386/i386.h (TARGET_MPX): New define.
2096 (TARGET_MPX_P): Ditto.
2097
2098 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2099
2100 * config/i386/constraints.md (B): New.
2101 (Ti): New.
2102 (Tb): New.
2103 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
2104 * config/i386/i386-modes.def (BND32): New.
2105 (BND64): New.
2106 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
2107 * config/i386/i386.c (isa_opts): Add mmpx.
2108 (regclass_map): Add bound registers.
2109 (dbx_register_map): Likewise.
2110 (dbx64_register_map): Likewise.
2111 (svr4_dbx_register_map): Likewise.
2112 (PTA_MPX): New.
2113 (ix86_option_override_internal): Support MPX ISA.
2114 (ix86_conditional_register_usage): Support bound registers.
2115 (print_reg): Likewise.
2116 (ix86_code_end): Add MPX bnd prefix.
2117 (output_set_got): Likewise.
2118 (ix86_output_call_insn): Likewise.
2119 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
2120 (ix86_print_operand_punct_valid_p): Likewise.
2121 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
2122 UNSPEC_BNDMK_ADDR.
2123 (ix86_class_likely_spilled_p): Add bound regs support.
2124 (ix86_hard_regno_mode_ok): Likewise.
2125 (x86_order_regs_for_local_alloc): Likewise.
2126 (ix86_bnd_prefixed_insn_p): New.
2127 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
2128 (FIXED_REGISTERS): Add bound registers.
2129 (CALL_USED_REGISTERS): Likewise.
2130 (REG_ALLOC_ORDER): Likewise.
2131 (HARD_REGNO_NREGS): Likewise.
2132 (TARGET_MPX): New.
2133 (VALID_BND_REG_MODE): New.
2134 (FIRST_BND_REG): New.
2135 (LAST_BND_REG): New.
2136 (reg_class): Add BND_REGS.
2137 (REG_CLASS_NAMES): Likewise.
2138 (REG_CLASS_CONTENTS): Likewise.
2139 (BND_REGNO_P): New.
2140 (ANY_BND_REG_P): New.
2141 (BNDmode): New.
2142 (HI_REGISTER_NAMES): Add bound registers.
2143 * config/i386/i386.md (UNSPEC_BNDMK): New.
2144 (UNSPEC_BNDMK_ADDR): New.
2145 (UNSPEC_BNDSTX): New.
2146 (UNSPEC_BNDLDX): New.
2147 (UNSPEC_BNDLDX_ADDR): New.
2148 (UNSPEC_BNDCL): New.
2149 (UNSPEC_BNDCU): New.
2150 (UNSPEC_BNDCN): New.
2151 (UNSPEC_MPX_FENCE): New.
2152 (BND0_REG): New.
2153 (BND1_REG): New.
2154 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
2155 (length_immediate): Likewise.
2156 (prefix_0f): Likewise.
2157 (memory): Likewise.
2158 (prefix_rep): Check for bnd prefix.
2159 (length_nobnd): New.
2160 (length): Use length_nobnd if specified.
2161 (BND): New.
2162 (bnd_ptr): New.
2163 (BNDCHECK): New.
2164 (bndcheck): New.
2165 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
2166 (*jcc_2): Likewise.
2167 (jump): Likewise.
2168 (simple_return_internal): Likewise.
2169 (simple_return_pop_internal): Likewise.
2170 (*indirect_jump): Add MPX bnd prefix.
2171 (*tablejump_1): Likewise.
2172 (simple_return_internal_long): Likewise.
2173 (simple_return_indirect_internal): Likewise.
2174 (<mode>_mk): New.
2175 (*<mode>_mk): New.
2176 (mov<mode>): New.
2177 (*mov<mode>_internal_mpx): New.
2178 (<mode>_<bndcheck>): New.
2179 (*<mode>_<bndcheck>): New.
2180 (<mode>_ldx): New.
2181 (*<mode>_ldx): New.
2182 (<mode>_stx): New.
2183 (*<mode>_stx): New.
2184 * config/i386/predicates.md (lea_address_operand): Rename to...
2185 (address_no_seg_operand): ... this.
2186 (address_mpx_no_base_operand): New.
2187 (address_mpx_no_index_operand): New.
2188 (bnd_mem_operator): New.
2189 * config/i386/i386.opt (mmpx): New.
2190 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
2191 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
2192
2193 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2194
2195 * mode-classes.def (MODE_POINTER_BOUNDS): New.
2196 * tree.def (POINTER_BOUNDS_TYPE): New.
2197 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
2198 (POINTER_BOUNDS_MODE): New.
2199 (make_pointer_bounds_mode): New.
2200 * machmode.h (POINTER_BOUNDS_MODE_P): New.
2201 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
2202 (layout_type): Support POINTER_BOUNDS_TYPE.
2203 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
2204 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
2205 (type_contains_placeholder_1): Likewise.
2206 * tree.h (POINTER_BOUNDS_TYPE_P): New.
2207 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
2208 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
2209
2210 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2211
2212 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
2213 mode of address computation when calling memory_address_addr_space.
2214
2215 2013-08-24 Richard Henderson <rth@twiddle.net>
2216
2217 PR rtl/58542
2218 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
2219 instead of create_convert_operand_to.
2220 (maybe_emit_sync_lock_test_and_set): Likewise.
2221 (expand_atomic_compare_and_swap): Likewise.
2222 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
2223
2224 2013-08-24 Sriraman Tallam <tmsriram@google.com>
2225
2226 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
2227 Change param_is to use the struct and not the pointer to the struct.
2228
2229 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
2230
2231 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
2232 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2233 Move to targhooks.c.
2234 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
2235 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2236 Relocate from gimplify.c.
2237 * targhooks.h: Add 2 prototypes.
2238 * tree.h: Delete 2 prototypes.
2239
2240 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2241
2242 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
2243 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
2244
2245 2013-10-24 Cong Hou <congh@google.com>
2246
2247 * convert.c (convert_to_real): Guard those unsafe math function
2248 convertions with flag_unsafe_math_optimizations. Handle sqrt()
2249 specially.
2250
2251 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
2252
2253 PR ipa/58712
2254 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
2255 as argument.
2256 (cgraph_create_edge): Use the new argument.
2257 (cgraph_create_indirect_edge): Likewise.
2258
2259 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
2260
2261 * gcc/config/arc/arc.c (arc_ccfsm_post_advance): Also handle
2262 TYPE_UNCOND_BRANCH.
2263 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
2264 changes statep->state.
2265
2266 2013-10-24 Tobias Burnus <burnus@net-b.de>
2267
2268 PR other/33426
2269 * tree-cfg.c (replace_loop_annotate): New function.
2270 (execute_build_cfg): Call it.
2271 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
2272 * internal-fn.c (expand_ANNOTATE): New function.
2273 * internal-fn.def (ANNOTATE): Define as new internal function.
2274 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
2275 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
2276 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
2277 * doc/extend.texi (Pragmas): Document #pragma ivdep.
2278 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
2279
2280 2013-10-17 Ian Bolton <ian.bolton@arm.com>
2281 Marcus Shawcroft <marcus.shawcroft@arm.com>
2282
2283 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
2284 Special case reload SP+C into none GENERAL_REGS.
2285
2286 2013-10-24 Michael Matz <matz@suse.de>
2287
2288 * gengtype.c (is_file_equal): Check that files will be same length.
2289
2290 2013-10-25 Christian Bruel <christian.bruel@st.com>
2291
2292 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
2293 * gcc/config/sh/t-sh (sh-mem.o): New rule.
2294 * gcc/config/sh/sh-mem.cc (expand_block_move): Moved here.
2295 (sh_expand_cmpstr): New function.
2296 * gcc/config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
2297 * gcc/config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
2298 * gcc/config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
2299 (rotlhi3_8): Rename.
2300
2301 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2302
2303 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
2304 * Makefile.in (CXXDEPMODE): Assign and change users.
2305 (CCDEPMODE): Delete.
2306 * configure: Regenerate.
2307
2308 2013-10-23 David Malcolm <dmalcolm@redhat.com>
2309
2310 * gengtype-parse.c (require_without_advance): New.
2311 (type): For GTY-marked types that are not GTY((user)), parse any
2312 base classes, requiring them to be single-inheritance, and not
2313 be templates. For non-GTY-marked types and GTY((user)),
2314 continue to skip over any C++ inheritance specification.
2315 * gengtype-state.c (state_writer::write_state_struct_type):
2316 Write base class of type (if any).
2317 (read_state_struct_type): Read base class of type (if any).
2318 * gengtype.c (new_structure): Add a "base_class" parameter.
2319 (create_optional_field_): Update for new parameter to new_structure.
2320 (adjust_field_rtx_def): Likewise.
2321 (adjust_field_tree_exp): Likewise.
2322 * gengtype.h (struct type): Add "base_class" field to the s
2323 union field.
2324 (new_structure): Add "base" parameter.
2325
2326 2013-10-23 Sriraman Tallam <tmsriram@google.com>
2327
2328 PR target/57756
2329 * config/i386/i386.c (ix86_option_override_internal):
2330 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
2331 (ix86_valid_target_attribute_tree):
2332 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
2333 Change TARGET_SSE to TARGET_SSE_P (opts->...)
2334
2335 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2336
2337 * tree-ssa-loop.h: Remove include files.
2338 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2339 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
2340 * graphite-clast-to-gimple.c: Likewise.
2341 * graphite-scop-detection.c: Likewise.
2342 * graphite-sese-to-poly.c: Likewise.
2343 * ipa-inline-analysis.c: Likewise.
2344 * ipa-pure-const.c: Likewise.
2345 * loop-init.c: Likewise.
2346 * passes.c: Likewise.
2347 * predict.c: Likewise.
2348 * tree-cfg.c: Likewise.
2349 * tree-cfgcleanup.c: Likewise.
2350 * tree-chrec.c: Likewise.
2351 * tree-data-ref.c: Likewise.
2352 * tree-loop-distribution.c: Likewise.
2353 * tree-parloops.c: Likewise.
2354 * tree-predcom.c: Likewise.
2355 * tree-scalar-evolution.c: Likewise.
2356 * tree-ssa-address.c: Likewise.
2357 * tree-ssa.c: Likewise.
2358 * tree-ssa-dce.c: Likewise.
2359 * tree-ssa-loop.c: Likewise.
2360 * tree-ssa-loop-im.c: Likewise.
2361 * tree-ssa-loop-ivcanon.c: Likewise.
2362 * tree-ssa-loop-ivopts.c: Likewise.
2363 * tree-ssa-loop-manip.c: Likewise.
2364 * tree-ssa-loop-niter.c: Likewise.
2365 * tree-ssa-loop-prefetch.c: Likewise.
2366 * tree-ssa-loop-unswitch.c: Likewise.
2367 * tree-ssa-reassoc.c: Likewise.
2368 * tree-vect-data-refs.c: Likewise.
2369 * tree-vect-loop.c: Likewise.
2370 * tree-vect-loop-manip.c: Likewise.
2371 * tree-vectorizer.c: Likewise.
2372 * tree-vect-stmts.c: Likewise.
2373 * tree-vrp.c: Likewise.
2374
2375 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2376
2377 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
2378
2379 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2380
2381 PR tree-optimization/58775
2382 PR tree-optimization/58791
2383 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
2384 (insert_stmt_after): Rewritten, don't move the stmt, but really
2385 insert it.
2386 (get_stmt_uid_with_default): Remove.
2387 (build_and_add_sum): Use insert_stmt_after and
2388 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
2389 (update_range_test): Set uid on stmts added by
2390 force_gimple_operand_gsi. Don't immediately modify statements
2391 in inter-bb optimization, just update oe->op values.
2392 (optimize_range_tests): Return bool whether any changed have been made.
2393 (update_ops): New function.
2394 (struct inter_bb_range_test_entry): New type.
2395 (maybe_optimize_range_tests): Perform statement changes here.
2396 (not_dominated_by, appears_later_in_bb, get_def_stmt,
2397 ensure_ops_are_available): Remove.
2398 (find_insert_point): Rewritten.
2399 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
2400 return LHS of the (new resp. old) stmt. Don't call
2401 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
2402 instead of last, move new stmt at the right place.
2403 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
2404 (negate_value): Likewise. Set uids.
2405 (break_up_subtract_bb): Initialize uids.
2406 (reassociate_bb): Adjust rewrite_expr_tree caller.
2407 (do_reassoc): Don't call renumber_gimple_stmt_uids.
2408
2409 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
2410
2411 PR target/58838
2412 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
2413 TARGET_32BIT final condition.
2414 (mulsi3_internal2 and splitter): Same.
2415
2416 2013-10-23 Jeff Law <law@redhat.com>
2417
2418 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
2419 through joiner blocks with abnormal outgoing edges.
2420
2421 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
2422 Add parameter JOINERS, to allow/disallow threading through joiner
2423 blocks.
2424 (thread_block): New. Call thread_block_1.
2425 (mark_threaded_blocks): Remove code to filter out certain cases
2426 of threading through joiner blocks.
2427 (thread_through_all_blocks): Document how we can have a dangling
2428 edge AUX field and clear it.
2429
2430 2013-10-23 Ian Lance Taylor <iant@google.com>
2431
2432 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
2433 (C++ Dialect Options): Likewise.
2434 (Optimize Options): Likewise.
2435
2436 2013-10-23 Tom de Vries <tom@codesourcery.com>
2437
2438 PR tree-optimization/58805
2439 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
2440
2441 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2442
2443 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
2444 sequence based on get_range_info returned range.
2445
2446 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2447
2448 * tree-ssa.h: Remove all #include's
2449 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2450 * alias.c: Move required includes from tree-ssa.h.
2451 * asan.c: Likewise.
2452 * builtins.c: Likewise.
2453 * calls.c: Likewise.
2454 * cfgexpand.c: Likewise.
2455 * cfghooks.c: Likewise.
2456 * cfgloop.c: Likewise.
2457 * cfgloopmanip.c: Likewise.
2458 * cgraph.c: Likewise.
2459 * cgraphbuild.c: Likewise.
2460 * cgraphclones.c: Likewise.
2461 * cgraphunit.c: Likewise.
2462 * dse.c: Likewise.
2463 * except.c: Likewise.
2464 * expr.c: Likewise.
2465 * final.c: Likewise.
2466 * fold-const.c: Likewise.
2467 * ggc-page.c: Likewise.
2468 * gimple-builder.c: Likewise.
2469 * gimple-fold.c: Likewise.
2470 * gimple-iterator.c: Likewise.
2471 * gimple-low.c: Likewise.
2472 * gimple-pretty-print.c: Likewise.
2473 * gimple-ssa-strength-reduction.c: Likewise.
2474 * gimple-streamer-in.c: Likewise.
2475 * gimple-streamer-out.c: Likewise.
2476 * gimplify.c: Likewise.
2477 * graphite-blocking.c: Likewise.
2478 * graphite-clast-to-gimple.c: Likewise.
2479 * graphite-dependences.c: Likewise.
2480 * graphite-interchange.c: Likewise.
2481 * graphite-optimize-isl.c: Likewise.
2482 * graphite-poly.c: Likewise.
2483 * graphite-scop-detection.c: Likewise.
2484 * graphite-sese-to-poly.c: Likewise.
2485 * graphite.c: Likewise.
2486 * ipa-cp.c: Likewise.
2487 * ipa-inline-analysis.c: Likewise.
2488 * ipa-inline-transform.c: Likewise.
2489 * ipa-inline.c: Likewise.
2490 * ipa-prop.c: Likewise.
2491 * ipa-pure-const.c: Likewise.
2492 * ipa-reference.c: Likewise.
2493 * ipa-split.c: Likewise.
2494 * ipa-utils.c: Likewise.
2495 * loop-init.c: Likewise.
2496 * lto-cgraph.c: Likewise.
2497 * lto-section-in.c: Likewise.
2498 * lto-section-out.c: Likewise.
2499 * lto-streamer-in.c: Likewise.
2500 * lto-streamer-out.c: Likewise.
2501 * lto-streamer.c: Likewise.
2502 * omp-low.c: Likewise.
2503 * passes.c: Likewise.
2504 * predict.c: Likewise.
2505 * print-tree.c: Likewise.
2506 * profile.c: Likewise.
2507 * sese.c: Likewise.
2508 * targhooks.c: Likewise.
2509 * tracer.c: Likewise.
2510 * trans-mem.c: Likewise.
2511 * tree-call-cdce.c: Likewise.
2512 * tree-cfg.c: Likewise.
2513 * tree-cfgcleanup.c: Likewise.
2514 * tree-chrec.c: Likewise.
2515 * tree-complex.c: Likewise.
2516 * tree-data-ref.c: Likewise.
2517 * tree-dfa.c: Likewise.
2518 * tree-eh.c: Likewise.
2519 * tree-emutls.c: Likewise.
2520 * tree-if-conv.c: Likewise.
2521 * tree-inline.c: Likewise.
2522 * tree-into-ssa.c: Likewise.
2523 * tree-loop-distribution.c: Likewise.
2524 * tree-mudflap.c: Likewise.
2525 * tree-nested.c: Likewise.
2526 * tree-nrv.c: Likewise.
2527 * tree-object-size.c: Likewise.
2528 * tree-outof-ssa.c: Likewise.
2529 * tree-parloops.c: Likewise.
2530 * tree-phinodes.c: Likewise.
2531 * tree-predcom.c: Likewise.
2532 * tree-pretty-print.c: Likewise.
2533 * tree-profile.c: Likewise.
2534 * tree-scalar-evolution.c: Likewise.
2535 * tree-sra.c: Likewise.
2536 * tree-ssa-address.c: Likewise.
2537 * tree-ssa-alias.c: Likewise.
2538 * tree-ssa-ccp.c: Likewise.
2539 * tree-ssa-coalesce.c: Likewise.
2540 * tree-ssa-copy.c: Likewise.
2541 * tree-ssa-copyrename.c: Likewise.
2542 * tree-ssa-dce.c: Likewise.
2543 * tree-ssa-dom.c: Likewise.
2544 * tree-ssa-dse.c: Likewise.
2545 * tree-ssa-forwprop.c: Likewise.
2546 * tree-ssa-ifcombine.c: Likewise.
2547 * tree-ssa-live.c: Likewise.
2548 * tree-ssa-loop-ch.c: Likewise.
2549 * tree-ssa-loop-im.c: Likewise.
2550 * tree-ssa-loop-ivcanon.c: Likewise.
2551 * tree-ssa-loop-ivopts.c: Likewise.
2552 * tree-ssa-loop-manip.c: Likewise.
2553 * tree-ssa-loop-niter.c: Likewise.
2554 * tree-ssa-loop-prefetch.c: Likewise.
2555 * tree-ssa-loop-unswitch.c: Likewise.
2556 * tree-ssa-loop.c: Likewise.
2557 * tree-ssa-math-opts.c: Likewise.
2558 * tree-ssa-operands.c: Likewise.
2559 * tree-ssa-phiopt.c: Likewise.
2560 * tree-ssa-phiprop.c: Likewise.
2561 * tree-ssa-pre.c: Likewise.
2562 * tree-ssa-propagate.c: Likewise.
2563 * tree-ssa-reassoc.c: Likewise.
2564 * tree-ssa-sccvn.c: Likewise.
2565 * tree-ssa-sink.c: Likewise.
2566 * tree-ssa-strlen.c: Likewise.
2567 * tree-ssa-structalias.c: Likewise.
2568 * tree-ssa-tail-merge.c: Likewise.
2569 * tree-ssa-ter.c: Likewise.
2570 * tree-ssa-threadedge.c: Likewise.
2571 * tree-ssa-threadupdate.c: Likewise.
2572 * tree-ssa-uncprop.c: Likewise.
2573 * tree-ssa-uninit.c: Likewise.
2574 * tree-ssa.c: Likewise.
2575 * tree-ssanames.c: Likewise.
2576 * tree-stdarg.c: Likewise.
2577 * tree-streamer-in.c: Likewise.
2578 * tree-switch-conversion.c: Likewise.
2579 * tree-tailcall.c: Likewise.
2580 * tree-vect-data-refs.c: Likewise.
2581 * tree-vect-generic.c: Likewise.
2582 * tree-vect-loop-manip.c: Likewise.
2583 * tree-vect-loop.c: Likewise.
2584 * tree-vect-patterns.c: Likewise.
2585 * tree-vect-slp.c: Likewise.
2586 * tree-vect-stmts.c: Likewise.
2587 * tree-vectorizer.c: Likewise.
2588 * tree-vrp.c: Likewise.
2589 * tree.c: Likewise.
2590 * tsan.c: Likewise.
2591 * value-prof.c: Likewise.
2592 * var-tracking.c: Likewise.
2593 * varpool.c: Likewise.
2594 * vtable-verify.c: Likewise.
2595
2596 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2597
2598 * config/tilegx/tilegx.c: Include "tree.h".
2599
2600 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2601
2602 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
2603 the info, not after it.
2604 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
2605 dump_ssaname_info call.
2606 (pp_gimple_stmt_1): Adjust caller.
2607 (dump_phi_nodes): Likewise. Don't print "# " here.
2608
2609 2013-10-22 Jan Hubicka <jh@suse.cz>
2610
2611 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
2612 tuning flag.
2613 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
2614 * i386.c (expand_small_movmem_or_setmem): New function.
2615 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
2616 function.
2617 (alg_usable_p): Add support for value ranges; cleanup.
2618 (ix86_expand_set_or_movmem): Add support for misaligned moves.
2619
2620 2013-10-22 Sterling Augustine <saugustine@google.com>
2621
2622 * doc/invoke.texi: Document -ggnu-pubnames.
2623 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
2624 logic.
2625 * dwarf2out.c: Include gdb/gdb-index.h.
2626 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
2627 debug_generate_pub_sections.
2628 (is_java, output_pubtables, output_pubname): New functions.
2629 (include_pubname_in_output): Handle debug_generate_pub_sections at
2630 level 2.
2631 (size_of_pubnames): Use new local space_for_flags based on
2632 debug_generate_pub_sections.
2633 (output_pubnames): Unify pubnames and pubtypes output logic.
2634 Genericize comments. Call output_pubname.
2635 (dwarf2out_finish): Move logic to output_pubnames and call it.
2636
2637 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
2638
2639 PR target/58779
2640 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
2641 Remove CCCmode handling.
2642 <case LTU>: Return 'c' suffix for CCCmode.
2643 <case GEU>: Return 'nc' suffix for CCCmode.
2644 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
2645 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
2646 (*sub<mode>3_cc_overflow): Ditto.
2647 (*subsi3_zext_cc_overflow): Ditto.
2648
2649 2013-10-22 Steve Ellcey <sellcey@mips.com>
2650
2651 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
2652 (AND (NOT OP1) (NOT OP2)).
2653
2654 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
2655
2656 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2657 meaning of merge-high and merge-low masks for little endian; avoid
2658 use of vector-pack masks for little endian for mismatched modes.
2659
2660 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2661
2662 * config/tilepro/tilepro.c: Include "tree.h".
2663
2664 2013-10-22 Andreas Schwab <schwab@suse.de>
2665
2666 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
2667 with PRE_DEC.
2668
2669 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
2670
2671 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
2672 [Ed Smith-Rowland]): New entries.
2673 ([Stephen M. Webb]): Update.
2674
2675 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
2676
2677 * tree-ssa-ter.h: Remove duplicate copy of file contents.
2678
2679 2013-10-21 Marek Polacek <polacek@redhat.com>
2680
2681 PR middle-end/58809
2682 * fold-const.c (fold_range_test): Return 0 if the type is not
2683 an integral type.
2684
2685 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
2686
2687 * system.h: Move hwint.h include further down.
2688 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
2689 gcc_checking_asserts.
2690 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
2691
2692 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2693
2694 Fix volatile issues in optimize_bit_field_compare.
2695 * fold-const.c (optimize_bit_field_compare): Bail out if
2696 lvolatilep or rvolatilep.
2697
2698 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2699
2700 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
2701 and get_inner_reference returning different pmode for non-volatile
2702 bit-field members dependent on flag_strict_volatile_bitfields.
2703 * stor-layout.c (layout_decl): Remove special handling of
2704 flag_strict_volatile_bitfields.
2705 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
2706 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
2707
2708 2013-10-21 Paulo Matos <pmatos@broadcom.com>
2709
2710 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
2711 calculations.
2712
2713 2013-10-21 Jeff Law <law@redhat.com>
2714
2715 * tree-ssa-threadedge.c (thread_through_normal_block): New
2716 argument VISITED. Remove VISISTED as a local variable. When we
2717 have a threadable jump, verify the destination of the jump has not
2718 been visised.
2719 (thread_across_edge): Allocate VISITED bitmap once at function
2720 scope and use it throughout. Make sure to set appropriate bits in
2721 VISITED for E (start of jump thread path).
2722 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
2723 through a joiner if any edge on the path has a recorded jump thread.
2724
2725 2013-10-21 Ian Lance Taylor <iant@google.com>
2726
2727 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
2728 don't imply that attributes can solve all problems.
2729 (Directory Options): Fix typo.
2730
2731 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2732
2733 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
2734 extend and extend_add.
2735
2736 2013-10-21 Richard Biener <rguenther@suse.de>
2737
2738 PR tree-optimization/58794
2739 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
2740 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
2741
2742 2013-10-21 Richard Biener <rguenther@suse.de>
2743
2744 PR middle-end/58742
2745 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
2746 to (T) X for sign-changing conversions (or no conversion).
2747
2748 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
2749
2750 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
2751
2752 2013-10-20 Jan Hubicka <jh@suse.cz>
2753
2754 * config/i386/i386-tune.def: Add comment; organize into categories
2755
2756 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2757
2758 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
2759 argument. Update function comment.
2760 (expand_set_or_movmem_via_rep): New function combining
2761 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
2762 (expand_movmem_via_rep_mov): Remove.
2763 expand_setmem_via_rep_stos): Remove.
2764 (expand_movmem_epilogue): Update calls correspondingly.
2765 (expand_setmem_epilogue_via_loop): Likewise.
2766 (emit_memset): New.
2767 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
2768 (expand_set_or_movmem_prologue): New function combining
2769 expand_movmem_prologue and expand_setmem_prologue.
2770 (expand_movmem_prologue): Remove.
2771 (expand_setmem_prologue): Remove.
2772 (expand_set_or_movmem_constant_prologue): New function combining
2773 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
2774 (expand_constant_movmem_prologue): Remove.
2775 (expand_constant_setmem_prologue): Remove.
2776 (promote_duplicated_reg): Allow vector-const0 value.
2777 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
2778 and ix86_expand_setmem.
2779 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
2780 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
2781
2782 2013-10-21 Diego Novillo <dnovillo@google.com>
2783
2784 * asan.c: Include tree.h
2785 * bb-reorder.c: Likewise.
2786 * cfgcleanup.c: Likewise.
2787 * cfgloopmanip.c: Likewise.
2788 * data-streamer-in.c: Likewise.
2789 * data-streamer-out.c: Likewise.
2790 * data-streamer.c: Likewise.
2791 * dwarf2cfi.c: Likewise.
2792 * graphite-blocking.c: Likewise.
2793 * graphite-clast-to-gimple.c: Likewise.
2794 * graphite-dependences.c: Likewise.
2795 * graphite-interchange.c: Likewise.
2796 * graphite-optimize-isl.c: Likewise.
2797 * graphite-poly.c: Likewise.
2798 * graphite-scop-detection.c: Likewise.
2799 * graphite-sese-to-poly.c: Likewise.
2800 * graphite.c: Likewise.
2801 * ipa-devirt.c: Likewise.
2802 * ipa-profile.c: Likewise.
2803 * ipa.c: Likewise.
2804 * ira.c: Likewise.
2805 * loop-init.c: Likewise.
2806 * loop-unroll.c: Likewise.
2807 * lower-subreg.c: Likewise.
2808 * lto/lto-object.c: Likewise.
2809 * recog.c: Likewise.
2810 * reginfo.c: Likewise.
2811 * tree-loop-distribution.c: Likewise.
2812 * tree-parloops.c: Likewise.
2813 * tree-ssa-strlen.c: Likewise.
2814 * tree-streamer.c: Likewise.
2815 * value-prof.c: Likewise.
2816 * target-globals.c: Likewise.
2817 * expr.h: Include tree-core.h instead of tree.h.
2818 * gimple.h: Likewise.
2819 * ipa-prop.h: Likewise.
2820 * ipa-utils.h: Likewise.
2821 * lto-streamer.h: Likewise.
2822 * streamer-hooks.h: Likewise.
2823 * ipa-reference.h: Include cgraph.h instead of tree.h.
2824 * cgraph.h: Include basic-block.h instead of tree.h.
2825 * tree-streamer.h: Do not include tree.h.
2826 * genattrtab.c (write_header): Generate inclusion of
2827 tree.h.
2828 * genautomata.c (main): Likewise.
2829 * genemit.c: Likewise.
2830 * genopinit.c: Likewise.
2831 * genoutput.c (output_prologue): Likewise.
2832 * genpeep.c: Likewise.
2833
2834 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2835
2836 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
2837 little endian.
2838 (vec_unpacku_hi_v8hi): Likewise.
2839 (vec_unpacku_lo_v16qi): Likewise.
2840 (vec_unpacku_lo_v8hi): Likewise.
2841
2842 2013-10-20 Jan Hubicka <jh@suse.cz>
2843
2844 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
2845 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
2846 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
2847
2848 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
2849
2850 * config/mips/mips.h (ISA_HAS_WSBH): Define.
2851 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
2852 constants.
2853 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
2854
2855 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
2856
2857 PR target/58603
2858 * system.h: Undef m_slot.
2859
2860 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2861
2862 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
2863 all elements for both endian flavors.
2864
2865 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2866
2867 PR target/58792
2868 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
2869 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
2870 and SI_REG for 64bit SYSV ABI targets.
2871
2872 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2873
2874 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
2875 to multi_reg_return. Clarify that we are skipping USEs of multiple
2876 return registers. Use bool type where appropriate.
2877
2878 2013-10-18 Jan Hubicka <jh@suse.cz>
2879
2880 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
2881 for cold functions.
2882 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
2883 (X86_TUNE_PUSH_MEMORY): Likewise.
2884 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
2885 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
2886 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
2887 New.
2888 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
2889 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
2890 Remove.
2891 (ix86_option_override_internal): Update to use tune features instead
2892 of variables.
2893
2894 2013-10-18 Cong Hou <congh@google.com>
2895
2896 PR tree-optimization/58508
2897 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
2898 statement that contains data refs with zero-step.
2899
2900 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2901
2902 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
2903 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
2904 * gimple-low.c (gimple_check_call_arg,
2905 gimple_check_call_matching_types): Move to cgraph.c.
2906 * gimple-low.h: Remove prototype.
2907 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
2908 Relocate from gimple-low.c.
2909 * cgraph.h: Add prototype. Don't include basic-block.h.
2910 * gimplify.c: Add gimple-low to include list.
2911 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
2912 * tree-eh.c: Add gimple-low to include list.
2913 * tree-nested.c: Likewise.
2914 * cfgexpand.c: Add tree-ssa-address.h to include list.
2915 * expr.c: Likewise.
2916 * gimple-fold.c: Likewise.
2917 * gimple-ssa-strength-reduction.c: Likewise.
2918 * trans-mem.c: Likewise.
2919 * tree-mudflap.c: Likewise.
2920 * tree-ssa-loop-ivopts.c: Likewise.
2921 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
2922 (degenerate_phi_result): Move to tree-phinodes.c.
2923 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
2924 * tree-ssa-threadedge.c: Likewise.
2925 * tree-vrp.c: Likewise.
2926 * tree-phinodes.c (degenerate_phi_result): Relocate here.
2927 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
2928 * tree-phinodes.h (degenerate_phi_result): Add prototype.
2929 * tree-ssa-copy.c: Include tree-ssa-dom.h.
2930 * tree-ssa-forwprop.c: Likewise.
2931 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
2932 pass_data_cleanup_cfg_post_optimizing,
2933 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
2934 * tree-optimize.c: Delete File.
2935 * graphite.c: Include tree-cfgcleanup.h.
2936 * passes.c: Likewise.
2937 * tree-cfg.c: Likewise.
2938 * tree-profile.c: Likewise.
2939 * tree-ssa-dse.c: Likewise.
2940 * tree-ssa-loop-ivcanon.c: Likewise.
2941 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
2942 * tree-outof-ssa.c: Include sbitmap.h.
2943 * tree-ssa-live.c: Likewise.
2944 * tree-ssa-propagate.c: Likewise.
2945 * tree-ssa-structalias.c: Likewise.
2946 * tree-stdarg.c: Likewise.
2947 * Makefile.in (OBJS): Delete tree-optimize.o.
2948 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
2949 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
2950 * varasm.c: Include basic-block.h.
2951 * cfgloop.h: Include function.h instead of basic-block.h
2952 (bb_loop_depth): Move to cfgloop.c.
2953 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
2954
2955 2013-10-18 Teresa Johnson <tejohnson@google.com>
2956
2957 * predict.c (probably_never_executed): Compare frequency-based
2958 count to number of training runs.
2959 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
2960
2961 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2962
2963 * config/arm/arm.c (cortexa9_extra_costs): New table.
2964 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
2965
2966 2013-10-18 Jeff Law <law@redhat.com>
2967
2968 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
2969
2970 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
2971 (register_jump_thread): Add "registered_jump_thread" debug
2972 counter support.
2973 * dbgcnt.def (registered_jump_thread): New debug counter.
2974
2975 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2976
2977 * config/rs6000/rs6000.c: Include cgraph.h.
2978
2979 2013-10-18 Teresa Johnson <tejohnson@google.com>
2980
2981 * tree-ssa-tail-merge.c (replace_block_by): Update edge
2982 weights during merging.
2983
2984 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2985
2986 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
2987 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
2988 * tree-ssa-operands.h: Remove prototype.
2989 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
2990 * gimple.c (virtual_operand_p): Relocate from gimple.c.
2991 * gimple.h: Add prototype.
2992 * gimple-ssa.h: Include tree-ssa-operands.h.
2993 * tree-dump.c: Add tree-cfg.h to include list.
2994 * tree-ssa-alias.c: Add ipa-reference.h to include list.
2995 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
2996 * config/i386/i386.c: Don't include tree-flow.h.
2997 * config/rs6000/rs6000.c: Likewise.
2998
2999 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3000
3001 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
3002
3003 2013-10-18 Richard Biener <rguenther@suse.de>
3004
3005 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
3006 or TYPE_UNSIGNED of integral types.
3007 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
3008 NULL_TREE for zero-precision integral types.
3009
3010 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
3011
3012 * config/aarch64/arm_neon.h
3013 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
3014
3015 2013-10-17 Sriraman Tallam <tmsriram@google.com>
3016
3017 PR target/57756
3018 * opth-gen.awk: Define target_flags_explicit.
3019
3020 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3021
3022 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
3023 fields to the reg_addr array that describes the valid addressing
3024 mode for any register, general purpose registers, floating point
3025 registers, and Altivec registers.
3026 (FIRST_RELOAD_REG_CLASS): Likewise.
3027 (LAST_RELOAD_REG_CLASS): Likewise.
3028 (struct reload_reg_map_type): Likewise.
3029 (reload_reg_map_type): Likewise.
3030 (RELOAD_REG_VALID): Likewise.
3031 (RELOAD_REG_MULTIPLE): Likewise.
3032 (RELOAD_REG_INDEXED): Likewise.
3033 (RELOAD_REG_OFFSET): Likewise.
3034 (RELOAD_REG_PRE_INCDEC): Likewise.
3035 (RELOAD_REG_PRE_MODIFY): Likewise.
3036 (reg_addr): Likewise.
3037 (mode_supports_pre_incdec_p): New helper functions to say whether
3038 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
3039 (mode_supports_pre_modify_p): Likewise.
3040 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
3041 print the valid address mode bits for each mode.
3042 (rs6000_debug_print_mode): Likewise.
3043 (rs6000_debug_reg_global): Likewise.
3044 (rs6000_setup_reg_addr_masks): New function to set up the address
3045 mask bits for each type.
3046 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
3047 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
3048 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
3049 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
3050 PRE_MODIFY are supported.
3051 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
3052 registers, instead of {src,dest}_av_p.
3053 (rs6000_print_options_internal): Tweak the debug output slightly.
3054
3055 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3056
3057 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
3058 isa attribute.
3059
3060 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
3061
3062 * tree-flow.h (struct omp_region): Move to omp-low.c.
3063 Remove omp_ prototypes and variables.
3064 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
3065 (copy_var_decl): Relocate prototype from tree-flow.h.
3066 * gimple.c (copy_var_decl): Relocate from omp-low.c.
3067 * tree.h: Move prototype to omp-low.h.
3068 * omp-low.h: New File. Relocate prototypes here.
3069 * omp-low.c (struct omp_region): Make local here.
3070 (root_omp_region): Make static.
3071 (copy_var_decl) Move to gimple.c.
3072 (new_omp_region): Make static.
3073 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
3074 * tree-cfg.c: Include omp-low.h.
3075 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
3076 * gimplify.c: Include omp-low.h.
3077 * tree-parloops.c: Likewise.
3078
3079 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3080
3081 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
3082 operands are in memory, prefer to force non-matched operand 1 to
3083 the register.
3084
3085 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3086
3087 PR target/58673
3088 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
3089 restrict TImode addresses to single indirect registers if both
3090 -mquad-memory and -mvsx-timode are used.
3091 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
3092 we should emit load/store quad. Remove using %y for quad memory
3093 addresses.
3094
3095 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
3096 constraints to allow load/store quad on machines where TImode is
3097 not allowed in VSX registers. Use 'n' instead of 'F' constraint
3098 for TImode to load integer constants.
3099
3100 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3101
3102 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
3103
3104 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
3105
3106 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
3107 handling of STACK_REG.
3108
3109 2013-10-17 Richard Biener <rguenther@suse.de>
3110
3111 PR tree-optimization/58143
3112 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
3113 New function.
3114 (rewrite_to_defined_overflow): Likewise.
3115 (move_computations_dom_walker::before_dom): Rewrite stmts
3116 with undefined signed overflow that are not always executed
3117 into unsigned arithmetic.
3118
3119 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3120
3121 PR target/57756
3122 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
3123 explicit isa flag to be an options variable, instead of using
3124 global_options_set. Remove define from rs6000.h.
3125 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
3126
3127 * config/rs6000/rs6000.c (rs6000_option_override_internal):
3128 Initialize rs6000_isa_flags_explicit.
3129 (rs6000_function_specific_save): Add gcc_options* parameter, so
3130 that the powerpc builds after the 2013-10-15 changes.
3131 (rs6000_function_specific_restore): Likewise.
3132
3133 2013-10-16 DJ Delorie <dj@redhat.com>
3134
3135 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
3136 op is a REG before checking REGNO.
3137 (rl78_alloc_physical_registers): Verify pattern is a SET before
3138 checking SET_SRC.
3139
3140 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3141
3142 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
3143 endianness.
3144 (vec_unpacks_lo_v4sf): Likewise.
3145 (vec_unpacks_float_hi_v4si): Likewise.
3146 (vec_unpacks_float_lo_v4si): Likewise.
3147 (vec_unpacku_float_hi_v4si): Likewise.
3148 (vec_unpacku_float_lo_v4si): Likewise.
3149
3150 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3151
3152 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
3153 (vsx_concat_v2sf): Likewise.
3154
3155 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
3156
3157 * config/aarch64/aarch64.md
3158 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
3159
3160 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
3161
3162 PR tree-optimization/58697
3163 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
3164 estimated_loop_iterations_int.
3165 (max_stmt_executions_int): Call get_max_loop_iterations_int.
3166 (get_max_loop_iterations_int): New. HWINT version of
3167 get_max_loop_iterations.
3168 * cfgloop.h: Add prototypes.
3169 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
3170 * loop-unroll.c (decide_peel_once_rolling): Call
3171 get_estimated_loop_iterations_int.
3172 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
3173 * tree-ssa-loop-niter.h: Tweak prototypes.
3174
3175 2013-10-16 David Malcolm <dmalcolm@redhat.com>
3176
3177 * gengtype-parse.c (struct_field_seq): Ignore access-control
3178 keywords ("public:" etc).
3179
3180 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
3181
3182 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
3183 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
3184
3185 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
3186
3187 * config/arm/arm.opt (mlra): New option.
3188 * config/arm/arm.c (arm_lra_p): New function.
3189 (TARGET_LRA_P): Define.
3190
3191 2013-10-16 Paulo Matos <pmatos@broadcom.com>
3192
3193 * tree-core.h (tree_code_name): Remove.
3194 * tree.h (get_tree_code_name): New prototype.
3195 * tree.c (tree_code_name): Make static.
3196 (get_tree_code_name): New function.
3197 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
3198 tree_class_check_failed, tree_range_check_failed,
3199 tree_not_class_check_failed, omp_clause_check_failed,
3200 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
3201 wrapper get_tree_code_name instead of calling tree_code_name directly.
3202 * tree-vrp.c (dump_asserts_for): Likewise.
3203 * tree-dump.c (dequeue_and_dump): Likewise.
3204 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
3205 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
3206 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
3207 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
3208 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
3209 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
3210 dump_gimple_omp_for): Likewise.
3211 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
3212 * tree-ssa-pre.c (print_pre_expr): Likewise.
3213 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
3214 * print-tree.c (print_node_brief, print_node): Likewise.
3215 * gimple.c (gimple_check_failed): Likewise.
3216 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
3217 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
3218 * config/mep/mep.c (mep_validate_vliw): Likewise.
3219 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
3220 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
3221
3222 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
3223
3224 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
3225 for AMD bdver3.
3226
3227 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
3228
3229 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
3230 (MULTILIB_MATCHES): Add multilib for -march=v8.
3231
3232 2013-10-15 Sriraman Tallam <tmsriram@google.com>
3233
3234 PR target/57756
3235 * optc-save-gen.awk: Add extra parameter to the save and restore
3236 target calls.
3237 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
3238 * tree.c (build_optimization_node): New parameter. Add extra parameter
3239 to call to cl_optimization_save.
3240 (build_target_option_node): New parameter. Add extra parameter
3241 to call to cl_target_option_save.
3242 * tree.h (build_optimization_node): New parameter.
3243 (build_target_option_node): New parameter.
3244 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
3245 build_optimization_node and build_target_option_node.
3246 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
3247 (handle_pragma_push_options): Ditto.
3248 * toplev.c (process_options): Ditto.
3249 * opts.c (init_options_struct): Check for opts_set non-null.
3250 * target.def (target_option.save): New parameter.
3251 (target_option.restore): New parameter.
3252 * tm.texi: Generate.
3253 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
3254 (ix86_pragma_target_parse): Ditto.
3255 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
3256 parameters.
3257 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
3258 to build_optimization_node and build_target_option_node.
3259 (rs6000_valid_attribute_p): Ditto.
3260 (rs6000_pragma_target_parse): Ditto.
3261 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
3262 data.
3263 * config/i386/i386.h:
3264 TARGET_64BIT_P: New Macro
3265 TARGET_MMX_P: New Macro.
3266 TARGET_3DNOW_P: New Macro.
3267 TARGET_3DNOW_A_P: New Macro.
3268 TARGET_SSE_P: New Macro.
3269 TARGET_SSE2_P: New Macro.
3270 TARGET_SSE3_P: New Macro.
3271 TARGET_SSSE3_P: New Macro.
3272 TARGET_SSE4_1_P: New Macro.
3273 TARGET_SSE4_2_P: New Macro.
3274 TARGET_AVX_P: New Macro.
3275 TARGET_AVX2_P: New Macro.
3276 TARGET_AVX512F_P: New Macro.
3277 TARGET_AVX512PF_P: New Macro.
3278 TARGET_AVX512ER_P: New Macro.
3279 TARGET_AVX512CD_P: New Macro.
3280 TARGET_FMA_P: New Macro.
3281 TARGET_SSE4A_P: New Macro.
3282 TARGET_FMA4_P: New Macro.
3283 TARGET_XOP_P: New Macro.
3284 TARGET_LWP_P: New Macro.
3285 TARGET_ABM_P: New Macro.
3286 TARGET_BMI_P: New Macro.
3287 TARGET_BMI2_P: New Macro.
3288 TARGET_LZCNT_P: New Macro.
3289 TARGET_TBM_P: New Macro.
3290 TARGET_POPCNT_P: New Macro.
3291 TARGET_SAHF_P: New Macro.
3292 TARGET_MOVBE_P: New Macro.
3293 TARGET_CRC32_P: New Macro.
3294 TARGET_AES_P: New Macro.
3295 TARGET_PCLMUL_P: New Macro.
3296 TARGET_CMPXCHG16B_P: New Macro.
3297 TARGET_FSGSBASE_P: New Macro.
3298 TARGET_RDRND_P: New Macro.
3299 TARGET_F16C_P: New Macro.
3300 TARGET_RTM_P: New Macro.
3301 TARGET_HLE_P: New Macro.
3302 TARGET_RDSEED_P: New Macro.
3303 TARGET_PRFCHW_P: New Macro.
3304 TARGET_ADX_P: New Macro.
3305 TARGET_FXSR_P: New Macro.
3306 TARGET_XSAVE_P: New Macro.
3307 TARGET_XSAVEOPT_P: New Macro.
3308 TARGET_LP64_P: New Macro.
3309 TARGET_X32_P: New Macro.
3310 TARGET_FPMATH_DEFAULT_P: New Macro.
3311 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
3312 * config/i386/i386.c (ix86_option_override_internal): New parameters.
3313 opts and opts_set.
3314 Change ix86_tune_string to access opts->x_ix86_tune_string.
3315 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
3316 Change ix86_arch_string to access opts->x_ix86_arch_string.
3317 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
3318 Change ix86_pmode to access opts->x_ix86_pmode.
3319 Change ix86_abi to access opts->x_ix86_abi.
3320 Change ix86_cmodel to access opts->x_ix86_cmodel.
3321 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
3322 Change ix86_isa_flags_explicit to access
3323 opts->x_ix86_isa_flags_explicit.
3324 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
3325 Change ix86_regparm to access opts->x_ix86_regparm.
3326 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
3327 Change ix86_preferred_stack_boundary_arg to access
3328 opts->x_ix86_preferred_stack_boundary_arg.
3329 Change ix86_force_align_arg_pointer to access
3330 opts->x_ix86_force_align_arg_pointer.
3331 Change ix86_incoming_stack_boundar_arg to access
3332 opts->x_ix86_incoming_stack_boundar_arg.
3333 Change ix86_fpmath to access opts->x_ix86_fpmath.
3334 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
3335 Change ix86_recip_name to access opts->x_ix86_recip_name.
3336 Change ix86_stack_protector_guard to access
3337 opts->x_ix86_stack_protector_guard.
3338 Change ix86_tune_memcpy_strategy to access
3339 opts->x_ix86_tune_memcpy_strategy.
3340 Change ix86_tune_memset_strategy to access
3341 opts->x_ix86_tune_memset_strategy.
3342 Change global_options to access opts.
3343 Change global_options_set to access opts_set.
3344 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
3345 Change TARGET_MMX to TARGET_MMX_P (opts->...).
3346 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
3347 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
3348 Change TARGET_SSE to TARGET_SSE_P (opts->...).
3349 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
3350 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
3351 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
3352 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
3353 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
3354 Change TARGET_AVX to TARGET_AVX_P (opts->...).
3355 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
3356 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
3357 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
3358 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
3359 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
3360 Change TARGET_FMA to TARGET_FMA_P (opts->...).
3361 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
3362 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
3363 Change TARGET_XOP to TARGET_XOP_P (opts->...).
3364 Change TARGET_LWP to TARGET_LWP_P (opts->...).
3365 Change TARGET_ABM to TARGET_ABM_P (opts->...).
3366 Change TARGET_BMI to TARGET_BMI_P (opts->...).
3367 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
3368 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
3369 Change TARGET_TBM to TARGET_TBM_P (opts->...).
3370 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
3371 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
3372 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
3373 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
3374 Change TARGET_AES to TARGET_AES_P (opts->...).
3375 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
3376 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
3377 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
3378 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
3379 Change TARGET_F16C to TARGET_F16C_P (opts->...).
3380 Change TARGET_RTM to TARGET_RTM_P (opts->...).
3381 Change TARGET_HLE to TARGET_HLE_P (opts->...).
3382 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
3383 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
3384 Change TARGET_ADX to TARGET_ADX_P (opts->...).
3385 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
3386 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
3387 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
3388 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
3389 Change TARGET_X32 to TARGET_X32_P (opts->...).
3390 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
3391 Change TARGET_FLOAT_RETURNS_IN_80387 to
3392 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
3393 (ix86_function_specific_save): New parameter. Use opts-> fields
3394 to replace global fields.
3395 (ix86_function_specific_restore): Ditto.
3396 (ix86_valid_target_attribute_inner_p): New parameters.
3397 Fix recursive call.
3398 Fix call to ix86_handle_option and set_option.
3399 (ix86_valid_target_attribute_tree): New parameters.
3400 Change global_options to access opts.
3401 Change global_options_set to access opts_set.
3402 Fix call to ix86_valid_target_attribute_inner_p.
3403 Change ix86_tune_string to access opts->x_ix86_tune_string.
3404 Change ix86_arch_string to access opts->x_ix86_arch_string.
3405 Change ix86_fpmath to access opts->x_ix86_fpmath
3406 Fix call to ix86_option_override_internal.
3407 Fix call to ix86_add_new_builtins.
3408 Fix calls to build_optimization_node and build_target_option_node.
3409 (ix86_valid_target_attribute_p): Remove access to global_options.
3410 Use new gcc_options structure func_options.
3411 Fix call to ix86_valid_target_attribute_tree.
3412 Fix call to build_optimization_node.
3413 (get_builtin_code_for_version): Fix call to
3414 ix86_valid_target_attribute_tree.
3415
3416 2013-10-15 David Malcolm <dmalcolm@redhat.com>
3417
3418 * Makefile.in (PICFLAG): New.
3419 (enable_host_shared): New.
3420 (INTERNAL_CFLAGS): Use PICFLAG.
3421 (LIBIBERTY): Use pic build of libiberty.a if configured with
3422 --enable-host-shared.
3423 * configure.ac: Add --enable-host-shared, setting up new
3424 PICFLAG variable.
3425 * configure: Regenerate.
3426 * doc/install.texi (--enable-shared): Add note contrasting it with ...
3427 (--enable-host-shared): New option.
3428
3429 2013-10-15 Richard Biener <rguenther@suse.de>
3430
3431 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
3432 for built-in functions.
3433
3434 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
3435
3436 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
3437 (optimize_range_tests_1): New function,
3438 extracted from optimize_range_tests.
3439 (optimize_range_tests_xor): Similarly.
3440 (optimize_range_tests_diff): New function.
3441 (optimize_range_tests): Use optimize_range_tests_1.
3442
3443 2013-10-15 Cong Hou <congh@google.com>
3444
3445 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
3446 requirement of the reduction pattern so that one operand of the
3447 reduction operation can come from outside of the loop.
3448
3449 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3450
3451 * config/arm/neon-schedgen.ml: Remove.
3452 * config/arm/cortex-a9-neon.md: Remove comment regarding
3453 neon-schedgen.ml.
3454
3455 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3456
3457 * config/arm/types: Remove old neon types.
3458
3459 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3460
3461 * config/arm/cortex-a7.md
3462 (cortex_a7_neon_type): New.
3463 (cortex_a7_neon_mul): Update for new types.
3464 (cortex_a7_neon_mla): Likewise.
3465 (cortex_a7_neon): Likewise.
3466
3467 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3468
3469 * config/arm/cortex-a15-neon.md
3470 (cortex_a15_neon_type): New,
3471
3472 (cortex_a15_neon_int_1): Remove.
3473 (cortex_a15_neon_int_2): Likewise.
3474 (cortex_a15_neon_int_3): Likewise.
3475 (cortex_a15_neon_int_4): Likewise.
3476 (cortex_a15_neon_int_5): Likewise.
3477 (cortex_a15_neon_vqneg_vqabs): Likewise.
3478 (cortex_a15_neon_vmov): Likewise.
3479 (cortex_a15_neon_vaba): Likewise.
3480 (cortex_a15_neon_vaba_qqq): Likewise.
3481 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3482 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3483 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3484 Likewise.
3485 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3486 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3487 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
3488 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3489 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3490 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3491 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3492 (cortex_a15_neon_shift_1): Likewise.
3493 (cortex_a15_neon_shift_2): Likewise.
3494 (cortex_a15_neon_shift_3): Likewise.
3495 (cortex_a15_neon_vshl_ddd): Likewise.
3496 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3497 (cortex_a15_neon_vsra_vrsra): Likewise.
3498 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3499 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3500 (cortex_a15_neon_bp_3cycle): Likewise.
3501 (cortex_a15_neon_ldm_2): Likewise.
3502 (cortex_a15_neon_stm_2): Likewise.
3503 (cortex_a15_neon_mcr): Likewise.
3504 (cortex_a15_neon_mrc): Likewise.
3505 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3506 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3507 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3508 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3509 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3510 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3511 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3512 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3513 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3514 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3515 (cortex_a15_neon_bp_simple): Likewise.
3516 (cortex_a15_neon_bp_2cycle): Likewise.
3517 (cortex_a15_neon_bp_3cycle): Likewise.
3518 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3519 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3520 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3521 (cortex_a15_neon_vld2_4_regs): Likewise.
3522 (cortex_a15_neon_vld3_vld4): Likewise.
3523 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3524 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3525 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
3526 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
3527 (cortex_a15_neon_vst3_vst4): Rename to...
3528 (cortex_a15_neon_vst4): This, update for new attributes.
3529 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
3530 (cortex_a15_neon_vld3_vld4_lane): Likewise.
3531 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3532 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3533 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3534 (cortex_a15_neon_ldm_2): Likewise.
3535 (cortex_a15_neon_stm_2): Likewise.
3536 (cortex_a15_neon_mcr): Likewise.
3537 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3538 (cortex_a15_neon_mrc): Likewise.
3539 (cortex_a15_neon_mrrc): Likewise.
3540
3541 (cortex_a15_neon_abd): New.
3542 (cortex_a15_neon_abd_q): Likewise.
3543 (cortex_a15_neon_aba): Likewise.
3544 (cortex_a15_neon_aba_q): Likewise.
3545 (cortex_a15_neon_acc): Likewise.
3546 (cortex_a15_neon_acc_q): Likewise.
3547 (cortex_a15_neon_arith_basic): Likewise.
3548 (cortex_a15_neon_arith_complex): Likewise.
3549 (cortex_a15_neon_multiply): Likewise.
3550 (cortex_a15_neon_multiply_q): Likewise.
3551 (cortex_a15_neon_mla): Likewise.
3552 (cortex_a15_neon_mla_q): Likewise.
3553 (cortex_a15_neon_sat_mla_long): Likewise.
3554 (cortex_a15_neon_shift_acc): Likewise.
3555 (cortex_a15_neon_shift_imm_basic): Likewise.
3556 (cortex_a15_neon_shift_imm_complex): Likewise.
3557 (cortex_a15_neon_shift_reg_basic): Likewise.
3558 (cortex_a15_neon_shift_reg_basic_q): Likewise.
3559 (cortex_a15_neon_shift_reg_complex): Likewise.
3560 (cortex_a15_neon_shift_reg_complex_q): Likewise.
3561 (cortex_a15_neon_fp_negabs): Likewise
3562 (cortex_a15_neon_fp_arith): Likewise
3563 (cortex_a15_neon_fp_arith_q): Likewise
3564 (cortex_a15_neon_fp_cvt_int): Likewise
3565 (cortex_a15_neon_fp_cvt_int_q): Likewise
3566 (cortex_a15_neon_fp_cvt_16): Likewise
3567 (cortex_a15_neon_fp_mul): Likewise
3568 (cortex_a15_neon_fp_mul_q): Likewise
3569 (cortex_a15_neon_fp_mla): Likewise
3570 (cortex_a15_neon_fp_mla_q): Likewise
3571 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
3572 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
3573 (cortex_a15_neon_bitops): Likewise.
3574 (cortex_a15_neon_bitops_q): Likewise.
3575 (cortex_a15_neon_from_gp): Likewise.
3576 (cortex_a15_neon_from_gp_q): Likewise.
3577 (cortex_a15_neon_tbl3_tbl4): Likewise.
3578 (cortex_a15_neon_zip_q): Likewise.
3579 (cortex_a15_neon_to_gp): Likewise.
3580 (cortex_a15_neon_load_a): Likewise.
3581 (cortex_a15_neon_load_b): Likewise.
3582 (cortex_a15_neon_load_c): Likewise.
3583 (cortex_a15_neon_load_d): Likewise.
3584 (cortex_a15_neon_load_e): Likewise.
3585 (cortex_a15_neon_load_f): Likewise.
3586 (cortex_a15_neon_store_a): Likewise.
3587 (cortex_a15_neon_store_b): Likewise.
3588 (cortex_a15_neon_store_c): Likewise.
3589 (cortex_a15_neon_store_d): Likewise.
3590 (cortex_a15_neon_store_e): Likewise.
3591 (cortex_a15_neon_store_f): Likewise.
3592 (cortex_a15_neon_store_g): Likewise.
3593 (cortex_a15_neon_store_h): Likewise.
3594 (cortex_a15_vfp_to_from_gp): Likewise.
3595
3596 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3597
3598 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
3599
3600 (cortex_a9_neon_vshl_ddd): Remove.
3601 (cortex_a9_neon_vst3_vst4): Likewise.
3602 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3603
3604 (cortex_a9_neon_bit_ops_q): New.
3605
3606 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
3607 (cortex_a9_neon_int_2): Likewise.
3608 (cortex_a9_neon_int_3): Likewise.
3609 (cortex_a9_neon_int_4): Likewise.
3610 (cortex_a9_neon_int_5): Likewise.
3611 (cortex_a9_neon_vqneg_vqabs): Likewise.
3612 (cortex_a9_neon_vmov): Likewise.
3613 (cortex_a9_neon_vaba): Likewise.
3614 (cortex_a9_neon_vaba_qqq): Likewise.
3615 (cortex_a9_neon_shift_1): Likewise.
3616 (cortex_a9_neon_shift_2): Likewise.
3617 (cortex_a9_neon_shift_3): Likewise.
3618 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3619 (cortex_a9_neon_vsra_vrsra): Likewise.
3620 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3621 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3622 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3623 Likewise.
3624 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3625 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3626 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3627 Likewise.
3628 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3629 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3630 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3631 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3632 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3633 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3634 (cortex_a9_neon_fp_vsum): Likewise.
3635 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3636 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3637 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3638 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3639 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3640 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3641 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3642 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3643 (cortex_a9_neon_bp_simple): Likewise.
3644 (cortex_a9_neon_bp_2cycle): Likewise.
3645 (cortex_a9_neon_bp_3cycle): Likewise.
3646 (cortex_a9_neon_ldr): Likewise.
3647 (cortex_a9_neon_str): Likewise.
3648 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3649 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3650 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3651 (cortex_a9_neon_vld2_4_regs): Likewise.
3652 (cortex_a9_neon_vld3_vld4): Likewise.
3653 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3654 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3655 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3656 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3657 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3658 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3659 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3660 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3661 (cortex_a9_neon_mcr): Likewise.
3662 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3663 (cortex_a9_neon_mrc): Likewise.
3664 (cortex_a9_neon_mrrc): Likewise.
3665
3666 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3667
3668 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
3669
3670 (cortex_a8_neon_vshl_ddd): Remove.
3671 (cortex_a8_neon_vst3_vst4): Likewise.
3672 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3673
3674 (cortex_a8_neon_bit_ops_q): New.
3675
3676 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
3677 (cortex_a8_neon_int_2): Likewise..
3678 (cortex_a8_neon_int_3): Likewise.
3679 (cortex_a8_neon_int_5): Likewise.
3680 (cortex_a8_neon_vqneg_vqabs): Likewise.
3681 (cortex_a8_neon_int_4): Likewise.
3682 (cortex_a8_neon_vaba): Likewise.
3683 (cortex_a8_neon_vaba_qqq): Likewise.
3684 (cortex_a8_neon_shift_1): Likewise.
3685 (cortex_a8_neon_shift_2): Likewise.
3686 (cortex_a8_neon_shift_3): Likewise.
3687 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3688 (cortex_a8_neon_vsra_vrsra): Likewise.
3689 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3690 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3691 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3692 Likewise.
3693 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3694 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3695 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3696 Likewise.
3697 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3698 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3699 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3700 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3701 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3702 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3703 (cortex_a8_neon_fp_vsum): Likewise.
3704 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3705 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3706 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3707 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3708 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3709 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3710 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3711 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3712 (cortex_a8_neon_bp_simple): Likewise.
3713 (cortex_a8_neon_bp_2cycle): Likewise.
3714 (cortex_a8_neon_bp_3cycle): Likewise.
3715 (cortex_a8_neon_ldr): Likewise.
3716 (cortex_a8_neon_str): Likewise.
3717 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3718 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3719 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3720 (cortex_a8_neon_vld2_4_regs): Likewise.
3721 (cortex_a8_neon_vld3_vld4): Likewise.
3722 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3723 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3724 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3725 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3726 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3727 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3728 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3729 (cortex_a8_neon_mcr): Likewise.
3730 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3731 (cortex_a8_neon_mrc): Likewise.
3732 (cortex_a8_neon_mrrc): Likewise.
3733
3734 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3735
3736 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
3737 (fp): New.
3738 * config/aarch64/aarch64-simd.md (neon_type): Remove.
3739 (aarch64_simd_dup<mode>): Add "type" attribute.
3740 (aarch64_dup_lane<mode>): Likewise.
3741 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
3742 (*aarch64_simd_mov<mode>): Likewise.
3743 (aarch64_simd_mov_from_<mode>low): Likewise.
3744 (aarch64_simd_mov_from_<mode>high): Likewise.
3745 (orn<mode>3): Likewise.
3746 (bic<mode>3): Likewise.
3747 (add<mode>3): Likewise.
3748 (sub<mode>3): Likewise.
3749 (mul<mode>3): Likewise.
3750 (*aarch64_mul3_elt<mode>): Likewise.
3751 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
3752 (*aarch64_mul3_elt_to_128df): Likewise.
3753 (*aarch64_mul3_elt_to_64v2df): Likewise.
3754 (neg<mode>2): Likewise.
3755 (abs<mode>2): Likewise.
3756 (abd<mode>_3): Likewise.
3757 (aba<mode>_3): Likewise.
3758 (fabd<mode>_3): Likewise.
3759 (*fabd_scalar<mode>3): Likewise.
3760 (and<mode>3): Likewise.
3761 (ior<mode>3): Likewise.
3762 (xor<mode>3): Likewise.
3763 (one_cmpl<mode>2): Likewise.
3764 (aarch64_simd_vec_set<mode>): Likewise.
3765 (aarch64_simd_lshr<mode>): Likewise.
3766 (aarch64_simd_ashr<mode>): Likewise.
3767 (aarch64_simd_imm_shl<mode>): Likewise.
3768 (aarch64_simd_reg_sshl<mode): Likewise.
3769 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
3770 (aarch64_simd_reg_shl<mode>_signed): Likewise.
3771 (aarch64_simd_vec_setv2di): Likewise.
3772 (aarch64_simd_vec_set<mode>): Likewise.
3773 (aarch64_mla<mode>): Likewise.
3774 (*aarch64_mla_elt<mode>): Likewise.
3775 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
3776 (aarch64_mls<mode>): Likewise.
3777 (*aarch64_mls_elt<mode>): Likewise.
3778 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
3779 (<su><maxmin><mode>3): Likewise.
3780 (move_lo_quad_<mode>): Likewise.
3781 (aarch64_simd_move_hi_quad_<mode>): Likewise.
3782 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
3783 (vec_pack_trunc_<mode>): Likewise.
3784 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
3785 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
3786 (*aarch64_<su>mlal_lo<mode>): Likewise.
3787 (*aarch64_<su>mlal_hi<mode>): Likewise.
3788 (*aarch64_<su>mlsl_lo<mode>): Likewise.
3789 (*aarch64_<su>mlsl_hi<mode>): Likewise.
3790 (*aarch64_<su>mlal<mode>): Likewise.
3791 (*aarch64_<su>mlsl<mode>): Likewise.
3792 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
3793 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
3794 (add<mode>3): Likewise.
3795 (sub<mode>3): Likewise.
3796 (mul<mode>3): Likewise.
3797 (div<mode>3): Likewise.
3798 (neg<mode>2): Likewise.
3799 (abs<mode>2): Likewise.
3800 (fma<mode>4): Likewise.
3801 (*aarch64_fma4_elt<mode>): Likewise.
3802 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
3803 (*aarch64_fma4_elt_to_128df): Likewise.
3804 (*aarch64_fma4_elt_to_64v2df): Likewise.
3805 (fnma<mode>4): Likewise.
3806 (*aarch64_fnma4_elt<mode>): Likewise.
3807 (*aarch64_fnma4_elt_<vswap_width_name><mode>
3808 (*aarch64_fnma4_elt_to_128df): Likewise.
3809 (*aarch64_fnma4_elt_to_64v2df): Likewise.
3810 (<frint_pattern><mode>2): Likewise.
3811 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
3812 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
3813 (vec_unpacks_lo_v4sf): Likewise.
3814 (aarch64_float_extend_lo_v2df): Likewise.
3815 (vec_unpacks_hi_v4sf): Likewise.
3816 (aarch64_float_truncate_lo_v2sf): Likewise.
3817 (aarch64_float_truncate_hi_v4sf): Likewise.
3818 (aarch64_vmls<mode>): Likewise.
3819 (<su><maxmin><mode>3): Likewise.
3820 (<maxmin_uns><mode>3): Likewise.
3821 (reduc_<sur>plus_<mode>): Likewise.
3822 (reduc_<sur>plus_v2di): Likewise.
3823 (reduc_<sur>plus_v2si): Likewise.
3824 (reduc_<sur>plus_<mode>): Likewise.
3825 (aarch64_addpv4sf): Likewise.
3826 (clz<mode>2): Likewise.
3827 (reduc_<maxmin_uns>_<mode>): Likewise.
3828 (reduc_<maxmin_uns>_v2di): Likewise.
3829 (reduc_<maxmin_uns>_v2si): Likewise.
3830 (reduc_<maxmin_uns>_<mode>): Likewise.
3831 (reduc_<maxmin_uns>_v4sf): Likewise.
3832 (aarch64_simd_bsl<mode>_internal): Likewise.
3833 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
3834 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
3835 (aarch64_get_lane<mode>): Likewise.
3836 (*aarch64_combinez<mode>): Likewise.
3837 (aarch64_combine<mode>): Likewise.
3838 (aarch64_simd_combine<mode>): Likewise.
3839 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
3840 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
3841 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
3842 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
3843 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
3844 (aarch64_<sur>h<addsub><mode>): Likewise.
3845 (aarch64_<sur><addsub>hn<mode>): Likewise.
3846 (aarch64_<sur><addsub>hn2<mode>): Likewise.
3847 (aarch64_pmul<mode>): Likewise.
3848 (aarch64_<su_optab><optab><mode>): Likewise.
3849 (aarch64_<sur>qadd<mode>): Likewise.
3850 (aarch64_sqmovun<mode>): Likewise.
3851 (aarch64_<sur>qmovn<mode>): Likewise.
3852 (aarch64_s<optab><mode>): Likewise.
3853 (aarch64_sq<r>dmulh<mode>): Likewise.
3854 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3855 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
3856 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3857 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
3858 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3859 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3860 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
3861 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
3862 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
3863 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
3864 (aarch64_sqdmull<mode>): Likewise.
3865 (aarch64_sqdmull_lane<mode>_internal): Likewise.
3866 (aarch64_sqdmull_n<mode>): Likewise.
3867 (aarch64_sqdmull2<mode>_internal): Likewise.
3868 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
3869 (aarch64_sqdmull2_n<mode>_internal): Likewise.
3870 (aarch64_<sur>shl<mode>): Likewise.
3871 (aarch64_<sur>q<r>shl<mode>
3872 (aarch64_<sur>shll_n<mode>): Likewise.
3873 (aarch64_<sur>shll2_n<mode>): Likewise.
3874 (aarch64_<sur>shr_n<mode>): Likewise.
3875 (aarch64_<sur>sra_n<mode>): Likewise.
3876 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
3877 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
3878 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
3879 (aarch64_cm<optab><mode>): Likewise.
3880 (aarch64_cm<optab>di): Likewise.
3881 (aarch64_cm<optab><mode>): Likewise.
3882 (aarch64_cm<optab>di): Likewise.
3883 (aarch64_cmtst<mode>): Likewise.
3884 (aarch64_cmtstdi): Likewise.
3885 (aarch64_cm<optab><mode>): Likewise.
3886 (*aarch64_fac<optab><mode>): Likewise.
3887 (aarch64_addp<mode>): Likewise.
3888 (aarch64_addpdi): Likewise.
3889 (sqrt<mode>2): Likewise.
3890 (vec_load_lanesoi<mode>): Likewise.
3891 (vec_store_lanesoi<mode>): Likewise.
3892 (vec_load_lanesci<mode>): Likewise.
3893 (vec_store_lanesci<mode>): Likewise.
3894 (vec_load_lanesxi<mode>): Likewise.
3895 (vec_store_lanesxi<mode>): Likewise.
3896 (*aarch64_mov<mode>): Likewise.
3897 (aarch64_ld2<mode>_dreg): Likewise.
3898 (aarch64_ld2<mode>_dreg): Likewise.
3899 (aarch64_ld3<mode>_dreg): Likewise.
3900 (aarch64_ld3<mode>_dreg): Likewise.
3901 (aarch64_ld4<mode>_dreg): Likewise.
3902 (aarch64_ld4<mode>_dreg): Likewise.
3903 (aarch64_tbl1<mode>): Likewise.
3904 (aarch64_tbl2v16qi): Likewise.
3905 (aarch64_combinev16qi): Likewise.
3906 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
3907 (aarch64_st2<mode>_dreg): Likewise.
3908 (aarch64_st2<mode>_dreg): Likewise.
3909 (aarch64_st3<mode>_dreg): Likewise.
3910 (aarch64_st3<mode>_dreg): Likewise.
3911 (aarch64_st4<mode>_dreg): Likewise.
3912 (aarch64_st4<mode>_dreg): Likewise.
3913 (*aarch64_simd_ld1r<mode>): Likewise.
3914 (aarch64_frecpe<mode>): Likewise.
3915 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
3916 (aarch64_frecps<mode>): Likewise.
3917
3918 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3919
3920 * config/arm/iterators.md (V_elem_ch): New.
3921 (q): Likewise.
3922 (VQH_type): Likewise.
3923 * config/arm/arm.md (is_neon_type): New.
3924 (conds): Use is_neon_type.
3925 (anddi3_insn): Update type attribute.
3926 (xordi3_insn): Likewise.
3927 (one_cmpldi2): Likewise.
3928 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
3929 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
3930 (*movmisalign<mode>_neon_store): Likewise.
3931 (*movmisalign<mode>_neon_load): Likewise.
3932 (vec_set<mode>_internal): Likewise.
3933 (vec_set<mode>_internal): Likewise.
3934 (vec_setv2di_internal): Likewise.
3935 (vec_extract<mode>): Likewise.
3936 (vec_extract<mode>): Likewise.
3937 (vec_extractv2di): Likewise.
3938 (*add<mode>3_neon): Likewise.
3939 (adddi3_neon): Likewise.
3940 (*sub<mode>3_neon): Likewise.
3941 (subdi3_neon): Likewise.
3942 (fma<VCVTF:mode>4): Likewise.
3943 (fma<VCVTF:mode>4_intrinsic): Likewise.
3944 (*fmsub<VCVTF:mode>4): Likewise.
3945 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3946 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3947 (ior<mode>3): Likewise.
3948 (and<mode>3): Likewise.
3949 (orn<mode>3_neon): Likewise.
3950 (orndi3_neon): Likewise.
3951 (bic<mode>3_neon): Likewise.
3952 (bicdi3_neon): Likewise.
3953 (xor<mode>3): Likewise.
3954 (one_cmpl<mode>2): Likewise.
3955 (abs<mode>2): Likewise.
3956 (neg<mode>2): Likewise.
3957 (negdi2_neon): Likewise.
3958 (*umin<mode>3_neon): Likewise.
3959 (*umax<mode>3_neon): Likewise.
3960 (*smin<mode>3_neon): Likewise.
3961 (*smax<mode>3_neon): Likewise.
3962 (vashl<mode>3): Likewise.
3963 (vashr<mode>3_imm): Likewise.
3964 (vlshr<mode>3_imm): Likewise.
3965 (ashl<mode>3_signed): Likewise.
3966 (ashl<mode>3_unsigned): Likewise.
3967 (neon_load_count): Likewise.
3968 (ashldi3_neon_noclobber): Likewise.
3969 (ashldi3_neon): Likewise.
3970 (signed_shift_di3_neon): Likewise.
3971 (unsigned_shift_di3_neon): Likewise.
3972 (ashrdi3_neon_imm_noclobber): Likewise.
3973 (lshrdi3_neon_imm_noclobber): Likewise.
3974 (<shift>di3_neon): Likewise.
3975 (widen_ssum<mode>3): Likewise.
3976 (widen_usum<mode>3): Likewise.
3977 (quad_halves_<code>v4si): Likewise.
3978 (quad_halves_<code>v4sf): Likewise.
3979 (quad_halves_<code>v8hi): Likewise.
3980 (quad_halves_<code>v16qi): Likewise.
3981 (reduc_splus_v2di): Likewise.
3982 (neon_vpadd_internal<mode>): Likewise.
3983 (neon_vpsmin<mode>): Likewise.
3984 (neon_vpsmax<mode>): Likewise.
3985 (neon_vpumin<mode>): Likewise.
3986 (neon_vpumax<mode>): Likewise.
3987 (*ss_add<mode>_neon): Likewise.
3988 (*us_add<mode>_neon): Likewise.
3989 (*ss_sub<mode>_neon): Likewise.
3990 (*us_sub<mode>_neon): Likewise.
3991 (neon_vadd<mode>_unspec): Likewise.
3992 (neon_vaddl<mode>): Likewise.
3993 (neon_vaddw<mode>): Likewise.
3994 (neon_vhadd<mode>): Likewise.
3995 (neon_vqadd<mode>): Likewise.
3996 (neon_vaddhn<mode>): Likewise.
3997 (neon_vmul<mode>): Likewise.
3998 (neon_vfms<VCVTF:mode>): Likewise.
3999 (neon_vmlal<mode>): Likewise.
4000 (neon_vmls<mode>): Likewise.
4001 (neon_vmlsl<mode>): Likewise.
4002 (neon_vqdmulh<mode>): Likewise.
4003 (neon_vqdmlal<mode>): Likewise.
4004 (neon_vqdmlsl<mode>): Likewise.
4005 (neon_vmull<mode>): Likewise.
4006 (neon_vqdmull<mode>): Likewise.
4007 (neon_vsub<mode>_unspec): Likewise.
4008 (neon_vsubl<mode>): Likewise.
4009 (neon_vsubw<mode>): Likewise.
4010 (neon_vqsub<mode>): Likewise.
4011 (neon_vhsub<mode>): Likewise.
4012 (neon_vsubhn<mode>): Likewise.
4013 (neon_vceq<mode>): Likewise.
4014 (neon_vcge<mode>): Likewise.
4015 (neon_vcgeu<mode>): Likewise.
4016 (neon_vcgt<mode>): Likewise.
4017 (neon_vcgtu<mode>): Likewise.
4018 (neon_vcle<mode>): Likewise.
4019 (neon_vclt<mode>): Likewise.
4020 (neon_vcage<mode>): Likewise.
4021 (neon_vcagt<mode>): Likewise.
4022 (neon_vtst<mode>): Likewise.
4023 (neon_vabd<mode>): Likewise.
4024 (neon_vabdl<mode>): Likewise.
4025 (neon_vaba<mode>): Likewise.
4026 (neon_vabal<mode>): Likewise.
4027 (neon_vmax<mode>): Likewise.
4028 (neon_vmin<mode>): Likewise.
4029 (neon_vpaddl<mode>): Likewise.
4030 (neon_vpadal<mode>): Likewise.
4031 (neon_vpmax<mode>): Likewise.
4032 (neon_vpmin<mode>): Likewise.
4033 (neon_vrecps<mode>): Likewise.
4034 (neon_vrsqrts<mode>): Likewise.
4035 (neon_vqabs<mode>): Likewise.
4036 (neon_vqneg<mode>): Likewise.
4037 (neon_vcls<mode>): Likewise.
4038 (clz<mode>2): Likewise.
4039 (popcount<mode>2): Likewise.
4040 (neon_vrecpe<mode>): Likewise.
4041 (neon_vrsqrte<mode>): Likewise.
4042 (neon_vget_lane<mode>_sext_internal): Likewise.
4043 (neon_vget_lane<mode>_zext_internal): Likewise.
4044 (neon_vdup_n<mode>): Likewise.
4045 (neon_vdup_n<mode>): Likewise.
4046 (neon_vdup_nv2di): Likewise.
4047 (neon_vdup_lane<mode>_interal): Likewise.
4048 (*neon_vswp<mode>): Likewise.
4049 (neon_vcombine<mode>): Likewise.
4050 (float<mode><V_cvtto>2): Likewise.
4051 (floatuns<mode><V_cvtto>2): Likewise.
4052 (fix_trunc<mode><V_cvtto>2): Likewise.
4053 (fixuns_trunc<mode><V_cvtto>2
4054 (neon_vcvt<mode>): Likewise.
4055 (neon_vcvt<mode>): Likewise.
4056 (neon_vcvtv4sfv4hf): Likewise.
4057 (neon_vcvtv4hfv4sf): Likewise.
4058 (neon_vcvt_n<mode>): Likewise.
4059 (neon_vcvt_n<mode>): Likewise.
4060 (neon_vmovn<mode>): Likewise.
4061 (neon_vqmovn<mode>): Likewise.
4062 (neon_vqmovun<mode>): Likewise.
4063 (neon_vmovl<mode>): Likewise.
4064 (neon_vmul_lane<mode>): Likewise.
4065 (neon_vmul_lane<mode>): Likewise.
4066 (neon_vmull_lane<mode>): Likewise.
4067 (neon_vqdmull_lane<mode>): Likewise.
4068 (neon_vqdmulh_lane<mode>): Likewise.
4069 (neon_vqdmulh_lane<mode>): Likewise.
4070 (neon_vmla_lane<mode>): Likewise.
4071 (neon_vmla_lane<mode>): Likewise.
4072 (neon_vmlal_lane<mode>): Likewise.
4073 (neon_vqdmlal_lane<mode>): Likewise.
4074 (neon_vmls_lane<mode>): Likewise.
4075 (neon_vmls_lane<mode>): Likewise.
4076 (neon_vmlsl_lane<mode>): Likewise.
4077 (neon_vqdmlsl_lane<mode>): Likewise.
4078 (neon_vext<mode>): Likewise.
4079 (neon_vrev64<mode>): Likewise.
4080 (neon_vrev32<mode>): Likewise.
4081 (neon_vrev16<mode>): Likewise.
4082 (neon_vbsl<mode>_internal): Likewise.
4083 (neon_vshl<mode>): Likewise.
4084 (neon_vqshl<mode>): Likewise.
4085 (neon_vshr_n<mode>): Likewise.
4086 (neon_vshrn_n<mode>): Likewise.
4087 (neon_vqshrn_n<mode>): Likewise.
4088 (neon_vqshrun_n<mode>): Likewise.
4089 (neon_vshl_n<mode>): Likewise.
4090 (neon_vqshl_n<mode>): Likewise.
4091 (neon_vqshlu_n<mode>): Likewise.
4092 (neon_vshll_n<mode>): Likewise.
4093 (neon_vsra_n<mode>): Likewise.
4094 (neon_vsri_n<mode>): Likewise.
4095 (neon_vsli_n<mode>): Likewise.
4096 (neon_vtbl1v8qi): Likewise.
4097 (neon_vtbl2v8qi): Likewise.
4098 (neon_vtbl3v8qi): Likewise.
4099 (neon_vtbl4v8qi): Likewise.
4100 (neon_vtbl1v16qi): Likewise.
4101 (neon_vtbl2v16qi): Likewise.
4102 (neon_vcombinev16qi): Likewise.
4103 (neon_vtbx1v8qi): Likewise.
4104 (neon_vtbx2v8qi): Likewise.
4105 (neon_vtbx3v8qi): Likewise.
4106 (neon_vtbx4v8qi): Likewise.
4107 (*neon_vtrn<mode>_insn): Likewise.
4108 (*neon_vzip<mode>_insn): Likewise.
4109 (*neon_vuzp<mode>_insn): Likewise.
4110 (neon_vld1<mode>): Likewise.
4111 (neon_vld1_lane<mode>): Likewise.
4112 (neon_vld1_lane<mode>): Likewise.
4113 (neon_vld1_dup<mode>): Likewise.
4114 (neon_vld1_dup<mode>): Likewise.
4115 (neon_vld1_dupv2di): Likewise.
4116 (neon_vst1<mode>): Likewise.
4117 (neon_vst1_lane<mode>): Likewise.
4118 (neon_vst1_lane<mode>): Likewise.
4119 (neon_vld2<mode>): Likewise.
4120 (neon_vld2<mode>): Likewise.
4121 (neon_vld2_lane<mode>): Likewise.
4122 (neon_vld2_lane<mode>): Likewise.
4123 (neon_vld2_dup<mode>): Likewise.
4124 (neon_vst2<mode>): Likewise.
4125 (neon_vst2<mode>): Likewise.
4126 (neon_vst2_lane<mode>): Likewise.
4127 (neon_vst2_lane<mode>): Likewise.
4128 (neon_vld3<mode>): Likewise.
4129 (neon_vld3qa<mode>): Likewise.
4130 (neon_vld3qb<mode>): Likewise.
4131 (neon_vld3_lane<mode>): Likewise.
4132 (neon_vld3_lane<mode>): Likewise.
4133 (neon_vld3_dup<mode>): Likewise.
4134 (neon_vst3<mode>): Likewise.
4135 (neon_vst3qa<mode>): Likewise.
4136 (neon_vst3qb<mode>): Likewise.
4137 (neon_vst3_lane<mode>): Likewise.
4138 (neon_vst3_lane<mode>): Likewise.
4139 (neon_vld4<mode>): Likewise.
4140 (neon_vld4qa<mode>): Likewise.
4141 (neon_vld4qb<mode>): Likewise.
4142 (neon_vld4_lane<mode>): Likewise.
4143 (neon_vld4_lane<mode>): Likewise.
4144 (neon_vld4_dup<mode>): Likewise.
4145 (neon_vst4<mode>): Likewise.
4146 (neon_vst4qa<mode>): Likewise.
4147 (neon_vst4qb<mode>): Likewise.
4148 (neon_vst4_lane<mode>): Likewise.
4149 (neon_vst4_lane<mode>): Likewise.
4150 (neon_vec_unpack<US>_lo_<mode>): Likewise.
4151 (neon_vec_unpack<US>_hi_<mode>): Likewise.
4152 (neon_vec_<US>mult_lo_<mode>): Likewise.
4153 (neon_vec_<US>mult_hi_<mode>): Likewise.
4154 (neon_vec_<US>shiftl_<mode>): Likewise.
4155 (neon_unpack<US>_<mode>): Likewise.
4156 (neon_vec_<US>mult_<mode>): Likewise.
4157 (vec_pack_trunc_<mode>): Likewise.
4158 (neon_vec_pack_trunc_<mode>): Likewise.
4159 (neon_vabd<mode>_2): Likewise.
4160 (neon_vabd<mode>_3): Likewise.
4161
4162 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4163
4164 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
4165 (load_pair): Update type attribute.
4166 (store_pair): Update type attribute.
4167 * config/aarch64/iterators.md (q): New.
4168
4169 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4170
4171 * config/arm/types.md: Add new types for Neon insns.
4172
4173 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4174 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4175 Sergey Lega <sergey.s.lega@intel.com>
4176 Anna Tikhonova <anna.tikhonova@intel.com>
4177 Ilya Tocar <ilya.tocar@intel.com>
4178 Andrey Turetskiy <andrey.turetskiy@intel.com>
4179 Ilya Verbin <ilya.verbin@intel.com>
4180 Kirill Yukhin <kirill.yukhin@intel.com>
4181 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4182
4183 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
4184 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
4185 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
4186 (rcp14<mode>): New.
4187 (srcp14<mode>): Ditto.
4188 (rsqrt14<mode>): Ditto.
4189 (rsqrt14<mode>): Ditto.
4190 (avx512f_vmscalef<mode>): Ditto.
4191 (avx512f_scalef<mode>): Ditto.
4192 (avx512f_getexp<mode>): Ditto.
4193 (avx512f_sgetexp<mode>): Ditto.
4194 (avx512f_fixupimm<mode>): Ditto.
4195 (avx512f_sfixupimm<mode>): Ditto.
4196 (avx512f_rndscale<mode>): Ditto.
4197 (*avx512er_exp2<mode>): Ditto.
4198 (*avx512er_rcp28<mode>): Ditto.
4199 (avx512er_rsqrt28<mode>): Ditto.
4200 (avx512f_getmant<mode>): Ditto.
4201 (avx512f_getmant<mode>): Ditto.
4202 (avx512f_rndscale<mode>): Fix formatting.
4203
4204 2013-10-15 Martin Jambor <mjambor@suse.cz>
4205
4206 * ipa-utils.h (ipa_edge_within_scc): Declare.
4207 * ipa-cp.c (edge_within_scc): Moved...
4208 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
4209
4210 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4211 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4212 Sergey Lega <sergey.s.lega@intel.com>
4213 Anna Tikhonova <anna.tikhonova@intel.com>
4214 Ilya Tocar <ilya.tocar@intel.com>
4215 Andrey Turetskiy <andrey.turetskiy@intel.com>
4216 Ilya Verbin <ilya.verbin@intel.com>
4217 Kirill Yukhin <kirill.yukhin@intel.com>
4218 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4219
4220 * config/i386/predicates.md (const_8_to_15_operand): New.
4221 (const_16_to_31_operand): Ditto.
4222 * config/i386/sse.md (V8FI): New.
4223 (V16FI): Ditto.
4224 (reduc_splus_v8df): Ditto.
4225 (reduc_splus_v16sf): Ditto.
4226 (avx512f_vextract<shuffletype>32x4_1): Ditto.
4227 (vec_extract_hi_<mode>): Ditto.
4228 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
4229 (vec_set_lo_<mode>): Ditto.
4230 (vec_set_hi_<mode>): Ditto.
4231 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
4232 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
4233 (avx512f_pshufd_1): Ditto.
4234 (avx512f_broadcast<mode>): Ditto.
4235 (avx512f_broadcast<mode>): Ditto.
4236 (define_split): Split vec_extract_lo into move.
4237 (ssequartermode): Ditto.
4238 (ssedoublemode): Extened with wider modes.
4239 (vec_extract_lo_<mode>): Ditto.
4240
4241 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4242 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4243 Sergey Lega <sergey.s.lega@intel.com>
4244 Anna Tikhonova <anna.tikhonova@intel.com>
4245 Ilya Tocar <ilya.tocar@intel.com>
4246 Andrey Turetskiy <andrey.turetskiy@intel.com>
4247 Ilya Verbin <ilya.verbin@intel.com>
4248 Kirill Yukhin <kirill.yukhin@intel.com>
4249 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4250
4251 * config/i386/predicates.md (register_or_constm1_operand): New.
4252 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
4253 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
4254 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
4255 UNSPEC_SCATTER_PREFETCH
4256 (VI48_512): New.
4257 (avx512f_ucmp<mode>3): Ditto.
4258 (avx512f_vternlog<mode>): Ditto.
4259 (avx512f_align<mode>): Ditto.
4260 (<shift_insn><mode>3): Ditto.
4261 (avx512f_<rotate>v<mode>): Ditto.
4262 (avx512f_<rotate><mode>): Ditto.
4263 (avx512f_eq<mode>3): Ditto.
4264 (avx512f_eq<mode>3_1): Ditto.
4265 (avx512f_gt<mode>3): Ditto.
4266 (avx512f_testm<mode>3): Ditto.
4267 (avx512f_testnm<mode>3): Ditto.
4268 (avx512pf_gatherpf<mode>): Ditto.
4269 (*avx512pf_gatherpf<mode>_mask): Ditto.
4270 (*avx512pf_gatherpf<mode>): Ditto.
4271 (avx512pf_scatterpf<mode>): Ditto.
4272 (*avx512pf_scatterpf<mode>_mask): Ditto.
4273 (*avx512pf_scatterpf<mode>): Ditto.
4274 (avx512f_vec_dup_gpr<mode>): Ditto.
4275 (clz<mode>2): Ditto.
4276 (conflict<mode>): Ditto.
4277 (REDUC_SMINMAX_MODE): Extened with wider modes.
4278 (reduc_<code>_<mode>): Ditto.
4279 (vlshr<mode>3): Ditto.
4280 (vashl<mode>3): Ditto.
4281
4282 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4283 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4284 Sergey Lega <sergey.s.lega@intel.com>
4285 Anna Tikhonova <anna.tikhonova@intel.com>
4286 Ilya Tocar <ilya.tocar@intel.com>
4287 Andrey Turetskiy <andrey.turetskiy@intel.com>
4288 Ilya Verbin <ilya.verbin@intel.com>
4289 Kirill Yukhin <kirill.yukhin@intel.com>
4290 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4291
4292 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
4293 UNSPEC_SCATTER.
4294 (VI48F_512): New.
4295 (avx512fmaskmode): Ditto.
4296 (bcstscalarsuff): Ditto.
4297 (avx512f_blendm<mode>): Ditto.
4298 (cmp_imm_predicate): Ditto.
4299 (avx512f_cmp<mode>3): Ditto.
4300 (avx512f_vec_dup<mode>): Ditto.
4301 (avx512f_vec_dup_mem<mode>): Ditto.
4302 (avx512f_vpermi2var<mode>3): Ditto.
4303 (avx512f_vpermt2var<mode>3): Ditto.
4304 (vec_init<mode>): Ditto.
4305 (avx512f_gathersi<mode>): Ditto.
4306 (*avx512f_gathersi<mode>): Ditto.
4307 (*avx512f_gathersi<mode>_2): Ditto.
4308 (avx512f_gatherdi<mode>): Ditto.
4309 (*avx512f_gatherdi<mode>): Ditto.
4310 (*avx512f_gatherdi<mode>_2): Ditto.
4311 (avx512f_scattersi<mode>): Ditto.
4312 (*avx512f_scattersi<mode>): Ditto.
4313 (avx512f_scatterdi<mode>): Ditto.
4314 (*avx512f_scatterdi<mode>): Ditto.
4315 (sseintprefix): Extened with wider modes.
4316 (VEC_GATHER_IDXSI): Ditto.
4317 (VEC_GATHER_IDXDI): Ditto.
4318 (VEC_GATHER_SRCDI): Ditto.
4319
4320 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
4321 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4322
4323 * config/arm/t-aprofile: New file.
4324 * config.gcc: Handle --with-multilib-list option.
4325
4326 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
4327
4328 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
4329 the input for r2.
4330
4331 2013-10-15 Richard Biener <rguenther@suse.de>
4332
4333 * tree-loop-distribution.c (build_empty_rdg): Inline into
4334 single user.
4335 (rdg_flag_vertex): Inline into single user.
4336 (rdg_flag_vertex_and_dependent): Likewise.
4337 (build_rdg_partition_for_vertex): Remove processed bitmap.
4338 (rdg_build_partitions): Simplify.
4339
4340 2013-10-15 Richard Biener <rguenther@suse.de>
4341
4342 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
4343 Restructure forwarding through conversions and copies to
4344 avoid performing copy-propagation the wrong way. Adjust
4345 recursion invocations.
4346 (forward_propagate_addr_expr): Add argument stating if we
4347 are recursing from a single-use.
4348 (ssa_forward_propagate_and_combine): Adjust.
4349
4350 2013-10-14 David Malcolm <dmalcolm@redhat.com>
4351
4352 * dumpfile.h (gcc::dump_manager): New class, to hold state
4353 relating to dumpfile management.
4354 (get_dump_file_name): Remove in favor of method of dump_manager.
4355 (dump_initialized_p): Likewise.
4356 (dump_start): Likewise.
4357 (dump_finish): Likewise.
4358 (dump_switch_p): Likewise.
4359 (dump_register): Likewise.
4360 (get_dump_file_info): Likewise.
4361 * context.c (gcc::context::context): Construct the dump_manager
4362 instance.
4363 * context.h (gcc::context::get_dumps): New.
4364 (gcc::context::m_dumps): New.
4365 * coverage.c (coverage_init): Port to dump_manager API.
4366 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
4367 (extra_dump_files_in_use): Likewise.
4368 (extra_dump_files_alloced): Likewise.
4369 (gcc::dump_manager::dump_manager): New.
4370 (dump_register): Convert to...
4371 (gcc::dump_manager::dump_register): ...method, replacing
4372 function-static next_dump with m_next_dump field.
4373 (get_dump_file_info): Convert to...
4374 (gcc::dump_manager::get_dump_file_info): ...method.
4375 (get_dump_file_name): Convert to...
4376 (gcc::dump_manager::get_dump_file_name): ...method.
4377 (dump_start): Convert to...
4378 (gcc::dump_manager::dump_start): ...method.
4379 (dump_finish): Convert to...
4380 (gcc::dump_manager::dump_finish): ...method.
4381 (dump_begin): Replace body with...
4382 (gcc::dump_manager::dump_begin): ...new method.
4383 (dump_phase_enabled_p): Convert to...
4384 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4385 (dump_phase_enabled_p): Convert to...
4386 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4387 (dump_initialized_p): Convert to...
4388 (gcc::dump_manager::dump_initialized_p): ...method.
4389 (dump_flag_name): Replace body with...
4390 (gcc::dump_manager::dump_flag_name): ...new method.
4391 (dump_enable_all): Convert to...
4392 (gcc::dump_manager::dump_enable_all): ...new method.
4393 (opt_info_enable_passes): Convert to...
4394 (gcc::dump_manager::opt_info_enable_passes): ...new method.
4395 (dump_switch_p_1): Convert to...
4396 (gcc::dump_manager::dump_switch_p_1): ...new method.
4397 (dump_switch_p): Convert to...
4398 (gcc::dump_manager::dump_switch_p): ...new method.
4399 (opt_info_switch_p): Port to dump_manager API.
4400 (enable_rtl_dump_file): Likewise.
4401 * opts-global.c (handle_common_deferred_options): Port to new
4402 dump_manager API.
4403 * passes.c (pass_manager::finish_optimization_passes): Likewise.
4404 (pass_manager::register_one_dump_file): Likewise.
4405 (pass_manager::register_pass): Likewise.
4406 (pass_init_dump_file): Likewise.
4407 (pass_fini_dump_file): Likewise.
4408 * statistics.c (statistics_early_init): Likewise.
4409
4410 2013-10-14 Richard Biener <rguenther@suse.de>
4411
4412 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
4413 iterative_hash_canonical_type, gimple_canonical_type_hash,
4414 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
4415 gimple_register_canonical_type, print_gimple_types_stats,
4416 free_gimple_type_tables): Move to lto/lto.c
4417 (gt-gimple.h): Do not include.
4418 * gimple.h (gimple_register_canonical_type,
4419 print_gimple_types_stats, free_gimple_type_tables): Remove.
4420 * Makefile.in (GTFILES): Remove gimple.c.
4421
4422 2013-10-14 Travis Snoozy <quandary@remstate.com>
4423
4424 PR target/58716
4425 * config/msp430/msp430.c (msp430_option_override): Correct thinko
4426 scanning for msp430x targets.
4427
4428 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
4429
4430 PR bootstrap/58509
4431 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
4432 (registers_ok_for_ldd_peep): Move around.
4433 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
4434 * config/sparc/sparc.md (widening peepholes): Use it.
4435
4436 2013-10-14 Richard Biener <rguenther@suse.de>
4437
4438 PR middle-end/58712
4439 PR middle-end/55358
4440 * gimple.c (iterative_hash_canonical_type): Make sure to
4441 record the hash into the correct hashtable slot.
4442
4443 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
4444
4445 PR rtl-optimization/58662
4446 * combine.c (try_combine): Take into account death nodes on I2 when
4447 splitting a PARALLEL of two independent SETs. Fix dump message.
4448
4449 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
4450
4451 PR target/51244
4452 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
4453 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
4454 * config/sh/t-sh (sh_treg_combine.o): New entry.
4455 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
4456 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
4457 (register_sh_passes): New function. Register sh_treg_combine pass.
4458 (sh_option_override): Invoke it.
4459 (sh_canonicalize_comparison): Handle op0_preserve_value.
4460 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
4461 opportunities. Canonicalize branch condition.
4462 (nott): Allow only if pseudos can be created for non-SH2A.
4463
4464 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
4465
4466 PR target/58690
4467 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
4468 (ix86_expand_movmem): Replace copy_addr_to_reg with
4469 ix86_copy_addr_to_reg.
4470 (ix86_expand_setmem): Likewise.
4471
4472 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
4473
4474 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
4475 provided by ix86_fp_compare_mode instead of CCFPUmode.
4476
4477 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
4478
4479 * config/aarch64/arm_neon.h
4480 (vtbx<1,3>_<psu>8): Fix register constriants.
4481
4482 2013-10-11 Jeff Law <law@redhat.com>
4483
4484 PR tree-optimization/58640
4485 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
4486 threading paths that cross over two loop entry points.
4487
4488 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4489
4490 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
4491 handle vector float as well.
4492 (*vsx_le_perm_load_v4si): Likewise.
4493 (*vsx_le_perm_store_v2di): Likewise.
4494 (*vsx_le_perm_store_v4si): Likewise.
4495
4496 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4497
4498 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
4499 directly to circumvent subtract from splat{31} workaround.
4500 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
4501 prototype.
4502 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
4503 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
4504 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
4505 (altivec_vperm_<mode>): Convert to define_insn_and_split to
4506 separate big and little endian logic.
4507 (*altivec_vperm_<mode>_internal): New define_insn.
4508 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
4509 separate big and little endian logic.
4510 (*altivec_vperm_<mode>_uns_internal): New define_insn.
4511 (vec_permv16qi): Add little endian logic.
4512
4513 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
4514
4515 * doc/extend.texi (returns_nonnull): Remove arguments.
4516
4517 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4518 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4519 Sergey Lega <sergey.s.lega@intel.com>
4520 Anna Tikhonova <anna.tikhonova@intel.com>
4521 Ilya Tocar <ilya.tocar@intel.com>
4522 Andrey Turetskiy <andrey.turetskiy@intel.com>
4523 Ilya Verbin <ilya.verbin@intel.com>
4524 Kirill Yukhin <kirill.yukhin@intel.com>
4525 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4526
4527 * config/i386/sse.md (VI48F_256_512): New.
4528 (avx2_permvar<mode>): Change to ...
4529 (<avx2_avx512f>_permvar<mode>): This.
4530
4531 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4532 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4533 Sergey Lega <sergey.s.lega@intel.com>
4534 Anna Tikhonova <anna.tikhonova@intel.com>
4535 Ilya Tocar <ilya.tocar@intel.com>
4536 Andrey Turetskiy <andrey.turetskiy@intel.com>
4537 Ilya Verbin <ilya.verbin@intel.com>
4538 Kirill Yukhin <kirill.yukhin@intel.com>
4539 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4540
4541 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
4542 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
4543 * config/i386/sse.md (VI4_AVX): New.
4544 (sf2simodelower): Ditto.
4545 (sse2_cvtps2dq): Change to ...
4546 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
4547
4548 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4549 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4550 Sergey Lega <sergey.s.lega@intel.com>
4551 Anna Tikhonova <anna.tikhonova@intel.com>
4552 Ilya Tocar <ilya.tocar@intel.com>
4553 Andrey Turetskiy <andrey.turetskiy@intel.com>
4554 Ilya Verbin <ilya.verbin@intel.com>
4555 Kirill Yukhin <kirill.yukhin@intel.com>
4556 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4557
4558 * config/i386/sse.md (V_512): New.
4559 (VI_512): Ditto.
4560 (vcond<V_512:mode><VF_512:mode>): Ditto.
4561 (vcond<V_512:mode><VI_512:mode>): Ditto.
4562 (vcondu<V_512:mode><VI_512:mode>): Ditto.
4563
4564 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4565 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4566 Sergey Lega <sergey.s.lega@intel.com>
4567 Anna Tikhonova <anna.tikhonova@intel.com>
4568 Ilya Tocar <ilya.tocar@intel.com>
4569 Andrey Turetskiy <andrey.turetskiy@intel.com>
4570 Ilya Verbin <ilya.verbin@intel.com>
4571 Kirill Yukhin <kirill.yukhin@intel.com>
4572 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4573
4574 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
4575 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
4576 (FMAMODE): Ditto.
4577 (fma<mode>4): Removed condition.
4578 (fms<mode>4): Ditto.
4579 (fnma<mode>4): Ditto.
4580 (fnms<mode>4): Ditto.
4581 (fma4i_fmadd_<mode>): Ditto.
4582 (*fma_fmadd_<mode>): Ditto.
4583 (*fma_fmsub_<mode>): Ditto.
4584 (*fma_fnmadd_<mode>): Ditto.
4585 (*fma_fnmsub_<mode>): Ditto.
4586 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
4587 (*fma_fmaddsub_<mode>): Ditto.
4588 (*fma_fmsubadd_<mode>): Ditto.
4589 (*fmai_fmadd_<mode>): Ditto.
4590 (*fmai_fmsub_<mode>): Ditto.
4591 (*fmai_fnmadd_<mode>): Ditto.
4592 (*fmai_fnmsub_<mode>): Ditto.
4593
4594 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4595 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4596 Sergey Lega <sergey.s.lega@intel.com>
4597 Anna Tikhonova <anna.tikhonova@intel.com>
4598 Ilya Tocar <ilya.tocar@intel.com>
4599 Andrey Turetskiy <andrey.turetskiy@intel.com>
4600 Ilya Verbin <ilya.verbin@intel.com>
4601 Kirill Yukhin <kirill.yukhin@intel.com>
4602 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4603
4604 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
4605 (VI124_256): Changed to ...
4606 (VI124_256_48_512): This.
4607 (ssepackmode): Extended with wider modes.
4608 (<code><mode>3): Changed iterator.
4609 (*avx2_<code><mode>3): Ditto.
4610 (vec_pack_trunc_<mode>): Ditto.
4611
4612 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4613 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4614 Sergey Lega <sergey.s.lega@intel.com>
4615 Anna Tikhonova <anna.tikhonova@intel.com>
4616 Ilya Tocar <ilya.tocar@intel.com>
4617 Andrey Turetskiy <andrey.turetskiy@intel.com>
4618 Ilya Verbin <ilya.verbin@intel.com>
4619 Kirill Yukhin <kirill.yukhin@intel.com>
4620 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4621
4622 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
4623 (VI8F_256_512): Ditto.
4624 (abs<mode>2): Changed iterator.
4625 (avx2_perm<mode>): Changed to ...
4626 (<avx2_avx512f>_perm<mode>): This.
4627 (avx2_perm<mode>_1): Changed to ...
4628 (<avx2_avx512f>_perm<mode>_1): This.
4629
4630 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4631 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4632 Sergey Lega <sergey.s.lega@intel.com>
4633 Anna Tikhonova <anna.tikhonova@intel.com>
4634 Ilya Tocar <ilya.tocar@intel.com>
4635 Andrey Turetskiy <andrey.turetskiy@intel.com>
4636 Ilya Verbin <ilya.verbin@intel.com>
4637 Kirill Yukhin <kirill.yukhin@intel.com>
4638 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4639
4640 * config/i386/sse.md (VI48_AVX512F): New.
4641 (VI48_AVX2): Changed to ...
4642 (VI48_AVX2_48_AVX512F): This.
4643 (avx2_ashrv<mode>): Changed to ...
4644 (<avx2_avx512f>_ashrv<mode>): This.
4645 (avx2_<shift_insn>v<mode>): Changed to ...
4646 (<avx2_avx512f>_<shift_insn>v<mode>): This.
4647
4648 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4649 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4650 Sergey Lega <sergey.s.lega@intel.com>
4651 Anna Tikhonova <anna.tikhonova@intel.com>
4652 Ilya Tocar <ilya.tocar@intel.com>
4653 Andrey Turetskiy <andrey.turetskiy@intel.com>
4654 Ilya Verbin <ilya.verbin@intel.com>
4655 Kirill Yukhin <kirill.yukhin@intel.com>
4656 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4657
4658 * config/i386/sse.md (VI4_AVX512F): New.
4659 (VI8_AVX2_AVX512F): Ditto.
4660 (mul<mode>3): Extended with wider modes.
4661 (*<sse4_1_avx2>_mul<mode>3): Ditto.
4662 (mul<mode>3): Ditto.
4663 (vec_widen_<s>mult_odd_<mode>): Ditto.
4664
4665 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4666 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4667 Sergey Lega <sergey.s.lega@intel.com>
4668 Anna Tikhonova <anna.tikhonova@intel.com>
4669 Ilya Tocar <ilya.tocar@intel.com>
4670 Andrey Turetskiy <andrey.turetskiy@intel.com>
4671 Ilya Verbin <ilya.verbin@intel.com>
4672 Kirill Yukhin <kirill.yukhin@intel.com>
4673 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4674
4675 * config/i386/sse.md (VI2_AVX512F): New.
4676 (VI124_AVX512F): Ditto.
4677 (sseunpackmode): Extended with wider modes.
4678 (sseunpackfltmode): Ditto.
4679 (vec_unpacks_float_hi_<mode>): Ditto.
4680 (vec_unpacks_float_lo_<mode>): Ditto.
4681 (vec_unpacku_float_hi_<mode>): Ditto.
4682 (vec_unpacku_float_lo_<mode>): Ditto.
4683 (vec_unpacks_lo_<mode>): Ditto.
4684 (vec_unpacks_hi_<mode>): Ditto.
4685 (vec_unpacku_lo_<mode>): Ditto.
4686 (vec_unpacku_hi_<mode>): Ditto.
4687
4688 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4689 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4690 Sergey Lega <sergey.s.lega@intel.com>
4691 Anna Tikhonova <anna.tikhonova@intel.com>
4692 Ilya Tocar <ilya.tocar@intel.com>
4693 Andrey Turetskiy <andrey.turetskiy@intel.com>
4694 Ilya Verbin <ilya.verbin@intel.com>
4695 Kirill Yukhin <kirill.yukhin@intel.com>
4696 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4697
4698 * config/i386/i386.md (multdiv): New.
4699 (multdiv_mnemonic): Ditto.
4700 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
4701 (<sse>_vm<multdiv_mnemonic><mode>3): This.
4702 (<sse>_vmdiv<mode>3): Removed.
4703
4704 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4705 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4706 Sergey Lega <sergey.s.lega@intel.com>
4707 Anna Tikhonova <anna.tikhonova@intel.com>
4708 Ilya Tocar <ilya.tocar@intel.com>
4709 Andrey Turetskiy <andrey.turetskiy@intel.com>
4710 Ilya Verbin <ilya.verbin@intel.com>
4711 Kirill Yukhin <kirill.yukhin@intel.com>
4712 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4713
4714 * config/i386/sse.md (V): Extended with wider modes.
4715 (VF2): Ditto.
4716 (ssehalfvecmode): Ditto.
4717 (i128): Ditto.
4718 (ssepackfltmode): Ditto.
4719 (avx_vec_concat<mode>): Ditto.
4720 (V_256_512): New iterator.
4721 (VF2_512_256): Ditto.
4722 (si2dfmode): New attribute.
4723 (si2dfmodelower): Ditto.
4724 (sf2dfmode): Ditto.
4725 (concat_tg_mode): Ditto.
4726 (floatv4siv4df2): Changed to ...
4727 (float<si2dfmodelower><mode>2): This.
4728 (avx_cvtps2pd256): Changed to ...
4729 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
4730 (vec_pack_trunc_v4df): Changed to ...
4731 (vec_pack_trunc_<mode>): This.
4732 (avx_vpermil<mode>): Changed to ...
4733 (<sse2_avx_avx512f>_vpermil<mode>): This.
4734 (<fixsuffix>fix_truncv8dfv8si2): New.
4735 (vec_pack_sfix_trunc_v8df): Ditto.
4736 (avx512f_rndscale<mode>): Ditto.
4737 (avx512f_roundpd512): Ditto.
4738 (vec_pack_ufix_trunc_<mode>): Updated iterator.
4739
4740 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4741 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4742 Sergey Lega <sergey.s.lega@intel.com>
4743 Anna Tikhonova <anna.tikhonova@intel.com>
4744 Ilya Tocar <ilya.tocar@intel.com>
4745 Andrey Turetskiy <andrey.turetskiy@intel.com>
4746 Ilya Verbin <ilya.verbin@intel.com>
4747 Kirill Yukhin <kirill.yukhin@intel.com>
4748 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4749
4750 * config/i386/i386.md (any_fix): New iterator.
4751 (fixsuffix): New attribute.
4752 * config/i386/sse.md (VF1): Extened with wider modes.
4753 (VI): Ditto.
4754 (VI_AVX2): Ditto.
4755 (VI8): Ditto.
4756 (sseintvecmodelower): Ditto.
4757 (ssescalarmode): Ditto.
4758 (ssescalarnum): Ditto.
4759 (VF1_128_256): New.
4760 (ssexmmmode): Ditto.
4761 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
4762 (<sse>_rcp<mode>2): Change iterator.
4763 (rsqrt<mode>2): Ditto.
4764 (<sse>_rsqrt<mode>2): Ditto.
4765 (avx2_vec_dup<mode>): Ditto.
4766 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
4767 (round<mode>2_sfix): Ditto.
4768 (avx2_pbroadcast<mode>): Ditto.
4769 (*andnot<mode>3): Handle XI mode.
4770 (*<code><mode>3): Ditto.
4771 (AVXTOSSEMODE): Removed.
4772 (avx_vpermil<mode>): Changed to ...
4773 (<sse2_avx_avx512f>_vpermil<mode>): This.
4774
4775 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4776 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4777 Sergey Lega <sergey.s.lega@intel.com>
4778 Anna Tikhonova <anna.tikhonova@intel.com>
4779 Ilya Tocar <ilya.tocar@intel.com>
4780 Andrey Turetskiy <andrey.turetskiy@intel.com>
4781 Ilya Verbin <ilya.verbin@intel.com>
4782 Kirill Yukhin <kirill.yukhin@intel.com>
4783 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4784
4785 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
4786 (<sse>_comi): Ditto.
4787 (<sse>_ucomi): Ditto.
4788 (sse_cvtss2siq_2): Ditto.
4789 (sse2_cvtsd2si): Ditto.
4790 (sse2_cvtsd2siq): Ditto.
4791 (sse2_cvttsd2si): Ditto.
4792 (sse2_cvttsd2siq): Ditto.
4793 (<shift_insn><mode>3): Ditto.
4794 (sse2_cvtsi2sdq): Update constraint and prefix.
4795 (sse_cvtsi2ss): Update prefix.
4796 (sse_cvtsi2ssq): Ditto.
4797
4798 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4799
4800 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
4801 ignore internal calls.
4802
4803 2013-10-11 Richard Biener <rguenther@suse.de>
4804
4805 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
4806 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
4807 TREE_OVERFLOW is set.
4808
4809 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
4810
4811 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
4812
4813 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
4814 * gimple.h: Likewise.
4815
4816 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
4817
4818 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
4819 macros.
4820
4821 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
4822
4823 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
4824 * configure: Regenerate.
4825
4826 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4827
4828 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
4829 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
4830 formatting fixes, use pp_colon instead of pp_character (..., ':'),
4831 similarly pp_right_paren.
4832 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
4833 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
4834 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
4835 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
4836 clauses.
4837 (omp_declare_simd_clauses_equal,
4838 omp_remove_redundant_declare_simd_attrs): New functions.
4839 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
4840 (walk_tree_1): Handle new OpenMP 4.0 clauses.
4841 * tree.h (OMP_LOOP_CHECK): Define.
4842 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
4843 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
4844 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
4845 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
4846 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
4847 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
4848 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
4849 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
4850 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
4851 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
4852 OMP_CLAUSE_SIMDLEN_EXPR): Define.
4853 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
4854 (omp_remove_redundant_declare_simd_attrs): New prototype.
4855 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
4856 GIMPLE_OMP_TEAMS): New codes.
4857 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
4858 * omp-low.c (struct omp_context): Add cancel_label and cancellable
4859 fields.
4860 (target_nesting_level): New variable.
4861 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
4862 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
4863 for collapse > 1 static schedule unless ordered.
4864 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
4865 (determine_parallel_type): Adjust get_ws_args_for caller.
4866 (install_var_field): Handle mask & 4 for double indirection.
4867 (scan_sharing_clauses): Ignore shared clause on teams construct.
4868 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
4869 (create_omp_child_function): If inside target or declare target
4870 constructs, set "omp declare target" attribute on the child function.
4871 (find_combined_for): New function.
4872 (scan_omp_parallel): Handle combined loops.
4873 (scan_omp_target, scan_omp_teams): New functions.
4874 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
4875 restrictions and set ctx->cancellable for cancellable constructs.
4876 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
4877 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
4878 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
4879 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
4880 (omp_clause_aligned_alignment): New function.
4881 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
4882 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
4883 on teams constructs. Handle user defined reductions and new
4884 OpenMP 4.0 clauses.
4885 (lower_reduction_clauses): Don't set placeholder to address of ref
4886 if it has already the right type.
4887 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
4888 (expand_parallel_call): Use the new non-_start suffixed builtins,
4889 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
4890 and GOMP_parallel_end after the call.
4891 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
4892 (expand_omp_for_init_counts): Handle combined loops.
4893 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
4894 loops.
4895 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
4896 end of cancellable loops.
4897 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
4898 Likewise. Handle collapse > 1 loops.
4899 (expand_omp_simd): Handle combined loops.
4900 (expand_omp_for): Add inner_stmt argument, adjust callers of
4901 expand_omp_for* functions, use expand_omp_for_static*chunk even
4902 for collapse > 1 unless ordered.
4903 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
4904 of cancellable sections.
4905 (expand_omp_single): Remove need_barrier variable, just rely on
4906 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
4907 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
4908 (expand_omp_atomic_load, expand_omp_atomic_store,
4909 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
4910 (expand_omp_target): New function.
4911 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
4912 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
4913 (build_omp_regions_1): Immediately close region for
4914 GF_OMP_TARGET_KIND_UPDATE.
4915 (maybe_add_implicit_barrier_cancel): New function.
4916 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
4917 cancellation.
4918 (lower_omp_single): Likewise. Add clobber after the barrier.
4919 (lower_omp_taskgroup): New function.
4920 (lower_omp_for): Handle combined loops. Adjust
4921 lower_rec_input_clauses caller. Handle cancellation.
4922 (lower_depend_clauses): New function.
4923 (lower_omp_taskreg): Lower depend clauses. Adjust
4924 lower_rec_input_clauses caller. Add clobber after the call. Handle
4925 cancellation.
4926 (lower_omp_target, lower_omp_teams): New functions.
4927 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
4928 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
4929 and GOMP_cancellation_point calls.
4930 (lower_omp): Fold stmts inside of target region.
4931 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
4932 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4933 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
4934 (BT_FN_VOID_OMPFN_PTR_UINT,
4935 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
4936 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
4937 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
4938 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
4939 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
4940 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
4941 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
4942 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
4943 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
4944 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
4945 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
4946 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
4947 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
4948 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
4949 BUILT_IN_GOMP_PARALLEL_END.
4950 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
4951 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4952 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
4953 GF_OMP_FOR_KIND_DISTRIBUTE.
4954 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
4955 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
4956 (dump_gimple_omp_return): Print lhs if it has any.
4957 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
4958 gimple_omp_atomic_seq_cst_p.
4959 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
4960 and GIMPLE_OMP_TEAMS.
4961 * langhooks.c (lhd_omp_mappable_type): New function.
4962 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
4963 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
4964 hook.
4965 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
4966 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
4967 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
4968 (struct gimplify_omp_ctx): Add combined_loop field.
4969 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
4970 on stmts inside of target region.
4971 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
4972 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
4973 ORT_TARGET and ORT_TARGET_DATA.
4974 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
4975 Handle GOVD_MAP.
4976 (omp_notice_threadprivate_variable): Complain about threadprivate
4977 variables in target region.
4978 (omp_notice_variable): Complain about vars with non-mappable type
4979 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
4980 (omp_check_private): Ignore ORT_TARGET* regions.
4981 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
4982 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
4983 (find_combined_omp_for): New function.
4984 (gimplify_omp_for): Handle gimplification of combined loops.
4985 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
4986 OMP_TEAMS.
4987 (gimplify_omp_target_update): New function.
4988 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
4989 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
4990 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
4991 (gimplify_body): If fndecl has "omp declare target" attribute, add
4992 implicit ORT_TARGET context around it.
4993 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
4994 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
4995 * tree-nested.c (convert_nonlocal_reference_stmt,
4996 convert_local_reference_stmt, convert_gimple_call): Handle
4997 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4998 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
4999 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
5000 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
5001 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
5002 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
5003 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
5004 BUILT_IN_GOMP_LOOP_END_CANCEL,
5005 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
5006 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
5007 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
5008 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
5009 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
5010 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
5011 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
5012 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
5013 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
5014 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
5015 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
5016 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
5017 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
5018 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
5019 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
5020 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5021 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
5022 gimple_build_omp_teams): New functions.
5023 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
5024 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
5025 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
5026 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5027 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
5028 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
5029 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
5030 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
5031 GF_OMP_ATOMIC_SEQ_CST): New.
5032 (gimple_build_omp_taskgroup, gimple_build_omp_target,
5033 gimple_build_omp_teams): New prototypes.
5034 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
5035 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5036 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
5037 GIMPLE_OMP_SINGLE as end of range.
5038 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
5039 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
5040 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
5041 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
5042 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
5043 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
5044 gimple_omp_target_kind, gimple_omp_target_set_kind,
5045 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
5046 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
5047 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
5048 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
5049 gimple_omp_teams_set_clauses): New inlines.
5050 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
5051 and GIMPLE_OMP_TASKGROUP.
5052 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
5053 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
5054 enum omp_clause_proc_bind_kind): New.
5055 (union omp_clause_subcode): Add depend_kind, map_kind and
5056 proc_bind_kind fields.
5057 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
5058 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5059 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
5060 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
5061 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
5062
5063 2013-10-10 Teresa Johnson <tejohnson@google.com>
5064
5065 * predict.c (tree_estimate_probability): Add new parameter
5066 for estimate_bb_frequencies.
5067 (estimate_bb_frequencies): Add new parameter to force estimation.
5068 (rebuild_frequencies): When max frequency in function is small,
5069 recompute counts from frequencies.
5070 * predict.h (estimate_bb_frequencies): New parameter.
5071
5072 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5073
5074 * ipa-inline.c (ipa_inline): Fix leak of "order" when
5075 optimizations are disabled.
5076
5077 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5078
5079 * coverage.c (coverage_finish): Fix leak of da_file_name.
5080
5081 2013-10-10 Jan Hubicka <jh@suse.cz>
5082
5083 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
5084 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
5085 Bobcat and generic.
5086
5087 2013-10-10 Jakub Jelinek <jakub@redhat.com>
5088
5089 PR middle-end/58670
5090 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
5091 if any labels are in FALLTHRU_BB, use a special label emitted
5092 immediately after the asm goto insn rather than label_rtx
5093 of the LABEL_DECL.
5094 (expand_asm_stmt): Adjust caller.
5095 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
5096 edge if the last insn in predecessor is a jump with single successor,
5097 but it isn't simplejump_p.
5098
5099 2013-10-10 Richard Biener <rguenther@suse.de>
5100
5101 PR tree-optimization/58656
5102 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
5103
5104 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5105
5106 * gimplify.c: Include expr.h and tm_p.h for targets with special
5107 va-arg padding requirements.
5108
5109 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5110
5111 * tree-flow.h: Move some prototypes to gimple.h.
5112 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
5113 * gimple.h: Relocate some prototypes from tree-flow.h
5114 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
5115 Move to gimplify.c.
5116 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
5117 (build_va_arg_indirect_ref): Relocate and make static.
5118 (std_gimplify_va_arg_expr): Relocate here.
5119 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
5120 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
5121
5122 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5123
5124 * doc/md.texi: Document the mnemonic attribute.
5125
5126 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5127
5128 PR target/57377
5129 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
5130 (set_attr_alternative x ...) when searching for user defined
5131 mnemonic attribute.
5132
5133 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5134
5135 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
5136
5137 2013-10-09 Easwaran Raman <eraman@google.com>
5138
5139 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
5140 * cfgexpand.c (defer_stack_allocation): ...use here
5141 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
5142
5143 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
5144
5145 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
5146 (operand_equal_for_value_replacement): New function, extracted from
5147 value_replacement and enhanced to catch more cases.
5148 (value_replacement): Use operand_equal_for_value_replacement.
5149
5150 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5151
5152 * loop-doloop.c (doloop_modify, doloop_optimize): Use
5153 get_max_loop_iterations.
5154
5155 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5156
5157 * config/arm/aarch-common.c (arm_early_load_addr_dep):
5158 Place comment above function.
5159
5160 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5161
5162 * tree-flow.h: Remove all remaining prototypes, enums and structs that
5163 are not related to tree-cfg.c.
5164 * tree-ssa-address.h: New file. Relocate prototypes.
5165 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
5166 (addr_for_mem_ref): New. Combine call to get_address_description and
5167 return addr_for_mem_ref.
5168 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
5169 * tree-ssa-live.h: Adjust prototypes.
5170 * passes.c: Include tree-ssa-live.h.
5171 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
5172 * graphite.c (graphite_transform_loops): Make static.
5173 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5174 make_pass_graphite, pass_data_graphite_transforms,
5175 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
5176 * ipa-pure-const.c (warn_function_noreturn): Make static.
5177 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5178 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5179 Relocate from tree-cfg.c
5180 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
5181 static.
5182 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5183 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5184 Move to ipa-pure-const.c.
5185 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
5186 Relocate from tree-optimize.c.
5187 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
5188 make_pass_fixup_cfg): Move to tree-cfg.c.
5189 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
5190 Relocate some prototypes.
5191 * tree-data-ref.h (tree_check_data_deps) Add prototype.
5192 * tree-dump.c (dump_function_to_file): Remove prototype.
5193 Add tree-flow.h to the include file.
5194 * tree-dump.h: Remove prototype.
5195 * tree-parloops.h: New File. Add prototypes.
5196 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
5197 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
5198 from tree-ssa-loop.c.
5199 * tree-predcom.c (run_tree_predictive_commoning,
5200 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
5201 Relocate here from tree-ssa-loop.c.
5202 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
5203 ssa_name_values.release ().
5204 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
5205 (ssa_name_values): Relocate from tree-flow.h.
5206 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
5207 * tree-ssa-loop.c (run_tree_predictive_commoning,
5208 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
5209 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5210 make_pass_graphite, pass_data_graphite_transforms,
5211 make_pass_graphite_transforms, gate_tree_parallelize_loops,
5212 tree_parallelize_loops, pass_data_parallelize_loops,
5213 make_pass_parallelize_loops): Move to other files.
5214 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
5215 moved here.
5216 * tree.h: Remove prototypes from tree-address.c.
5217
5218 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5219
5220 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
5221 (struct int_tree_map): Move to tree-hasher.h
5222 (SCALE, LABEL, PERCENT): Move to gimple.h
5223 * tree-flow-inline.h: Delete. Move functions to other files.
5224 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
5225 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
5226 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
5227 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
5228 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
5229 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
5230 * tree-hasher.h: Don't include tree-flow.h.
5231 (struct int_tree_map): Relocate from tree-flow.h.
5232 * tree-sra.c (contains_view_convert_expr_p): Relocate from
5233 tree-flow-inline.h and make static.
5234 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
5235 tree-flow-inline.h.
5236 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
5237 tree-flow-inline.h and make static.
5238 * tree.h (is_global_var, may_be_aliased): Relocate from
5239 tree-flow-inline.h.
5240 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
5241 * value-prof.c: No longer include tree-flow-inline.h.
5242 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
5243
5244 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5245
5246 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
5247 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
5248 (enum move_pos): Move to tree-ssa-loop-im.h
5249 * cfgloop.h: Move some prototypes.
5250 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
5251 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
5252 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
5253 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
5254 (loop_containing_stmt): Relocate from tree-flow-inline.h.
5255 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
5256 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
5257 (enum move_pos): Relocate here.
5258 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
5259 tree-ssa-loop.c.
5260 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
5261 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
5262 make_pass_lim): Relocate here from tree-ssa-loop.c.
5263 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
5264 tree-ssa-loop.c.
5265 (loop_edge_to_cancel, unloop_loops): Make static.
5266 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5267 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
5268 (tree_complete_unroll, gate_tree_complete_unroll,
5269 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
5270 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5271 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
5272 * tree-ssa-loop-ivopts.c: Remove local prototypes.
5273 (stmt_invariant_in_loop_p): Remove unused function.
5274 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
5275 * tree-ssa-loop-manip.h: New file. Add prototypes.
5276 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
5277 (gcov_type_to_double_int): Move to cfgloop.h.
5278 (double_int_cmp, bound_index,
5279 estimate_numbers_of_iterations_loop): Make static.
5280 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
5281 (max_loop_iterations): Factor out get_max_loop_iterations.
5282 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
5283 cfgloop.c.
5284 * tree-ssa-loop-niter.h: New file. Add prototypes.
5285 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
5286 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
5287 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
5288 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
5289 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
5290 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
5291 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
5292 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
5293 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
5294 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
5295 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5296 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
5297 pass_data_complete_unroll, make_pass_complete_unroll,
5298 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5299 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
5300 tree-ssa-loop-ivcanon.c.
5301 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
5302 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
5303 tree-ssa-loop-prefetch.c.
5304 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
5305 tree-ssa-loop-im.c.
5306 (get_lsm_tmp_name): Relocate and add suffix parameter.
5307 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
5308 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
5309 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
5310 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
5311 (get_estimated_loop_iterations): Factor out accessor from
5312 estimated_loop_iterations in tree-ssa-loop-niter.c.
5313 (get_max_loop_iterations): Factor out accessor from
5314 _max_loop_iterations in tree-ssa-niter.c.
5315 * loop-unroll.c (decide_unroll_constant_iterations,
5316 decide_unroll_runtime_iterations, decide_peel_simple,
5317 decide_unroll_stupid): Use new get_* accessors.
5318
5319 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
5320
5321 PR tree-optimization/20318
5322 * doc/extend.texi (returns_nonnull): New function attribute.
5323 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
5324 attribute.
5325 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
5326 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
5327
5328 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
5329
5330 PR middle-end/58570
5331 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
5332 false if both components are bitfields.
5333
5334 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5335
5336 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
5337 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
5338 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
5339 * config/aarch64/aarch64.h
5340 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
5341 * config/aarch64/aarch64-simd-builtins.def
5342 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
5343
5344 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5345
5346 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
5347
5348 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5349
5350 * config/aarch64/arm_neon.h (vdiv_f64): Added.
5351
5352 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5353
5354 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
5355 (vneg_f64): New intrinsic.
5356 (vneg_s8): Asm replaced with C.
5357 (vneg_s16): Likewise.
5358 (vneg_s32): Likewise.
5359 (vneg_s64): New intrinsic.
5360 (vnegq_f32): Asm replaced with C.
5361 (vnegq_f64): Likewise.
5362 (vnegq_s8): Likewise.
5363 (vnegq_s16): Likewise.
5364 (vnegq_s32): Likewise.
5365 (vnegq_s64): Likewise.
5366
5367 2013-10-09 Renlin Li <Renlin.Li@arm.com>
5368
5369 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
5370
5371 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5372
5373 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
5374 packed stack special handling.
5375 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
5376 back to fixed stack slots for FPRs saved due to stdarg.
5377
5378 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5379
5380 * config/s390/s390.c (s390_frame_info): Restructure function.
5381
5382 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5383
5384 * config/s390/s390.c (struct s390_frame_layout): New field
5385 gpr_save_slots.
5386 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
5387 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
5388 regs_ever_clobbered to char*.
5389 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
5390 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
5391 (s390_register_info): Enable FPR save slots. Move/Copy some
5392 functionality into ...
5393 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
5394 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
5395 function.
5396 (s390_frame_info): Do gpr slot allocation here now. stdarg does
5397 not imply a stack frame.
5398 (s390_init_frame_layout): Remove variable clobbered_regs.
5399 (s390_update_register_info): Remove function.
5400 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
5401 cannot be used for register renaming.
5402 (s390_hard_regno_scratch_ok): New function.
5403 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
5404 (s390_initial_elimination_offset): Change offset calculation of
5405 the return address pointer.
5406 (save_gprs): Deal with only r6 being saved from the call-saved regs.
5407 (restore_gprs): Set frame related flag.
5408 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
5409 (s390_emit_prologue): Call s390_register_info instead of
5410 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
5411 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
5412 (s390_optimize_prologue): Call s390_optimize_register_info.
5413 Try to remove also FPR slot save/restore INSNs. Remove frame
5414 related flags from restore INSNs.
5415
5416 2013-10-08 DJ Delorie <dj@redhat.com>
5417
5418 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
5419 (movhi): Likewise.
5420
5421 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
5422 avoid conflict with the MI use of %c.
5423 * config/rl78/rl78-real.md: change %c to %C throughout.
5424 * config/rl78/rl78-virt.md: Likewise.
5425
5426 2013-10-08 Jan Hubicka <jh@suse.cz>
5427
5428 * config/i386/i386.c (ix86_option_override_internal): Switch
5429 to SSE math for -ffast-math when target ISA supports SSE2.
5430
5431 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
5432
5433 * tree-flow.h: Remove some prototypes.
5434 * tree.h: Remove some protypes, add a couple.
5435 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
5436 using_eh_for_cleanups_p): Add interface routines for front ends.
5437 * tree-eh.h: New file. Add protoptyes.
5438 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
5439 (add_stmt_to_eh_lp_fn): Make static.
5440 (lower_try_finally): Use new using_eh_for_cleanups_p.
5441 * emit-rtl.c: Include tree-eh.h.
5442 * gimple.h: Include tree-eh.h.
5443
5444 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
5445
5446 PR tree-optimization/58480
5447 * tree-vrp.c (infer_nonnull_range): New function.
5448 (infer_value_range): Call infer_nonnull_range.
5449
5450 2013-10-08 Dehao Chen <dehao@google.com>
5451
5452 PR tree-optimization/58619
5453 * tree-inline.c (copy_phis_for_bb): Combine location data
5454 only if non-null.
5455
5456 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5457
5458 PR target/58423
5459 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
5460 RTX_FRAME_RELATED_P on INSN.
5461
5462 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5463
5464 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
5465 (altivec_expand_vec_perm_const): Call it.
5466
5467 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5468
5469 * config/rs6000/vector.md (mov<mode>): Emit permuted move
5470 sequences for LE VSX loads and stores at expand time.
5471 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
5472 prototype.
5473 * config/rs6000/rs6000.c (rs6000_const_vec): New.
5474 (rs6000_gen_le_vsx_permute): New.
5475 (rs6000_gen_le_vsx_load): New.
5476 (rs6000_gen_le_vsx_store): New.
5477 (rs6000_gen_le_vsx_move): New.
5478 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
5479 (*vsx_le_perm_load_v4si): New.
5480 (*vsx_le_perm_load_v8hi): New.
5481 (*vsx_le_perm_load_v16qi): New.
5482 (*vsx_le_perm_store_v2di): New.
5483 (*vsx_le_perm_store_v4si): New.
5484 (*vsx_le_perm_store_v8hi): New.
5485 (*vsx_le_perm_store_v16qi): New.
5486 (*vsx_xxpermdi2_le_<mode>): New.
5487 (*vsx_xxpermdi4_le_<mode>): New.
5488 (*vsx_xxpermdi8_le_V8HI): New.
5489 (*vsx_xxpermdi16_le_V16QI): New.
5490 (*vsx_lxvd2x2_le_<mode>): New.
5491 (*vsx_lxvd2x4_le_<mode>): New.
5492 (*vsx_lxvd2x8_le_V8HI): New.
5493 (*vsx_lxvd2x16_le_V16QI): New.
5494 (*vsx_stxvd2x2_le_<mode>): New.
5495 (*vsx_stxvd2x4_le_<mode>): New.
5496 (*vsx_stxvd2x8_le_V8HI): New.
5497 (*vsx_stxvd2x16_le_V16QI): New.
5498
5499 2013-10-07 Renlin Li <Renlin.Li@arm.com>
5500
5501 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
5502
5503 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5504
5505 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
5506 loop to work also for 31bit ABI.
5507 Save the stack pointer for frame_size > 0.
5508
5509 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5510
5511 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
5512 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
5513 constraint letters from expanders.
5514 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
5515 retry count to general_operand.
5516 ("tabort"): Give operand 0 a mode.
5517 ("tabort_1"): Add mode and constraint letter for operand 0.
5518 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
5519
5520 2013-10-04 Jeff Law <law@redhat.com>
5521
5522 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
5523
5524 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
5525 out of ...
5526 (thread_across_edge): Here. Call it.
5527
5528 2013-10-04 Cary Coutant <ccoutant@google.com>
5529
5530 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
5531 discarding a location list expression (or a piece of one).
5532
5533 2013-10-03 Jan Hubicka <jh@suse.cz>
5534
5535 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
5536 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
5537 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
5538
5539 2013-10-03 Jan Hubicka <jh@suse.cz>
5540
5541 * config/i386/i386.c (ix86_option_override_internal): Do not enable
5542 accumulate-outgoing-args when producing unwind info.
5543
5544 2013-10-03 Wei Mi <wmi@google.com>
5545
5546 * lra-constraints.c (insert_move_for_subreg): New function
5547 extracted from simplify_operand_subreg.
5548 (simplify_operand_subreg): Add reload for paradoxical subreg.
5549
5550 2013-10-03 Rong Xu <xur@google.com>
5551
5552 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
5553 the candidate of builtin_expect such that we should fix the
5554 size/time estimation.
5555 (estimate_function_body_sizes): Do the acutally size/time fix-up
5556 for builtin_expect.
5557
5558 2013-10-03 Rong Xu <xur@google.com>
5559
5560 * predict.c (tree_predict_by_opcode): Get the probability
5561 for builtin_expect from param builtin_expect_probability.
5562 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
5563 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
5564 * doc/invoke.texi: Add documentation for builtin-expect-probability.
5565
5566 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
5567
5568 PR c++/19476
5569 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
5570 * calls.c (alloca_call_p): Use get_callee_fndecl.
5571 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
5572 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
5573 Likewise.
5574 (vrp_visit_stmt): Remove duplicated code.
5575
5576 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
5577
5578 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
5579 ceildf2, btruncdf2, instead of vsx_* name.
5580
5581 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
5582 iterators to only do V2DF and V4SF here. Move the DF code to
5583 rs6000.md where it is combined with SF mode. Replace <VSv> with
5584 just 'v' since only vector operations are handled with these insns
5585 after moving the DF support to rs6000.md.
5586 (vsx_sub<mode>3): Likewise.
5587 (vsx_mul<mode>3): Likewise.
5588 (vsx_div<mode>3): Likewise.
5589 (vsx_fre<mode>2): Likewise.
5590 (vsx_neg<mode>2): Likewise.
5591 (vsx_abs<mode>2): Likewise.
5592 (vsx_nabs<mode>2): Likewise.
5593 (vsx_smax<mode>3): Likewise.
5594 (vsx_smin<mode>3): Likewise.
5595 (vsx_sqrt<mode>2): Likewise.
5596 (vsx_rsqrte<mode>2): Likewise.
5597 (vsx_fms<mode>4): Likewise.
5598 (vsx_nfma<mode>4): Likewise.
5599 (vsx_copysign<mode>3): Likewise.
5600 (vsx_btrunc<mode>2): Likewise.
5601 (vsx_floor<mode>2): Likewise.
5602 (vsx_ceil<mode>2): Likewise.
5603 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
5604 (vsx_sminsf3): Likewise.
5605 (vsx_fmadf4): Likewise.
5606 (vsx_fmsdf4): Likewise.
5607 (vsx_nfmadf4): Likewise.
5608 (vsx_nfmsdf4): Likewise.
5609 (vsx_cmpdf_internal1): Likewise.
5610
5611 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
5612 simpler to select whether a target has SPE or traditional floating
5613 point support in iterators.
5614 (TARGET_DF_SPE): Likewise.
5615 (TARGET_SF_FPR): Likewise.
5616 (TARGET_DF_FPR): Likewise.
5617 (TARGET_SF_INSN): Macros to say whether floating point support
5618 exists for a given operation for expanders.
5619 (TARGET_DF_INSN): Likewise.
5620
5621 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
5622 combining of SF/DF mode operations, using both traditional and VSX
5623 registers.
5624 (Fvsx): Likewise.
5625 (Ff): Likewise.
5626 (Fv): Likewise.
5627 (Fs): Likewise.
5628 (Ffre): Likewise.
5629 (FFRE): Likewise.
5630 (abs<mode>2): Combine SF/DF modes using traditional floating point
5631 instructions. Add support for using the upper DF registers with
5632 VSX support, and SF registers with power8-vector support. Update
5633 expanders for operations supported by both the SPE and traditional
5634 floating point units.
5635 (abs<mode>2_fpr): Likewise.
5636 (nabs<mode>2): Likewise.
5637 (nabs<mode>2_fpr): Likewise.
5638 (neg<mode>2): Likewise.
5639 (neg<mode>2_fpr): Likewise.
5640 (add<mode>3): Likewise.
5641 (add<mode>3_fpr): Likewise.
5642 (sub<mode>3): Likewise.
5643 (sub<mode>3_fpr): Likewise.
5644 (mul<mode>3): Likewise.
5645 (mul<mode>3_fpr): Likewise.
5646 (div<mode>3): Likewise.
5647 (div<mode>3_fpr): Likewise.
5648 (sqrt<mode>3): Likewise.
5649 (sqrt<mode>3_fpr): Likewise.
5650 (fre<Fs>): Likewise.
5651 (rsqrt<mode>2): Likewise.
5652 (cmp<mode>_fpr): Likewise.
5653 (smax<mode>3): Likewise.
5654 (smin<mode>3): Likewise.
5655 (smax<mode>3_vsx): Likewise.
5656 (smin<mode>3_vsx): Likewise.
5657 (negsf2): Delete SF operations that are merged with DF.
5658 (abssf2): Likewise.
5659 (addsf3): Likewise.
5660 (subsf3): Likewise.
5661 (mulsf3): Likewise.
5662 (divsf3): Likewise.
5663 (fres): Likewise.
5664 (fmasf4_fpr): Likewise.
5665 (fmssf4_fpr): Likewise.
5666 (nfmasf4_fpr): Likewise.
5667 (nfmssf4_fpr): Likewise.
5668 (sqrtsf2): Likewise.
5669 (rsqrtsf_internal1): Likewise.
5670 (smaxsf3): Likewise.
5671 (sminsf3): Likewise.
5672 (cmpsf_internal1): Likewise.
5673 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
5674 (negdf2): Delete DF operations that are merged with SF.
5675 (absdf2): Likewise.
5676 (nabsdf2): Likewise.
5677 (adddf3): Likewise.
5678 (subdf3): Likewise.
5679 (muldf3): Likewise.
5680 (divdf3): Likewise.
5681 (fred): Likewise.
5682 (rsqrtdf_internal1): Likewise.
5683 (fmadf4_fpr): Likewise.
5684 (fmsdf4_fpr): Likewise.
5685 (nfmadf4_fpr): Likewise.
5686 (nfmsdf4_fpr): Likewise.
5687 (sqrtdf2): Likewise.
5688 (smaxdf3): Likewise.
5689 (smindf3): Likewise.
5690 (cmpdf_internal1): Likewise.
5691 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
5692 (btrunc<mode>2): Delete separate expander, and combine with the
5693 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
5694 (btrunc<mode>2_fpr): Likewise.
5695 (ceil<mode>2): Likewise.
5696 (ceil<mode>2_fpr): Likewise.
5697 (floor<mode>2): Likewise.
5698 (floor<mode>2_fpr): Likewise.
5699 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
5700 Add support for using the upper registers with VSX and
5701 power8-vector. Move insns to be closer to the define_expands. On
5702 VSX systems, prefer the traditional form of FMA over the VSX
5703 version, since the traditional form allows the target not to
5704 overlap with the inputs.
5705 (fms<mode>4_fpr): Likewise.
5706 (nfma<mode>4_fpr): Likewise.
5707 (nfms<mode>4_fpr): Likewise.
5708
5709 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5710 Richard Earnshaw <richard.earnshaw@arm.com>
5711
5712 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
5713 (struct mult_cost_table): Likewise.
5714 (struct mem_cost_table): Likewise.
5715 (struct fp_cost_table): Likewise.
5716 (struct vector_cost_table): Likewise.
5717 (cpu_cost_table): Likewise.
5718 * config/arm/arm.opt (mold-rts-costs): New option.
5719 (mnew-generic-costs): Likewise.
5720 * config/arm/arm.c (generic_extra_costs): New table.
5721 (cortexa15_extra_costs): Likewise.
5722 (arm_slowmul_tune): Use NULL as new costs.
5723 (arm_fastmul_tune): Likewise.
5724 (arm_strongarm_tune): Likewise.
5725 (arm_xscale_tune): Likewise.
5726 (arm_9e_tune): Likewise.
5727 (arm_v6t2_tune): Likewise.
5728 (arm_cortex_a5_tune): Likewise.
5729 (arm_cortex_a9_tune): Likewise.
5730 (arm_v6m_tune): Likewise.
5731 (arm_fa726te_tune): Likewise.
5732 (arm_cortex_a15_tune): Use cortex15_extra_costs.
5733 (arm_cortex_tune): Use generict_extra_costs.
5734 (shifter_op_p): New function.
5735 (arm_unspec_cost): Likewise.
5736 (LIBCALL_COST): Define.
5737 (arm_new_rtx_costs): New function.
5738 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
5739 table is available. Use old costs otherwise unless mnew-generic-costs
5740 is specified.
5741 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
5742 (cpu_cost_table): Declare.
5743
5744 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
5745
5746 PR target/58460
5747 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
5748 (*subs_mul_imm_<mode>)
5749 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
5750 (*sub_<shift>_<mode>)
5751 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
5752 Remove k constraint.
5753
5754 2013-10-03 Ian Bolton <ian.bolton@arm.com>
5755
5756 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
5757 code.
5758 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
5759
5760 2013-10-02 Teresa Johnson <tejohnson@google.com>
5761
5762 * predict.c (probably_never_executed): New function.
5763 (probably_never_executed_bb_p): Invoke probably_never_executed.
5764 (probably_never_executed_edge_p): Ditto.
5765 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
5766 Treat profile insanities conservatively.
5767
5768 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
5769
5770 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
5771
5772 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
5773
5774 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
5775 use it. Use smaller increase for scratch. Don't increase reject
5776 for early clobber scratch.
5777 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
5778 setting eliminated regs except setting fp from hfp.
5779 (lra_eliminate): Check lra_insn_recog_data on NULL.
5780
5781 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5782
5783 PR target/58587
5784 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
5785 setting -mvsx-timode by default until the underlying problem is fixed.
5786 (RS6000_CPU, power7 defaults): Likewise.
5787
5788 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5789
5790 * config/x-linux (host-linux.o): Remove header dependencies.
5791 Use $(COMPILE) and $(POSTCOMPILE).
5792 * config/t-linux-android (linux-android.o): Ditto.
5793
5794 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5795
5796 * Makefile.in (expmed.o-warn): Remove.
5797
5798 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5799
5800 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
5801 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
5802
5803 2013-10-02 Teresa Johnson <tejohnson@google.com>
5804
5805 * dojump.c (do_jump_1): Divide probability between
5806 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
5807
5808 2013-10-02 Tom Tromey <tromey@redhat.com>
5809
5810 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
5811
5812 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5813
5814 * tree-flow.h: Remove some prototypes.
5815 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
5816 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
5817 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
5818 mark_virtual_phi_result_for_renaming): Relocate here.
5819 * tree-into-ssa.h: Add prototypes.
5820 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
5821 single_pred_before_succ_order.
5822 (blocks_in_phiopt_order): Rename and move to cfganal.c.
5823 (nonfreeing_call_p) Move to gimple.c.
5824 * cfganal.c (single_pred_before_succ_order): Move and renamed from
5825 tree-ssa-phiopt.c.
5826 * basic-block.h (single_pred_before_succ_order): Add prototype.
5827 * gimple.c (nonfreeing_call_p): Relocate here.
5828 * gimple.h: Add prototype.
5829 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
5830 * tree-ssa-dom.h: New file. Relocate prototypes here.
5831 * tree-ssa.h: Include tree-ssa-dom.h.
5832
5833 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5834
5835 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
5836 Use $(COMPILE) and $(POSTCOMPILE).
5837
5838 * config/alpha/x-alpha (driver-alpha.o): Ditto.
5839
5840 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5841
5842 * tree-flow.h: Remove some prototypes.
5843 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
5844 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5845 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
5846 propagate_tree_value*): Move from here to...
5847 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
5848 propagate_tree_value*): Relocate here.
5849 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5850 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
5851 * gimple-fold.c: Remove gimple-fold.h from include list.
5852 * tree-vrp.c: Remove gimple-fold.h from include list.
5853 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
5854 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
5855 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
5856 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
5857 * sese.c: Add tree-ssa-propagate.h to include list.
5858
5859 2013-10-02 Richard Biener <rguenther@suse.de>
5860
5861 * tree-loop-distribution.c: Include tree-vectorizer.h for
5862 find_loop_location.
5863 (enum partition_kind): Remove PKIND_REDUCTION.
5864 (struct partition_s): Remove has_writes member, add reduction_p member.
5865 (partition_alloc): Adjust.
5866 (partition_builtin_p): Likewise.
5867 (partition_has_writes): Remove.
5868 (partition_reduction_p): New function.
5869 (partition_merge_into): Likewise.
5870 (generate_code_for_partition): Commonize builtin partition
5871 handling tail.
5872 (rdg_cannot_recompute_vertex_p): Remove.
5873 (already_processed_vertex_p): Likewise.
5874 (rdg_flag_vertex): Do not set has_writes.
5875 (classify_partition): Adjust.
5876 (rdg_build_partitions): Do not set has_writes, treat all
5877 partitions as useful.
5878 (distribute_loop): Record number of library calls generated. Adjust.
5879 (tree_loop_distribution): Report number of loops and library
5880 calls generated as opt-info.
5881
5882 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5883
5884 * tree-flow.h: Include new .h files. Move prototypes.
5885 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
5886 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
5887 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
5888 * tree-pretty-print.h: Add prototypes from tree-flow.h.
5889 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
5890 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
5891 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
5892 * tree.h (get_ref_base_and_extent): Move prototype out.
5893 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
5894 tree-dfa.h.
5895 * gimple-low.h: New File. Add prototypes from tree-flow.h.
5896 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
5897 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
5898 * tree-scalar-evolution.c: Include tree.h.
5899 * sese.c: Include tree.h.
5900 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
5901 * dwarf2out.c: Include tree-dfa.h.
5902 * tree-chrec.c: Include tree.h.
5903 * tree-data-ref.c: Include tree.h.
5904
5905 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
5906
5907 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
5908 Fix whitespace.
5909
5910 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5911
5912 * config/t-sol2 (sol2-c.o): Remove header dependencies.
5913 Use $(COMPILE) and $(POSTCOMPILE).
5914 (sol2-cxx.o): Likewise.
5915 (sol2-stubs.o): Likewise.
5916 (sol2.o): Likewise.
5917 * config/x-solaris (host-solaris.o): Likewise.
5918
5919 * config/sparc/t-sparc (sparc.o): Remove.
5920 (sparc-c.o): Remove header dependencies.
5921 Use $(COMPILE) and $(POSTCOMPILE).
5922 * config/sparc/x-sparc: Likewise.
5923
5924 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
5925
5926 * config/arc/arc-opts.h: Add 2013 to Copyright years.
5927 * config/arc/arc700.md: Likewise.
5928 * config/arc/arc-modes.def: Likewise.
5929 * config/arc/arc-simd.h: Likewise.
5930 * config/arc/t-arc-uClibc: Likewise.
5931 * config/arc/t-arc-newlib: Likewise.
5932
5933 2013-10-02 Renlin Li <renlin.li@arm.com>
5934
5935 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
5936 plus_constant.
5937 (aarch64_expand_epilogue): Likewise.
5938
5939 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5940 Yufeng Zhang <yufeng.zhang@arm.com>
5941
5942 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
5943 declaration.
5944 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
5945 'base_in' represent a conversion and legal_cast_p_1 holds; set
5946 'base_in' with the returned value from get_unwidened.
5947
5948 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5949
5950 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
5951 plus_constant is not used.
5952
5953 2013-10-01 Wei Mi <wmi@google.com>
5954
5955 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
5956 * config/i386/i386.md: Add define_peephole2 to
5957 break partial reg stall for cvtss2sd/cvtsd2ss.
5958
5959 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5960
5961 * config/arc/arc.c (pass_arc_ifcvt::clone):
5962 Update for ctxt_ -> m_ctxt change.
5963
5964 2013-10-01 Jeff Law <law@redhat.com>
5965
5966 * tree-ssa-threadupdate.c (struct redirection_data): Delete
5967 outgoing_edge and intermediate_edge fields. Instead store the path.
5968 (redirection_data::hash): Hash on the last edge's destination index.
5969 (redirection_data::equal): Check the entire thread path.
5970 (lookup_redirectio_data): Corresponding changes.
5971 (create_edge_and_update_destination_phis): Likewise.
5972 (thread_single_edge): Likewise.
5973
5974 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5975 Diego Novillo <dnovillo@google.com>
5976
5977 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
5978 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
5979 (vld32wh_insn, vld32wl_insn): Delete commented-out old
5980 versions of these patterns.
5981
5982 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
5983 (__builtin_arc_aligned): Likewise.
5984
5985 * config/arc/arc.md: Expand adc_0 comment stating the intended
5986 purpose and why it isn't ready.
5987 Replace commented out call_value_via_label_mixed with a
5988 plain comment about bl_s.
5989
5990 * config/arc/arc.c (stdio.h): Don't include directly.
5991 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
5992 Remove if (1) condition.
5993 (arc_encode_section_info): Fix comment.
5994
5995 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5996
5997 * config/arc/arc.c (arc_conditional_register_usage):
5998 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
5999 Also set reg_alloc_order for DMA config regs.
6000
6001 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6002 Jeremy Bennett <jeremy.bennett@embecosm.com>
6003
6004 * doc/install.texi (--with-cpu): Mention ARC.
6005 (arc-*-elf32): New paragraph.
6006 (arc-linux-uclibc): Likewise.
6007 * doc/md.texi (Machine Constraints): Add ARC part.
6008 * doc/invoke.texi: (menu): Add ARC Options.
6009 (Machine Dependent Options) <ARC Options>: Add synopsis.
6010 (node ARC Options): Add.
6011 * doc/extend.texi (long_call / short_call attribute): Add ARC.
6012 (ARC Built-in Functions): New section defining
6013 generic ARC built-in functions.
6014 (ARC SIMD Built-in Functions): New section defining SIMD specific
6015 built-in functions.
6016 (Declaring Attributes of Functions): Extended
6017 description of short_call and long_call attributes for ARC and
6018 added index entries.
6019
6020 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
6021 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
6022 Joern Rennecke <joern.rennecke@embecosm.com>
6023 Muhammad Khurram Riaz <khurram.riaz@arc.com>
6024 Brendan Kehoe <brendan@zen.org>
6025 Michael Eager <eager@eagercon.com>
6026 Simon Cook <simon.cook@embecosm.com>
6027 Jeremy Bennett <jeremy.bennett@embecosm.com>
6028
6029 * config/arc, common/config/arc: New directories.
6030
6031 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6032 Brendan Kehoe <brendan@zen.org>
6033 Simon Cook <simon.cook@embecosm.com>
6034
6035 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
6036
6037 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6038
6039 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
6040 * tree-ssa-coalesce.h: New. Move prototype to here.
6041 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
6042 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
6043 (gimple_can_coalesce_p): Move to...
6044 * gimple.c (gimple_can_coalesce_p): Here.
6045
6046 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6047
6048 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
6049 (dump_decl_set): Move to gimple.c.
6050 * gimple.h: Don't include tree-ssa-operands.h.
6051 (dump_decl_set): Add prototype.
6052 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6053 Move to gimple-ssa.h.
6054 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
6055 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
6056 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
6057 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
6058 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
6059 than PHI_ARG_DEF.
6060 (dump_decl_set): Relocate here.
6061 * gimple-ssa.h: New file.
6062 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6063 Relocate from gimple.h.
6064 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
6065 * tree-ssa-operands.c (swap_ssa_operands): Rename from
6066 swap_tree_operands and remove non-ssa path.
6067 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
6068 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
6069 swap_ssa_operands.
6070 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
6071 vect_is_simple_reduction_1): Use swap_ssa_operands.
6072 * tree-flow.h: Move various prototypes to tree-phinodes.h.
6073 (enum need_phi_state): Move to tree-into-ssa.c.
6074 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6075 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
6076 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
6077 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
6078 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6079 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6080 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6081 num_imm_uses): Move to ssa-iterators.h.
6082 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
6083 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
6084 tree-phinodes.h.
6085 (op_iter_done, op_iter_next_def, op_iter_next_tree,
6086 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
6087 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
6088 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
6089 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
6090 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
6091 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
6092 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
6093 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
6094 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
6095 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
6096 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
6097 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
6098 (set_phi_nodes): Move to tree-phinodes.h.
6099 * tree-ssa-operands.h (enum ssa_op_iter_type,
6100 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
6101 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
6102 (dump_decl_set): Remove prototype.
6103 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
6104 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
6105 (set_phi_nodes): Relocate from tree-flow-inline.h.
6106 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
6107 tree-flow-inline.h
6108 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
6109 include list. Temporarily add gimple.h to include list.
6110 * ssa-iterators.h: New file.
6111 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6112 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
6113 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
6114 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
6115 Relocate from tree-ssa-operands.h.
6116 (delink_imm_use, link_imm_use_to_list, link_imm_use,
6117 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6118 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6119 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6120 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
6121 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
6122 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
6123 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
6124 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
6125 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
6126 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
6127 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
6128 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
6129 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
6130 Relocate from tree-flow-inline.h.
6131 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
6132
6133 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
6134
6135 * aarch64-simd.md
6136 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
6137 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
6138 Insert '\t' to output template.
6139 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
6140 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
6141 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
6142 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
6143
6144 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
6145
6146 * doc/install.texi (Host/target specific installation notes for GCC):
6147 Put @anchor before @heading.
6148 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
6149 Use @email for email addresses.
6150
6151 2013-10-01 Jeff Law <law@redhat.com>
6152
6153 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
6154 a vec. Only delete the path if we create one without successfully
6155 registering a jump thread.
6156 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
6157 as a pointer.
6158 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
6159 (paths): New vector of jump threading paths.
6160 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
6161 (THREAD_PATH): New accessor macro for the entire thread path.
6162 (lookup_redirection_data): Get intermediate and final outgoing edge
6163 from the thread path.
6164 (create_edge_and_update_destination_phis): Copy the threading path.
6165 (ssa_fix_duplicate_block_edges): Get edges and block types from the
6166 jump threading path.
6167 (ssa_redirect_edges): Get edges and block types from the jump threading
6168 path. Free the path vector.
6169 (thread_block): Get edges from the jump threading path. Look at the
6170 entire path to see if we thread to a loop exit. If we cancel a jump
6171 thread request, then free the path vector.
6172 (thread_single_edge): Get edges and block types from the jump threading
6173 path. Free the path vector.
6174 (thread_through_loop_header): Get edges and block types from the jump
6175 threading path. Free the path vector.
6176 (mark_threaded_blocks): Iterate over the vector of paths and store
6177 the path on the appropriate edge. Get edges and block types from the
6178 jump threading path.
6179 (mark_threaded_blocks): Get edges and block types from the jump
6180 threading path. Free the path vector.
6181 (thread_through_all_blocks): Use the vector of paths rather than
6182 a vector of 3-edge sets.
6183 (register_jump_thread): Accept pointer to a path vector rather
6184 than the path vector itself. Store the path vector for later use.
6185 Simplify.
6186
6187 2013-10-01 Jakub Jelinek <jakub@redhat.com>
6188 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6189
6190 PR target/58574
6191 * config/s390/s390.c (s390_split_branches): Modify check for table
6192 jump insns.
6193 (s390_chunkify_start): Rearrange table jump insn check in order to
6194 deal with compare and branch insns correctly.
6195
6196 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
6197
6198 PR target/58578
6199 Revert
6200 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
6201 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
6202 define_insn_and_split.
6203 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
6204 (shiftsi3_compare): New pattern.
6205 (rrx): New pattern.
6206 * config/arm/unspecs.md (UNSPEC_RRX): New.
6207
6208 2013-10-01 Alan Modra <amodra@gmail.com>
6209
6210 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
6211 casing inout operands.
6212
6213 2013-10-01 Richard Biener <rguenther@suse.de>
6214
6215 PR tree-optimization/58553
6216 * tree-loop-distribution.c (struct partition_s): Add niter member.
6217 (classify_partition): Populate niter member for the partition
6218 and properly identify whether the relevant store happens before
6219 or after the loop exit.
6220 (generate_memset_builtin): Use niter member from the partition.
6221 (generate_memcpy_builtin): Likewise.
6222
6223 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6224
6225 * vec.h (vec_prefix, vec): Prefix member names with "m_".
6226 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
6227
6228 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6229
6230 * basic-block.h (edge_list): Prefix member names with "m_".
6231 * context.h (context): Likewise.
6232 * domwalk.h (dom_walker): Likewise.
6233 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
6234 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
6235 * hash-table.h (hash_table): Likewise.
6236 * machmode.h (bit_field_mode_iterator): Likewise.
6237 * pass_manager.h (pass_list): Likewise.
6238 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
6239 * tree-pass.h (pass_data): Likewise.
6240 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
6241 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
6242 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
6243 * asan.c (pass_data_asan): Update accordingly.
6244 * cfganal.c (control_dependences::find_control_dependence): Likewise.
6245 (control_dependences::control_dependences): Likewise.
6246 (control_dependences::~control_dependences): Likewise.
6247 (control_dependences::~control_dependences): Likewise.
6248 (control_dependences::get_edges_dependent_on): Likewise.
6249 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
6250 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
6251 * context.c (gcc::context::context): Likewise.
6252 * cprop.c (pass_rtl_cprop::clone): Likewise.
6253 * domwalk.c (dom_walker::walk): Likewise.
6254 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
6255 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
6256 * mode-switching.c (pass_mode_switching::clone): Likewise.
6257 * passes.c (opt_pass::opt_pass): Likewise.
6258 (pass_manager::pass_manager): Likewise.
6259 * predict.c (pass_strip_predict_hints::clone): Likewise.
6260 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
6261 (pass_split_all_insns::clone): Likewise.
6262 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
6263 Likewise.
6264 (bit_field_mode_iterator::next_mode): Likewise.
6265 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
6266 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
6267 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
6268 * tree-complex.c (pass_lower_complex::clone): Likewise.
6269 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
6270 * tree-object-size.c (pass_object_sizes::clone): Likewise.
6271 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
6272 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
6273 (pass_fold_builtins::clone): Likewise.
6274 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
6275 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
6276 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
6277 (pass_cd_dce::clone): Likewise.
6278 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
6279 (pass_phi_only_cprop::clone): Likewise.
6280 * tree-ssa-dse.c (pass_dse::clone): Likewise.
6281 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
6282 * tree-ssa-loop.c (pass_lim::clone): Likewise.
6283 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
6284 * tree-ssa-pre.c (pass_fre::clone): Likewise.
6285 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
6286 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
6287 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
6288 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
6289 * tree-vrp.c (pass_vrp::clone): Likewise.
6290 * tsan.c (pass_tsan::clone): Likewise.
6291
6292 2013-09-30 Jakub Jelinek <jakub@redhat.com>
6293
6294 PR middle-end/58564
6295 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
6296 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
6297
6298 PR middle-end/58564
6299 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
6300 optimization, punt if sign_bit_p looked through any zero extension.
6301
6302 2013-09-30 Teresa Johnson <tejohnson@google.com>
6303
6304 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
6305 Update redirected out edge count in joiner case.
6306 (ssa_redirect_edges): Common the joiner and non-joiner cases
6307 so that joiner case gets profile updates.
6308
6309 2013-09-30 Richard Biener <rguenther@suse.de>
6310
6311 PR tree-optimization/58554
6312 * tree-loop-distribution.c (classify_partition): Require
6313 unconditionally executed stores for memcpy and memset recognition.
6314 (tree_loop_distribution): Calculate dominance info.
6315
6316 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
6317
6318 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
6319 (NO_PROFILE_COUNTERS): Likewise.
6320 (PROFILE_HOOK): Likewise.
6321 (FUNCTION_PROFILER): Likewise.
6322 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
6323
6324 2013-09-30 Iain Sandoe <iain@codesourcery.com>
6325
6326 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
6327 calls and defines in TARGET_MACHO conditional.
6328 (load_macho_picbase_di): Likewise.
6329 (reload_macho_picbase): Likewise.
6330 (reload_macho_picbase_si): Likewise.
6331 (reload_macho_picbase_di): Likewise.
6332 (nonlocal_goto_receiver): Likewise.
6333
6334 2013-09-30 Nick Clifton <nickc@redhat.com>
6335
6336 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
6337 that use the MSP430X ISA.
6338 (msp430_option_override): Scan -mmcu command line option for any
6339 MCU name that supports the MSP430X ISA.
6340 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
6341 -mmcu options which enable the MSP430X ISA.
6342
6343 2013-09-30 Richard Biener <rguenther@suse.de>
6344
6345 PR middle-end/58532
6346 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
6347 before looking for setjmp-like calls.
6348
6349 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6350
6351 PR target/10901
6352 * config/darwin-protos.h (machopic_get_function_picbase): New.
6353 * config/darwin.c (machopic_get_function_picbase): New.
6354 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
6355 label for a new func. (load_macho_picbase_di): Likewise.
6356 (reload_macho_picbase): New expand.
6357 (reload_macho_picbase_si): New insn.
6358 (reload_macho_picbase_di): New insn.
6359 (nonlocal_goto_receiver): New define and split.
6360 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
6361 (unspecv enum): Add UNSPECV_NLGR.
6362
6363 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6364
6365 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
6366 that altivec registers are correctly sized on Darwin.
6367
6368 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6369
6370 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
6371 darwin-driver.o): Use COMPILE and POSTCOMPILE.
6372 * config/x-darwin (host-darwin.o): Likewise.
6373 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
6374 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
6375 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
6376
6377 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
6378
6379 * doc/invoke.texi: Fix usage of @tie{} command.
6380
6381 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
6382
6383 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
6384
6385 2013-09-28 Jan Hubicka <jh@suse.cz>
6386
6387 * config/i386/x86-tune.def: Add documentation for each of the options;
6388 add whitespace.
6389
6390 2013-09-28 Jan Hubicka <jh@suse.cz>
6391
6392 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
6393 generic.
6394 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
6395 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
6396 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
6397 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
6398 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
6399
6400 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
6401
6402 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
6403 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
6404 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
6405 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
6406 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
6407 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
6408 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
6409 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
6410 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
6411 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
6412 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
6413 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
6414 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
6415 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
6416 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
6417 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
6418 gimple.h, godump.c, graphite-clast-to-gimple.c,
6419 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
6420 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
6421 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
6422 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6423 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
6424 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
6425 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
6426 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
6427 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
6428 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
6429 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
6430 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
6431 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
6432 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
6433 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
6434 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
6435 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
6436 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
6437 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
6438 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
6439 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
6440 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
6441 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
6442 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
6443 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6444 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6445 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
6446 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
6447 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
6448 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6449 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6450 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
6451 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
6452 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6453 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6454 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
6455 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
6456 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
6457 whitespace before "(".
6458
6459 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
6460
6461 * expr.h (extract_bit_field): Remove packedp parameter.
6462 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
6463 from forward declaration.
6464 (store_split_bit_field): Remove packedp arg from calls to
6465 extract_fixed_bit_field.
6466 (extract_bit_field_1): Remove packedp parameter and packedp
6467 argument from recursive calls and calls to extract_fixed_bit_field.
6468 (extract_bit_field): Remove packedp parameter and corresponding
6469 arg to extract_bit_field_1.
6470 (extract_fixed_bit_field): Remove packedp parameter. Remove code
6471 to issue warnings.
6472 (extract_split_bit_field): Remove packedp arg from call to
6473 extract_fixed_bit_field.
6474 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
6475 (copy_blkmode_from_reg): Likewise.
6476 (copy_blkmode_to_reg): Likewise.
6477 (read_complex_part): Likewise.
6478 (store_field): Likewise.
6479 (expand_expr_real_1): Likewise.
6480 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
6481 to extract_bit_field.
6482 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
6483 call to extract_bit_field.
6484 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
6485 call to extract_bit_field.
6486 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
6487 and special packedp behavior from -fstrict-volatile-bitfields
6488 documentation.
6489
6490 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6491
6492 * lra-eliminations.c (init_elim_table): Guard value_p.
6493
6494 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6495
6496 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
6497 DFmode, DImode, and SFmode in the upper VSX registers based on the
6498 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
6499 if -mpower8-vector. Combine -mvsx-timode handling with the rest
6500 of the VSX register handling.
6501
6502 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
6503 (f32_sv): Likewise.
6504 (zero_extendsidi2_lfiwzx): Add support for loading into the
6505 Altivec registers with -mpower8-vector. Use wu/wv constraints to
6506 only do VSX memory options on Altivec registers.
6507 (extendsidi2_lfiwax): Likewise.
6508 (extendsfdf2_fpr): Likewise.
6509 (mov<mode>_hardfloat, SF/SD modes): Likewise.
6510 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
6511 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
6512 (movdi_internal64): Likewise.
6513
6514 2013-09-27 Xinliang David Li <davidxl@google.com>
6515
6516 * opts.c (finish_options): Adjust parameters
6517 according to vect cost model.
6518 (common_handle_option): Set dynamic vect cost
6519 model for FDO.
6520 targhooks.c (default_add_stmt_cost): Compute stmt cost
6521 unconditionally.
6522 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
6523 Use helper function.
6524 * tree-vectorizer.h (unlimited_cost_model): New function.
6525 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
6526 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
6527 function.
6528 (vect_enhance_data_refs_alignment): Ditto.
6529 * flag-types.h: New enum.
6530 * common/config/i386/i386-common.c (ix86_option_init_struct):
6531 No need to initialize vect_cost_model flag.
6532 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
6533 unconditionally.
6534
6535 2013-09-27 Diego Novillo <dnovillo@google.com>
6536
6537 * gimple.h (enum ssa_mode): Remove.
6538
6539 2013-09-27 Paulo Matos <pmatos@broadcom.com>
6540
6541 * cfgloop.h (number_of_loops): Fix typo in check for null.
6542
6543 2013-09-27 Jakub Jelinek <jakub@redhat.com>
6544
6545 PR middle-end/58551
6546 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
6547 are children of outermost saved_cfun's loop, and set it up to
6548 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
6549 if loop != loop0 and SESE region contains bbs that belong to loop0.
6550
6551 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6552
6553 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
6554 (binary_scale_code_p, get_base_term, get_index_term): New functions.
6555 (set_address_segment, set_address_base, set_address_index)
6556 (set_address_disp): Accept the argument unconditionally.
6557 (baseness): Remove must_be_base_p and must_be_index_p checks.
6558 (decompose_normal_address): Classify as much as possible in the
6559 main loop.
6560
6561 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6562
6563 * cse.c (count_reg_usage): Handle INT_LIST.
6564 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
6565 * reginfo.c (reg_scan_mark_refs): Likewise.
6566 * reload1.c (eliminate_regs_1): Likewise.
6567
6568 2013-09-27 Iain Sandoe <iain@codesourcery.com>
6569
6570 PR middle-end/58547
6571 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
6572 signed.
6573
6574 2013-09-27 Richard Biener <rguenther@suse.de>
6575
6576 PR tree-optimization/58459
6577 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
6578 restriction not propagating into loops.
6579
6580 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
6581
6582 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
6583 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
6584 * doc/tree-ssa.texi (Walking use-def chains): Delete.
6585
6586 2013-09-26 Richard Biener <rguenther@suse.de>
6587
6588 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
6589
6590 2013-09-26 Richard Biener <rguenther@suse.de>
6591
6592 * alias.h (component_uses_parent_alias_set): Rename to ...
6593 (component_uses_parent_alias_set_from): ... this.
6594 * alias.c (component_uses_parent_alias_set): Rename to ...
6595 (component_uses_parent_alias_set_from): ... this and return
6596 the desired parent.
6597 (reference_alias_ptr_type_1): Use the result from
6598 component_uses_parent_alias_set_from instead of stripping
6599 components one at a time.
6600 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
6601
6602 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6603
6604 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
6605 Move prototypes to...
6606 * tree-ssa-ter.h: New File. Move prototypes here.
6607 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
6608 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
6609 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
6610 from is_replaceable_p.
6611 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
6612 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
6613 (process_replaceable): Use ter_is_replaceable_p.
6614 (find_replaceable_in_bb): Use ter_is_replaceable_p.
6615 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
6616 newly refactored ssa_is_replaceable_p.
6617 * cfgexpand.c: Include tree-outof-ssa.h.
6618 * ssaexpand.h: Delete.
6619
6620 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6621
6622 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
6623 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
6624 tree-ssa.c
6625 (create_gimple_tmp): Delete.
6626 (get_expr_type, build_assign, build_type_cast): Move to...
6627 * gimple-builder.c: New File.
6628 (get_expr_type): Relocate from gimple.c.
6629 (build_assign, build_type_cast): Change to only create ssanames.
6630 * gimple.h: Move prototypes to...
6631 * gimple-builder.h: New File. Here.
6632 * tree-ssa.h: And here.
6633 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
6634 count_uses_and_derefs): Relocate from gimple.c.
6635 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
6636 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
6637 * tree-ssa-math-opts (execute_cse_reciprocals): Use
6638 gimple_replace_ssa_lhs.
6639 * asan.c: Include gimple-builder.h.
6640 * Makefile.in: Add gimple-builder.o.
6641
6642 2013-09-26 Richard Biener <rguenther@suse.de>
6643
6644 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
6645 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
6646 (loe_visit_block): Use gcc_checking_assert.
6647 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
6648 gimple_assign_ssa_name_copy_p.
6649 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
6650 change.
6651
6652 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
6653
6654 * config/rs6000/t-rs6000 (rs6000.o): Remove.
6655 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
6656
6657 2013-09-26 Richard Biener <rguenther@suse.de>
6658
6659 PR tree-optimization/58539
6660 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
6661 the fact that debug statements are not taking part in loop-closed
6662 SSA construction.
6663
6664 2013-09-26 Nick Clifton <nickc@redhat.com>
6665
6666 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
6667 time warning message.
6668 (msp430_print_operand_raw): Delete unused letter parameter.
6669 (TARGET_PRINT_OPERAND_ADDRESS): Define.
6670 (msp430_print_operand_address): New function.
6671 (msp430_print_operand): Move address printing code from here to
6672 new function.
6673 * config/msp430/msp430.md (movsipsi2): Add comment in generated
6674 assembler.
6675 (zero_extendpsisi2): Likewise.
6676 (extendpsisi2): New pattern.
6677 (andneghi3): New pattern.
6678
6679 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
6680
6681 * config/aarch64/aarch64.opt (mlra): New option.
6682 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
6683 (TARGET_LRA_P): Define.
6684
6685 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
6686
6687 * expr.c (expand_assignment): Remove obsolete comment.
6688
6689 2013-09-25 Jeff Law <law@redhat.com>
6690
6691 * tree-flow.h (thread_through_all_blocks): Prototype moved into
6692 tree-ssa-threadupdate.h.
6693 (register_jump_thread): Similarly.
6694 * tree-ssa-threadupdate.h: New header file.
6695 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
6696 * tree-vrp.c: Likewise.
6697 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
6698 (thread_around_empty_blocks): Change type of path vector argument to
6699 an edge,type pair from just an edge. Initialize both elements when
6700 appending to a jump threading path. Tweak references to elements
6701 appropriately.
6702 (thread_across_edge): Similarly. Release memory for the elements
6703 as needed.
6704 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
6705 (dump_jump_thread_path): New function broken out from
6706 register_jump_thread.
6707 (register_jump_thread): Use dump_jump_thread_path. Change type of
6708 path vector entries. Search the path for NULL edges and dump
6709 the path if one is found. Tweak the conversion of path to 3-edge
6710 form to use the block copy type information embedded in the path.
6711
6712 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6713
6714 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
6715
6716 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6717 Vladimir Makarov <vmakarov@redhat.com>
6718
6719 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
6720 from the least significant bit.
6721 (strip_address_mutations): Add bitfield operations handling.
6722 (must_be_index_p): Add shifting and rotate operations handling.
6723 (set_address_base): Use must_be_base_p predicate.
6724 (set_address_index): Use must_be_index_p predicate.
6725
6726 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
6727 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6728 Sergey Lega <sergey.s.lega@intel.com>
6729 Anna Tikhonova <anna.tikhonova@intel.com>
6730 Ilya Tocar <ilya.tocar@intel.com>
6731 Andrey Turetskiy <andrey.turetskiy@intel.com>
6732 Ilya Verbin <ilya.verbin@intel.com>
6733 Kirill Yukhin <kirill.yukhin@intel.com>
6734 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6735
6736 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
6737 Use new names.
6738 (ix86_expand_vector_move_misalign): Support new unaligned load and
6739 stores and use new names.
6740 (CODE_FOR_sse2_storedqu): Rename to ...
6741 (CODE_FOR_sse2_storedquv16qi): ... this.
6742 (CODE_FOR_sse2_loaddqu): Rename to ...
6743 (CODE_FOR_sse2_loaddquv16qi): ... this.
6744 (CODE_FOR_avx_loaddqu256): Rename to ...
6745 (CODE_FOR_avx_loaddquv32qi): ... this.
6746 (CODE_FOR_avx_storedqu256): Rename to ...
6747 (CODE_FOR_avx_storedquv32qi): ... this.
6748 * config/i386/i386.md (fpint_logic): New.
6749 * config/i386/sse.md (VMOVE): Extend for AVX512.
6750 (VF): Ditto.
6751 (VF_128_256): New.
6752 (VF_512): Ditto.
6753 (VI_UNALIGNED_LOADSTORE): Ditto.
6754 (sse2_avx_avx512f): Ditto.
6755 (sse2_avx2): Extend for AVX512.
6756 (sse4_1_avx2): Ditto.
6757 (avx2_avx512f): New.
6758 (sse): Extend for AVX512.
6759 (sse2): Ditto.
6760 (sse4_1): Ditto.
6761 (avxsizesuffix): Ditto.
6762 (sseintvecmode): Ditto.
6763 (ssePSmode): Ditto.
6764 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
6765 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
6766 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6767 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
6768 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6769 (<sse2_avx_avx512f>_storedqu<mode): ... this.
6770 (<sse>_movnt<mode>): Replace constraint "x" with "v".
6771 (STORENT_MODE): Extend for AVX512.
6772 (*absneg<mode>2): Replace constraint "x" with "v".
6773 (*mul<mode>3): Ditto.
6774 (*ieee_smin<mode>3): Ditto.
6775 (*ieee_smax<mode>3): Ditto.
6776 (avx_cmp<mode>3): Replace VF with VF_128_256.
6777 (*<sse>_maskcmp<mode>3_comm): Ditto.
6778 (<sse>_maskcmp<mode>3): Ditto.
6779 (<sse>_andnot<mode>3): Extend for AVX512.
6780 (<code><mode>3, anylogic): Replace VF with VF_128_256.
6781 (<code><mode>3, fpint_logic): New.
6782 (*<code><mode>3): Extend for AVX512.
6783 (avx512flogicsuff): New.
6784 (avx512f_<logic><mode>): Ditto.
6785 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
6786 VF_128_256.
6787 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
6788 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
6789 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
6790 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
6791 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
6792 (xop_vpermil2<mode>3): Ditto.
6793 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
6794 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
6795 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
6796 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
6797
6798 2013-09-25 Tom Tromey <tromey@redhat.com>
6799
6800 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
6801 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
6802 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
6803 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
6804 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
6805 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
6806 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
6807 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
6808 (GRAPHITE_HTAB_H): Remove.
6809
6810 2013-09-25 Tom Tromey <tromey@redhat.com>
6811
6812 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
6813
6814 2013-09-25 Tom Tromey <tromey@redhat.com>
6815
6816 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
6817
6818 2013-09-25 Tom Tromey <tromey@redhat.com>
6819
6820 * config/i386/t-i386 (i386.o): Remove.
6821 (i386-c.o): Use COMPILE and POSTCOMPILE.
6822
6823 2013-09-25 Tom Tromey <tromey@redhat.com>
6824
6825 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
6826
6827 2013-09-25 Tom Tromey <tromey@redhat.com>
6828
6829 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
6830 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
6831 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
6832 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
6833 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
6834 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
6835 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
6836 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
6837 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
6838 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
6839 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
6840 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
6841 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
6842 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
6843 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
6844 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
6845 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
6846 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
6847 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
6848 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
6849 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
6850 (tree-ssa-pre.o, tree-ssa-sccvn.o)
6851 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
6852 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
6853 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
6854 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
6855 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
6856 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
6857 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
6858 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
6859 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
6860 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
6861 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
6862 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
6863 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
6864 (graphite-clast-to-gimple.o, graphite-dependences.o)
6865 (graphite-interchange.o, graphite-poly.o)
6866 (graphite-scop-detection.o, graphite-sese-to-poly.o)
6867 (graphite-optimize-isl.o, tree-vect-loop.o)
6868 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
6869 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
6870 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
6871 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
6872 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
6873 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
6874 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
6875 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
6876 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
6877 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
6878 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
6879 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
6880 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
6881 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
6882 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
6883 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
6884 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
6885 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
6886 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
6887 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
6888 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
6889 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
6890 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
6891 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
6892 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
6893 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
6894 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
6895 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
6896 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
6897 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
6898 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
6899 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
6900 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
6901 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
6902 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
6903 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
6904 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
6905 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
6906 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
6907 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
6908 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
6909 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
6910 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
6911 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
6912 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
6913 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
6914 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
6915 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
6916 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
6917 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
6918 (gcov-dump.o): Remove.
6919 (default-c.o): Use COMPILE and POSTCOMPILE.
6920 (CFLAGS-gcc.o): New variable.
6921 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
6922
6923 2013-09-25 Tom Tromey <tromey@redhat.com>
6924
6925 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
6926 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
6927 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
6928 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
6929 (c-family/c-pragma.o, c-family/c-pretty-print.o)
6930 (c-family/c-semantics.o, c-family/c-ada-spec.o)
6931 (c-family/array-notation-common.o, c-family/stub-objc.o)
6932 (c-family/c-ubsan.o): Remove.
6933
6934 2013-09-25 Tom Tromey <tromey@redhat.com>
6935
6936 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
6937
6938 2013-09-25 Tom Tromey <tromey@redhat.com>
6939
6940 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
6941 to add -DENABLE_SHARED_LIBGCC.
6942 (gcc.o): Don't use subshell.
6943
6944 2013-09-25 Tom Tromey <tromey@redhat.com>
6945
6946 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
6947 * configure.ac: Don't invoke AM_PROG_CC_C_O.
6948 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
6949 * configure, config.in: Rebuild.
6950
6951 2013-09-25 Tom Tromey <tromey@redhat.com>
6952
6953 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
6954 (COMPILE, POSTCOMPILE): New variables.
6955 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
6956 (DEPFILES): New variable.
6957 Include ".Po" files.
6958 * configure.ac: Add checks for dependency checking.
6959 * configure, aclocal.m4: Regenerate.
6960
6961 2013-09-25 Tom Tromey <tromey@redhat.com>
6962
6963 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
6964 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
6965
6966 2013-09-25 Tom Tromey <tromey@redhat.com>
6967
6968 * Makefile.in (generated_files): Add options.h,
6969 target-hooks-def.h, insn-opinit.h,
6970 common/common-target-hooks-def.h, pass-instances.def,
6971 c-family/c-target-hooks-def.h.
6972
6973 2013-09-25 Jeff Law <law@redhat.com>
6974
6975 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
6976 than foo[foo.length () - 1] to access last member in a vec.
6977 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
6978
6979 2013-09-25 Richard Biener <rguenther@suse.de>
6980
6981 PR middle-end/58521
6982 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
6983
6984 2013-09-25 Jan Hubicka <jh@suse.cz>
6985
6986 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
6987 test.
6988
6989 2013-09-25 Marek Polacek <polacek@redhat.com>
6990
6991 PR sanitizer/58420
6992 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
6993 when determining the type name.
6994
6995 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
6996
6997 * config/sh/sh.md: Fix formatting.
6998
6999 2013-09-24 Xinliang David Li <davidxl@google.com>
7000
7001 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
7002 max peel iterations parameter.
7003 * param.def: New parameter.
7004 * doc/invoke.texi: Document New parameter.
7005
7006 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
7007
7008 * gimple-pretty-print.c: Various whitespace tweaks.
7009 * tree-core.h: Likewise.
7010 * tree-pretty-print.c: Likewise.
7011 * tree-ssa-alias.c: Likewise.
7012 * tree-ssa-copy.c: Likewise.
7013 * tree-ssanames.c: Likewise.
7014 * tree-ssanames.h: Likewise.
7015 * tree-vrp.c: Likewise.
7016
7017 2013-09-24 Alan Modra <amodra@gmail.com>
7018
7019 PR middle-end/57134
7020 PR middle-end/57586
7021 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
7022 for output operands that disallow regs. Don't use EXPAND_WRITE on
7023 inout operands.
7024
7025 2013-09-24 Richard Biener <rguenther@suse.de>
7026
7027 PR middle-end/58513
7028 * tree.c (reference_alias_ptr_type): Move ...
7029 * alias.c (reference_alias_ptr_type): ... here and implement
7030 in terms of the new reference_alias_ptr_type_1.
7031 (ref_all_alias_ptr_type_p): New helper.
7032 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
7033 use ref_all_alias_ptr_type_p.
7034 (get_deref_alias_set): Add flag_strict_aliasing check here.
7035 (reference_alias_ptr_type_1): New function, split out from ...
7036 (get_alias_set): ... here.
7037 (alias_ptr_types_compatible_p): New function.
7038 * alias.h (reference_alias_ptr_type): Declare.
7039 (alias_ptr_types_compatible_p): Likewise.
7040 * tree.h (reference_alias_ptr_type): Remove.
7041 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
7042 to compare MEM_REF alias types.
7043
7044 2013-09-24 Richard Biener <rguenther@suse.de>
7045
7046 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
7047
7048 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7049
7050 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
7051 reload helper function arrays into a single array reg_addr.
7052 (reload_fpr_gpr): Likewise.
7053 (reload_gpr_vsx): Likewise.
7054 (reload_vsx_gpr): Likewise.
7055 (struct rs6000_reg_addr): Likewise.
7056 (reg_addr): Likewise.
7057 (rs6000_debug_reg_global): Change rs6000_vector_reload,
7058 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
7059 (rs6000_init_hard_regno_mode_ok): Likewise.
7060 (rs6000_secondary_reload_direct_move): Likewise.
7061 (rs6000_secondary_reload): Likewise.
7062
7063 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
7064 constraints: wu, ww, and wy. Repurpose wv constraint added during
7065 power8 changes. Put wg constraint in alphabetical order.
7066
7067 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
7068 for future work to add ISA 2.07 VSX single precision support.
7069 (-mvsx-scalar-double): Change default from -1 to 1, update
7070 documentation comment.
7071 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
7072 (-mupper-regs-df): New debug switch to control whether DF values
7073 can go in the traditional Altivec registers.
7074 (-mupper-regs-sf): New debug switch to control whether SF values
7075 can go in the traditional Altivec registers.
7076
7077 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
7078 and wy constraints.
7079 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
7080 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
7081 Add new constraints, wu/ww/wy. Repurpose wv constraint.
7082 (rs6000_debug_legitimate_address_p): Print if we are running
7083 before, during, or after reload.
7084 (rs6000_secondary_reload): Add a comment.
7085 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
7086
7087 * config/rs6000/constraints.md (wa constraint): Sort w<x>
7088 constraints. Update documentation string.
7089 (wd constraint): Likewise.
7090 (wf constraint): Likewise.
7091 (wg constraint): Likewise.
7092 (wn constraint): Likewise.
7093 (ws constraint): Likewise.
7094 (wt constraint): Likewise.
7095 (wx constraint): Likewise.
7096 (wz constraint): Likewise.
7097 (wu constraint): New constraint for ISA 2.07 SFmode scalar
7098 instructions.
7099 (ww constraint): Likewise.
7100 (wy constraint): Likewise.
7101 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
7102 in the previous submissions.
7103 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
7104
7105 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
7106
7107 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
7108 (REG_BR_PROB): Say that the probability is stored in an int_list.
7109 * reg-notes.def: Update commentary to mention INT_LIST.
7110 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
7111 (INT_LIST): New rtx.
7112 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
7113 * rtlanal.c (int_reg_note_p): New function.
7114 (alloc_reg_note): Assert that the note does not have an int argument.
7115 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
7116 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
7117 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
7118 rather than an INSN_LIST. Handle INT_LIST.
7119 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
7120 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
7121 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
7122 Manipulate them as ints rather than rtxes.
7123 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
7124 * regmove.c (copy_src_to_dest): Likewise.
7125 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
7126
7127 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
7128 into the cases that need it.
7129 * config/arm/arm.c (arm_unwind_emit): Likewise.
7130
7131 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
7132 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
7133 * loop-doloop.c (add_test, doloop_modify): Likewise.
7134 * loop-unswitch.c (compare_and_jump_seq): Likewise.
7135 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
7136 * predict.c (combine_predictions_for_insn): Likewise.
7137 * print-rtl.c (print_rtx): Handle INT_LIST.
7138 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
7139 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
7140 * config/arm/arm.c (emit_unlikely_jump): Likewise.
7141 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
7142 (ix86_split_fp_branch, predict_jump): Likewise.
7143 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
7144 * config/sh/sh.c (expand_cbranchsi4): Likewise.
7145 * config/spu/spu.c (ea_load_store_inline): Likewise.
7146
7147 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
7148 value of a REG_BR_PROB note.
7149 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
7150 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
7151 * emit-rtl.c (try_split): Likewise.
7152 * predict.c (br_prob_note_reliable_p): Likewise.
7153 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
7154 * reorg.c (mostly_true_jump): Likewise.
7155 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
7156 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
7157 * config/i386/i386.c (ix86_print_operand): Likewise.
7158 * config/ia64/ia64.c (ia64_print_operand): Likewise.
7159 * config/mmix/mmix.c (mmix_print_operand): Likewise.
7160 * config/rs6000/rs6000.c (output_cbranch): Likewise.
7161 * config/s390/s390.c (s390_expand_tbegin): Likewise.
7162 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
7163 * config/sparc/sparc.c (output_cbranch): Likewise.
7164 * config/spu/spu.c (get_branch_target): Likewise.
7165 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
7166 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
7167
7168 2013-09-23 Jan Hubicka <jh@suse.cz>
7169
7170 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
7171 for ipa-devirt.
7172 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
7173 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
7174 of external calls
7175 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
7176 (gimple_fold_call): Dump inconsistent devirtualizations; add
7177 sanity check for type based devirtualizations.
7178 * ipa-prop.c: Include ipa-utils.h
7179 (ipa_intraprocedural_devirtualization): Add sanity check.
7180 (try_make_edge_direct_virtual_call): Likewise.
7181
7182 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
7183
7184 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
7185 assignment statements.
7186
7187 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
7188
7189 * gimple-pretty-print.c (dump_ssaname_info): New function.
7190 (dump_gimple_phi): Call it.
7191 (pp_gimple_stmt_1): Likewise.
7192 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
7193 (range_info_def): Declare.
7194 * tree-pretty-print.c (pp_double_int): New function.
7195 (dump_generic_node): Call it.
7196 * tree-pretty-print.h (pp_double_int): Declare.
7197 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
7198 * tree-ssanames.h (range_info_def): New structure.
7199 (value_range_type): Move definition here.
7200 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
7201 Declare.
7202 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
7203 initialization.
7204 (set_range_info): New function.
7205 (get_range_info): Likewise.
7206 (duplicate_ssa_name_range_info): Likewise.
7207 (duplicate_ssa_name_fn): Check pointer type and call
7208 duplicate_ssa_name_range_info.
7209 * tree-ssa-copy.c (fini_copy_prop): Likewise.
7210 * tree-vrp.c (value_range_type): Remove definition, now in
7211 tree-ssanames.h.
7212 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
7213 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
7214 (SSA_NAME_RANGE_INFO): New macro.
7215
7216 2013-09-23 Richard Biener <rguenther@suse.de>
7217
7218 PR tree-optimization/58464
7219 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7220 (phi_trans_add): Change to add conditionally on being not present.
7221 (phi_translate_1): Remove recursion detection here.
7222 (phi_translate): Pre-seed the cache with NULL to catch
7223 recursion here in a more generic way.
7224 (bitmap_find_leader): Adjust comment.
7225 (get_representative_for): Dump value-numbers.
7226 (create_expression_by_pieces): Likewise.
7227 (insert_into_preds_of_block): Likewise.
7228
7229 2013-09-23 Christian Bruel <christian.bruel@st.com>
7230
7231 PR target/58475
7232 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
7233 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
7234
7235 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
7236
7237 Revert r202780:
7238 2013-09-20 Renlin Li <renlin.li@arm.com>
7239
7240 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
7241 plus_constant.
7242 (aarch64_expand_epilogue): Likewise.
7243 (aarch64_legitimize_reload_address): Likewise.
7244
7245 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
7246
7247 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
7248 NULL_TREE before pushing them onto the vector. Likewise for labels.
7249
7250 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
7251
7252 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
7253 comparison operators when -fno-trapping-math is in effect.
7254 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
7255 comparison operators in TFmode and assert that unsupported operators
7256 cannot reach here.
7257 (ia64_print_operand): Likewise.
7258
7259 2013-09-21 Jan Hubicka <jh@suse.cz>
7260
7261 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
7262 (sse_typeless_stores): Enable for core
7263 (sse_load0_by_pxor): Likewise.
7264 (four_jump_limit): Disable for core.
7265 (pad_returns): Likewise.
7266 (avoid_vector_decode): Likewise.
7267 (fuse_cmp_and_branch): Enable for cores.
7268 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
7269
7270 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7271
7272 PR middle-end/56791
7273 * config/pa/pa.c (pa_option_override): Disable auto increment and
7274 decrement instructions until reload is completed.
7275
7276 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
7277 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
7278 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
7279
7280 2013-09-20 DJ Delorie <dj@redhat.com>
7281 Nick Clifton <nickc@redhat.com>
7282
7283 * config/rl78/rl78.c: Various whitespace and comment tweaks.
7284 (need_to_save): Save bank 0 on interrupts.
7285 (characterize_address): Strip far address wrappers.
7286 (rl78_as_legitimate_address): Likewise.
7287 (transcode_memory_rtx): Likewise.
7288 (rl78_peep_movhi_p): Disable this peephole after devirt.
7289 (rl78_propogate_register_origins): Forget all origins when a
7290 CLOBBER is seen.
7291 * config/rl78/rl78-virt.md: Various whitespace tweaks.
7292 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
7293 peephole2's.
7294 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
7295 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
7296 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
7297 relocating.
7298 * config/rl78/constraints.md: Various whitespace and paren tweaks.
7299
7300 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7301
7302 * config/pa/pa.md: In "scc" insn patterns, change output template to
7303 handle const0_rtx in reg_or_0_operand operands.
7304
7305 2013-09-20 Martin Husemann <martin@NetBSD.org>
7306
7307 PR target/56875
7308 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
7309 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
7310
7311 2013-09-20 Richard Biener <rguenther@suse.de>
7312
7313 PR middle-end/58484
7314 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
7315 remembering SSA name version and block index.
7316 (new_scev_info_str): Adjust.
7317 (hash_scev_info): Likewise. Also hash the block index.
7318 (eq_scev_info): Adjust.
7319 (find_var_scev_info): Likewise.
7320 (struct instantiate_cache_entry): Remove.
7321 (struct instantiate_cache_type): Use a htab to map name, block
7322 to chrec.
7323 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7324 (get_instantiated_value_entry): Likewise.
7325 (hash_idx_scev_info, eq_idx_scev_info): New functions.
7326 (instantiate_scev_name): Adjust.
7327
7328 2013-09-20 Jeff Law <law@redhat.com>
7329
7330 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
7331
7332 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
7333
7334 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
7335 Call aarch64_simd_expand_args to update op[argc].
7336
7337 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7338
7339 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
7340 plugin argument.
7341
7342 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7343
7344 * gengtype.c (file_rules): Added rule for *.cc files.
7345 (get_output_file_with_visibility): Give fatal message when no
7346 rules found.
7347
7348 2013-09-20 Renlin Li <renlin.li@arm.com>
7349
7350 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
7351 (aarch64_expand_epilogue): Likewise.
7352 (aarch64_legitimize_reload_address): Likewise.
7353
7354 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
7355
7356 PR middle-end/57748
7357 * expr.c (expand_assignment): Remove misalignp code path.
7358
7359 2013-09-20 Marek Polacek <polacek@redhat.com>
7360
7361 PR sanitizer/58413
7362 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
7363 TYPE_PRECISION. Add asserts.
7364
7365 2013-09-20 Richard Biener <rguenther@suse.de>
7366
7367 PR tree-optimization/58453
7368 * tree-loop-distribution.c (distribute_loop): Apply the cost
7369 model for -ftree-loop-distribute-patterns, too.
7370
7371 2013-09-20 Richard Biener <rguenther@suse.de>
7372
7373 PR middle-end/58473
7374 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
7375 make type comparison less strict.
7376
7377 2013-09-20 Alan Modra <amodra@gmail.com>
7378
7379 * configure: Regenerate.
7380 * aclocal.m4: Regenerate.
7381
7382 2013-09-20 Marek Polacek <polacek@redhat.com>
7383
7384 PR other/58467
7385 * doc/extend.texi: Document that attribute used is meant to be used
7386 on variables with static storage duration.
7387
7388 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7389
7390 PR tree-optimization/58472
7391 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
7392 simd_lane_access set inv_p = false.
7393 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
7394 the simduid magic VAR_DECL.
7395
7396 2013-09-19 Jan Hubicka <jh@suse.cz>
7397
7398 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
7399
7400 2013-09-17 Jeff Law <law@redhat.com>
7401
7402 * tree-ssa-dom.c (record_temporary_equivalences): New function
7403 split out of dom_opt_dom_walker::after_dom_children.
7404 (dom_opt_dom_walker::thread_across_edge): Move common code
7405 in here from dom_opt_dom_walker::after_dom_children.
7406 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
7407
7408 2013-09-19 Jan Hubicka <jh@suse.cz>
7409
7410 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
7411 (TARGET_GENERIC): Use PROCESOR_GENERIC
7412 (enum processor_type): Unify generic32 and 64.
7413 * i386.md (cpu): Likewise.
7414 * x86-tune.def (use_leave): Enable for generic32.
7415 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
7416 * athlon.md: Change generic64 to generic in all occurences.
7417 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
7418 (ix86_target_macros_internal): Likewise.
7419 * driver-i386.c (host_detect_local_cpu): Likewise.
7420 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
7421 to ..
7422 (generic_memcpy, generic_memset, generic_cost): This one.
7423 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
7424 (m_GENERIC32, m_GENERIC64): Remove.
7425 (m_GENERIC): Turn into one flag.
7426 (processor_target): Unify generic tunnings.
7427 (ix86_option_override_internal): Replace generic32/64 by generic.
7428 (ix86_issue_rate): Likewise.
7429 (ix86_adjust_cost): Likewise.
7430
7431 2013-09-19 Jan Hubicka <jh@suse.cz>
7432
7433 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
7434 of speculative flag.
7435
7436 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7437
7438 * omp-low.c (expand_omp_sections): Always pass len - 1 to
7439 GOMP_sections_start, even if !exit_reachable.
7440
7441 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7442
7443 * lra-constraints.c (need_for_all_save_p): Use macro
7444 HARD_REGNO_CALL_PART_CLOBBERED.
7445 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
7446 set up pseudo conflict hard regs.
7447
7448 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7449
7450 PR target/58452
7451 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
7452 operands.
7453
7454 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7455
7456 PR rtl-optimization/58438
7457 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
7458 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
7459 from previous subpasses.
7460
7461 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
7462
7463 * arm.c (arm_get_frame_offsets): Validate architecture supports
7464 LDRD/STRD before accepting the tuning preference.
7465 (arm_expand_prologue): Likewise.
7466 (arm_expand_epilogue): Likewise.
7467
7468 2013-09-18 Richard Biener <rguenther@suse.de>
7469
7470 PR tree-optimization/58417
7471 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
7472 have chrecs with symbols defined in the loop as operands.
7473 (chrec_fold_multiply): Likewise.
7474 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
7475 parameters before folding binary operations.
7476 (struct instantiate_cache_entry_hasher): Remove.
7477 (struct instantiate_cache_type): Use a pointer-map.
7478 (instantiate_cache_type::instantiate_cache_type): New function.
7479 (instantiate_cache_type::get): Likewise.
7480 (instantiate_cache_type::set): Likewise.
7481 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7482 (get_instantiated_value_entry): Likewise.
7483 (global_cache): New global.
7484 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
7485 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
7486 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
7487 (instantiate_scev_name): Adjust.
7488 (instantiate_scev): Construct global instead of local cache.
7489 (resolve_mixers): Likewise.
7490
7491 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
7492 Paolo Carlini <paolo.carlini@oracle.com>
7493
7494 PR c++/58458
7495 * doc/implement-cxx.texi: Fix references to the C++ standards.
7496
7497 2013-09-18 Jakub Jelinek <jakub@redhat.com>
7498
7499 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
7500 * tree-vect-data-refs.c (vect_analyze_data_refs): For
7501 simd_lane_access drs, update also DR_ALIGNED_TO.
7502
7503 2013-09-18 Marek Polacek <polacek@redhat.com>
7504
7505 PR sanitizer/58411
7506 * doc/extend.texi: Document no_sanitize_undefined attribute.
7507 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
7508 no_sanitize_undefined attribute.
7509
7510 2013-09-18 Nick Clifton <nickc@redhat.com>
7511
7512 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
7513 (ASM_DECLARE_FUNCTION_NAME): Define.
7514
7515 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
7516
7517 * compare-elim.c (find_comparison_dom_walker): New class
7518 (find_comparisons_in_bb): Rename to
7519 find_comparison_dom_walker::before_dom_children
7520 (find_comparisons): Adjust
7521 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
7522 adjust.
7523 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
7524 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
7525 (init_walk_dominator_tree): Remove declaration.
7526 (fini_walk_dominator_tree): Remove declaration.
7527 * fwprop.c (single_def_use_dom_walker): New class
7528 (single_def_use_enter_block): Convert to
7529 single_def_use_dom_walker::before_dom_children.
7530 (single_def_use_leave_block): Convert to
7531 single_def_use_dom_walker::after_dom_children.
7532 (build_single_def_use_links): Adjust.
7533 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
7534 class.
7535 (find_candidates_in_block): Convert to
7536 find_candidates_dom_walker::before_dom_children.
7537 (execute_strength_reduction): Adjust.
7538 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
7539 (sese_dom_walker): New class.
7540 (sese_dom_walker::sese_dom_walker): New constructor.
7541 (sese_dom_walker::~sese_dom_walker): New destructor.
7542 (build_sese_conditions_before): Convert to
7543 sese_dom_walker::before_dom_children.
7544 (build_sese_conditions_after): Convert to
7545 sese_dom_walker::after_dom_children.
7546 (build_poly_scop): Adjust
7547 * tree-into-ssa.c (rewrite_dom_walker): New class
7548 (rewrite_enter_block): Convert to
7549 rewrite_dom_walker::before_dom_children.
7550 (rewrite_leave_block): Convert to
7551 rewrite_dom_walker::after_dom_children.
7552 (rewrite_update_dom_walker): New class.
7553 (rewrite_update_enter_block): Convert to
7554 rewrite_update_dom_walker::before_dom_children.
7555 (rewrite_update_leave_block): Convert to
7556 rewrite_update_dom_walker::after_dom_children.
7557 (rewrite_blocks, rewrite_into_ssa): Adjust.
7558 (mark_def_dom_walker): New class.
7559 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
7560 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
7561 (mark_def_sites_blocks): Convert to
7562 mark_def_dom_walker::before_dom_children.
7563 (mark_def_site_blocks): Remove.
7564 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
7565 (tree_ssa_dominator_optimize): Adjust.
7566 (dom_thread_across_edge): Convert to method
7567 dom_opt_dom_walker::thread_across_edge.
7568 (dom_opt_enter_block): Convert to member function
7569 dom_opt_dom_walker::before_dom_children.
7570 (dom_opt_leave_block): Convert to member function
7571 dom_opt_dom_walker::after_dom_children.
7572 * tree-ssa-dse.c (dse_dom_walker): New class.
7573 (dse_enter_block): Convert to member function
7574 dse_dom_walker::before_dom_children.
7575 (tree_ssa_dse): Adjust.
7576 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
7577 (determine_invariantness_stmt): Convert to method
7578 invariantness_dom_walker::before_dom_children.
7579 (determine_invariantness): Remove
7580 (move_computations_dom_walker): New class.
7581 (move_computations_stmt): Convert to method
7582 move_computations_dom_walker::before_dom_children.
7583 (move_computations, tree_ssa_lim): Adjust.
7584 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
7585 (nt_init_block): Convert to method
7586 notrappping_dom_walker::before_dom_children.
7587 (nt_fini_block): Convert to method
7588 method nontrapping_dom_walker::after_dom_children.
7589 (get_non_trapping): Adjust.
7590 * tree-ssa-pre.c (eliminate_dom_walker): New class.
7591 (eliminate_bb): Convert to method
7592 eliminate_dom_walker::before_dom_children.
7593 (eliminate_leave_block): Convert to method
7594 eliminate_dom_walker::after_dom_children.
7595 (eliminate): Adjust.
7596 * tree-ssa-strlen.c (strlen_dom_walker): New class.
7597 (strlen_enter_block): Convert to method
7598 strlen_dom_walker::before_dom_children.
7599 (strlen_leave_block): Convert to method
7600 method strlen_dom_walker::after_dom_children.
7601 (tree_ssa_strlen): Adjust.
7602 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
7603 (tree_ssa_uncprop): Adjust.
7604 (uncprop_leave_block): Convert to method
7605 uncprop_dom_walker::after_dom_children.
7606 (uncprop_leave_block): Convert to method
7607 uncprop_dom_walker::before_dom_children.
7608
7609 2013-09-18 Bin Cheng <bin.cheng@arm.com>
7610
7611 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
7612 branch in same basic block. Check both src and dest of the move insn.
7613
7614 2013-09-17 Nick Clifton <nickc@redhat.com>
7615
7616 * config/rl78/rl78-real.md (bf): New pattern.
7617 (bt): New pattern.
7618 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
7619 (rl78_print_operand): Do not put a # before a %B.
7620 * config/rl78/rl78.opt: Tweak doc strings.
7621
7622 2013-09-17 DJ Delorie <dj@redhat.com>
7623
7624 * config/rl78/constraints.md (Wcv): Allow up to $r31.
7625 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
7626 (rl78_option_override): Likewise, if -mallregs.
7627 (is_virtual_register): Likewise.
7628 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
7629 (REGNO_OK_FOR_BASE_P): Likewise.
7630 * config/rl78/rl78.opt (-mallregs): New.
7631
7632 2013-09-17 Nick Clifton <nickc@redhat.com>
7633
7634 * config/rl78/rl78.c (need_to_save): Change return type to bool.
7635 For interrupt functions: save all call clobbered registers if the
7636 interrupt handler is not a leaf function.
7637 (rl78_expand_prologue): Always recompute the frame information.
7638 For interrupt functions: only select bank 0 if one of the bank 0
7639 registers is going to be psuhed.
7640
7641 2013-09-17 DJ Delorie <dj@redhat.com>
7642
7643 * config/rl78/constraints.md: For each W* constraint, rename to C*
7644 and create a W* constraint that checks for an optional ES: prefix
7645 pattern also.
7646 * config/rl78/rl78.md (UNS_ES_ADDR): New.
7647 (es_addr): New. Used to wrap far addresses.
7648 * config/rl78/rl78-protos.h (rl78_es_addr): New.
7649 (rl78_es_base): New.
7650 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
7651 wrapped far addresses.
7652 (rl78_print_operand_1): Unwrap far addresses before processing.
7653 (rl78_lo16): Wrap far addresses in unspecs.
7654 (rl78_es_addr): New.
7655 (rl78_es_base): New.
7656 (insn_ok_now): Check for not-yet-wrapped far addresses.
7657 (transcode_memory_rtx): Properly re-wrap far addresses.
7658
7659 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
7660
7661 * config/sparc/t-rtems: Add leon3 multilibs.
7662
7663 2013-09-17 Cong Hou <congh@google.com>
7664
7665 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
7666 when checking the dot production pattern. The type of rhs operand
7667 of multiply is now checked correctly.
7668
7669 2013-09-17 Jeff Law <law@redhat.com>
7670
7671 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
7672 edge implied equivalences into successor phis.
7673 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
7674 here from tree-ssa-threadedge.c.
7675 (mark_threaded_blocks): When threading through a joiner, if both
7676 successors of the joiner's clone reach the same block, verify the
7677 PHI arguments are equal. If not, cancel the jump threading request.
7678 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
7679 tree-ssa-threadupdate.c
7680 (thread_across_edge): Don't check PHI argument equality when
7681 threading through joiner block here.
7682
7683 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
7684
7685 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
7686 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
7687 (warn_uninit, warn_uninitialized_vars,
7688 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7689 Move to tree-ssa-uninit.c.
7690 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
7691 (has_undefined_value_p): New. Pass dependant parts of
7692 ssa_undefined_value_p.
7693 (uninit_undefined_value_p): Use has_undefined_value_p.
7694 (warn_uninit, warn_uninitialized_vars,
7695 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7696 Move from tree-ssa.c.
7697 * tree-ssa.h: Adjust prototypes.
7698
7699 2013-09-17 Jan Hubicka <jh@suse.cz>
7700
7701 PR middle-end/58332
7702 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
7703 * ipa-inline.c (can_inline_edge_p): Do not downgrade
7704 FUNCTION_NOT_OPTIMIZED.
7705 * ipa-inline-analysis.c (compute_inline_parameters): Function
7706 not optimized is not inlinable unless it is alwaysinline.
7707 (inline_analyze_function): Force calls in not optimized
7708 function not inlinable.
7709
7710 2013-09-17 Jan Hubicka <jh@suse.cz>
7711
7712 PR middle-end/58329
7713 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
7714 to return NULL.
7715 * ipa.c (function_and_variable_visibility): Likewise.
7716 * ipa-profile.c (ipa_profile): Likewise.
7717
7718 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
7719
7720 PR ipa/58398
7721 * cgraph.c (cgraph_function_body_availability): Check for ifunc
7722 attribute, and don't inline the resolver in this case.
7723
7724 2013-09-17 Teresa Johnson <tejohnson@google.com>
7725
7726 * coverage.c (get_coverage_counts): Add missing newline.
7727
7728 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7729
7730 PR tree-optimization/58088
7731 * fold-const.c (mask_with_trailing_zeros): New function.
7732 (fold_binary_loc): Make sure we don't recurse infinitely
7733 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
7734 Use mask_with_trailing_zeros where appropriate.
7735
7736 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
7737
7738 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
7739 of 'prev' var to get better distance estimation.
7740
7741 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
7742
7743 * tree-inline.h (struct copy_body_data): Add transform_parameter.
7744 * tree-inline.c (is_parameter_of): New predicate.
7745 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
7746 a parameter has been remapped.
7747 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
7748 (optimize_inline_calls): Initialize transform_parameter.
7749 (copy_gimple_seq_and_replace_locals): Likewise.
7750 (tree_function_versioning): Likewise.
7751 (maybe_inline_call_in_expr): Likewise.
7752
7753 2013-09-17 Nick Clifton <nickc@redhat.com>
7754
7755 * config/msp430/msp430-protos.h: Add prototypes for new functions.
7756 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
7757 interrupt handlers.
7758 (is_attr_func): New function.
7759 (msp430_is_interrupt_func): New function.
7760 (is_naked_func): New function.
7761 (is_reentrant_func): New function.
7762 (is_critical_func): New function.
7763 (msp430_start_function): Add annotations for function attributes.
7764 (msp430_attr): New function.
7765 (msp430_attribute_table): New.
7766 (msp430_function_section): New function.
7767 (TARGET_ASM_FUNCTION_SECTION): Define.
7768 (msp430_builtin): New enum.
7769 (msp430_init_builtins): New function.
7770 (msp430_builtin_devl): New function.
7771 (msp430_expand_builtin): New function.
7772 (TARGET_INIT_BUILTINS): Define.
7773 (TARGET_EXPAND_BUILTINS): Define.
7774 (TARGET_BUILTIN_DECL): Define.
7775 (msp430_expand_prologue): Add support for naked, interrupt,
7776 critical and reentrant functions.
7777 (msp430_expand_epilogue): Likewise.
7778 (msp430_print_operand): Handle 'O' character.
7779 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
7780 NO_TRAMPOLINES.
7781 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
7782 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
7783 (pushm): Use a 'n' rather than an 'i' constraint.
7784 (msp_return): Add generation of the interrupt return instruction.
7785 (disable_interrupts): New pattern.
7786 (enable_interrupts): New pattern.
7787 (push_intr_state): New pattern.
7788 (pop_intr_state): New pattern.
7789 (bic_SR): New pattern.
7790 (bis_SR): New pattern.
7791 * doc/extend.texi: Document MSP430 function attributes and builtin
7792 functions.
7793
7794 2013-09-17 Richard Biener <rguenther@suse.de>
7795
7796 PR tree-optimization/58432
7797 * tree-loop-distribution.c (tree_loop_distribution): Also
7798 scan PHIs for outside loop uses and seed a partition from them.
7799
7800 2013-09-17 Bin Cheng <bin.cheng@arm.com>
7801
7802 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
7803 (restructure_reference): Call backtrace_base_for_ref.
7804
7805 2013-09-17 Alan Modra <amodra@gmail.com>
7806
7807 PR target/57589
7808 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
7809 patch.
7810
7811 2013-09-16 DJ Delorie <dj@redhat.com>
7812
7813 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
7814 vregs location for RL78/G10.
7815 (rl78_expand_prologue): Avoid SEL on G10.
7816 (rl78_expand_epilogue): Likewise.
7817 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
7818 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
7819 __RL78_G10__ when appropriate.
7820 (ASM_SPEC): Pass -mg10 along to the assembler.
7821 * config/rl78/rl78.md (sel_rb): Disable for G10.
7822 * config/rl78/rl78.opt: Add -mg10 option.
7823 * config/rl78/t-rl78: Add -mg10 multilib.
7824
7825 2013-09-16 Xinliang David Li <davidxl@google.com>
7826
7827 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
7828 * omp-low.c (omp_max_vf): Ditto.
7829 (expand_omp_simd): Ditto.
7830 * tree-vectorizer.c (vectorize_loops): Ditto.
7831 (gate_vect_slp): Ditto.
7832 (gate_increase_alignment): Ditto.
7833 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
7834 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
7835 (gate_tree_vectorize): Name change.
7836 (tree_vectorize): Ditto.
7837 (pass_vectorize::gate): Call new function.
7838 (pass_vectorize::execute): Ditto.
7839 * opts.c: O3 default setting change.
7840 (finish_options): Check new flag.
7841 * doc/invoke.texi: Document new flags.
7842 * common.opt: New flags.
7843
7844 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
7845
7846 * doc/tm.texi.in (Cond Exec Macros): Remove node.
7847 (Condition Code): Don't reference it.
7848 * doc/tm.texi: Regenerate.
7849
7850 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
7851
7852 PR middle-end/58418
7853 * lra-constraints.c (undo_optional_reloads): Consider all optional
7854 reload even if it did not get a hard reg.
7855
7856 2013-09-16 Teresa Johnson <tejohnson@google.com>
7857
7858 * dumpfile.c (dump_loc): Remove newline emission.
7859 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
7860 emission to dump_printf_loc calls where missing.
7861 (vect_mark_for_runtime_alias_test): Ditto.
7862 (vect_analyze_data_ref_dependence): Ditto.
7863 (vect_analyze_data_ref_dependences): Ditto.
7864 (vect_slp_analyze_data_ref_dependence): Ditto.
7865 (vect_slp_analyze_data_ref_dependences): Ditto.
7866 (vect_compute_data_ref_alignment): Ditto.
7867 (vect_update_misalignment_for_peel): Ditto.
7868 (vect_verify_datarefs_alignment): Ditto.
7869 (vector_alignment_reachable_p): Ditto.
7870 (vect_get_data_access_cost): Ditto.
7871 (vect_enhance_data_refs_alignment): Ditto.
7872 (vect_find_same_alignment_drs): Ditto.
7873 (vect_analyze_data_refs_alignment): Ditto.
7874 (vect_analyze_group_access): Ditto.
7875 (vect_analyze_data_ref_access): Ditto.
7876 (vect_analyze_data_ref_accesses): Ditto.
7877 (vect_prune_runtime_alias_test_list): Ditto.
7878 (vect_analyze_data_refs): Ditto.
7879 (vect_create_addr_base_for_vector_ref): Ditto.
7880 (vect_create_data_ref_ptr): Ditto.
7881 (vect_grouped_store_supported): Ditto.
7882 (vect_grouped_load_supported): Ditto.
7883 * value-prof.c (check_counter): Ditto.
7884 (check_ic_target): Ditto.
7885 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
7886 (vect_recog_widen_mult_pattern): Ditto.
7887 (vect_recog_widen_sum_pattern): Ditto.
7888 (vect_recog_over_widening_pattern): Ditto.
7889 (vect_recog_widen_shift_pattern): Ditto.
7890 (vect_recog_rotate_pattern): Ditto.
7891 (vect_recog_vector_vector_shift_pattern): Ditto.
7892 (vect_recog_divmod_pattern): Ditto.
7893 (vect_recog_mixed_size_cond_pattern): Ditto.
7894 (vect_recog_bool_pattern): Ditto.
7895 (vect_pattern_recog_1): Ditto.
7896 (vect_pattern_recog): Ditto.
7897 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
7898 (vect_is_simple_iv_evolution): Ditto.
7899 (vect_analyze_scalar_cycles_1): Ditto.
7900 (vect_get_loop_niters): Ditto.
7901 (vect_analyze_loop_1): Ditto.
7902 (vect_analyze_loop_form): Ditto.
7903 (vect_analyze_loop_operations): Ditto.
7904 (vect_analyze_loop_2): Ditto.
7905 (vect_analyze_loop): Ditto.
7906 (report_vect_op): Ditto.
7907 (vect_is_slp_reduction): Ditto.
7908 (vect_is_simple_reduction_1): Ditto.
7909 (vect_get_known_peeling_cost): Ditto.
7910 (vect_estimate_min_profitable_iters): Ditto.
7911 (vect_model_reduction_cost): Ditto.
7912 (vect_model_induction_cost): Ditto.
7913 (get_initial_def_for_induction): Ditto.
7914 (vect_create_epilog_for_reduction): Ditto.
7915 (vectorizable_reduction): Ditto.
7916 (vectorizable_induction): Ditto.
7917 (vectorizable_live_operation): Ditto.
7918 (vect_loop_kill_debug_uses): Ditto.
7919 (vect_transform_loop): Ditto.
7920 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
7921 (vect_stmt_relevant_p): Ditto.
7922 (process_use): Ditto.
7923 (vect_mark_stmts_to_be_vectorized): Ditto.
7924 (vect_model_simple_cost): Ditto.
7925 (vect_model_promotion_demotion_cost): Ditto.
7926 (vect_model_store_cost): Ditto.
7927 (vect_get_store_cost): Ditto.
7928 (vect_model_load_cost): Ditto.
7929 (vect_get_load_cost): Ditto.
7930 (vect_init_vector_1): Ditto.
7931 (vect_get_vec_def_for_operand): Ditto.
7932 (vect_finish_stmt_generation): Ditto.
7933 (vectorizable_call): Ditto.
7934 (vectorizable_conversion): Ditto.
7935 (vectorizable_assignment): Ditto.
7936 (vectorizable_shift): Ditto.
7937 (vectorizable_operation): Ditto.
7938 (vectorizable_store): Ditto.
7939 (vectorizable_load): Ditto.
7940 (vectorizable_condition): Ditto.
7941 (vect_analyze_stmt): Ditto.
7942 (vect_transform_stmt): Ditto.
7943 (vect_is_simple_use): Ditto.
7944 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
7945 (vect_can_advance_ivs_p): Ditto.
7946 (vect_update_ivs_after_vectorizer): Ditto.
7947 (vect_do_peeling_for_loop_bound): Ditto.
7948 (vect_gen_niters_for_prolog_loop): Ditto.
7949 (vect_update_inits_of_drs): Ditto.
7950 (vect_create_cond_for_alias_checks): Ditto.
7951 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
7952 (vect_build_slp_tree_1): Ditto.
7953 (vect_supported_load_permutation_p): Ditto.
7954 (vect_analyze_slp_instance): Ditto.
7955 (vect_analyze_slp): Ditto.
7956 (vect_make_slp_decision): Ditto.
7957 (vect_detect_hybrid_slp): Ditto.
7958 (vect_bb_vectorization_profitable_p): Ditto.
7959 (vect_slp_analyze_bb_1): Ditto.
7960 (vect_update_slp_costs_according_to_vf): Ditto.
7961 (vect_get_mask_element): Ditto.
7962 (vect_transform_slp_perm_load): Ditto.
7963 (vect_schedule_slp_instance): Ditto.
7964 (vect_schedule_slp): Ditto.
7965 (vect_slp_transform_bb): Ditto.
7966 * profile.c (read_profile_edge_counts): Ditto.
7967 (compute_branch_probabilities): Ditto.
7968 * coverage.c (get_coverage_counts): Ditto.
7969
7970 2013-09-16 Diego Novillo <dnovillo@google.com>
7971
7972 * tree-core.h: Add missing comment lines from refactoring of tree.h.
7973
7974 2013-09-16 Jan Hubicka <jh@suse.cz>
7975
7976 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
7977 abstract functions; for static functions check the presence of body.
7978
7979 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7980
7981 * config/aarch64/aarch64-simd-builtins.def (fma): New.
7982 * config/aarch64/aarch64-simd.md
7983 (aarch64_mla_elt<mode>): New.
7984 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7985 (aarch64_mls_elt<mode>): Likewise.
7986 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7987 (aarch64_fma4_elt<mode>): Likewise.
7988 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7989 (aarch64_fma4_elt_to_128v2df): Likewise.
7990 (aarch64_fma4_elt_to_64df): Likewise.
7991 (fnma<mode>4): Likewise.
7992 (aarch64_fnma4_elt<mode>): Likewise.
7993 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7994 (aarch64_fnma4_elt_to_128v2df): Likewise.
7995 (aarch64_fnma4_elt_to_64df): Likewise.
7996 * config/aarch64/iterators.md (VDQSF): New.
7997 * config/aarch64/arm_neon.h
7998 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
7999 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
8000
8001 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8002
8003 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
8004 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
8005 (aarch64_mul3_elt_to_128df): Likewise.
8006 (aarch64_mul3_elt_to_64v2df): Likewise.
8007 * config/aarch64/iterators.md (VEL): Also handle DFmode.
8008 (VMUL): New.
8009 (VMUL_CHANGE_NLANES) Likewise.
8010 (h_con): Likewise.
8011 (f): Likewise.
8012 * config/aarch64/arm_neon.h
8013 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
8014
8015 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8016
8017 * config/aarch64/arm_neon.h
8018 (vcvtx_high_f32_f64): Fix parameters.
8019
8020 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8021 Uros Bizjak <ubizjak@gmail.com>
8022
8023 * config/alpha.c: Include tree-ssanames.h.
8024
8025 2013-09-16 Richard Biener <rguenther@suse.de>
8026
8027 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
8028 (dot_rdg_1): Handle control_dd.
8029 (create_edge_for_control_dependence): New function.
8030 (create_rdg_edges): Add control dependences if asked for.
8031 (build_rdg): Likewise.
8032 (generate_loops_for_partition): If there are not necessary
8033 control stmts remove all their dependencies.
8034 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
8035 (distribute_loop): Pass on control dependences.
8036 (tree_loop_distribution): Compute control dependences and remove
8037 restriction on number of loop nodes.
8038
8039 2013-09-16 Jakub Jelinek <jakub@redhat.com>
8040
8041 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
8042 for internal calls.
8043
8044 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
8045
8046 * cse.c (try_const_anchors): Punt on CC modes.
8047
8048 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8049
8050 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
8051
8052 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
8053
8054 PR target/58382
8055 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
8056 calls to word_mode.
8057
8058 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8059
8060 PR target/48094
8061 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
8062 seen.
8063 (darwin_objc1_section): Likewise.
8064 (darwin_file_end): Emit Image Info section when required.
8065
8066 2013-09-14 Jan Hubicka <jh@suse.cz>
8067
8068 * tree-into-ssa.c (gate_into_ssa): New.
8069 (pass_data_build_ssa): Use it.
8070 * cgraph.h (expand_thunk): Update prototype.
8071 * cgraphunit.c (analyze_function): Expand thunks early.
8072 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
8073 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
8074 set lowered flag; do not add new function.
8075 (assemble_thunks_and_aliases): Update.
8076 * tree-ssa.c (gate_init_datastructures): New gate.
8077 (pass_data_init_datastructures): Use it.
8078
8079 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8080
8081 PR target/58269
8082 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
8083 xmm register set described in the psABI.
8084
8085 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
8086
8087 * dwarf2out.c (should_emit_struct_debug): Add check
8088 for type_decl variable is not NULL.
8089
8090 2013-09-13 Jacek Caban <jacek@codeweavers.com>
8091
8092 * config.gcc: Use new winnt-c.c target hooks
8093 * config/t-winnt: New file
8094 * config/winnt-c.c: New file
8095 * doc/tm.texi.in: Document new hook
8096 * doc/tm.texi: Regenerated
8097
8098 2013-09-13 Jan Hubicka <jh@suse.cz>
8099
8100 PR middle-end/58094
8101 * ipa-inline.c (check_callers): New function.
8102 (check_caller_edge): Remove.
8103 (want_inline_function_to_all_callers_p): Also permit alises that are
8104 called dirrectly.
8105 (inline_to_all_callers): Terminate the walk when devirtualization
8106 introduce new calls.
8107
8108 2013-09-13 Jan Hubicka <jh@suse.cz>
8109
8110 * ipa-inline-analysis.c (struct growth_data): Add node.
8111 (do_estimate_growth_1): Fix detection of recursion.
8112
8113 2013-09-13 Jakub Jelinek <jakub@redhat.com>
8114
8115 PR tree-optimization/58392
8116 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
8117 to avoid shadowing of outer loop variable. If
8118 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
8119 replace_by_duplicate_decl simduid of loops that have it set and
8120 set dest_cfun->has_simduid_loops and/or
8121 dest_cfun->has_force_vect_loops.
8122 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
8123 instead of maybe_lookup_decl.
8124 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
8125 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
8126 call. Copy over force_vect and copy and remap simduid. Set
8127 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
8128 (copy_cfg_body): Remove blocks_to_copy argument. Use
8129 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
8130 caller. Don't set cfun->has_simduid_loops and/or
8131 cfun->has_force_vect_loops here.
8132 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
8133 caller.
8134 (expand_call_inline, tree_function_versioning): Adjust copy_body
8135 callers.
8136
8137 2013-09-13 Martin Jambor <mjambor@suse.cz>
8138
8139 PR bootstrap/58388
8140 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
8141 the assert if the edge was a speculative one.
8142
8143 2013-09-13 Richard Biener <rguenther@suse.de>
8144
8145 * tree-data-ref.h (known_dependences_p): Move here ...
8146 * tree-loop-distribution.c (known_dependences_p): ... from here.
8147 (dump_rdg_component, debug_rdg_component): Remove.
8148 (dump_rdg): Adjust.
8149 (generate_loops_for_partition): Use gimple_uid instead of
8150 relying on matching stmt visit order.
8151 (rdg_build_partitions): Take starting stmt vector.
8152 (ldist_gen): Merge into ...
8153 (distribute_loop): ... this function. Do not compute starting
8154 vertices vector.
8155 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
8156
8157 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8158
8159 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
8160 Set type attribute correctly. Set predicable_short_it attribute.
8161 (cmpsi_shiftsi): Remove %? from output template.
8162
8163 2013-09-13 Richard Biener <rguenther@suse.de>
8164
8165 * tree-loop-distribution.c (struct rdg_component,
8166 rdg_defs_used_in_other_loops_p, free_rdg_components,
8167 rdg_build_components): Remove.
8168 (stmts_from_loop): Do not record virtual PHIs.
8169 (generate_loops_for_partition): Skip virtual PHIs.
8170 (build_rdg_partition_for_component): Rename to ...
8171 (build_rdg_partition_for_vertex): ... this and adjust.
8172 (rdg_build_partitions): Take a vector of starting vertices
8173 instead of components. Remove unnecessary leftover handling.
8174 (ldist_gen): Do not build components or record other stores.
8175 (distribute_loop): Do not distribute loops containing stmts
8176 with side-effects.
8177
8178 2013-09-13 Christian Bruel <christian.bruel@st.com>
8179
8180 PR target/58314
8181 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
8182
8183 2013-09-13 Kai Tietz <ktietz@redhat.com>
8184
8185 * config.gcc: Separate cases for mingw and cygwin targets,
8186 and add 64-bit cygwin target case.
8187
8188 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
8189 dll-export inline-functions.
8190 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
8191
8192
8193 2013-09-13 Jeff Law <law@redhat.com>
8194
8195 PR middle-end/58387
8196 Revert:
8197 2013-09-06 Jeff Law <law@redhat.com>
8198
8199 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8200 edge implied equivalences into successor phis.
8201
8202 2013-09-12 DJ Delorie <dj@redhat.com>
8203
8204 * config/rl78/rl78-virt.md: Change from | to \; for asm line
8205 separators.
8206
8207 2013-09-12 Brooks Moses <bmoses@google.com>
8208
8209 PR driver/42955
8210 * Makefile.in: Do not install driver binaries in $(target)/bin.
8211
8212 2013-09-12 DJ Delorie <dj@redhat.com>
8213
8214 * config/rl78/rl78.opt (mrelax): New.
8215 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
8216 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
8217
8218 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
8219 between SP and FP.
8220 (rl78_expand_epilogue): Likewise.
8221
8222 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
8223
8224 PR middle-end/58335
8225 * lra-eliminations.c (remove_reg_equal_offset_note): New.
8226 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
8227 pointer elimination with using remove_reg_equal_offset_note.
8228
8229 2013-09-12 DJ Delorie <dj@redhat.com>
8230
8231 * config/msp430/: New port.
8232 * config.gcc (msp430): Added.
8233 * doc/invoke.texi: Document MSP430 options.
8234 * doc/install.texi: Document msp430-elf
8235 * doc/md.texi: Document msp430-elf
8236 * doc/contrib.texi: Document msp430-elf
8237
8238 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
8239 PSImode.
8240
8241 2013-09-12 Martin Jambor <mjambor@suse.cz>
8242
8243 PR ipa/58389
8244 * ipa-prop.c (remove_described_reference): Give up if the edge in the
8245 reference descriptor is NULL.
8246 (ipa_edge_removal_hook): If owning a reference descriptor, set its
8247 edge to NULL.
8248
8249 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8250
8251 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
8252 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
8253 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
8254 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
8255 make_temp_ssa_name): move to tree-ssanames.h
8256 * tree-ssa-alias.h: Move prototype.
8257 * tree-ssa.h: Include tree-ssanames.h.
8258 * tree-ssanames.c (FREE_SSANAMES): Move to here.
8259 * tree-ssanames.h: New. Move items from tree-flow*.h
8260 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
8261
8262 2013-09-12 Richard Biener <rguenther@suse.de>
8263
8264 PR tree-optimization/58404
8265 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
8266 propagate non-invariant addresses into dereferences wrapped
8267 in component references.
8268
8269 2013-09-12 Richard Biener <rguenther@suse.de>
8270
8271 PR tree-optimization/58402
8272 * passes.def: Move pass_late_warn_uninitialized later.
8273
8274 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8275
8276 * tree-ssa.h: New. Move content from tree-flow.h and
8277 tree-flow-inline.h.
8278 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
8279 Move prototypes belonging to tree-ssa.c.
8280 * tree-flow-inline.h (redirect_edge_var_map_def,
8281 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
8282 tree-ssa.h.
8283 * gimple.h: Adjust prototypes.
8284 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
8285 to...
8286 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
8287 * tree.h: Move prototype to tree-ssa.h.
8288 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
8289 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
8290 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
8291 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
8292 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
8293 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
8294 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
8295 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
8296 graphite-blocking.c, graphite-clast-to-gimple.c,
8297 graphite-dependences.c, graphite-interchange.c,
8298 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
8299 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
8300 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
8301 ipa-reference.c, ipa-split.c, ipa-utils.c,
8302 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
8303 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
8304 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
8305 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
8306 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
8307 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
8308 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
8309 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
8310 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
8311 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
8312 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
8313 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
8314 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
8315 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
8316 value-prof.c, var-tracking.c,
8317 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
8318
8319 2013-09-12 Richard Biener <rguenther@suse.de>
8320
8321 PR tree-optimization/58396
8322 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
8323 (build_rdg): Take a loop-nest parameter, fix memleaks.
8324 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
8325
8326 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8327
8328 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
8329 for SLM.
8330
8331 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
8332
8333 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
8334 The FMA instruction names should have a 'v' prefix.
8335
8336 2013-09-12 Richard Biener <rguenther@suse.de>
8337
8338 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
8339 (dot_rdg): Use popen instead of system in optional code.
8340 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
8341 (already_processed_vertex_p): Adjust.
8342 (has_anti_or_output_dependence, predecessor_has_mem_write,
8343 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
8344 rdg_flag_uses): Remove.
8345 (rdg_flag_vertex): Simplify.
8346 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
8347 remove recursion.
8348 (build_rdg_partition_for_component): Process the first vertex
8349 of a component only.
8350 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
8351
8352 2013-09-12 Alan Modra <amodra@gmail.com>
8353
8354 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
8355
8356 2013-09-11 DJ Delorie <dj@redhat.com>
8357 Nick Clifton <nickc@redhat.com>
8358
8359 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
8360 (rl78_stack_based_mem): New.
8361 * config/rl78/constraints.md (Iv08): New.
8362 (Iv16): New.
8363 (Iv24): New.
8364 (Is09): New.
8365 (Is17): New.
8366 (Is25): New.
8367 (ISsi): New.
8368 (IShi): New.
8369 (ISqi): New.
8370 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
8371 (movhi): Likewise.
8372 (movsi): Change from expand to insn-and-split.
8373 (ashrsi3): Clobber AX.
8374 (lshrsi3): New.
8375 (ashlsi3): New.
8376 (cbranchsi4): New.
8377 * config/rl78/rl78.md (CC_REG): Fix.
8378 (addsi3): Allow memory and immediate operands.
8379 (addsi3_internal): Split into...
8380 (addsi3_internal_virt): ...new, and ...
8381 (addsi3_internal_real): ...new.
8382 (subsi): New.
8383 (subsi3_internal_virt): New.
8384 (subsi3_internal_real): New.
8385 (mulsi3): Add memory operand.
8386 (mulsi3_rl78): Likewise.
8387 (mulsi3_g13): Likewise.
8388 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
8389 (cbranchqi4_real): Add more constraint options.
8390 (cbranchhi4_real): Expand pattern.
8391 (cbranchhi4_real_signed): New.
8392 (cbranchhi4_real_inverted): New.
8393 (cbranchsi4_real_lt): New.
8394 (cbranchsi4_real_ge): New.
8395 (cbranchsi4_real_signed): New.
8396 (cbranchsi4_real): New.
8397 (peephole2): New.
8398 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
8399 constant shifts.
8400 (lshrsi3_virt): Likewise.
8401 (ashlsi3_virt): Likewise.
8402 (cbranchqi4_virt_signed): New.
8403 (cbranchhi4_virt_signed): New.
8404 (cbranchsi4_virt): New.
8405 * config/rl78/rl78.c: Whitespace fixes throughout.
8406 (move_elim_pass): New.
8407 (pass_data_rl78_move_elim): New.
8408 (pass_rl78_move_elim): New.
8409 (make_pass_rl78_move_elim): New.
8410 (rl78_devirt_info): Run devirt earlier.
8411 (rl78_move_elim_info): New.
8412 (rl78_asm_file_start): Register it.
8413 (rl78_split_movsi): New.
8414 (rl78_as_legitimate_address): Allow virtual base registers when
8415 appropriate.
8416 (rl78_addr_space_convert): Remove spurious debug stuff.
8417 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
8418 (rl78_print_operand): More cases for not printing '#'.
8419 (rl78_expand_compare): Remove most of the logic.
8420 (content_memory): New.
8421 (clear_content_memory): New.
8422 (get_content_index): New.
8423 (get_content_name): New.
8424 (display_content_memory): New.
8425 (update_content): New.
8426 (record_content): New.
8427 (already_contains): New.
8428 (insn_ok_now): Re-recog insns with virtual registers.
8429 (add_postponed_content_update): New.
8430 (process_postponed_content_update): New.
8431 (gen_and_emit_move): New.
8432 (transcode_memory_rtx): Record new location content.
8433 Use gen_and_emit_move.
8434 (force_into_acc): New.
8435 (move_to_acc): Use gen_and_emit_move.
8436 (move_from_acc): Likewise.
8437 (move_acc_to_reg): Likewise.
8438 (move_to_x): Likewise.
8439 (move_to_hl): Likewise.
8440 (move_to_de): Likewise.
8441 (rl78_alloc_physical_registers_op1): Record location content.
8442 (has_constraint): New.
8443 (rl78_alloc_physical_registers_op2): Record location content.
8444 Optimize use of HL.
8445 (rl78_alloc_physical_registers_ro1): Likewise.
8446 (rl78_alloc_physical_registers_cmp): Likewise.
8447 (rl78_alloc_physical_registers_umul): Likewise.
8448 (rl78_alloc_address_registers_macax): New.
8449 (rl78_alloc_physical_registers): Initialize and set location
8450 content memory as needed.
8451 (rl78_reorg): Make sure split2 is called.
8452 (rl78_rtx_costs): New.
8453
8454 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
8455
8456 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
8457 for (not (neg ...)) and (neg (not ...)) cases.
8458
8459 2013-09-11 Richard Biener <rguenther@suse.de>
8460
8461 PR middle-end/58377
8462 * passes.def: Split critical edges before late uninit warning passes.
8463 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
8464
8465 2013-09-11 Jakub Jelinek <jakub@redhat.com>
8466
8467 PR tree-optimization/58385
8468 * fold-const.c (build_range_check): If both low and high are NULL,
8469 use omit_one_operand_loc to preserve exp side-effects.
8470
8471 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8472
8473 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
8474
8475 2013-09-11 Richard Biener <rguenther@suse.de>
8476
8477 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
8478 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
8479 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
8480 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
8481 stmts_from_loop, known_dependences_p, build_empty_rdg,
8482 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
8483 * tree-loop-distribution.c: ... here.
8484 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
8485 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
8486 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
8487 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
8488 * tree-loop-distribution.c: ... here.
8489 * tree-loop-distribution.c: Include gimple-pretty-print.h.
8490 (struct partition_s): Add loops member.
8491 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
8492 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
8493 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
8494
8495 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8496 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8497 Sergey Lega <sergey.s.lega@intel.com>
8498 Anna Tikhonova <anna.tikhonova@intel.com>
8499 Ilya Tocar <ilya.tocar@intel.com>
8500 Andrey Turetskiy <andrey.turetskiy@intel.com>
8501 Ilya Verbin <ilya.verbin@intel.com>
8502 Kirill Yukhin <kirill.yukhin@intel.com>
8503 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8504
8505 * config/i386/constraints.md (k): New.
8506 (Yk): Ditto.
8507 * config/i386/i386.c (const regclass_map): Add new mask registers.
8508 (dbx_register_map): Ditto.
8509 (dbx64_register_map): Ditto.
8510 (svr4_dbx_register_map): Ditto.
8511 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
8512 disabled.
8513 (ix86_preferred_reload_class): Disable constants for mask registers.
8514 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
8515 (ix86_hard_regno_mode_ok): Support new mask registers.
8516 (x86_order_regs_for_local_alloc): Ditto.
8517 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
8518 (FIXED_REGISTERS): Add new mask registers.
8519 (CALL_USED_REGISTERS): Ditto.
8520 (REG_ALLOC_ORDER): Ditto.
8521 (VALID_MASK_REG_MODE): New.
8522 (FIRST_MASK_REG): Ditto.
8523 (LAST_MASK_REG): Ditto.
8524 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
8525 (MAYBE_MASK_CLASS_P): New.
8526 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
8527 (REG_CLASS_CONTENTS): Ditto.
8528 (MASK_REGNO_P): New.
8529 (ANY_MASK_REG_P): Ditto.
8530 (HI_REGISTER_NAMES): Add new mask registers.
8531 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
8532 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
8533 mask registers.
8534 (attribute "type"): Add mskmov, msklog.
8535 (attribute "length_immediate"): Support them.
8536 (attribute "memory"): Ditto.
8537 (attribute "prefix_0f"): Ditto.
8538 (*movhi_internal): Support new mask registers.
8539 (*movqi_internal): Ditto.
8540 (define_split): Split out clobber pattern is a logic
8541 insn on mask registers.
8542 (*k<logic><mode>): New.
8543 (*andhi_1): Extend to support mask regs.
8544 (*andqi_1): Extend to support mask regs.
8545 (kandn<mode>): New.
8546 (define_split): Split and-not to and and not if operands
8547 are not mask regs.
8548 (*<code><mode>_1): Separate HI mode to new pattern...
8549 (*<code>hi_1): This.
8550 (*<code>qi_1): Extend to support mask regs.
8551 (kxnor<mode>): New.
8552 (kortestzhi): Ditto.
8553 (kortestchi): Ditto.
8554 (kunpckhi): Ditto.
8555 (*one_cmpl<mode>2_1): Remove HImode and handle it...
8556 (*one_cmplhi2_1): ...Here, now with mask registers support.
8557 (*one_cmplqi2_1): Support new mask registers.
8558 (HI/QImode arithmetics splitter): Don't split if mask registers
8559 are used.
8560 (HI/QImode not splitter): Ditto.
8561 * config/i386/predicated.md (mask_reg_operand): New.
8562 (general_reg_operand): Ditto.
8563
8564 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8565
8566 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
8567 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
8568
8569 2013-09-10 Jeff Law <law@redhat.com>
8570
8571 PR tree-optimization/58380
8572 * tree-ssa-threadupdate.c (thread_block): Recognize another case
8573 of threading through a buried loop header.
8574
8575 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
8576 return value for single successor case.
8577
8578 2013-09-10 Jan Hubicka <jh@suse.cz>
8579
8580 * ipa-devirt.c (ipa_devirt): Enable with LTO.
8581
8582 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
8583
8584 PR target/58361
8585 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
8586 support conditional execution.
8587 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
8588
8589 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
8590
8591 * lra.c (lra): Clear lra_optional_reload_pseudos before every
8592 constraint pass.
8593 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
8594 Check destination too to check move insn.
8595 (undo_optional_reloads): Add check that the original peudo did not
8596 changed its allocation and the optional reload was inherited on last
8597 inheritance pass. Break loop after deciding to keep optional reload.
8598 (lra_undo_inheritance): Add check that inherited pseudo still in
8599 memory.
8600
8601 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
8602
8603 * config/aarch64/aarch64.md (generic_sched): New.
8604 * config/aarch64/aarch64-generic.md (load): Make conditional
8605 on generic_sched attribute.
8606 (nonload): Likewise.
8607
8608 2013-09-10 Jan Hubicka <jh@suse.cz>
8609
8610 * lto-cgraph.c: Include ipa-utils.h.
8611 (compute_ltrans_boundary): Also add possible targets into the boundary.
8612
8613 2013-09-10 Jan Hubicka <jh@suse.cz>
8614
8615 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
8616 VAR_DECL of vtable rather than full expression.
8617
8618 2013-09-10 Jan Hubicka <jh@suse.cz>
8619 Paolo Carlini <paolo.carlini@oracle.com>
8620
8621 * cgraphunit.c (analyze_functions): Save input_location, set it
8622 to UNKNOWN_LOCATION and restore it at the end.
8623
8624 2013-09-10 Martin Jambor <mjambor@suse.cz>
8625
8626 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
8627 represented by a thunk.
8628
8629 2013-09-10 Jeff Law <law@redhat.com>
8630
8631 PR tree-optimization/58343
8632 * tree-ssa-threadupdate.c (thread_block): Identify and disable
8633 jump threading requests through loop headers buried in the middle
8634 of a jump threading path.
8635
8636 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
8637 in return value/type.
8638
8639 2013-09-10 Jakub Jelinek <jakub@redhat.com>
8640
8641 PR rtl-optimization/58365
8642 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
8643 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
8644 it differs.
8645
8646 2013-09-10 Richard Biener <rguenther@suse.de>
8647
8648 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
8649 * tree-data-ref.c (create_rdg_vertices): Collect all data
8650 references, signal failure to the caller, use data-ref API.
8651 (build_rdg): Compute data references only once. Maintain lifetime
8652 of data references and data dependences from within RDG.
8653 (free_rdg): Free dependence relations.
8654 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
8655 inventing extra dependences.
8656 (distribute_loop): Update for RDG API changes.
8657
8658 2013-09-10 Kai Tietz <ktietz@redhat.com>
8659
8660 * doc/invoke.texi (fms-extensions): Document changed
8661 behavior for ms-abi targets.
8662 * config/i386/i386.c (ix86_option_override_internal):
8663 Set default value of option -fms-extension for ms-abi targets.
8664
8665 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
8666
8667 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
8668
8669 2013-09-10 Alan Modra <amodra@gmail.com>
8670
8671 PR target/58330
8672 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
8673
8674 2013-09-10 Alan Modra <amodra@gmail.com>
8675
8676 * config/rs6000/predicates.md (add_cint_operand): New.
8677 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
8678 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
8679 using add_cint_operand.
8680 (largetoc_high_plus_aix): Likewise.
8681
8682 2013-09-09 Jakub Jelinek <jakub@redhat.com>
8683
8684 PR tree-optimization/58364
8685 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
8686 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
8687 the current range can't be an unconditional true or false.
8688
8689 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
8690
8691 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
8692
8693 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
8694
8695 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
8696
8697 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
8698
8699 PR c++/43452
8700 * doc/invoke.texi (-Wdelete-incomplete): Document it.
8701
8702 2013-09-09 Ian Bolton <ian.bolton@arm.com>
8703
8704 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
8705 NO_REGS for immediate that can't be moved directly into FP_REGS.
8706
8707 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8708
8709 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
8710 comparison with negated operand.
8711 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
8712 RTL form.
8713
8714 2013-09-09 Richard Biener <rguenther@suse.de>
8715
8716 PR middle-end/58326
8717 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
8718 of a subloop record all its block as affecting loop-closed SSA form.
8719
8720 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8721
8722 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
8723 of an rtx/bitpos pair.
8724 (store_fixed_bit_field): Update accordingly.
8725
8726 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8727
8728 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
8729 GEN_INT.
8730 * builtins.c (expand_errno_check): Likewise.
8731 * dwarf2cfi.c (init_return_column_size): Likewise.
8732 * except.c (sjlj_mark_call_sites): Likewise.
8733 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
8734 * lra-constraints.c (emit_inc): Likewise.
8735 * ree.c (combine_set_extension): Likewise.
8736 * regmove.c (fixup_match_2): Likewise.
8737 * reload1.c (inc_for_reload): Likewise.
8738
8739 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8740
8741 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
8742 (force_to_mode, simplify_shift_const_1, simplify_comparison):
8743 Use gen_int_mode with the mode of the associated simplify_* call.
8744 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
8745 * expmed.c (expand_shift_1): Likewise.
8746 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8747 * loop-iv.c (iv_number_of_iterations): Likewise.
8748 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8749 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
8750
8751 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8752
8753 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
8754 of the associated expand_* call.
8755 (asan_emit_stack_protection): Likewise.
8756 * builtins.c (round_trampoline_addr): Likewise.
8757 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
8758 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
8759 (emit_store_flag): Likewise.
8760 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
8761 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
8762 Likewise.
8763 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8764 * ifcvt.c (noce_try_store_flag_constants): Likewise.
8765 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8766 * modulo-sched.c (generate_prolog_epilog): Likewise.
8767 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
8768 (expand_ctz, expand_ffs, expand_unop): Likewise.
8769
8770 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8771
8772 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
8773 of the associated gen_rtx_* call.
8774 * caller-save.c (init_caller_save): Likewise.
8775 * combine.c (find_split_point, make_extraction): Likewise.
8776 (make_compound_operation): Likewise.
8777 * dwarf2out.c (mem_loc_descriptor): Likewise.
8778 * explow.c (plus_constant, probe_stack_range): Likewise.
8779 * expmed.c (expand_mult_const): Likewise.
8780 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
8781 * reload1.c (init_reload): Likewise.
8782 * valtrack.c (cleanup_auto_inc_dec): Likewise.
8783 * var-tracking.c (adjust_mems): Likewise.
8784 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
8785 rather than gen_rtx_fmt_ee.
8786
8787 2013-09-09 Jan Hubicka <jh@suse.cz>
8788
8789 PR middle-end/58294
8790 * value-prof.c (gimple_ic): Copy also abnormal edges.
8791
8792 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8793
8794 * asan.c (asan_shadow_cst): Use gen_int_mode.
8795
8796 2013-09-08 Jan Hubicka <jh@suse.cz>
8797
8798 * ipa-profile.c: Add toplevel comment.
8799 (ipa_propagate_frequency_1): Be more conservative when profile is read.
8800 (contains_hot_call_p): New function.
8801 (ipa_propagate_frequency): Set frequencies based on counts when
8802 profile is read.
8803 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
8804 profile; do not tamper with profile after inlining if it is read.
8805
8806 2013-09-08 Jan Hubicka <jh@suse.cz>
8807
8808 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
8809 speculative edges.
8810
8811 2013-09-08 Jan Hubicka <jh@suse.cz>
8812
8813 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
8814 summary generation.
8815
8816 2013-09-08 Jeff Law <law@redhat.com>
8817
8818 PR bootstrap/58340
8819 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
8820 of 'found'.
8821
8822 2013-09-08 Andi Kleen <ak@linux.intel.com>
8823
8824 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
8825
8826 2013-09-08 Jan Hubicka <jh@suse.cz>
8827
8828 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
8829
8830 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
8831
8832 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
8833 for non-debug insns.
8834 * lra.c (new_insn_reg): Take the containing insn as a parameter.
8835 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
8836 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
8837 accordingly.
8838
8839 2013-09-08 Jan Hubicka <jh@suse.cz>
8840
8841 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
8842 targets and devirtualize to BUILT_IN_UNREACHABLE.
8843 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
8844 * ipa.c (walk_polymorphic_call_targets): New function.
8845 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
8846 functions; use the new timevar.
8847 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
8848 was removed from the program.
8849 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
8850 not consider it in the walk when its vtable is dead.
8851 (possible_polymorphic_call_targets_1): Pass anonymous flag to
8852 record_binfo.
8853 (devirt_variable_node_removal_hook): New function.
8854 (possible_polymorphic_call_targets): Also register
8855 devirt_variable_node_removal_hook.
8856 (ipa_devirt): Do not do non-speculative devirtualization.
8857 (gate_ipa_devirt): One execute if devirtualizing speculatively.
8858
8859 2013-09-08 Jan Hubicka <jh@suse.cz>
8860
8861 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
8862 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
8863 varpool_remove_variable_insertion_hook): Declare.
8864 * varpool.c (varpool_node_hook_list): New structure.
8865 (first_varpool_node_removal_hook,
8866 first_varpool_variable_insertion_hook): New variables.
8867 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
8868 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
8869 varpool_remove_variable_insertion_hook,
8870 varpool_call_variable_insertion_hooks): New functions.
8871 (varpool_remove_node): Use it.
8872
8873 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
8874
8875 PR c++/54941
8876 * diagnostic.c (diagnostic_build_prefix): When s.file is
8877 "<built-in>" don't output line and column numbers.
8878
8879 2013-09-06 Jan Hubicka <jh@suse.cz>
8880
8881 * cgraphunit.c (expand_thunk): Get body before touching arguments.
8882 * lto-streamer-out.c: Stream thunks, too.
8883 * lto-streamer-in.c (input_function): Pop cfun here
8884 (lto_read_body): Instead of here.
8885
8886 2013-09-06 Caroline Tice <cmtice@google.com>
8887
8888 * doc/install.texi: Add documentation for the --enable-vtable-verify
8889 and the --disable-libvtv configure options.
8890
8891 2013-09-06 Jeff Law <law@redhat.com>
8892
8893 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8894 edge implied equivalences into successor phis.
8895
8896 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
8897
8898 * resource.c (mark_referenced_resources): Handle COND_EXEC.
8899
8900 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
8901
8902 * resource.c (mark_target_live_regs): Compute resources taking
8903 into account if a call is predicated or not.
8904
8905 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
8906
8907 * toplev.c (output_stack_usage): Be prepared for suffixes created by
8908 the compiler in the function names.
8909
8910 2013-09-06 Jan Hubicka <jh@suse.cz>
8911
8912 PR middle-end/58094
8913 * ipa-inline.c (has_caller_p): New function.
8914 (want_inline_function_to_all_callers_p): Use it.
8915 (sum_callers, inline_to_all_callers): Break out from ...
8916 (ipa_inline): ... here.
8917
8918 2013-09-06 Jan Hubicka <jh@suse.cz>
8919
8920 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
8921 only when AVX is enabled.
8922
8923 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8924
8925 * config/aarch64/aarch64.md
8926 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
8927 is fpsimd_<load/store>2.
8928 (load_pair<mode>): Likewise.
8929 (store_pair<mode>): Likewise.
8930
8931 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8932
8933 * config/arm/types.md (type): Add "mrs" type.
8934 * config/aarch64/aarch64.md
8935 (aarch64_load_tp_hard): Make type "mrs".
8936 * config/arm/arm.md
8937 (load_tp_hard): Make type "mrs".
8938 * config/arm/cortex-a15.md: Update with new attributes.
8939 * config/arm/cortex-a5.md: Update with new attributes.
8940 * config/arm/cortex-a53.md: Update with new attributes.
8941 * config/arm/cortex-a7.md: Update with new attributes.
8942 * config/arm/cortex-a8.md: Update with new attributes.
8943 * config/arm/cortex-a9.md: Update with new attributes.
8944 * config/arm/cortex-m4.md: Update with new attributes.
8945 * config/arm/cortex-r4.md: Update with new attributes.
8946 * config/arm/fa526.md: Update with new attributes.
8947 * config/arm/fa606te.md: Update with new attributes.
8948 * config/arm/fa626te.md: Update with new attributes.
8949 * config/arm/fa726te.md: Update with new attributes.
8950
8951 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8952
8953 * config/aarch64/aarch64.md
8954 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
8955 (*movtf_aarch64): Likewise.
8956 * config/arm/arm.md
8957 (thumb1_movdi_insn): Use "multiple" for type where more than one
8958 instruction is used for a move.
8959 (*arm32_movhf): Likewise.
8960 (*thumb_movdf_insn): Likewise.
8961
8962 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8963
8964 * config/arm/types.md (type): Rename fcpys to fmov.
8965 * config/arm/vfp.md
8966 (*arm_movsi_vfp): Rename type fcpys as fmov.
8967 (*thumb2_movsi_vfp): Likewise
8968 (*movhf_vfp_neon): Likewise
8969 (*movhf_vfp): Likewise
8970 (*movsf_vfp): Likewise
8971 (*thumb2_movsf_vfp): Likewise
8972 (*movsfcc_vfp): Likewise
8973 (*thumb2_movsfcc_vfp): Likewise
8974 * config/aarch64/aarch64-simd.md
8975 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
8976 * config/aarch64/aarch64.md
8977 (*movsi_aarch64): Replace type mov_reg with fmovs.
8978 (*movdi_aarch64): Likewise
8979 (*movsf_aarch64): Likewise
8980 (*movdf_aarch64): Likewise
8981 * config/arm/arm.c
8982 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
8983 * config/arm/iwmmxt.md
8984 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
8985 * config/arm/arm1020e.md: Update with new attributes.
8986 * config/arm/cortex-a15-neon.md: Update with new attributes.
8987 * config/arm/cortex-a5.md: Update with new attributes.
8988 * config/arm/cortex-a53.md: Update with new attributes.
8989 * config/arm/cortex-a7.md: Update with new attributes.
8990 * config/arm/cortex-a8-neon.md: Update with new attributes.
8991 * config/arm/cortex-a9.md: Update with new attributes.
8992 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8993 * config/arm/cortex-r4f.md: Update with new attributes.
8994 * config/arm/marvell-pj4.md: Update with new attributes.
8995 * config/arm/vfp11.md: Update with new attributes.
8996
8997 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8998
8999 * config/aarch64/aarch64.md
9000 (*madd<mode>): Fix type attribute.
9001 (*maddsi_uxtw): Likewise.
9002 (*msub<mode>): Likewise.
9003 (*msubsi_uxtw): Likewise.
9004 (<su_optab>maddsidi4): Likewise.
9005 (<su_optab>msubsidi4): Likewise.
9006
9007 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9008
9009 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
9010 * config/arm/arm.md (core_cycles): Remove fdiv.
9011 * config/arm/vfp.md:
9012 (*sqrtsf2_vfp): Update for attribute changes.
9013 (*sqrtdf2_vfp): Likewise.
9014 * config/aarch64/aarch64.md:
9015 (sqrt<mode>2): Update for attribute changes.
9016 * config/arm/arm1020e.md: Update with new attributes.
9017 * config/arm/cortex-a15-neon.md: Update with new attributes.
9018 * config/arm/cortex-a5.md: Update with new attributes.
9019 * config/arm/cortex-a53.md: Update with new attributes.
9020 * config/arm/cortex-a7.md: Update with new attributes.
9021 * config/arm/cortex-a8-neon.md: Update with new attributes.
9022 * config/arm/cortex-a9.md: Update with new attributes.
9023 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9024 * config/arm/cortex-r4f.md: Update with new attributes.
9025 * config/arm/marvell-pj4.md: Update with new attributes.
9026 * config/arm/vfp11.md: Update with new attributes.
9027
9028 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9029
9030 * config/arm/types.md
9031 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
9032 * config/aarch64/aarch64.md
9033 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
9034 new attributes.
9035 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
9036 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
9037 (float<GPI:mode><GPF:mode>2): Likewise.
9038 * config/arm/vfp.md
9039 (*truncsisf2_vfp): Update with new attributes.
9040 (*truncsidf2_vfp): Likewise.
9041 (fixuns_truncsfsi2): Likewise.
9042 (fixuns_truncdfsi2): Likewise.
9043 (*floatsisf2_vfp): Likewise.
9044 (*floatsidf2_vfp): Likewise.
9045 (floatunssisf2): Likewise.
9046 (floatunssidf2): Likewise.
9047 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
9048 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
9049 * config/arm/arm1020e.md: Update with new attributes.
9050 * config/arm/cortex-a15-neon.md: Update with new attributes.
9051 * config/arm/cortex-a5.md: Update with new attributes.
9052 * config/arm/cortex-a53.md: Update with new attributes.
9053 * config/arm/cortex-a7.md: Update with new attributes.
9054 * config/arm/cortex-a8-neon.md: Update with new attributes.
9055 * config/arm/cortex-a9.md: Update with new attributes.
9056 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9057 * config/arm/cortex-r4f.md: Update with new attributes.
9058 * config/arm/marvell-pj4.md: Update with new attributes.
9059 * config/arm/vfp11.md: Update with new attributes.
9060
9061 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9062
9063 * config/aarch64/arm_neon.h
9064 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
9065 (vqtbx<1,2,3,4><q>_s8): Likewise.
9066
9067 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9068
9069 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
9070 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
9071 (add<mode>3): Add type attribute.
9072 (add<mode>3): Likewise.
9073 (usadd<mode>3): Likewise.
9074 (ssadd<mode>3): Likewise.
9075 (sub<mode>3): Likewise.
9076 (sub<mode>3): Likewise.
9077 (ussub<mode>3): Likewise.
9078 (sssub<mode>3): Likewise.
9079 (ssmulsa3): Likewise.
9080 (usmulusa3): Likewise.
9081 (arm_usatsihi): Likewise.
9082 * config/arm/vfp.md
9083 (*movdi_vfp): Add types for all instructions.
9084 (*movdi_vfp_cortexa8): Likewise.
9085 (*movhf_vfp_neon): Likewise.
9086 (*movhf_vfp): Likewise.
9087 (*movdf_vfp): Likewise.
9088 (*thumb2_movdf_vfp): Likewise.
9089 (*thumb2_movdfcc_vfp): Likewise.
9090 * config/arm/arm.md: Add type attribute to all insn patterns.
9091 (*thumb1_adddi3): Add type attribute.
9092 (*arm_adddi3): Likewise.
9093 (*adddi_sesidi_di): Likewise.
9094 (*adddi_zesidi_di): Likewise.
9095 (*thumb1_addsi3): Likewise.
9096 (addsi3_compare0): Likewise.
9097 (*addsi3_compare0_scratch): Likewise.
9098 (*compare_negsi_si): Likewise.
9099 (cmpsi2_addneg): Likewise.
9100 (*addsi3_carryin_<optab>): Likewise.
9101 (*addsi3_carryin_alt2_<optab>): Likewise.
9102 (*addsi3_carryin_clobercc_<optab>): Likewise.
9103 (*subsi3_carryin): Likewise.
9104 (*subsi3_carryin_const): Likewise.
9105 (*subsi3_carryin_compare): Likewise.
9106 (*subsi3_carryin_compare_const): Likewise.
9107 (*arm_subdi3): Likewise.
9108 (*thumb_subdi3): Likewise.
9109 (*subdi_di_zesidi): Likewise.
9110 (*subdi_di_sesidi): Likewise.
9111 (*subdi_zesidi_di): Likewise.
9112 (*subdi_sesidi_di): Likewise.
9113 (*subdi_zesidi_ze): Likewise.
9114 (thumb1_subsi3_insn): Likewise.
9115 (*arm_subsi3_insn): Likewise.
9116 (*anddi3_insn): Likewise.
9117 (*anddi_zesidi_di): Likewise.
9118 (*anddi_sesdi_di): Likewise.
9119 (*ne_zeroextracts): Likewise.
9120 (*ne_zeroextracts): Likewise.
9121 (*ite_ne_zeroextr): Likewise.
9122 (*ite_ne_zeroextr): Likewise.
9123 (*anddi_notdi_di): Likewise.
9124 (*anddi_notzesidi): Likewise.
9125 (*anddi_notsesidi): Likewise.
9126 (andsi_notsi_si): Likewise.
9127 (thumb1_bicsi3): Likewise.
9128 (*iordi3_insn): Likewise.
9129 (*iordi_zesidi_di): Likewise.
9130 (*iordi_sesidi_di): Likewise.
9131 (*thumb1_iorsi3_insn): Likewise.
9132 (*xordi3_insn): Likewise.
9133 (*xordi_zesidi_di): Likewise.
9134 (*xordi_sesidi_di): Likewise.
9135 (*arm_xorsi3): Likewise.
9136 (*andsi_iorsi3_no): Likewise.
9137 (*smax_0): Likewise.
9138 (*smax_m1): Likewise.
9139 (*arm_smax_insn): Likewise.
9140 (*smin_0): Likewise.
9141 (*arm_smin_insn): Likewise.
9142 (*arm_umaxsi3): Likewise.
9143 (*arm_uminsi3): Likewise.
9144 (*minmax_arithsi): Likewise.
9145 (*minmax_arithsi_): Likewise.
9146 (*satsi_<SAT:code>): Likewise.
9147 (arm_ashldi3_1bit): Likewise.
9148 (arm_ashrdi3_1bit): Likewise.
9149 (arm_lshrdi3_1bit): Likewise.
9150 (*arm_negdi2): Likewise.
9151 (*thumb1_negdi2): Likewise.
9152 (*arm_negsi2): Likewise.
9153 (*thumb1_negsi2): Likewise.
9154 (*negdi_extendsid): Likewise.
9155 (*negdi_zero_extend): Likewise.
9156 (*arm_abssi2): Likewise.
9157 (*thumb1_abssi2): Likewise.
9158 (*arm_neg_abssi2): Likewise.
9159 (*thumb1_neg_abss): Likewise.
9160 (one_cmpldi2): Likewise.
9161 (extend<mode>di2): Likewise.
9162 (*compareqi_eq0): Likewise.
9163 (*arm_extendhisi2addsi): Likewise.
9164 (*arm_movdi): Likewise.
9165 (*thumb1_movdi_insn): Likewise.
9166 (*arm_movt): Likewise.
9167 (*thumb1_movsi_insn): Likewise.
9168 (pic_add_dot_plus_four): Likewise.
9169 (pic_add_dot_plus_eight): Likewise.
9170 (tls_load_dot_plus_eight): Likewise.
9171 (*thumb1_movhi_insn): Likewise.
9172 (*thumb1_movsf_insn): Likewise.
9173 (*movdf_soft_insn): Likewise.
9174 (*thumb_movdf_insn): Likewise.
9175 (cbranchsi4_insn): Likewise.
9176 (cbranchsi4_scratch): Likewise.
9177 (*negated_cbranchsi4): Likewise.
9178 (*tbit_cbranch): Likewise.
9179 (*tlobits_cbranch): Likewise.
9180 (*tstsi3_cbranch): Likewise.
9181 (*cbranchne_decr1): Likewise.
9182 (*addsi3_cbranch): Likewise.
9183 (*addsi3_cbranch_scratch): Likewise.
9184 (*arm_cmpdi_insn): Likewise.
9185 (*arm_cmpdi_unsig): Likewise.
9186 (*arm_cmpdi_zero): Likewise.
9187 (*thumb_cmpdi_zero): Likewise.
9188 (*deleted_compare): Likewise.
9189 (*mov_scc): Likewise.
9190 (*mov_negscc): Likewise.
9191 (*mov_notscc): Likewise.
9192 (*cstoresi_eq0_thumb1_insn): Likewise.
9193 (cstoresi_nltu_thumb1): Likewise.
9194 (cstoresi_ltu_thu): Likewise.
9195 (thumb1_addsi3_addgeu): Likewise.
9196 (*arm_jump): Likewise.
9197 (*thumb_jump): Likewise.
9198 (*check_arch2): Likewise.
9199 (arm_casesi_internal): Likewise.
9200 (thumb1_casesi_dispatch): Likewise.
9201 (*arm_indirect_jump): Likewise.
9202 (*thumb1_indirect_jump): Likewise.
9203 (nop): Likewise.
9204 (*and_scc): Likewise.
9205 (*ior_scc): Likewise.
9206 (*compare_scc): Likewise.
9207 (*cond_move): Likewise.
9208 (*cond_arith): Likewise.
9209 (*cond_sub): Likewise.
9210 (*cmp_ite0): Likewise.
9211 (*cmp_ite1): Likewise.
9212 (*cmp_and): Likewise.
9213 (*cmp_ior): Likewise.
9214 (*ior_scc_scc): Likewise.
9215 (*ior_scc_scc_cmp): Likewise.
9216 (*and_scc_scc): Likewise.
9217 (*and_scc_scc_cmp): Likewise.
9218 (*and_scc_scc_nod): Likewise.
9219 (*negscc): Likewise.
9220 (movcond_addsi): Likewise.
9221 (movcond): Likewise.
9222 (*ifcompare_plus_move): Likewise.
9223 (*if_plus_move): Likewise.
9224 (*ifcompare_move_plus): Likewise.
9225 (*if_move_plus): Likewise.
9226 (*ifcompare_arith_arith): Likewise.
9227 (*if_arith_arith): Likewise.
9228 (*ifcompare_arith_move): Likewise.
9229 (*if_arith_move): Likewise.
9230 (*ifcompare_move_arith): Likewise.
9231 (*if_move_arith): Likewise.
9232 (*ifcompare_move_not): Likewise.
9233 (*if_move_not): Likewise.
9234 (*ifcompare_not_move): Likewise.
9235 (*if_not_move): Likewise.
9236 (*ifcompare_shift_move): Likewise.
9237 (*if_shift_move): Likewise.
9238 (*ifcompare_move_shift): Likewise.
9239 (*if_move_shift): Likewise.
9240 (*ifcompare_shift_shift): Likewise.
9241 (*ifcompare_not_arith): Likewise.
9242 (*ifcompare_arith_not): Likewise.
9243 (*if_arith_not): Likewise.
9244 (*ifcompare_neg_move): Likewise.
9245 (*if_neg_move): Likewise.
9246 (*ifcompare_move_neg): Likewise.
9247 (*if_move_neg): Likewise.
9248 (prologue_thumb1_interwork): Likewise.
9249 (*cond_move_not): Likewise.
9250 (*sign_extract_onebit): Likewise.
9251 (*not_signextract_onebit): Likewise.
9252 (stack_tie): Likewise.
9253 (align_4): Likewise.
9254 (align_8): Likewise.
9255 (consttable_end): Likewise.
9256 (consttable_1): Likewise.
9257 (consttable_2): Likewise.
9258 (consttable_4): Likewise.
9259 (consttable_8): Likewise.
9260 (consttable_16): Likewise.
9261 (*thumb1_tablejump): Likewise.
9262 (prefetch): Likewise.
9263 (force_register_use): Likewise.
9264 (thumb_eh_return): Likewise.
9265 (load_tp_hard): Likewise.
9266 (load_tp_soft): Likewise.
9267 (tlscall): Likewise.
9268 (*arm_movtas_ze): Likewise.
9269 (*arm_rev): Likewise.
9270 (*arm_revsh): Likewise.
9271 (*arm_rev16): Likewise.
9272 * config/arm/thumb2.md
9273 (*thumb2_smaxsi3): Likewise.
9274 (*thumb2_sminsi3): Likewise.
9275 (*thumb32_umaxsi3): Likewise.
9276 (*thumb2_uminsi3): Likewise.
9277 (*thumb2_negdi2): Likewise.
9278 (*thumb2_abssi2): Likewise.
9279 (*thumb2_neg_abss): Likewise.
9280 (*thumb2_movsi_insn): Likewise.
9281 (tls_load_dot_plus_four): Likewise.
9282 (*thumb2_movhi_insn): Likewise.
9283 (*thumb2_mov_scc): Likewise.
9284 (*thumb2_mov_negs): Likewise.
9285 (*thumb2_mov_negs): Likewise.
9286 (*thumb2_mov_nots): Likewise.
9287 (*thumb2_mov_nots): Likewise.
9288 (*thumb2_movsicc_): Likewise.
9289 (*thumb2_movsfcc_soft_insn): Likewise.
9290 (*thumb2_indirect_jump): Likewise.
9291 (*thumb2_and_scc): Likewise.
9292 (*thumb2_ior_scc): Likewise.
9293 (*thumb2_ior_scc_strict_it): Likewise.
9294 (*thumb2_cond_move): Likewise.
9295 (*thumb2_cond_arith): Likewise.
9296 (*thumb2_cond_ari): Likewise.
9297 (*thumb2_cond_sub): Likewise.
9298 (*thumb2_negscc): Likewise.
9299 (*thumb2_movcond): Likewise.
9300 (thumb2_casesi_internal): Likewise.
9301 (thumb2_casesi_internal_pic): Likewise.
9302 (*thumb2_alusi3_short): Likewise.
9303 (*thumb2_mov<mode>_shortim): Likewise.
9304 (*thumb2_addsi_short): Likewise.
9305 (*thumb2_subsi_short): Likewise.
9306 (thumb2_addsi3_compare0): Likewise.
9307 (*thumb2_cbz): Likewise.
9308 (*thumb2_cbnz): Likewise.
9309 (*thumb2_one_cmplsi2_short): Likewise.
9310 (*thumb2_negsi2_short): Likewise.
9311 (*orsi_notsi_si): Likewise.
9312 * config/arm/arm1020e.md: Update with new attributes.
9313 * config/arm/arm1026ejs.md: Update with new attributes.
9314 * config/arm/arm1136jfs.md: Update with new attributes.
9315 * config/arm/arm926ejs.md: Update with new attributes.
9316 * config/arm/cortex-a15.md: Update with new attributes.
9317 * config/arm/cortex-a5.md: Update with new attributes.
9318 * config/arm/cortex-a53.md: Update with new attributes.
9319 * config/arm/cortex-a7.md: Update with new attributes.
9320 * config/arm/cortex-a8.md: Update with new attributes.
9321 * config/arm/cortex-a9.md: Update with new attributes.
9322 * config/arm/cortex-m4.md: Update with new attributes.
9323 * config/arm/cortex-r4.md: Update with new attributes.
9324 * config/arm/fa526.md: Update with new attributes.
9325 * config/arm/fa606te.md: Update with new attributes.
9326 * config/arm/fa626te.md: Update with new attributes.
9327 * config/arm/fa726te.md: Update with new attributes.
9328
9329 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9330
9331 * config/aarch64/aarch64-simd.md
9332 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
9333 <vwx> iterator to ensure correct register choice.
9334 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
9335 (aarch64_sqdmull_n<mode>): Likewise.
9336 (aarch64_sqdmull2_n<mode>_internal): Likewise.
9337 * config/aarch64/arm_neon.h
9338 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
9339 (vml<as><q>_n_<su>16): Likewise.
9340 (vml<as>l_high_lane<q>_<su>16): Likewise.
9341 (vml<as>l_high_n_<su>16): Likewise.
9342 (vml<as>l_lane<q>_<su>16): Likewise.
9343 (vml<as>l_n_<su>16): Likewise.
9344 (vmul<q>_lane<q>_<su>16): Likewise.
9345 (vmul<q>_n_<su>16): Likewise.
9346 (vmull_lane<q>_<su>16): Likewise.
9347 (vmull_n_<su>16): Likewise.
9348 (vmull_high_lane<q>_<su>16): Likewise.
9349 (vmull_high_n_<su>16): Likewise.
9350 (vqrdmulh<q>_n_s16): Likewise.
9351
9352 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
9353
9354 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
9355 have the correct lane parameter.
9356
9357 2013-09-06 Richard Biener <rguenther@suse.de>
9358
9359 * cfganal.c (control_dependences::~control_dependences):
9360 Properly free all of the vector.
9361
9362 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
9363
9364 PR target/58269
9365 * config/i386/i386.c (ix86_conditional_register_usage):
9366 Proper initialize extended SSE registers.
9367
9368 2013-09-06 Jan Hubicka <jh@suse.cz>
9369
9370 PR tree-optimization/58311
9371 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
9372
9373 2013-09-06 Jan Hubicka <jh@suse.cz>
9374
9375 * Makefile.in (tree-sra.o): Update dependencies.
9376 * tree-sra.c: Include ipa-utils.h
9377 (scan_function): Use recursive_call_p.
9378 (has_caller_p): New function.
9379 (cgraph_for_node_and_aliases): Count also callers of aliases.
9380
9381 2013-09-06 Jan Hubicka <jh@suse.cz>
9382
9383 PR middle-end/58094
9384 * cgraph.h (symtab_semantically_equivalent_p): Declare.
9385 * tree-tailcall.c: Include ipa-utils.h.
9386 (find_tail_calls): Use it.
9387 * ipa-pure-const.c (check_call): Likewise.
9388 * ipa-utils.c (recursive_call_p): New function.
9389 * ipa-utils.h (recursive_call_p): Dclare.
9390 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
9391 (symtab_semantically_equivalent_p): New function.
9392 * Makefile.in (tree-tailcall.o): Update dependencies.
9393
9394 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9395
9396 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
9397 non-inlinable part.
9398
9399 2013-09-06 Richard Biener <rguenther@suse.de>
9400
9401 * lto-streamer.h (lto_global_var_decls): Remove.
9402 * Makefile.in (OBJS): Remove lto-symtab.o.
9403 (lto-symtab.o): Remove.
9404 (GTFILES): Remove lto-symtab.c
9405 * lto-symtab.c: Move to lto/
9406
9407 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9408
9409 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
9410 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
9411 (UNSPEC_FPINT_RINT): New constant definitions.
9412 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
9413 definition with 2 attributes.
9414 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
9415 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
9416 definitions.
9417
9418 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9419
9420 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
9421 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
9422 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
9423 attribute to "z196_alone".
9424 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
9425 "zEC12_simple".
9426
9427 2013-09-06 Richard Biener <rguenther@suse.de>
9428
9429 * basic-block.h (class control_dependences): New.
9430 * tree-ssa-dce.c (control_dependence_map): Remove.
9431 (cd): New global.
9432 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
9433 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
9434 find_pdom, find_control_dependence, find_all_control_dependences):
9435 Move to cfganal.c.
9436 (mark_control_dependent_edges_necessary,
9437 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
9438 tree_dce_done, perform_tree_ssa_dce): Adjust.
9439 * cfganal.c (set_control_dependence_map_bit,
9440 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
9441 find_all_control_dependences): Move from tree-ssa-dce.c and
9442 implement as methods of control_dependences class.
9443 (control_dependences::control_dependences): New.
9444 (control_dependences::~control_dependences): Likewise.
9445 (control_dependences::get_edges_dependent_on): Likewise.
9446 (control_dependences::get_edge): Likewise.
9447
9448 2013-09-04 Jan Hubicka <jh@suse.cz>
9449
9450 * tree.c (types_same_for_odr): Drop overactive check.
9451 * ipa-devirt.c (hash_type_name): Likewise.
9452
9453 2013-09-04 Jan Hubicka <jh@suse.cz>
9454
9455 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
9456 (analyze_functions): ... here.
9457
9458 2013-09-04 Jan Hubicka <jh@suse.cz>
9459
9460 PR middle-end/58201
9461 * cgraphunit.c (analyze_functions): Clear AUX fields
9462 after processing; initialize assembler name has.
9463
9464 2013-09-05 Jeff Law <law@redhat.com>
9465
9466 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
9467 from thread_around_empty_block. Record threading path into PATH.
9468 Recurse if threading through the initial block is successful.
9469 (thread_across_edge): Corresponding changes to slightly simplify.
9470
9471 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9472
9473 * config/aarch64/aarch64.md
9474 (type): Remove frecpe, frecps, frecpx.
9475 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
9476 fix to be a TARGET_SIMD instruction.
9477 (aarch64_frecps): Remove.
9478 * config/aarch64/aarch64-simd.md
9479 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
9480 (aarch64_frecps<mode>): Handle all float/vector of float modes.
9481
9482 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9483 Sofiane Naci <sofiane.naci@arm.com>
9484
9485 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
9486 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
9487 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
9488 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
9489 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
9490 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
9491 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
9492 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
9493 into "clz, "rbit". Rename "shift" to "shift_imm".
9494 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
9495 changes. Update for attribute changes all occurrences of arlo_* and
9496 shift* types.
9497 * config/arm/arm-fixed.md: Update for attribute changes
9498 all occurrences of arlo_* types.
9499 * config/arm/thumb2.md: Update for attribute changes all occurrences
9500 of arlo_* types.
9501 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
9502 (cortexa7_older_only): Likewise.
9503 (cortexa7_younger): Likewise.
9504 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9505 (1020alu_shift_op): Likewise.
9506 (1020alu_shift_reg_op): Likewise.
9507 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9508 (alu_shift_op): Likewise.
9509 (alu_shift_reg_op): Likewise.
9510 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9511 (11_alu_shift_op): Likewise.
9512 (11_alu_shift_reg_op): Likewise.
9513 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9514 (9_alu_shift_reg_op): Likewise.
9515 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
9516 attribute changes.
9517 (cortex_a15_alu_shift): Likewise.
9518 (cortex_a15_alu_shift_reg): Likewise.
9519 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
9520 attribute changes.
9521 (cortex_a5_alu_shift): Likewise.
9522 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
9523 attribute changes.
9524 (cortex_a53_alu_shift): Likewise.
9525 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
9526 attribute changes.
9527 (cortex_a7_alu_reg): Likewise.
9528 (cortex_a7_alu_shift): Likewise.
9529 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
9530 attribute changes.
9531 (cortex_a8_alu_shift): Likewise.
9532 (cortex_a8_alu_shift_reg): Likewise.
9533 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9534 (cortex_a9_dp_shift): Likewise.
9535 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
9536 attribute changes.
9537 * config/arm/cortex-r4.md
9538 (cortex_r4_alu): Update for attribute changes.
9539 (cortex_r4_mov): Likewise.
9540 (cortex_r4_alu_shift_reg): Likewise.
9541 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9542 (526_alu_shift_op): Likewise.
9543 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9544 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9545 (626te_alu_shift_op): Likewise.
9546 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
9547 (726te_alu_shift_op): Likewise.
9548 (726te_alu_shift_reg_op): Likewise.
9549 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9550 (mp626_alu_shift_op): Likewise.
9551 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
9552 (pj4_alu_conds): Likewise.
9553 (pj4_shift): Likewise.
9554 (pj4_shift_conds): Likewise.
9555 (pj4_alu_shift): Likewise.
9556 (pj4_alu_shift_conds): Likewise.
9557 * config/aarch64/aarch64.md: Update for attribute change
9558 all occurrences of arlo_* and shift* types.
9559
9560 2013-09-05 Mike Stump <mikestump@comcast.net>
9561
9562 * tree.h: Move documentation for tree_function_decl to tree-core.h
9563 with the declaration.
9564
9565 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
9566
9567 PR target/58139
9568 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
9569 looking for widest mode.
9570
9571 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
9572
9573 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
9574
9575 2013-09-05 Richard Biener <rguenther@suse.de>
9576
9577 PR tree-optimization/58137
9578 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
9579 Do not create vectors of pointers.
9580 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
9581 types for the components of the vector initializer.
9582 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
9583 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
9584
9585 2013-09-05 Martin Jambor <mjambor@suse.cz>
9586
9587 * ipa-prop.c (remove_described_reference): Accept missing references,
9588 return false if that hppens, otherwise return true.
9589 (cgraph_node_for_jfunc): New function.
9590 (try_decrement_rdesc_refcount): Likewise.
9591 (try_make_edge_direct_simple_call): Use them.
9592 (ipa_edge_removal_hook): Remove references from rdescs.
9593 (ipa_edge_duplication_hook): Clone rdescs and their references
9594 when the new edge has the same caller as the old one.
9595 * cgraph.c (cgraph_resolve_speculation): Remove speculative
9596 reference before removing any edges.
9597
9598 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
9599
9600 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
9601 initial store.
9602 * thumb2.md (thumb2_storewb_parisi): New pattern.
9603
9604 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
9605
9606 * config/aarch64/aarch64-option-extensions.def: Add
9607 AARCH64_OPT_EXTENSION of 'crc'.
9608 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
9609 (AARCH64_ISA_CRC): Ditto.
9610 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
9611 description of the CRC extension.
9612
9613 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
9614
9615 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
9616 * config/rs6000/linux.h: Ditto.
9617 * alpha/linux.h: Ditto.
9618 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
9619 no_c99_libc_has_function.
9620 * config/c6x/uclinux-elf.h: Ditto.
9621 * config/lm32/uclinux-elf.h: Ditto.
9622 * config/m68k/uclinux.h: Ditto.
9623 * config/moxie/uclinux.h: Ditto.
9624 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
9625 (crisv32-*-linux*, cris-*-linux*): Ditto.
9626 * config/bfin/bfin.c: Include "tm_p.h".
9627
9628 2013-09-05 Richard Biener <rguenther@suse.de>
9629
9630 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
9631 check for a definition without a basic-block.
9632
9633 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9634 Sofiane Naci <sofiane.naci@arm.com>
9635
9636 * config/aarch64/aarch64.md
9637 (*movti_aarch64): Rename r_2_f and f_2_r.
9638 (*movsf_aarch64): Likewise.
9639 (*movdf_aarch64): Likewise.
9640 (*movtf_aarch64): Likewise.
9641 (aarch64_movdi_<mode>low): Likewise.
9642 (aarch64_movdi_<mode>high): Likewise.
9643 (aarch64_mov<mode>high_di): Likewise.
9644 (aarch64_mov<mode>low_di): Likewise.
9645 (aarch64_movtilow_tilow): Likewise.
9646 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
9647 values to config/arm/types.md
9648 (attribute "conds"): Update for attribute change.
9649 (anddi3_insn): Likewise.
9650 (iordi3_insn): Likewise.
9651 (xordi3_insn): Likewise.
9652 (one_cmpldi2): Likewise.
9653 * config/arm/types.md (type): Add Neon types.
9654 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
9655 use "type" attribute.
9656 (movmisalign<mode>_neon_store): Likewise.
9657 (movmisalign<mode>_neon_load): Likewise.
9658 (vec_set<mode>_internal): Likewise.
9659 (vec_setv2di_internal): Likewise.
9660 (vec_extract<mode>): Likewise.
9661 (vec_extractv2di): Likewise.
9662 (add<mode>3_neon): Likewise.
9663 (adddi3_neon): Likewise.
9664 (sub<mode>3_neon): Likewise.
9665 (subdi3_neon): Likewise.
9666 (mul<mode>3_neon): Likewise.
9667 (mul<mode>3add<mode>_neon): Likewise.
9668 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
9669 (fma<VCVTF:mode>4)): Likewise.
9670 (fma<VCVTF:mode>4_intrinsic): Likewise.
9671 (fmsub<VCVTF:mode>4)): Likewise.
9672 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
9673 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
9674 (ior<mode>3): Likewise.
9675 (and<mode>3): Likewise.
9676 (anddi3_neon): Likewise.
9677 (orn<mode>3_neon): Likewise.
9678 (orndi3_neon): Likewise.
9679 (bic<mode>3_neon): Likewise.
9680 (bicdi3_neon): Likewise.
9681 (xor<mode>3): Likewise.
9682 (one_cmpl<mode>2): Likewise.
9683 (abs<mode>2): Likewise.
9684 (neg<mode>2): Likewise.
9685 (umin<mode>3_neon): Likewise.
9686 (umax<mode>3_neon): Likewise.
9687 (smin<mode>3_neon): Likewise.
9688 (smax<mode>3_neon): Likewise.
9689 (vashl<mode>3): Likewise.
9690 (vashr<mode>3_imm): Likewise.
9691 (vlshr<mode>3_imm): Likewise.
9692 (ashl<mode>3_signed): Likewise.
9693 (ashl<mode>3_unsigned): Likewise.
9694 (neon_load_count): Likewise.
9695 (ashldi3_neon_noclobber): Likewise.
9696 (signed_shift_di3_neon): Likewise.
9697 (unsigned_shift_di3_neon): Likewise.
9698 (ashrdi3_neon_imm_noclobber): Likewise.
9699 (lshrdi3_neon_imm_noclobber): Likewise.
9700 (widen_ssum<mode>3): Likewise.
9701 (widen_usum<mode>3): Likewise.
9702 (quad_halves_<code>v4si): Likewise.
9703 (quad_halves_<code>v4sf): Likewise.
9704 (quad_halves_<code>v8hi): Likewise.
9705 (quad_halves_<code>v16qi): Likewise.
9706 (reduc_splus_v2di): Likewise.
9707 (neon_vpadd_internal<mode>): Likewise.
9708 (neon_vpsmin<mode>): Likewise.
9709 (neon_vpsmax<mode>): Likewise.
9710 (neon_vpumin<mode>): Likewise.
9711 (neon_vpumax<mode>): Likewise.
9712 (ss_add<mode>_neon): Likewise.
9713 (us_add<mode>_neon): Likewise.
9714 (ss_sub<mode>_neon): Likewise.
9715 (us_sub<mode>_neon): Likewise.
9716 (neon_vadd<mode>_unspec): Likewise.
9717 (neon_vaddl<mode>): Likewise.
9718 (neon_vaddw<mode>): Likewise.
9719 (neon_vhadd<mode>): Likewise.
9720 (neon_vqadd<mode>): Likewise.
9721 (neon_vaddhn<mode>): Likewise.
9722 (neon_vmul<mode>): Likewise.
9723 (neon_vmla<mode>): Likewise.
9724 (neon_vmlal<mode>): Likewise.
9725 (neon_vmls<mode>): Likewise.
9726 (neon_vmlsl<mode>): Likewise.
9727 (neon_vqdmulh<mode>): Likewise.
9728 (neon_vqdmlal<mode>): Likewise.
9729 (neon_vqdmlsl<mode>): Likewise.
9730 (neon_vmull<mode>): Likewise.
9731 (neon_vqdmull<mode>): Likewise.
9732 (neon_vsub<mode>_unspec): Likewise.
9733 (neon_vsubl<mode>): Likewise.
9734 (neon_vsubw<mode>): Likewise.
9735 (neon_vqsub<mode>): Likewise.
9736 (neon_vhsub<mode>): Likewise.
9737 (neon_vsubhn<mode>): Likewise.
9738 (neon_vceq<mode>): Likewise.
9739 (neon_vcge<mode>): Likewise.
9740 (neon_vcgeu<mode>): Likewise.
9741 (neon_vcgt<mode>): Likewise.
9742 (neon_vcgtu<mode>): Likewise.
9743 (neon_vcle<mode>): Likewise.
9744 (neon_vclt<mode>): Likewise.
9745 (neon_vcage<mode>): Likewise.
9746 (neon_vcagt<mode>): Likewise.
9747 (neon_vtst<mode>): Likewise.
9748 (neon_vabd<mode>): Likewise.
9749 (neon_vabdl<mode>): Likewise.
9750 (neon_vaba<mode>): Likewise.
9751 (neon_vabal<mode>): Likewise.
9752 (neon_vmax<mode>): Likewise.
9753 (neon_vmin<mode>): Likewise.
9754 (neon_vpaddl<mode>): Likewise.
9755 (neon_vpadal<mode>): Likewise.
9756 (neon_vpmax<mode>): Likewise.
9757 (neon_vpmin<mode>): Likewise.
9758 (neon_vrecps<mode>): Likewise.
9759 (neon_vrsqrts<mode>): Likewise.
9760 (neon_vqabs<mode>): Likewise.
9761 (neon_vqneg<mode>): Likewise.
9762 (neon_vcls<mode>): Likewise.
9763 (clz<mode>2): Likewise.
9764 (popcount<mode>2): Likewise.
9765 (neon_vrecpe): Likewise.
9766 (neon_vrsqrte): Likewise.
9767 (neon_vget_lane<mode>_sext_internal): Likewise.
9768 (neon_vget_lane<mode>_zext_internal): Likewise.
9769 (neon_vdup_n<mode>): Likewise.
9770 (neon_vdup_nv2di): Likewise.
9771 (neon_vdpu_lane<mode>_internal): Likewise.
9772 (neon_vswp<mode>): Likewise.
9773 (float<mode><V_cvtto>2): Likewise.
9774 (floatuns<mode><V_cvtto>2): Likewise.
9775 (fix_trunc<mode><V_cvtto>)2): Likewise
9776 (fixuns_trunc<mode><V_cvtto)2): Likewise.
9777 (neon_vcvt<mode>): Likewise.
9778 (neon_vcvtv4sfv4hf): Likewise.
9779 (neon_vcvtv4hfv4sf): Likewise.
9780 (neon_vcvt_n<mode>): Likewise.
9781 (neon_vmovn<mode>): Likewise.
9782 (neon_vqmovn<mode>): Likewise.
9783 (neon_vqmovun<mode>): Likewise.
9784 (neon_vmovl<mode>): Likewise.
9785 (neon_vmul_lane<mode>): Likewise.
9786 (neon_vmull_lane<mode>): Likewise.
9787 (neon_vqdmull_lane<mode>): Likewise.
9788 (neon_vqdmulh_lane<mode>): Likewise.
9789 (neon_vmla_lane<mode>): Likewise.
9790 (neon_vmlal_lane<mode>): Likewise.
9791 (neon_vqdmlal_lane<mode>): Likewise.
9792 (neon_vmls_lane<mode>): Likewise.
9793 (neon_vmlsl_lane<mode>): Likewise.
9794 (neon_vqdmlsl_lane<mode>): Likewise.
9795 (neon_vext<mode>): Likewise.
9796 (neon_vrev64<mode>): Likewise.
9797 (neon_vrev32<mode>): Likewise.
9798 (neon_vrev16<mode>): Likewise.
9799 (neon_vbsl<mode>_internal): Likewise.
9800 (neon_vshl<mode>): Likewise.
9801 (neon_vqshl<mode>): Likewise.
9802 (neon_vshr_n<mode>): Likewise.
9803 (neon_vshrn_n<mode>): Likewise.
9804 (neon_vqshrn_n<mode>): Likewise.
9805 (neon_vqshrun_n<mode>): Likewise.
9806 (neon_vshl_n<mode>): Likewise.
9807 (neon_vqshl_n<mode>): Likewise.
9808 (neon_vqshlu_n<mode>): Likewise.
9809 (neon_vshll_n<mode>): Likewise.
9810 (neon_vsra_n<mode>): Likewise.
9811 (neon_vsri_n<mode>): Likewise.
9812 (neon_vsli_n<mode>): Likewise.
9813 (neon_vtbl1v8qi): Likewise.
9814 (neon_vtbl2v8qi): Likewise.
9815 (neon_vtbl3v8qi): Likewise.
9816 (neon_vtbl4v8qi): Likewise.
9817 (neon_vtbx1v8qi): Likewise.
9818 (neon_vtbx2v8qi): Likewise.
9819 (neon_vtbx3v8qi): Likewise.
9820 (neon_vtbx4v8qi): Likewise.
9821 (neon_vtrn<mode>_internal): Likewise.
9822 (neon_vzip<mode>_internal): Likewise.
9823 (neon_vuzp<mode>_internal): Likewise.
9824 (neon_vld1<mode>): Likewise.
9825 (neon_vld1_lane<mode>): Likewise.
9826 (neon_vld1_dup<mode>): Likewise.
9827 (neon_vld1_dupv2di): Likewise.
9828 (neon_vst1<mode>): Likewise.
9829 (neon_vst1_lane<mode>): Likewise.
9830 (neon_vld2<mode>): Likewise.
9831 (neon_vld2_lane<mode>): Likewise.
9832 (neon_vld2_dup<mode>): Likewise.
9833 (neon_vst2<mode>): Likewise.
9834 (neon_vst2_lane<mode>): Likewise.
9835 (neon_vld3<mode>): Likewise.
9836 (neon_vld3qa<mode>): Likewise.
9837 (neon_vld3qb<mode>): Likewise.
9838 (neon_vld3_lane<mode>): Likewise.
9839 (neon_vld3_dup<mode>): Likewise.
9840 (neon_vst3<mode>): Likewise.
9841 (neon_vst3qa<mode>): Likewise.
9842 (neon_vst3qb<mode>): Likewise.
9843 (neon_vst3_lane<mode>): Likewise.
9844 (neon_vld4<mode>): Likewise.
9845 (neon_vld4qa<mode>): Likewise.
9846 (neon_vld4qb<mode>): Likewise.
9847 (neon_vld4_lane<mode>): Likewise.
9848 (neon_vld4_dup<mode>): Likewise.
9849 (neon_vst4<mode>): Likewise.
9850 (neon_vst4qa<mode>): Likewise.
9851 (neon_vst4qb<mode>): Likewise.
9852 (neon_vst4_lane<mode>): Likewise.
9853 (neon_vec_unpack<US>_lo_<mode>): Likewise.
9854 (neon_vec_unpack<US>_hi_<mode>): Likewise.
9855 (neon_vec_<US>mult_lo_<mode>): Likewise.
9856 (neon_vec_<US>mult_hi_<mode>): Likewise.
9857 (neon_vec_<US>shiftl_<mode>): Likewise.
9858 (neon_unpack<US>_<mode>): Likewise.
9859 (neon_vec_<US>mult_<mode>): Likewise.
9860 (vec_pack_trunc_<mode>): Likewise.
9861 (neon_vec_pack_trunk_<mode>): Likewise.
9862 (neon_vabd<mode>_2): Likewise.
9863 (neon_vabd<mode>_3): Likewise.
9864 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
9865 (thumb2_movsi_vfp): Likewise.
9866 (movdi_vfp): Likewise.
9867 (movdi_vfp_cortexa8): Likewise.
9868 (movhf_vfp_neon): Likewise.
9869 (movhf_vfp): Likewiwse.
9870 (movsf_vfp): Likewiwse.
9871 (thumb2_movsf_vfp): Likewiwse.
9872 (movdf_vfp): Likewise.
9873 (thumb2_movdf_vfp): Likewise.
9874 (movsfcc_vfp): Likewise.
9875 (thumb2_movsfcc_vfp): Likewise.
9876 (movdfcc_vfp): Likewise.
9877 (thumb2_movdfcc_vfp): Likewise.
9878 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
9879 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
9880 (v10_v2c): Likewise.
9881 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
9882 attribute change.
9883 (cortex_a15_neon_int_2): Likewise.
9884 (cortex_a15_neon_int_3): Likewise.
9885 (cortex_a15_neon_int_4): Likewise.
9886 (cortex_a15_neon_int_5): Likewise.
9887 (cortex_a15_neon_vqneg_vqabs): Likewise.
9888 (cortex_a15_neon_vmov): Likewise.
9889 (cortex_a15_neon_vaba): Likewise.
9890 (cortex_a15_neon_vaba_qqq): Likewise.
9891 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9892 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9893 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
9894 scalar_64_32_long_scalar): Likewise.
9895 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9896 (cortex_a15_neon_mla_qqq_8_16): Likewise.
9897 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9898 lotype_qdd_64_32_long): Likewise.
9899 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9900 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9901 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
9902 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9903 (cortex_a15_neon_shift_1): Likewise.
9904 (cortex_a15_neon_shift_2): Likewise.
9905 (cortex_a15_neon_shift_3): Likewise.
9906 (cortex_a15_neon_vshl_ddd): Likewise.
9907 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9908 (cortex_a15_neon_vsra_vrsra): Likewise.
9909 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
9910 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
9911 (cortex_a15_neon_fp_vmul_ddd): Likewise.
9912 (cortex_a15_neon_fp_vmul_qqd): Likewise.
9913 (cortex_a15_neon_fp_vmla_ddd): Likewise.
9914 (cortex_a15_neon_fp_vmla_qqq): Likewise.
9915 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
9916 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
9917 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9918 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9919 (cortex_a15_neon_bp_simple): Likewise.
9920 (cortex_a15_neon_bp_2cycle): Likewise.
9921 (cortex_a15_neon_bp_3cycle): Likewise.
9922 (cortex_a15_neon_vld1_1_2_regs): Likewise.
9923 (cortex_a15_neon_vld1_3_4_regs): Likewise.
9924 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9925 (cortex_a15_neon_vld2_4_regs): Likewise.
9926 (cortex_a15_neon_vld3_vld4): Likewise.
9927 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9928 (cortex_a15_neon_vst1_3_4_regs): Likewise.
9929 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
9930 (cortex_a15_neon_vst3_vst4): Likewise.
9931 (cortex_a15_neon_vld1_vld2_lane): Likewise.
9932 (cortex_a15_neon_vld3_vld4_lane" 10
9933 (cortex_a15_neon_vst1_vst2_lane): Likewise.
9934 (cortex_a15_neon_vst3_vst4_lane): Likewise.
9935 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
9936 (cortex_a15_neon_ldm_2): Likewise.0
9937 (cortex_a15_neon_stm_2): Likewise.
9938 (cortex_a15_neon_mcr): Likewise.
9939 (cortex_a15_neon_mcr_2_mcrr): Likewise.
9940 (cortex_a15_neon_mrc): Likewise.
9941 (cortex_a15_neon_mrrc): Likewise.
9942 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9943 change.
9944 (cortex_a15_alu_shift): Likewise.
9945 (cortex_a15_alu_shift_reg): Likewise.
9946 (cortex_a15_mult32): Likewise.
9947 (cortex_a15_mult64): Likewise.
9948 (cortex_a15_block): Likewise.
9949 (cortex_a15_branch): Likewise.
9950 (cortex_a15_load1): Likewise.
9951 (cortex_a15_load3): Likewise.
9952 (cortex_a15_store1): Likewise.
9953 (cortex_a15_store3): Likewise.
9954 (cortex_a15_call): Likewise.
9955 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
9956 (cortex_a5_f2r): Likewise.
9957 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
9958 change.
9959 (cortex_a53_f2r): Likewise.
9960 * config/arm/cortex-a7.md
9961 (cortex_a7_branch): Update for attribute change.
9962 (cortex_a7_call): Likewise.
9963 (cortex_a7_alu_imm): Likewise.
9964 (cortex_a7_alu_reg): Likewise.
9965 (cortex_a7_alu_shift): Likewise.
9966 (cortex_a7_mul): Likewise.
9967 (cortex_a7_load1): Likewise.
9968 (cortex_a7_store1): Likewise.
9969 (cortex_a7_load2): Likewise.
9970 (cortex_a7_store2): Likewise.
9971 (cortex_a7_load3): Likewise.
9972 (cortex_a7_store3): Likewise.
9973 (cortex_a7_load4): Likewise.
9974 (cortex_a7_store4): Likewise.
9975 (cortex_a7_fpalu): Likewise.
9976 (cortex_a7_fconst): Likewise.
9977 (cortex_a7_fpmuls): Likewise.
9978 (cortex_a7_neon_mul): Likewise.
9979 (cortex_a7_fpmacs): Likewise.
9980 (cortex_a7_neon_mla: Likewise.
9981 (cortex_a7_fpmuld: Likewise.
9982 (cortex_a7_fpmacd: Likewise.
9983 (cortex_a7_fpfmad: Likewise.
9984 (cortex_a7_fdivs: Likewise.
9985 (cortex_a7_fdivd: Likewise.
9986 (cortex_a7_r2f: Likewise.
9987 (cortex_a7_f2r: Likewise.
9988 (cortex_a7_f_flags: Likewise.
9989 (cortex_a7_f_loads: Likewise.
9990 (cortex_a7_f_loadd: Likewise.
9991 (cortex_a7_f_stores: Likewise.
9992 (cortex_a7_f_stored: Likewise.
9993 (cortex_a7_neon): Likewise.
9994 * config/arm/cortex-a8-neon.md
9995 (cortex_a8_neon_mrc): Update for attribute change.
9996 (cortex_a8_neon_mrrc): Likewise.
9997 (cortex_a8_neon_int_1): Likewise.
9998 (cortex_a8_neon_int_2): Likewise.
9999 (cortex_a8_neon_int_3): Likewise.
10000 (cortex_a8_neon_int_4): Likewise.
10001 (cortex_a8_neon_int_5): Likewise.
10002 (cortex_a8_neon_vqneg_vqabs): Likewise.
10003 (cortex_a8_neon_vmov): Likewise.
10004 (cortex_a8_neon_vaba): Likewise.
10005 (cortex_a8_neon_vaba_qqq): Likewise.
10006 (cortex_a8_neon_vsma): Likewise.
10007 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10008 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10009 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10010 long_scalar): Likewise.
10011 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10012 (cortex_a8_neon_mla_qqq_8_16): Likewise.
10013 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10014 long_scalar_qdd_64_32_long): Likewise.
10015 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10016 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10017 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
10018 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10019 (cortex_a8_neon_shift_1): Likewise.
10020 (cortex_a8_neon_shift_2): Likewise.
10021 (cortex_a8_neon_shift_3): Likewise.
10022 (cortex_a8_neon_vshl_ddd): Likewise.
10023 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10024 (cortex_a8_neon_vsra_vrsra): Likewise.
10025 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
10026 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
10027 (cortex_a8_neon_fp_vsum): Likewise.
10028 (cortex_a8_neon_fp_vmul_ddd): Likewise.
10029 (cortex_a8_neon_fp_vmul_qqd): Likewise.
10030 (cortex_a8_neon_fp_vmla_ddd): Likewise.
10031 (cortex_a8_neon_fp_vmla_qqq): Likewise.
10032 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
10033 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
10034 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10035 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10036 (cortex_a8_neon_bp_simple): Likewise.
10037 (cortex_a8_neon_bp_2cycle): Likewise.
10038 (cortex_a8_neon_bp_3cycle): Likewise.
10039 (cortex_a8_neon_ldr): Likewise.
10040 (cortex_a8_neon_str): Likewise.
10041 (cortex_a8_neon_vld1_1_2_regs): Likewise.
10042 (cortex_a8_neon_vld1_3_4_regs): Likewise.
10043 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10044 (cortex_a8_neon_vld2_4_regs): Likewise.
10045 (cortex_a8_neon_vld3_vld4): Likewise.
10046 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10047 (cortex_a8_neon_vst1_3_4_regs): Likewise.
10048 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
10049 (cortex_a8_neon_vst3_vst4): Likewise.
10050 (cortex_a8_neon_vld1_vld2_lane): Likewise.
10051 (cortex_a8_neon_vld3_vld4_lane): Likewise.
10052 (cortex_a8_neon_vst1_vst2_lane): Likewise.
10053 (cortex_a8_neon_vst3_vst4_lane): Likewise.
10054 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
10055 (cortex_a8_neon_mcr): Likewise.
10056 (cortex_a8_neon_mcr_2_mcrr): Likewise.
10057 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10058 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
10059 change.
10060 (ca9_neon_mrrc): Likewise.
10061 (cortex_a9_neon_int_1): Likewise.
10062 (cortex_a9_neon_int_2): Likewise.
10063 (cortex_a9_neon_int_3): Likewise.
10064 (cortex_a9_neon_int_4): Likewise.
10065 (cortex_a9_neon_int_5): Likewise.
10066 (cortex_a9_neon_vqneg_vqabs): Likewise.
10067 (cortex_a9_neon_vmov): Likewise.
10068 (cortex_a9_neon_vaba): Likewise.
10069 (cortex_a9_neon_vaba_qqq): Likewise.
10070 (cortex_a9_neon_vsma): Likewise.
10071 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10072 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10073 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10074 long_scalar): Likewise.
10075 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10076 (cortex_a9_neon_mla_qqq_8_16): Likewise.
10077 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10078 long_scalar_qdd_64_32_long): Likewise.
10079 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10080 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10081 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
10082 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10083 (cortex_a9_neon_shift_1): Likewise.
10084 (cortex_a9_neon_shift_2): Likewise.
10085 (cortex_a9_neon_shift_3): Likewise.
10086 (cortex_a9_neon_vshl_ddd): Likewise.
10087 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10088 (cortex_a9_neon_vsra_vrsra): Likewise.
10089 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
10090 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
10091 (cortex_a9_neon_fp_vsum): Likewise.
10092 (cortex_a9_neon_fp_vmul_ddd): Likewise.
10093 (cortex_a9_neon_fp_vmul_qqd): Likewise.
10094 (cortex_a9_neon_fp_vmla_ddd): Likewise.
10095 (cortex_a9_neon_fp_vmla_qqq): Likewise.
10096 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
10097 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
10098 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10099 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10100 (cortex_a9_neon_bp_simple): Likewise.
10101 (cortex_a9_neon_bp_2cycle): Likewise.
10102 (cortex_a9_neon_bp_3cycle): Likewise.
10103 (cortex_a9_neon_ldr): Likewise.
10104 (cortex_a9_neon_str): Likewise.
10105 (cortex_a9_neon_vld1_1_2_regs): Likewise.
10106 (cortex_a9_neon_vld1_3_4_regs): Likewise.
10107 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10108 (cortex_a9_neon_vld2_4_regs): Likewise.
10109 (cortex_a9_neon_vld3_vld4): Likewise.
10110 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10111 (cortex_a9_neon_vst1_3_4_regs): Likewise.
10112 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
10113 (cortex_a9_neon_vst3_vst4): Likewise.
10114 (cortex_a9_neon_vld1_vld2_lane): Likewise.
10115 (cortex_a9_neon_vld3_vld4_lane): Likewise.
10116 (cortex_a9_neon_vst1_vst2_lane): Likewise.
10117 (cortex_a9_neon_vst3_vst4_lane): Likewise.
10118 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
10119 (cortex_a9_neon_mcr): Likewise.
10120 (cortex_a9_neon_mcr_2_mcrr): Likewise.
10121 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
10122 (cortex_a9_fps): Likewise.
10123 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
10124 change.
10125 (cortex_m4_fmuls): Likewise.
10126 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
10127 change.
10128 (cortex_r4_mrc): Likewise.
10129 * config/arm/iterators.md: Update comment referring to neon_type.
10130 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
10131 (iwmmxt_movsi_insn): Likewise.
10132 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
10133 attribute change.
10134 (pj4_core_to_vfp): Likewise.
10135 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
10136 attribute change.
10137 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
10138 (vfp_fstore): Likewise.
10139 * doc/md.texi: Change references to neon_type to refer to type.
10140
10141 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10142
10143 * tree.h (DECL_BUILT_IN): Fix typo in comment.
10144
10145 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
10146
10147 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
10148 lglobl if not weak.
10149
10150 2013-09-04 Easwaran Raman <eraman@google.com>
10151
10152 PR middle-end/57370
10153 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
10154 (build_and_add_sum): Use it.
10155 (appears_later_in_bb): Simplify code.
10156
10157 2013-09-04 Teresa Johnson <tejohnson@google.com>
10158
10159 * dumpfile.c (dump_finish): Don't close stderr/stdout.
10160
10161 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
10162
10163 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
10164
10165 2013-09-04 Jan Hubicka <jh@suse.cz>
10166
10167 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
10168 * ipa-devirt.c: Include diganostic.h
10169 (odr_type_d): Add types and types_set.
10170 (hash_type_name): Work for types with vtables during LTO.
10171 (odr_hasher::remove): Fix comment; destroy types_set.
10172 (add_type_duplicate): New function,
10173 (get_odr_type): Use it.
10174 (dump_type_inheritance_graph): Dump type duplicates.
10175 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
10176 graph.
10177 * tree.c (types_same_for_odr): Give exact answers on types with
10178 virtual tables.
10179
10180 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10181
10182 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
10183 explaining their differences.
10184
10185 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
10186
10187 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
10188
10189 2013-09-03 Jeff Law <law@redhat.com>
10190
10191 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
10192 when not threading through a joiner block. Pass joiner/no joiner
10193 state to register_jump_thread.
10194 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
10195 state from argument rather than implying on path length.
10196 Dump the entire jump thread path into debugging dump.
10197 * tree-flow.h (register_jump_thread): Update prototype.
10198
10199 2013-08-29 Xinliang David Li <davidxl@google.com>
10200
10201 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10202 Remove a trivial gcc_assert.
10203
10204 2013-08-29 Xinliang David Li <davidxl@google.com>
10205
10206 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
10207 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
10208 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10209 Delay base decl alignment adjustment.
10210 * tree-vectorizer.c (vect_destroy_datarefs): New function.
10211 * tree-vectorizer.h: New data structure.
10212 (set_dr_misalignment): New function.
10213 (dr_misalignment): Ditto.
10214 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
10215 (vectorizable_load): Ditto.
10216 (ensure_base_align): New function.
10217 (vectorize_loops): Add dbg_cnt support.
10218 (execute_vect_slp): Ditto.
10219 * dbgcnt.def: New debug counter.
10220 * Makefile: New dependency.
10221
10222 2013-09-03 Meador Inge <meadori@codesourcery.com>
10223
10224 Revert:
10225
10226 2013-08-30 Meador Inge <meadori@codesourcery.com>
10227
10228 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10229
10230 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
10231
10232 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
10233 function descriptor.
10234
10235 2013-09-03 Richard Biener <rguenther@suse.de>
10236
10237 * tree-affine.c (add_elt_to_tree): Fix association issue,
10238 avoid useless converts and make sure to always return a
10239 properly typed result.
10240
10241 2013-09-03 Richard Biener <rguenther@suse.de>
10242
10243 PR middle-end/57656
10244 * fold-const.c (negate_expr_p): Fix division case.
10245 (negate_expr): Likewise.
10246
10247 2013-09-03 Richard Biener <rguenther@suse.de>
10248
10249 PR lto/58285
10250 * tree-streamer-out.c: Include tm.h.
10251 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
10252
10253 2013-09-03 Jan Hubicka <jh@suse.cz>
10254
10255 * tree-profile.c (tree_profiling): Cleanup CFG when done.
10256
10257 2013-09-03 Alan Modra <amodra@gmail.com>
10258
10259 * config.gcc (powerpc*-*-linux*): Add support for little-endian
10260 multilibs to big-endian target and vice versa.
10261 * config/rs6000/t-linux64: Use := assignment on all vars.
10262 (MULTILIB_EXTRA_OPTS): Remove fPIC.
10263 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
10264 * config/rs6000/t-linux64le: New file.
10265 * config/rs6000/t-linux64bele: New file.
10266 * config/rs6000/t-linux64lebe: New file.
10267
10268 2013-09-02 Jan Hubicka <jh@suse.cz>
10269
10270 * ipa-inline-transform.c (inline_transform): Do not
10271 optimize_inline_calls when not optimizing.
10272
10273 2013-09-02 Jan Hubicka <jh@suse.cz>
10274
10275 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
10276 duplicated nodes for assembler names.
10277 * symtab.c (symtab_unregister_node): Do not attempt to unlink
10278 hard registers from assembler name hash.
10279
10280 2013-09-02 Jan Hubicka <jh@suse.cz>
10281
10282 * ipa-split.c (execute_split_functions): Split externally visible
10283 functions called once.
10284
10285 2013-09-02 Martin Jambor <mjambor@suse.cz>
10286
10287 PR ipa/58106
10288 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
10289 linked list. When finding the correct duplicate, also consider also
10290 the caller in additon to its inlined_to node.
10291
10292 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
10293
10294 * config/aarch64/aarch64-simd-builtins.def
10295 (dup_lane_scalar): Remove.
10296 * config/aarch64/aarch64-simd.md
10297 (aarch64_simd_dup): Add 'w->w' alternative.
10298 (aarch64_dup_lane<mode>): Allow for VALL.
10299 (aarch64_dup_lane_scalar<mode>): Remove.
10300 (aarch64_dup_lane_<vswap_width_name><mode>): New.
10301 (aarch64_get_lane_signed<mode>): Add w->w altenative.
10302 (aarch64_get_lane_unsigned<mode>): Likewise.
10303 (aarch64_get_lane<mode>): Likewise.
10304 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
10305 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
10306 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
10307 (VCON): Change container of V2SF.
10308 (vswap_width_name): Likewise.
10309 * config/aarch64/arm_neon.h
10310 (__aarch64_vdup_lane_any): New.
10311 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10312 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
10313 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10314
10315 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
10316
10317 PR middle-end/56382
10318 * expr.c (emit_move_complex): Do not move complex FP values as parts if
10319 the source or the destination is a single hard register.
10320
10321 2013-09-02 Richard Biener <rguenther@suse.de>
10322
10323 PR middle-end/57511
10324 * tree-scalar-evolution.c (instantiate_scev_name): Allow
10325 non-linear SCEVs.
10326
10327 2013-09-02 Richard Biener <rguenther@suse.de>
10328
10329 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
10330 arithmetic to sizetype.
10331
10332 2013-09-02 Bin Cheng <bin.cheng@arm.com>
10333
10334 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
10335 Find auto-increment use both before and after candidate.
10336
10337 2013-09-02 Marek Polacek <polacek@redhat.com>
10338
10339 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
10340
10341 2013-09-01 Jan Hubicka <jh@suse.cz>
10342
10343 * Makefile.in: Add ipa-profile.o
10344 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
10345 * cgraph.c (struct cgraph_propagate_frequency_data,
10346 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
10347 ipa-profile.c; replace cgraph_ by ipa_ prefix.
10348 * cgraph.h (cgraph_propagate_frequency): Remove.
10349 * ipa-inline-analysis.c: Include ipa-utils.h;
10350 drop duplicated cfgloop.h.
10351 (inline_update_callee_summaries): Update.
10352 * ipa-profile.c: New file.
10353 * ipa-utils.h (ipa_propagate_frequency): Declare.
10354 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
10355 data-streamer.h, value-prof.h.
10356 (symtab_remove_unreachable_nodes): Update profile.
10357 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
10358 account_time_size, cmp_counts, dump_histogram,
10359 ipa_profile_generate_summary, ipa_profile_write_summary,
10360 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
10361 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
10362 Move to ipa-profile.c.
10363
10364 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
10365
10366 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
10367
10368 2013-09-01 Jan Hubicka <jh@suse.cz>
10369
10370 * common.opt (fdevirtualize-speculatively): New function.
10371 * invoke.texi (fdevirtualize-speculatively): Document.
10372 * ipa-devirt.c: Include ipa-inline.h
10373 (likely_target_p): New function.
10374 (ipa_devirt): New function.
10375 (gate_ipa_devirt): New function.
10376 (pass_data_ipa_devirt): New static var.
10377 (pass_ipa_devirt): Likewise.
10378 (make_pass_ipa_devirt): New function.
10379 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
10380 (common_handle_option): Disable devirtualization when
10381 value range profiling is available.
10382 * passes.def (pass_ipa_devirt): Add.
10383 * timever.def (TV_IPA_DEVIRT): New timevar.
10384 * tree-pass.h (make_pass_ipa_devirt):
10385
10386 2013-09-01 Iain Sandoe <iain@codesourcery.com>
10387
10388 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
10389 include sanitize(undefined).
10390
10391 2013-08-31 Diego Novillo <dnovillo@google.com>
10392
10393 * Makefile.in (TREE_CORE_H): Define.
10394 (TREE_H): Use.
10395 (GTFILES): Add tree-core.h.
10396 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
10397 size the array.
10398 * tree-core.h: New file.
10399 Move all data structures, enum, typedefs, global
10400 declarations and constants from ...
10401 * tree.h: ... here.
10402
10403 2013-08-31 Jan Hubicka <jh@suse.cz>
10404
10405 * bulitins.c (expand_builtin): Do not early exit for gcov
10406 instrumented functions.
10407
10408 2013-08-31 Marek Polacek <polacek@redhat.com>
10409
10410 * ubsan.c: Include tm_p.h.
10411
10412 2013-08-31 Jan Hubicka <jh@suse.cz>
10413
10414 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
10415 warning.
10416
10417 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
10418 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
10419 * tree-cfg.c (verify_gimple_label): ... here.
10420 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
10421 (ipa_merge_profiles): New function.
10422 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
10423 (lto_input_function_body): Likewise.
10424 * ipa-utils.h (ipa_merge_profiles): Declare.
10425 * lto-streamer.h (lto_input_function_body): Update prototype.
10426 (emit_label_in_global_context_p): Remove.
10427 * lto-symtab.c: Include ipa-utils.h
10428 (lto_cgraph_replace_node): Use ipa_merge_profiles.
10429
10430 2013-08-31 Jan Hubicka <jh@suse.cz>
10431
10432 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
10433
10434 2013-08-31 Jan Hubicka <jh@suse.cz>
10435
10436 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
10437
10438 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
10439
10440 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
10441 "cmp" RTX before signed_comparison_operator check to account
10442 for "code" changes.
10443
10444 2013-08-30 Jan Hubicka <jh@suse.cz>
10445
10446 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
10447 (detect_type_change_1): Rename to ...
10448 (detect_type_change): ... this one; early return on non-polymorphic
10449 types.
10450 (detect_type_change_ssa): Add comp_type parameter; update
10451 use of detect_type_change.
10452 (compute_complex_assign_jump_func): Add param_type parameter;
10453 update use of detect_type_change_ssa.
10454 (compute_complex_ancestor_jump_func): Likewise.
10455 (ipa_get_callee_param_type): New function.
10456 (ipa_compute_jump_functions_for_edge): Compute parameter type;
10457 update calls to the jump function computation functions.
10458
10459 2013-08-30 Teresa Johnson <tejohnson@google.com>
10460 Steven Bosscher <steven@gcc.gnu.org>
10461
10462 * cfgrtl.c (fixup_new_cold_bb): New routine.
10463 (commit_edge_insertions): Invoke fixup_partitions.
10464 (find_partition_fixes): New routine.
10465 (fixup_partitions): Ditto.
10466 (verify_hot_cold_block_grouping): Update comments.
10467 (rtl_verify_edges): Invoke find_partition_fixes.
10468 (rtl_verify_bb_pointers): Update comments.
10469 (rtl_verify_bb_layout): Ditto.
10470 * basic-block.h (probably_never_executed_edge_p): Declare.
10471 (fixup_partitions): Ditto.
10472 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
10473 * bb-reorder.c (sanitize_hot_paths): New function.
10474 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
10475 sanitize_hot_paths.
10476 * predict.c (probably_never_executed_edge_p): New routine.
10477 * cfg.c (check_bb_profile): Add partition insanity warnings.
10478
10479 2013-08-30 Meador Inge <meadori@codesourcery.com>
10480
10481 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10482
10483 2013-08-30 Marek Polacek <polacek@redhat.com>
10484
10485 * Makefile.in (ubsan.o): Add.
10486 (c-family/c-ubsan.o): Add.
10487 (builtins.o): Add ubsan.h dependency.
10488 * ubsan.h: New file.
10489 * ubsan.c: New file.
10490 * common.opt: Add -fsanitize=undefined option.
10491 (flag_sanitize): Add variable.
10492 (fsanitize=): Add option. Add Driver.
10493 (fsanitize=thread): Remove option.
10494 (fsanitize=address): Likewise.
10495 (static-libubsan): New option.
10496 * doc/invoke.texi: Document the new flag and -static-libubsan.
10497 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
10498 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
10499 * builtin-attrs.def (ATTR_COLD): Define.
10500 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10501 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
10502 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
10503 * flag-types.h (sanitize_code): New enum.
10504 * opts.c (common_handle_option): Parse command line arguments
10505 of -fsanitize=. Add -fsanitize=unreachable option.
10506 * varasm.c (get_variable_section): Adjust.
10507 (assemble_noswitch_variable): Likewise.
10508 (assemble_variable): Likewise.
10509 (output_constant_def_contents): Likewise.
10510 (categorize_decl_for_section): Likewise.
10511 (place_block_symbol): Likewise.
10512 (output_object_block): Likewise.
10513 * builtins.def: Likewise.
10514 * toplev.c (compile_file): Likewise.
10515 (process_options): Likewise.
10516 * cppbuiltin.c: Likewise.
10517 * tsan.c (tsan_pass): Likewise.
10518 (tsan_gate): Likewise.
10519 (tsan_gate_O0): Likewise.
10520 * cfgexpand.c (partition_stack_vars): Likewise.
10521 (expand_stack_vars): Likewise.
10522 (defer_stack_allocation): Likewise.
10523 (expand_used_vars): Likewise.
10524 * cfgcleanup.c (old_insns_match_p): Likewise.
10525 * asan.c (asan_finish_file): Likewise.
10526 (asan_instrument): Likewise.
10527 (gate_asan): Likewise.
10528 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
10529 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10530 (asan_global_struct): Use pointer_sized_int_node instead
10531 calling build_nonstandard_integer_type.
10532 (initialize_sanitizer_builtins): Likewise.
10533 (asan_finish_file): Likewise.
10534 * gcc.c: Document %{%:function(args):X}.
10535 (static_spec_functions): Add sanitize.
10536 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
10537 store funcval != NULL there.
10538 (do_spec_1): Adjust handle_spec_function caller.
10539 (handle_braces): Allow %:function(args) as condition.
10540 (sanitize_spec_function): New function.
10541 (ADD_STATIC_LIBUBSAN_LIBS): Define.
10542 (LIBUBSAN_SPEC): Likewise.
10543 (LIBUBSAN_EARLY_SPEC): Likewise.
10544 (SANITIZER_SPEC): Handle libubsan.
10545 (SANITIZER_EARLY_SPEC): Likewise.
10546 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
10547 instead of fsanitize=address.
10548 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
10549 instead of fsanitize=address*.
10550 * builtins.c: Include ubsan.h.
10551 (fold_builtin_0): Instrument __builtin_unreachable.
10552 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
10553 instead of flag_asan.
10554 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
10555 (pointer_sized_int_node): Define.
10556 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
10557
10558 2013-08-30 Mike Stump <mikestump@comcast.net>
10559
10560 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
10561 with RE patterns.
10562
10563 2013-08-29 Jan Hubicka <jh@suse.cz>
10564
10565 * cgraph.c (cgraph_function_body_availability): Handle weakref
10566 correctly.
10567 * passes.def: Remove pass_fixup_cfg.
10568 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
10569 track when we need to remove functions.
10570 (gate_ipa_inline): Execute inlining always; add comment why.
10571 (pass_data_ipa_inline): Remove TODO_remove_functions.
10572 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
10573 do not produce summaries.
10574 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
10575 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
10576 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
10577 constant pool and vtable.
10578
10579 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
10580
10581 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
10582 New arm_neon.h's internal macros to specify 64-bit constants.
10583 Avoid using stdint.h's macros.
10584
10585 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
10586
10587 * recog.c (verify_changes): Verify that changes[i].old is non-zero
10588 before applying REG_P.
10589
10590 2013-08-30 Jakub Jelinek <jakub@redhat.com>
10591
10592 PR tree-optimization/58277
10593 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
10594 after seeing too many stmts with vdef in between dombb and current
10595 bb, invalidate everything.
10596
10597 2013-08-30 Richard Biener <rguenther@suse.de>
10598
10599 * fold-const.c (fold_single_bit_test): Fix overflow test.
10600
10601 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
10602
10603 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
10604 and which can live in a register, always retrieve the value on entry.
10605 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
10606 passed by invisible reference specially.
10607 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
10608 (vt_add_function_parameter): Correctly deal with a parameter passed by
10609 invisible reference.
10610
10611 2013-08-30 Jan Hubicka <jh@suse.cz>
10612
10613 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
10614
10615 2013-08-30 Richard Biener <rguenther@suse.de>
10616
10617 PR tree-optimization/58228
10618 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
10619 allow invariant loads in nested loop vectorization.
10620
10621 2013-08-30 Richard Biener <rguenther@suse.de>
10622
10623 PR tree-optimization/58223
10624 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
10625 (has_anti_or_output_dependence): ... this and adjust to also
10626 look for output dependences.
10627 (mark_nodes_having_upstream_mem_writes): Adjust.
10628 (rdg_flag_uses): Likewise.
10629
10630 2013-08-30 Richard Biener <rguenther@suse.de>
10631
10632 PR tree-optimization/58010
10633 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
10634 assert that we have a loop-closed PHI.
10635
10636 2013-08-29 Jan Hubicka <jh@suse.cz>
10637
10638 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
10639 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
10640 * lto-section-in.c (lto_free_function_in_decl_state): New function.
10641 (lto_free_function_in_decl_state_for_node): New function.
10642
10643 2013-08-29 Xinliang David Li <davidxl@google.com>
10644
10645 * loop-unroll.c (report_unroll_peel): Minor message change.
10646 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
10647 Emit alignment peeling message with default -fopt-info.
10648 (vect_loop_versioning): Emit loop version info message.
10649 * tree-vectorizer.c (vectorize_loops): Minor message change.
10650 (execute_vect_slp): Ditto.
10651
10652 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
10653
10654 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
10655 of the clone from the DECL_NAME of the original function.
10656
10657 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
10658
10659 * passes.c (register_pass): Add overload.
10660 * tree-pass.h (register_pass): Forward declare it. Add comment.
10661
10662 2013-08-29 Jan Hubicka <jh@suse.cz>
10663
10664 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
10665 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
10666 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
10667 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10668 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
10669 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
10670 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10671 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
10672
10673 2013-08-29 Teresa Johnson <tejohnson@google.com>
10674
10675 * dumpfile.c (dump_loc): Output column number.
10676 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
10677 * doc/invoke.texi: Document optall -fopt-info flag.
10678 * profile.c (read_profile_edge_counts): Use new dump framework.
10679 (compute_branch_probabilities): Ditto.
10680 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
10681 when pass not in any opt group.
10682 * pass_manager.h (pass_manager::get_pass_profile): New method.
10683 * value-prof.c (check_counter): Use new dump framework.
10684 (check_ic_target): Ditto.
10685 * coverage.c (get_coverage_counts): Ditto.
10686 (coverage_init): Setup new dump framework.
10687
10688 2013-08-29 Richard Biener <rguenther@suse.de>
10689
10690 PR tree-optimization/58246
10691 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
10692 handle the dominance check inside a basic-block.
10693
10694 2013-08-29 Richard Biener <rguenther@suse.de>
10695
10696 PR middle-end/57287
10697 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
10698 of default defs that appear in abnormal PHI nodes.
10699
10700 2013-08-29 Richard Biener <rguenther@suse.de>
10701
10702 PR tree-optimization/57685
10703 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
10704 single-use operands to avoid exponential complexity.
10705
10706 2013-08-28 Dehao Chen <dehao@google.com>
10707
10708 * ipa-inline.c (edge_badness): Fix integer underflow.
10709
10710 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
10711
10712 * gtm-builtins.def (_ITM_free): Declare leaf.
10713
10714 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10715
10716 PR target/58067
10717 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
10718 (*tls_local_dynamic_base_64_largepic): Likewise.
10719 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
10720 Remove predicate from call operand.
10721 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
10722 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
10723
10724 2013-08-28 Jeff Law <law@redhat.com>
10725
10726 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
10727 checks for the number of predecessors and successors allowed.
10728 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
10729 which require copying a joiner block if there is a request which
10730 is a subpath that requires no joiner block copying.
10731
10732 2013-08-28 Jan Hubicka <jh@suse.cz>
10733
10734 * lto-streamer-out.c (DFS_write_tree_body): Drop
10735 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
10736 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
10737 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
10738 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
10739 Do not read DECL_ERROR_ISSUED.
10740 (unpack_ts_decl_with_vis_value_fields): Do not read
10741 DECL_DEFER_OUTPUT.
10742 (lto_input_ts_binfo_tree_pointers): Do not read
10743 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
10744 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
10745 write DECL_ERROR_ISSUED..
10746 (pack_ts_decl_with_vis_value_fields): Do not write
10747 DECL_DEFER_OUTPUT.
10748 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
10749 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
10750 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
10751 * tree.h (tree_decl_common): Update comment.
10752 (DECL_ERROR_ISSUED): Remove.
10753
10754 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10755
10756 PR middle-end/58257
10757 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
10758
10759 2013-08-28 Jan Hubicka <jh@suse.cz>
10760
10761 * builtins.def (free): Declare leaf.
10762
10763 2013-08-27 David Malcolm <dmalcolm@redhat.com>
10764
10765 * gdbhooks.py: New.
10766 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
10767 * configure: Regenerate.
10768
10769 2013-08-27 Martin Jambor <mjambor@suse.cz>
10770
10771 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
10772 (ipa_ancestor_jf_data): Likewise.
10773 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
10774 (ipa_get_jf_pass_through_type_preserved): New function.
10775 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
10776 (ipa_get_jf_ancestor_type_preserved): New function.
10777 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
10778 (ipa_get_jf_ancestor_result): Likewise.
10779 (propagate_vals_accross_pass_through): Use
10780 ipa_get_jf_pass_through_result to do all the value mappings.
10781 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
10782 type_preserved flag.
10783 (ipa_set_jf_cst_copy): New function.
10784 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
10785 (ipa_set_jf_arith_pass_through): Likewise.
10786 (ipa_set_ancestor_jf): Likewise.
10787 (compute_complex_assign_jump_func): Set type_preserved instead of
10788 punting.
10789 (ipa_compute_jump_functions_for_edge): Likewise.
10790 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
10791 (update_jump_functions_after_inlining): Update type_preserved.
10792 Explicitely create jump functions when combining one with pass_through.
10793 (ipa_write_jump_function): Stream the type_preserved flags.
10794 (ipa_read_jump_function): Likewise.
10795
10796 2013-08-27 Jakub Jelinek <jakub@redhat.com>
10797 Aldy Hernandez <aldyh@redhat.com>
10798
10799 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
10800 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
10801 * function.h (struct function): Add has_force_vect_loops and
10802 has_simduid_loops.
10803 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
10804 * gimple.c (gimple_build_omp_critical): Add KIND argument and
10805 handle it.
10806 * gimple.def: Update CLAUSES comments.
10807 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
10808 (gimple_build_omp_for): Add argument to prototype.
10809 (gimple_omp_for_kind): New.
10810 (gimple_omp_for_set_kind): New.
10811 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
10812 GOVD_DATA_SHARE_CLASS.
10813 (enum omp_region_type): Add ORT_SIMD.
10814 (gimple_add_tmp_var): Handle ORT_SIMD.
10815 (gimplify_var_or_parm_decl): Same.
10816 (is_gimple_stmt): Same.
10817 (omp_firstprivatize_variable): Same.
10818 (omp_add_variable): Only use splay_tree_insert if lookup failed.
10819 (omp_notice_variable): Handle ORT_SIMD.
10820 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
10821 (omp_check_private): Handle ORT_SIMD.
10822 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10823 OMP_CLAUSE_SAFELEN.
10824 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
10825 Handle OMP_CLAUSE_LASTPRIVATE.
10826 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10827 OMP_CLAUSE_SAFELEN.
10828 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
10829 (gimplify_expr): Handle OMP_SIMD.
10830 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
10831 (expand_GOMP_SIMD_VF): New.
10832 (expand_GOMP_SIMD_LAST_LANE): New.
10833 * internal-fn.def (GOMP_SIMD_LANE): New.
10834 (GOMP_SIMD_VF): New.
10835 (GOMP_SIMD_LAST_LANE): New.
10836 * omp-low.c: Include target.h.
10837 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
10838 OMP_CLAUSE_SAFELEN.
10839 (check_omp_nesting_restrictions): Same.
10840 (omp_max_vf): New.
10841 (lower_rec_simd_input_clauses): New.
10842 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
10843 OMP_CLAUSE_LINEAR.
10844 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
10845 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
10846 (expand_omp_build_assign): New.
10847 (expand_omp_for_init_counts): New.
10848 (expand_omp_for_init_vars): New.
10849 (extract_omp_for_update_vars): New.
10850 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
10851 and rewrite accordingly.
10852 (expand_omp_simd): New.
10853 (expand_omp_for): Use expand_omp_simd.
10854 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
10855 (lower_omp_for): Do not lower the body.
10856 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
10857 in their own loops.
10858 * tree-flow.h (find_omp_clause): Remove prototype.
10859 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
10860 forcing vectorization of the loop, or if flag_tree_vectorize.
10861 (gate_tree_if_conversion): Similarly.
10862 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
10863 gimple_build_omp_for.
10864 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
10865 * tree-parloops (create_parallel_loop): Pass kind argument to
10866 gimple_build_omp_for.
10867 * tree-pretty-print.c (dump_omp_clause): Add cases for
10868 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
10869 OMP_CLAUSE__SIMDUID_.
10870 (dump_generic_node): Handle OMP_SIMD.
10871 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
10872 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
10873 unroll OMP_SIMD loops here.
10874 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
10875 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
10876 loop->safelen.
10877 (vect_analyze_data_refs): Handle simd loops.
10878 * tree-vect-loop.c (vectorizable_live_operation): Handle
10879 IFN_GOMP_SIMD*.
10880 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
10881 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
10882 (vectorizable_load): Same.
10883 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
10884 (struct simduid_to_vf): New.
10885 (simduid_to_vf::hash): New.
10886 (simduid_to-vf::equal): New.
10887 (struct simd_array_to_simduid): New.
10888 (simd_array_to_simduid::hash): New.
10889 (simd_array_to_simduid::equal): New.
10890 (adjust_simduid_builtins): New.
10891 (struct note_simd_array_uses_struct): New.
10892 (note_simd_array_uses_cb): New.
10893 (note_simd_array_uses): New.
10894 (vectorize_loops): Handle simd hints and adjust simd builtins
10895 accordingly.
10896 * tree-vectorizer.h (struct _stmt_vec_info): Add
10897 simd_lane_access_p field.
10898 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
10899 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
10900 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
10901 (omp_clause_code_name): Same.
10902 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
10903 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
10904 * tree.def (OMP_SIMD): New entry.
10905 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
10906 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
10907 (OMP_CLAUSE_DECL): Adjust range for new clauses.
10908 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
10909 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
10910 (OMP_CLAUSE_LINEAR_STEP): New.
10911 (OMP_CLAUSE_SAFELEN_EXPR): New.
10912 (OMP_CLAUSE__SIMDUID__DECL): New.
10913 (find_omp_clause): New prototype.
10914
10915 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
10916
10917 * config/i386/driver-i386.c (host_detect_local_cpu): Update
10918 Haswell processor detection.
10919
10920 2013-08-27 Christian Widmer <shadow@umbrox.de>
10921
10922 PR target/57927
10923 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
10924 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
10925 AVX2 capable processors.
10926
10927 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
10928
10929 * config/aarch64/arm_neon.h: Replace all inline asm implementations
10930 of vget_low_* with implementations in terms of other intrinsics.
10931
10932 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
10933
10934 PR middle-end/57219
10935 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
10936 values to -1, 0 and 1.
10937
10938 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
10939
10940 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
10941 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
10942 (<optab><mode>3_insn): Remove.
10943 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
10944 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
10945 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
10946 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
10947 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
10948 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
10949 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
10950 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
10951 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
10952 (ror<mode>3_insn): Likewise.
10953 * config/aarch64/predicates.md (aarch64_simd_register): New.
10954
10955 2013-08-27 Richard Biener <rguenther@suse.de>
10956
10957 PR tree-optimization/57521
10958 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
10959 one edge is non-critical.
10960 (find_phi_replacement_condition): Make sure to use a non-critical
10961 edge. Cleanup and remove old bug workarounds.
10962 (bb_postdominates_preds): Remove.
10963 (if_convertible_loop_p_1): Do not compute post-dominators.
10964 (combine_blocks): Do not free post-dominators.
10965 (main_tree_if_conversion): Likewise.
10966 (pass_data_if_conversion): Add TODO_verify_ssa.
10967
10968 2013-08-27 DJ Delorie <dj@redhat.com>
10969
10970 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
10971
10972 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
10973
10974 * function.c (assign_parm_find_data_types): Set passed_mode and
10975 nominal_mode to the TYPE_MODE of nominal_type for the built
10976 pointer type in case of the struct-pass-by-reference.
10977
10978 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10979
10980 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
10981 (UINT16_TYPE): Change default to "unsigned int".
10982
10983 * config/avr/avr.opt (mfract-convert-truncate): New option.
10984 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
10985 is set, round negative fractional integers according to n1169
10986 when converting to integer types.
10987
10988 2013-08-26 Jan Hubicka <jh@suse.cz>
10989
10990 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
10991 methods can not be called indirectly when their address is not taken.
10992
10993 2013-08-26 Jan Hubicka <jh@suse.cz>
10994
10995 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
10996 ctor_for_folding.
10997
10998 2013-08-26 Jan Hubicka <jh@suse.cz>
10999
11000 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
11001 can be unshared.
11002
11003 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11004
11005 * reload.c (find_valid_class): Allow classes that do not include
11006 FIRST_PSEUDO_REGISTER - 1.
11007
11008 2013-08-26 Jan Hubicka <jh@suse.cz>
11009
11010 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
11011 fix edge count/frequency when speculation failed; fix type check
11012 for the direct call.
11013
11014 2013-08-26 Jan Hubicka <jh@suse.cz>
11015
11016 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
11017
11018 2013-08-26 Jan Hubicka <jh@suse.cz>
11019
11020 * ipa-inline-transform.c (inline_transform): Be ready for basic block
11021 to be changed by edge redirection.
11022
11023 2013-08-26 Jan Hubicka <jh@suse.cz>
11024
11025 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
11026 formating; add sanity check.
11027 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
11028 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
11029 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
11030 (resolve_noninline_speculation): Update callee keys, too.
11031
11032 2013-08-26 Jan Hubicka <jh@suse.cz>
11033
11034 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
11035 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
11036
11037 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11038
11039 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
11040 into proper place.
11041
11042 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
11043
11044 * config/i386/i386.c (ix86_debug_options): Remove prototype.
11045 (x86_64_elf_select_section): Ditto.
11046 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
11047 arguments.
11048 (ix86_pass_by_reference): Ditto.
11049 (output_set_got): Ditto.
11050 (ix86_unary_operator_ok): Ditto.
11051 (ix86_expand_builtin): Ditto.
11052
11053 2013-08-23 Jan Hubicka <jh@suse.cz>
11054
11055 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
11056
11057 2013-08-23 Jan Hubicka <jh@suse.cz>
11058
11059 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
11060 (tree_decl_with_vis): Add FINAL field.
11061
11062 2013-08-23 Jeff Law <law@redhat.com>
11063
11064 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
11065 the debugging dump when the expression is fully redundant.
11066
11067 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11068
11069 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
11070 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
11071 * pretty-print.c (pp_formatted_text_data): Likewise.
11072 (pp_write_text_to_stream): Likewise.
11073 (pp_write_text_as_dot_label_to_stream): Likewise.
11074 (pp_append_r): Likewise.
11075 (pp_format): Likewise.
11076 (pp_flush): Likewise.
11077 (pp_clear_output_area): Likewise.
11078 (pp_append_text): Likewise.
11079 (pp_formatted_text): Likewise.
11080 (pp_remaining_character_count_for_line): Likewise.
11081 (pp_newline): Likewise.
11082 (pp_character): Likewise.
11083 (output_buffer::~output_buffer): Define.
11084 (pretty_printer::~pretty_printer): Destruct output buffer.
11085 * pretty-print.h (output_buffer::~output_buffer): Declare.
11086 (pretty_printer::~pretty_printer): Declare virtual.
11087
11088 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
11089
11090 PR other/57324
11091 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
11092 HOST_WIDE_INT_M1U): New macros.
11093 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
11094 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
11095 unsigned -1 for lshift.
11096 * cse.c (cse_insn): Likewise.
11097 * double-int.c (rshift_double, lshift_double): Likewise.
11098 * builtins.c (fold_builtin_bitop): Likewise.
11099 * combine.c (force_to_mode): Likewise.
11100 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
11101 * simplify-rtx.c (simplify_const_unary_operation,
11102 simplify_const_binary_operation): Likewise.
11103 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
11104 check_va_list_escapes): Likewise.
11105 * rtlanal.c (nonzero_bits1): Likewise.
11106 * expmed.c (expand_smod_pow2): Likewise.
11107 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
11108
11109 2013-08-23 Jan Hubicka <jh@suse.cz>
11110
11111 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
11112 as having address taken.
11113
11114 2013-08-23 Jan Hubicka <jh@suse.cz>
11115
11116 * ipa-utils.h (method_class_type): Declare.
11117 * ipa-devirt.c (method_class_type): Export.
11118
11119 * cgraphunit.c (analyze_functions): Do basic devirtualization;
11120 do not walk base classes of anonymous types.
11121
11122 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
11123
11124 PR rtl-optimization/58220
11125 PR regression/58221
11126 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
11127 handle SEQUENCE insns properly.
11128
11129 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11130
11131 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
11132 definition.
11133 (pp_newline_and_indent): Likewise.
11134 (pp_separate_with): Likewise.
11135 * pretty-print.c (pp_newline_and_flush): Define.
11136 (pp_newline_and_indent): Likewise.
11137 (pp_separate_with): Likewise.
11138
11139 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11140
11141 PR target/58218
11142 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
11143 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
11144
11145 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
11146
11147 * config/i386/predicates.md (ext_sse_reg_operand): New.
11148 * config/i386/i386.md (*movti_internal): Use
11149 predicate to determine if EVEX is needed.
11150 (*movsi_internal): Ditto.
11151 (*movdf_internal): Ditto.
11152 (*movsf_internal): Ditto.
11153 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
11154
11155 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11156
11157 PR tree-optimization/58209
11158 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
11159 (find_tail_calls): Give up for pointer result types if m is non-NULL.
11160 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
11161 emit POINTER_PLUS_EXPR.
11162 (create_tailcall_accumulator): For pointer result type accumulate in
11163 sizetype type.
11164
11165 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
11166
11167 * configure.ac: Add backslashes missing from the last change.
11168 * configure: Regenerate.
11169
11170 2013-08-22 Jan Hubicka <jh@susue.cz>
11171
11172 * ipa.c (function_and_variable_visibility): First remember function
11173 was global and then make it local.
11174
11175 2013-08-22 Julian Brown <julian@codesourcery.com>
11176
11177 * configure.ac: Add aarch64 to list of arches which use "nop" in
11178 debug_line test.
11179 * configure: Regenerate.
11180
11181 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11182
11183 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
11184 gnu_libc_has_function.
11185 * config/s390/tpf.h: Likewise.
11186
11187 2013-08-22 Jan Hubicka <jh@susue.cz>
11188
11189 * timevar.c (validate_phases): Add cast.
11190
11191 2013-08-22 Jan Hubicka <jh@susue.cz>
11192
11193 * timevar.c (validate_phases): Use size_t for memory.
11194 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
11195
11196 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
11197
11198 * pretty-print.h (output_buffer::output_buffer): Declare.
11199 (pretty_printer::pretty_printer): Likewise.
11200 (pp_construct): Remove.
11201 * pretty-print.c (output_buffer::output_buffer): Define.
11202 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
11203 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
11204 (print_gimple_expr): Likewise.
11205 (print_gimple_seq): Likewise.
11206 (gimple_dump_bb): Likewise.
11207 * sched-vis.c (dump_value_slim): Likewise.
11208 (dump_insn_slim): Likewise.
11209 (dump_rtl_slim): Likewise.
11210 (str_pattern_slim): Likewise.
11211 * tree-mudflap.c (mf_varname_tree): Likewise.
11212 * graph.c (print_graph_cfg): Likewise.
11213 (start_graph_dump): Likewise.
11214 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
11215 placement-new.
11216 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
11217 pretty printer initialization.
11218 * coretypes.h (diagnostic_context): Remove superflous type alias
11219 declaration.
11220 (pretty_printer): Likewise. Declare directly as a class.
11221 (pretty_print_info): Remove declaration as class.
11222 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
11223 and pp_clear_output_area.
11224 (asan_add_global): Likewise.
11225
11226 2013-08-22 Jan Hubicka <jh@suse.cz>
11227
11228 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
11229 * ipa-utils.h (update_type_inheritance_graph): Declare.
11230 (possible_polymorphic_call_target_p): Declare.
11231 (possible_polymorphic_call_target_p): New.
11232 * ipa-devirt.c: Update toplevel comments.
11233 (cached_polymorphic_call_targets): Move up.
11234 (odr_type_d): Move ID down.
11235 (polymorphic_type_binfo_p): Update comment.
11236 (odr_hasher::remove): Likewise;
11237 (get_odr_type): Set anonymous_namespace.
11238 (dump_odr_type): Dump it.
11239 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
11240 (maybe_record_node): Record node in cached_polymorphic_call_targets.
11241 (record_binfo): Add comment.
11242 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
11243 built.
11244 (devirt_node_removal_hook): Do not iCE when cache is freed.
11245 (possible_polymorphic_call_target_p): New predicate.
11246 (update_type_inheritance_graph): New function.
11247
11248 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
11249 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
11250 Sergey Lega <sergey.s.lega@intel.com>
11251 Anna Tikhonova <anna.tikhonova@intel.com>
11252 Ilya Tocar <ilya.tocar@intel.com>
11253 Andrey Turetskiy <andrey.turetskiy@intel.com>
11254 Ilya Verbin <ilya.verbin@intel.com>
11255 Kirill Yukhin <kirill.yukhin@intel.com>
11256 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11257
11258 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
11259 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
11260 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
11261 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
11262 (OPTION_MASK_ISA_AVX2_UNSET): Update.
11263 (OPTION_MASK_ISA_AVX512F_UNSET): New.
11264 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
11265 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
11266 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
11267 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
11268 OPT_mavx512pf, OPT_mavx512er cases.
11269 * config/i386/constraints.md (v): New constraint.
11270 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
11271 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
11272 (bit_AVX512CD): New.
11273 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11274 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
11275 * config/i386/i386-c.c (ix86_target_macros_internal):
11276 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
11277 __AVX512PF__.
11278 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
11279 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
11280 * config/i386/i386.c (regclass_map, dbx_register_map)
11281 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
11282 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
11283 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
11284 -mavx512pf options.
11285 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
11286 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
11287 -mavx512pf options. Fix formatting.
11288 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
11289 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
11290 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
11291 -mavx512cd, -mavx512pf options.
11292 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
11293 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
11294 (ix86_preferred_output_reload_class): Replace SSE_REGS with
11295 ALL_SSE_REGS.
11296 (ix86_hard_regno_mode_ok): Support 512-bit registers.
11297 (ix86_set_reg_reg_cost): Ditto.
11298 (x86_order_regs_for_local_alloc): Ditto.
11299 (MAX_VECT_LEN): Extend to 64-byte.
11300 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
11301 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
11302 (TARGET_AVX512ER, TARGET_AVX512CD): New.
11303 (BIGGEST_ALIGNMENT): Extend to 512-bits.
11304 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
11305 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
11306 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
11307 (SSE_REG_MODE_P): Support new modes.
11308 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
11309 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
11310 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
11311 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
11312 (REG_CLASS_CONTENTS): Add new registers.
11313 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
11314 (EXT_REX_SSE_REGNO_P): New.
11315 (HI_REGISTER_NAMES): Add new registers.
11316 * config/i386/i386.md: Define constants for new registers.
11317 (mode): Add new 512-bit modes.
11318 (prefix): Support evex prefix.
11319 (isa): Support avx512f, noavx512f, fma_avx512f.
11320 (ssemodesuffix): Add new 512-bit modes.
11321 (movxi): New.
11322 (*movxi_internal_avx512f): Ditto.
11323 (*movdi_internal): Replace constraint "x" with the new constraint "v".
11324 Support MODE_XI.
11325 (*movsi_internal): Likewise.
11326 (*movdf_internal): Likewise.
11327 (*movsf_internal): Likewise.
11328 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
11329 (<code><mode>3): Likewise.
11330 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
11331 New.
11332 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
11333 with the new constraint "v".
11334 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
11335 modes.
11336 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
11337 with the new constraint "v".
11338 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
11339 (<sse2>_storedqu<avxsizesuffix>): Likewise.
11340 (*<plusminus_insn><mode>3): Likewise.
11341 (<sse>_vm<plusminus_insn><mode>3): Likewise.
11342 (*mul<mode>3): Likewise.
11343 (<sse>_vmmul<mode>3): Likewise.
11344 (<sse>_div<mode>3): Likewise.
11345 (<sse>_vmdiv<mode>3): Likewise.
11346 (<sse>_sqrt<mode>2): Likewise.
11347 (<sse>_vmsqrt<mode>2): Likewise.
11348 (*<code><mode>3_finite): Likewise.
11349 (*<code><mode>3) <smaxmin>: Likewise.
11350 (<sse>_vm<code><mode>3): Likewise.
11351 (*<code><mode>3) <any_logic>: Likewise.
11352 (*fma_fmadd_<mode>): Likewise.
11353 (*fma_fmsub_<mode>): Likewise.
11354 (*fma_fnmadd_<mode>): Likewise.
11355 (*fma_fnmsub_<mode>): Likewise.
11356 (*fma_fmaddsub_<mode>): Likewise.
11357 (*fma_fmsubadd_<mode>): Likewise.
11358 (*fmai_fmadd_<mode>): Likewise.
11359 (*fmai_fmsub_<mode>): Likewise.
11360 (*fmai_fnmadd_<mode>): Likewise.
11361 (*fmai_fnmsub_<mode>): Likewise.
11362 (sse_cvtsi2ss): Likewise.
11363 (sse_cvtsi2ssq): Likewise.
11364 (sse_cvtss2si): Likewise.
11365 (sse_cvtss2si_2): Likewise.
11366 (sse_cvtss2siq): Likewise.
11367 (sse_cvtss2siq_2): Likewise.
11368 (sse_cvttss2si): Likewise.
11369 (sse_cvtss2siq_2): Likewise.
11370 (float<sseintvecmodelower><mode>2): Likewise.
11371 (sse2_cvtsd2si_2): Likewise.
11372 (sse2_cvtsd2siq_2): Likewise.
11373 (*<plusminus_insn><mode>3): Likewise.
11374 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
11375 (*<sse4_1_avx2>_mul<mode>3): Likewise.
11376 (ashr<mode>3): Likewise.
11377 (<shift_insn><mode>3): Likewise.
11378 (avx2_<code><mode>3): Likewise.
11379 (*avx2_<code><mode>3): Likewise.
11380 (*andnot<mode>3): Likewise.
11381 (*<code><mode>3) <any_logic>: Likewise.
11382 (abs<mode>2): Likewise.
11383 (avx2_permvar<mode>): Likewise.
11384 (avx2_perm<mode>_1): Likewise.
11385 (*avx_vpermilp<mode>): Likewise.
11386 (avx_vpermilvar<mode>3): Likewise.
11387 (avx2_ashrv<mode>): Likewise.
11388 (avx2_<shift_insn>v<mode>): Likewise.
11389 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
11390 -mavx512cd.
11391 * doc/rtl.texi: Document XImode.
11392
11393 2013-08-21 Jeff Law <law@redhat.com>
11394
11395 * tree-flow.h (register_jump_thread): Pass vector of edges
11396 instead of each important edge.
11397 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
11398 thread path into a vector and pass that to register_jump_thread.
11399 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
11400 passed in edge vector to the current 3-edge form.
11401
11402 Revert:
11403 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11404
11405 * dce.c (fini_dce): Call df_analyze again just in case
11406 delete_unmarked_insns removed anything.
11407
11408 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
11409
11410 * reload.h (struct reg_equivs): Rename to ..
11411 (struct reg_equivs_s): .. this.
11412
11413 2013-08-20 Martin Liska <marxin.liska@gmail.com>
11414
11415 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
11416
11417 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11418
11419 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
11420
11421 2013-08-21 Jeff Law <law@redhat.com>
11422
11423 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
11424 simplify assignments too. If the RHS collapses to a singleton
11425 range, then return the value for the range.
11426
11427 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
11428
11429 * config/i386/sse.md (V16): Rename to...
11430 (VMOVE): this.
11431 (mov<mode>): Update iterator name.
11432 (*mov<mode>_internal): Ditto.
11433 (push<mode>1): Ditto.
11434 (movmisalign<mode>): Ditto.
11435
11436 2013-08-20 Jan Hubicka <jh@suse.cz>
11437
11438 PR bootstrap/58186
11439 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
11440 entry for direct edges.
11441 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
11442
11443 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11444
11445 Revert my last two changes, r201865 and r201864:
11446
11447 Revert r201865:
11448 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11449
11450 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11451 instances can own GC refs.
11452
11453 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11454 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11455 (gcc::context::gt_pch_nx): Likewise.
11456 (gcc::context::gt_pch_nx): Likewise.
11457 * ggc.h (gt_ggc_mx <T>): New.
11458 (gt_pch_nx_with_op <T>): New.
11459 (gt_pch_nx <T>): New.
11460 * passes.c (opt_pass::gt_ggc_mx): New.
11461 (opt_pass::gt_pch_nx): New.
11462 (opt_pass::gt_pch_nx_with_op): New.
11463 (pass_manager::gt_ggc_mx): New.
11464 (pass_manager::gt_pch_nx): New.
11465 (pass_manager::gt_pch_nx_with_op): New.
11466 (pass_manager::operator new): Use
11467 ggc_internal_cleared_alloc_stat rather than xcalloc.
11468 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11469 (pass_manager::gt_ggc_mx): New.
11470 (pass_manager::gt_pch_nx): New.
11471 (pass_manager::gt_pch_nx_with_op): New.
11472 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11473 (opt_pass::operator new): New.
11474 (opt_pass::gt_ggc_mx): New.
11475 (opt_pass::gt_pch_nx): New.
11476 (opt_pass::gt_pch_nx_with_op): New.
11477
11478 Revert r201864:
11479 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11480
11481 * Makefile.in (GTFILES): Add context.h.
11482 * context.c (gcc::context::operator new): New.
11483 (gcc::context::gt_ggc_mx): New.
11484 (gcc::context::gt_pch_nx): New.
11485 (gcc::context::gt_pch_nx): New.
11486 * context.h (gcc::context): Add GTY((user)) marking.
11487 (gcc::context::operator new): New.
11488 (gcc::context::gt_ggc_mx): New.
11489 (gcc::context::gt_pch_nx): New.
11490 (gcc::context::gt_pch_nx): New.
11491 (g): Add GTY marking.
11492 (gt_ggc_mx (gcc::context *)): New.
11493 (gt_pch_nx (gcc::context *)): New.
11494 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11495 void *cookie)): New.
11496 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11497
11498 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11499
11500 * dce.c (fini_dce): Call df_analyze again just in case
11501 delete_unmarked_insns removed anything.
11502
11503 2013-08-20 Teresa Johnson <tejohnson@google.com>
11504
11505 PR rtl-optimizations/57451
11506 * final.c (reemit_insn_block_notes): Prevent lexical blocks
11507 from crossing split section boundaries.
11508
11509 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11510
11511 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
11512 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
11513 with MULTLIB_DEFAULTS.
11514
11515 2013-08-20 Nick Clifton <nickc@redhat.com>
11516
11517 * target.def (narrow_volatile_bitfield): Note that the default
11518 value is false, not !TARGET_STRICT_ALIGN.
11519 * doc/tm.texi: Regenerate.
11520
11521 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
11522
11523 Fix LIB_SPEC for systems without libpthread.
11524
11525 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
11526 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
11527 for Android.
11528 * config/i386/linux-common.h: Likewise.
11529 * config/mips/linux-common.h: Likewise.
11530
11531 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11532
11533 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
11534 checks.
11535
11536 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11537
11538 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11539 instances can own GC refs.
11540
11541 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11542 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11543 (gcc::context::gt_pch_nx): Likewise.
11544 (gcc::context::gt_pch_nx): Likewise.
11545 * ggc.h (gt_ggc_mx <T>): New.
11546 (gt_pch_nx_with_op <T>): New.
11547 (gt_pch_nx <T>): New.
11548 * passes.c (opt_pass::gt_ggc_mx): New.
11549 (opt_pass::gt_pch_nx): New.
11550 (opt_pass::gt_pch_nx_with_op): New.
11551 (pass_manager::gt_ggc_mx): New.
11552 (pass_manager::gt_pch_nx): New.
11553 (pass_manager::gt_pch_nx_with_op): New.
11554 (pass_manager::operator new): Use
11555 ggc_internal_cleared_alloc_stat rather than xcalloc.
11556 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11557 (pass_manager::gt_ggc_mx): New.
11558 (pass_manager::gt_pch_nx): New.
11559 (pass_manager::gt_pch_nx_with_op): New.
11560 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11561 (opt_pass::operator new): New.
11562 (opt_pass::gt_ggc_mx): New.
11563 (opt_pass::gt_pch_nx): New.
11564 (opt_pass::gt_pch_nx_with_op): New.
11565
11566 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11567
11568 * Makefile.in (GTFILES): Add context.h.
11569 * context.c (gcc::context::operator new): New.
11570 (gcc::context::gt_ggc_mx): New.
11571 (gcc::context::gt_pch_nx): New.
11572 (gcc::context::gt_pch_nx): New.
11573 * context.h (gcc::context): Add GTY((user)) marking.
11574 (gcc::context::operator new): New.
11575 (gcc::context::gt_ggc_mx): New.
11576 (gcc::context::gt_pch_nx): New.
11577 (gcc::context::gt_pch_nx): New.
11578 (g): Add GTY marking.
11579 (gt_ggc_mx (gcc::context *)): New.
11580 (gt_pch_nx (gcc::context *)): New.
11581 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11582 void *cookie)): New.
11583 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11584
11585 2013-08-20 Alan Modra <amodra@gmail.com>
11586
11587 PR target/57865
11588 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
11589 (rs6000_emit_epilogue): Likewise.
11590
11591 2013-08-19 Dehao Chen <dehao@google.com>
11592
11593 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
11594
11595 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
11596 Jakub Jelinek <jakub@redhat.com>
11597
11598 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
11599 (BUILT_IN_FABSD64): Likewise.
11600 (BUILT_IN_FABSD128): Likewise.
11601 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
11602 (fold_builtin_1): Likewise.
11603 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
11604 destination and source operands.
11605 (*abstd2_fpr): Likewise.
11606 (*nabstd2_fpr): Likewise.
11607
11608 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
11609
11610 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
11611 JUMP_P and INSN_P.
11612
11613 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
11614
11615 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
11616 incomplete.
11617
11618 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
11619
11620 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
11621 * builtins.c (default_libc_has_function): New.
11622 (gnu_libc_has_function): Ditto.
11623 (no_c99_libc_has_function): Ditto.
11624 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
11625 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
11626 (fold_builtin_sincos): Likewise.
11627 (fold_builtin_cexp): Likewise.
11628 * builtins.def (DEF_C94_BUILTIN): Likewise.
11629 (DEF_C99_BUILTIN): Likewise.
11630 (DEF_C99_C90RES_BUILTIN): Likewise.
11631 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
11632 definitions to using this define.
11633 * config/darwin-protos.h (darwin_libc_has_function): New.
11634 * config/darwin.c (darwin_libc_has_function): Ditto.
11635 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
11636 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
11637 * config/darwin.h: Ditto.
11638 * config/elfos.h: Ditto.
11639 * config/freebsd.h: Ditto.
11640 * config/i386/cygming.h: Ditto.
11641 * config/i386/djgpp.h: Ditto.
11642 * config/i386/i386-interix.h: Ditto.
11643 * config/microblaze/microblaze.h: Ditto.
11644 * config/mmix/mmix.h: Ditto.
11645 * config/gnu-user.h: Ditto.
11646 * config/ia64/hpux.h: Ditto.
11647 * config/pa/pa-hpux.h: Ditto.
11648 * config/pdp11/pdp11.h: Ditto.
11649 * config/picochip/picochip.h: Ditto.
11650 * config/linux.h: Ditto.
11651 * config/netbsd.h: Ditto.
11652 * config/openbsd.h: Ditto.
11653 * config/rs6000/aix43.h: Ditto.
11654 * config/rs6000/aix51.h: Ditto.
11655 * config/rs6000/aix52.h: Ditto.
11656 * config/rs6000/aix53.h: Ditto.
11657 * config/rs6000/aix61.h: Ditto.
11658 * config/rs6000/darwin.h: Ditto.
11659 * config/rs6000/linux.h: Ditto.
11660 * config/rs6000/linux64.h: Ditto.
11661 * config/s390/tpf.h: Ditto.
11662 * config/sol2-10.h: Ditto.
11663 * config/sol2.h: Ditto.
11664 * config/vms/vms.h: Ditto.
11665 * config/vxworks.h: Ditto.
11666 * config/linux-android.c (linux_android_libc_has_function):
11667 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
11668 * config/linux-protos.h (linux_android_libc_has_function):
11669 New declaration.
11670 * config/i386/i386.c (ix86_libc_has_function): New.
11671 * config/i386/i386-protos.h
11672 (ix86_libc_has_function): New declaration.
11673 * config/i386/i386.md
11674 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
11675 ("isinf<mode>2): Likewise.
11676 * convert.c (convert_to_integer): Using new target hook
11677 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
11678 TARGET_C99_FUNCTIONS.
11679 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
11680 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
11681 * coretypes.h (function_class): New enum for different
11682 classes of functions.
11683 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
11684 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
11685 (TARGET_HAS_SINCOS): Likewise.
11686 (TARGET_LIBC_HAS_FUNCTION): New.
11687 * doc/tm.texi: Regenerated.
11688 * targhooks.h (default_libc_has_function): New declaration.
11689 (no_c99_libc_has_function): Ditto.
11690 (gnu_libc_has_function): Ditto.
11691 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
11692 and TARGET_HAS_SINCOS.
11693
11694 2013-08-18 Jan Hubicka <jh@suse.cz>
11695
11696 * Makeifle-in (ipa-devirt.o): New.
11697 (GTFILES): Add ipa-utils.h and ipa-devirt.c
11698 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
11699 (analyze_functions): Look into possible targets of polymorphic call.
11700 * dumpfile.c (dump_files): Add type-inheritance dump.
11701 * dumpfile.h (TDI_inheritance): New.
11702 * ipa-devirt.c: New file.
11703 * ipa-utils.h (odr_type_d): Forward declare.
11704 (odr_type): New type.
11705 (build_type_inheritance_graph): Declare.
11706 (possible_polymorphic_call_targets): Declare and introduce inline
11707 variant when only edge is pased.
11708 (dump_possible_polymorphic_call_targets): Likewise.
11709 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
11710 * tree.c (type_in_anonymous_namespace_p): Break out from ...
11711 (types_same_for_odr): ... here.
11712 * tree.h (type_in_anonymous_namespace_p): Declare.
11713
11714 2013-08-18 Jakub Jelinek <jakub@redhat.com>
11715
11716 PR tree-optimization/58006
11717 * tree-parloops.c (take_address_of): Don't ICE if get_name
11718 returns NULL.
11719 (eliminate_local_variables_stmt): Remove clobber stmts.
11720
11721 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
11722
11723 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
11724 error message is issued for an alias to undefined symbol.
11725
11726 2013-08-18 Jan Hubicka <jh@suse.cz>
11727
11728 * cgraph.c (cgraph_create_indirect_edge): Discover
11729 polymorphic calls and record basic info into indirect_info.
11730 * gimple-fold.c (gimple_fold_call): When doing BINFO based
11731 devirtualization, ignore objc function calls.
11732 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
11733 call with no parm index info.
11734 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
11735 * tree.c (virtual_method_call_p): New function.
11736 * tree.h (virtual_method_call_p): Declare.
11737
11738 2013-08-16 Jan Hubicka <jh@suse.cz>
11739
11740 PR middle-end/58179
11741 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
11742
11743 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
11744
11745 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
11746 attribute.
11747
11748 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11749
11750 * gengtype.c (type_for_name): Add special-case support for
11751 locating types within the "gcc::" namespace.
11752 (open_base_files): Emit a "using namespace gcc" directive.
11753
11754 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11755
11756 PR target/58160
11757 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
11758 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
11759
11760 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
11761 array instead of each individual operand as a separate argument.
11762 (emit_fusion_gpr_load): Likewise.
11763 (expand_fusion_gpr_load): Add new function declaration.
11764
11765 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
11766 signature to have the operands passed as an array, instead of as
11767 separate arguments. Allow ZERO_EXTEND to be in the memory
11768 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
11769 depend on the register live/dead flags when peepholes are run.
11770 (expand_fusion_gpr_load): New function to be called from the
11771 peephole2 pass, to change the register that addis sets to be the
11772 target register.
11773 (emit_fusion_gpr_load): Change the calling signature to have the
11774 operands passed as an array, instead of as separate arguments.
11775 Allow ZERO_EXTEND to be in the memory address, and also
11776 SIGN_EXTEND if -mpower8-fusion-sign.
11777
11778 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
11779 unspec enumeration.
11780 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
11781 adjust the register addis loads up in the peephole2 pass. Do not
11782 depend on the register live/dead state when the peephole pass is done.
11783
11784 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11785
11786 * gengtype.c (create_user_defined_type): Ensure that the kind
11787 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
11788 declaration is seen before the GTY((user)) marking.
11789
11790 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
11791
11792 PR target/58105
11793 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
11794
11795 2013-08-16 Jan Hubicka <jh@suse.cz>
11796
11797 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
11798 arugment expected_type.
11799 (gimple_fold_call): Use it.
11800 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
11801 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
11802 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
11803 (try_make_edge_direct_virtual_call): Likewise.
11804 * tree.c (obj_type_ref_class): New.
11805 * tree.h (obj_type_ref_class): Use it.
11806
11807 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
11808
11809 * sched-vis.c (rtl_slim_pp_initialized): Remove.
11810 (rtl_slim_pp): Likewise.
11811 (init_rtl_slim_pretty_print): Likewise.
11812 (dump_value_slim): Don't call it. Use local pretty printer.
11813 (dump_insn_slim): Likewise.
11814 (dump_rtl_slim): Likewise.
11815 (str_pattern_slim): Likewise.
11816 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
11817 Simplify.
11818
11819 2013-08-16 Jakub Jelinek <jakub@redhat.com>
11820
11821 PR tree-optimization/58164
11822 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
11823 walk gimple_goto_dest of GIMPLE_GOTO.
11824
11825 PR tree-optimization/58165
11826 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
11827 bi_call must be the last stmt in a bb, don't split_block, instead
11828 use fallthru edge from it and give up if there is none.
11829 Release conds vector when returning early.
11830
11831 2013-08-14 Xinliang David Li <davidxl@google.com>
11832
11833 * config/i386/i386.c (ix86_option_override_internal):
11834 Remove unused variable and field.
11835
11836 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11837
11838 PR target/57949
11839 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
11840 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
11841 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
11842 and Linux, correct BLKmode alignment when 128-bit alignment is
11843 required and compatibility flag is not set.
11844 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
11845 for zero-size arguments when compatibility flag is not set.
11846
11847 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11848
11849 PR tree-optimization/58145
11850 * tree-sra.c (build_ref_for_offset): If prev_base has
11851 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
11852
11853 2013-08-14 Xinliang David Li <davidxl@google.com>
11854
11855 * config/i386/i386.c (ix86_option_override_internal):
11856 Fix uninitialized variable error.
11857
11858 2013-08-14 Xinliang David Li <davidxl@google.com>
11859
11860 * config/i386/i386.opt: Define two new options.
11861 * config/i386/x86-tune.def: Add arch selector field in macros.
11862 * config/i386/i386.h: Adjust macro definition.
11863 * config/i386/i386.c (ix86_option_override_internal):
11864 Refactor the code.
11865 (parse_mtune_ctrl_str): New function.
11866 (set_ix86_tune_features): New function.
11867 (ix86_function_specific_restore): Call the new helper function.
11868
11869 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
11870
11871 PR rtl-optimization/57662
11872 * sel-sched.c (code_motion_process_successors): When the current insn
11873 is removed after the recursive traversal, break from the loop.
11874 Add comments and debug printouts.
11875
11876 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11877 Alexandre Oliva <aoliva@redhat.com>
11878
11879 PR target/58067
11880 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
11881 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
11882 there also UNSPEC_PLTOFF.
11883
11884 2013-08-14 Marek Polacek <polacek@redhat.com>
11885
11886 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
11887 NUM_CONDITIONS bit positions.
11888
11889 2013-08-13 Cary Coutant <ccoutant@google.com>
11890
11891 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
11892 (attr_checksum): Hash vector contents instead of pointer.
11893 (attr_checksum_ordered): Likewise.
11894
11895 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
11896
11897 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
11898 when Pmode != word_mode. Add length_address attribute.
11899 (sse3_monitor_<mode>): Merge from sse3_monitor and
11900 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
11901 Pmode != word_mode. Update insn length attribute.
11902 * config/i386/i386.c (ix86_option_override_internal): Update
11903 ix86_gen_monitor selection for merged sse3_monitor insn.
11904
11905 2013-08-13 Julian Brown <julian@codesourcery.com>
11906
11907 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
11908 perform invalid legitimization on greater-than-word-size modes for
11909 TARGET_E500_DOUBLE.
11910
11911 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
11912
11913 * ira.c (setup_class_translate_array): Use aclass instead of cl
11914 for classes not fully covered by allocno classes.
11915
11916 2013-08-13 Jakub Jelinek <jakub@redhat.com>
11917
11918 PR tree-optimization/57661
11919 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
11920 * tree-inline.c (tree_function_versioning): Initialize it.
11921 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
11922 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
11923 that is not being copied.
11924
11925 PR sanitizer/56417
11926 * asan.c (instrument_strlen_call): Fix typo in comment.
11927 Use char * type even for the lhs of POINTER_PLUS_EXPR.
11928
11929 2013-08-13 Steve Ellcey <sellcey@mips.com>
11930
11931 * config/mips/mips.md (prefetch): Use lw instead of ld on
11932 loongson in 32bit mode.
11933
11934 2013-08-13 Nick Clifton <nickc@redhat.com>
11935
11936 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
11937
11938 2013-08-13 Jan Hubicka <jh@suse.cz>
11939
11940 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
11941 introduced edge; fix typo in sanity check.
11942 (cgraph_resolve_speculation): Export; improve diagnostic.
11943 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
11944 speculation at type mismatch.
11945 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
11946 (cgraph_resolve_speculation): Declare.
11947 (symtab_can_be_discarded): New function.
11948 * value-prof.c (gimple_ic_transform): Remove actual transform code.
11949 * ipa-inline-transform.c (speculation_removed): New global var.
11950 (clone_inlined_nodes): See if speculation can be removed.
11951 (inline_call): If speculations was removed, we growths may not match.
11952 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
11953 (speculation_useful_p): New function.
11954 (resolve_noninline_speculation): New function.
11955 (inline_small_functions): Resolve useless speculations.
11956 * ipa-inline.h (speculation_useful_p): Declare
11957 * ipa.c (can_replace_by_local_alias): Simplify.
11958 (ipa_profile): Produce speculative calls in non-lto, too;
11959 add simple cost model; produce local aliases.
11960
11961 2013-08-13 David Malcolm <dmalcolm@redhat.com>
11962
11963 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
11964 PASS_MANAGER_H.
11965
11966 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
11967
11968 * config/i386/i386.c (ix86_function_versions): Use error + inform.
11969
11970 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
11971
11972 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
11973 iterator instead of X87MODEF.
11974
11975 2013-08-12 Perez Read <netfirewall@gmail.com>
11976
11977 PR target/58132
11978 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
11979 operand 0 for intel asm alternative.
11980 (*movabs<mode>_2): Ditto for operand 1.
11981
11982 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
11983
11984 * config/aarch64/arm_none.h
11985 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
11986
11987 2013-08-12 Nick Clifton <nickc@redhat.com>
11988
11989 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
11990
11991 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
11992
11993 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
11994 expand for QI/HImode operand to produce more effictive code for
11995 unsigned char(short) --> float(double) conversion.
11996
11997 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
11998
11999 * doc/invoke.texi: Mention that -ftls-model does not force the final
12000 model.
12001
12002 2013-08-12 Marek Polacek <polacek@redhat.com>
12003 Marc Glisse <marc.glisse@inria.fr>
12004
12005 PR tree-optimization/57980
12006 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
12007 when creating -1 constant.
12008
12009 2013-08-10 Jan Hubicka <jh@suse.cz>
12010
12011 Workaround binutils PR14342.
12012 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
12013 (gimple_init_edge_profiler): Likewise.
12014 (gimple_gen_ic_func_profiler): Likewise.
12015
12016 2013-08-09 Jan Hubicka <jh@suse.cz>
12017
12018 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
12019
12020 2013-08-09 Xinliang David Li <davidxl@google.com>
12021
12022 * config/i386/stringop.def: New file.
12023 * config/i386/stringop.opt: New file.
12024 * config/i386/i386-opts.h: Include stringopt.def.
12025 * config/i386/i386.opt: Include stringopt.opt.
12026 * config/i386/i386.c (ix86_option_override_internal):
12027 Override default size based stringop inline strategies with options.
12028 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
12029 New function.
12030
12031 2013-08-09 Jan Hubicka <jh@suse.cz>
12032
12033 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
12034
12035 2013-08-09 Jan Hubicka <jh@suse.cz>
12036
12037 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
12038 CGRAPH_FREQ_MAX.
12039 (dump_cgraph_node): Dump profile-id.
12040 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
12041 and common_target_probability.
12042 * lto-cgraph.c (lto_output_edge): Stream common targets.
12043 (lto_output_node): Stream profile ids.
12044 (input_node): Stream profile ids.
12045 (input_edge): Stream common targets.
12046 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
12047 * ipa.c: Include value-prof.h
12048 (ipa_profile_generate_summary): Turn indirect call statement histograms
12049 into common targets.
12050 (ipa_profile): Turn common targets into speculative edges.
12051
12052 2013-08-09 Jan Hubicka <jh@suse.cz>
12053
12054 * cgraph.h (cgraph_node): Add profile_id.
12055 * value-prof.c (cgraph_node_map): Turn into pointer_map.
12056 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
12057 (del_node_map): Update.
12058 (find_func_by_funcdef_no): Replace by ...
12059 (find_func_by_profile_id): ... this one.
12060 (gimple_ic_transform): Do not remove useful histograms when
12061 speculation is not done; dump info when indirect call removal
12062 can happen at LTO.
12063 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
12064 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
12065 (__gcov_indirect_call_profiler_v2): .. this one.
12066 * profile.h (init_node_map): Update.
12067 * coverage.c (coverage_compute_profile_id): New function.
12068 * coverage.h (coverage_compute_profile_id): Declare.
12069 * tree-profile.c (init_ic_make_global_vars): Make
12070 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
12071 (gimple_init_edge_profiler): Update prototype of
12072 __gcov_indirect_call_profiler.
12073 (gimple_gen_ic_func_profiler): Simplify.
12074 (tree_profiling): Use init_node_map
12075
12076 2013-08-09 Jan Hubicka <jh@suse.cz>
12077
12078 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
12079 non-speculative refs.
12080 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
12081 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
12082 (cgraph_set_call_stmt): Likewise.
12083 (cgraph_create_edge_1): Fix release checking compilatoin;
12084 clear lto_stmt_uid.
12085 (cgraph_free_edge): Free indirect info.
12086 (cgraph_turn_edge_to_speculative): New function.
12087 (cgraph_speculative_call_info): New function.
12088 (cgraph_make_edge_direct): Return direct edge; handle speculation.
12089 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
12090 (dump_cgraph_node): Dump speculation.
12091 (verify_edge_count_and_frequency): Accept speculative edges.
12092 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
12093 (verify_cgraph_node): Handle speculation.
12094 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
12095 (cgraph_set_call_stmt): Update prototype.
12096 (cgraph_make_edge_direct): Update prototype.
12097 (cgraph_speculative_call_info): Declare.
12098 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
12099 to change; update call of ipa_find_references.
12100 * ipa-ref.c (ipa_record_reference): Fix return value; clear
12101 lto_stmt_uid and speculative flags.
12102 (ipa_dump_references): Dump speculation.
12103 (ipa_clone_references): Clone speculative flag.
12104 (ipa_clone_referring): Likewise.
12105 (ipa_clone_ref): New function.
12106 (ipa_find_reference): Look into lto_stmt_uids
12107 (ipa_clear_stmts_in_references): Do not clear speculative calls.
12108 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
12109 (ipa_find_reference): Update declaration.
12110 (ipa_clone_ref): Declare.
12111 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
12112 stream speculative flag.
12113 (lto_output_ref): Stream statements uids and speculation.
12114 (input_ref): Likewise.
12115 (input_edge): Stream speuclation.
12116 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
12117 (cgraph_set_call_stmt_including_clones): Handle speculation.
12118 * ipa-inline.c (heap_edge_removal_hook): New function.
12119 (inline_small_functions): Register it.
12120 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
12121 also initialize refs.
12122 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
12123 edge to change.
12124 (try_make_edge_direct_simple_call): Likewise.
12125 (try_make_edge_direct_simple_call): Likewise.
12126 (update_indirect_edges_after_inlining): Likewise.
12127 (remove_described_reference): Look proper lto_stmt_uid.
12128 (propagate_controlled_uses): Likewise.
12129 (propagate_controlled_uses): Liekwise.
12130 * tree-inline.c (copy_bb): Copy speculative edges.
12131 (redirect_all_calls): New function.
12132 (copy_cfg_body): Do redirection after loop info is updated.
12133 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
12134
12135 2013-08-09 Jan Hubicka <jh@suse.cz>
12136
12137 * lto-streamer-out.c (output_function): Renumber PHIs.
12138 * lto-streamer-in.c (input_function): Likewise.
12139
12140 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
12141
12142 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
12143 (get_lane_unsigned): Likewise.
12144 (dup_lane_scalar): Likewise.
12145 (get_lane): enable for VALL.
12146 * config/aarch64/aarch64-simd.md
12147 (aarch64_dup_lane_scalar<mode>): Remove.
12148 (aarch64_get_lane_signed<mode>): Likewise.
12149 (aarch64_get_lane_unsigned<mode>): Likewise.
12150 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
12151 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
12152 (aarch64_get_lane<mode>): Enable for all vector modes.
12153 (aarch64_get_lanedi): Remove misleading constraints.
12154 * config/aarch64/arm_neon.h
12155 (__aarch64_vget_lane_any): Define.
12156 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
12157 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
12158 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
12159 * config/aarch64/iterators.md (VDQQH): New.
12160 (VDQQHS): Likewise.
12161 (vwcore): Likewise.
12162
12163 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
12164
12165 * configure.ac: Add GAS check for LEON instructions on SPARC.
12166 * configure: Regenerate.
12167 * config.in: Likewise.
12168 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
12169 sparc*-*-* block.
12170 * config/sparc/sparc.opt (LEON, LEON3): New masks.
12171 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
12172 for LEON or LEON3.
12173 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
12174 (AS_LEON_FLAG): New macro.
12175 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
12176 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
12177 Deal with LEON and LEON3 for the memory model.
12178 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
12179 (atomic_compare_and_swap<mode>_1): Likewise.
12180 (*atomic_compare_and_swap<mode>_1): Likewise.
12181
12182 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12183
12184 * config/arm/neon.md (vcond): Fix floating-point vector
12185 comparisons against 0.
12186
12187 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
12188
12189 * lra-constraints.c (emit_spill_move): Remove assert.
12190 (process_alt_operands): Add more debugging
12191 output. Increase reject for spilling into memory. Decrease
12192 reject for reloading scratch.
12193 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
12194
12195 2013-08-08 Steve Ellcey <sellcey@mips.com>
12196
12197 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
12198 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
12199 micromips incompatible. Add nan2008.
12200 (MULTILIB_DIRNAMES): Add nan2008.
12201 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12202 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
12203 and micromips incompatible. Add nan2008.
12204 (MULTILIB_DIRNAMES): Add nan2008.
12205 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12206
12207 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
12208
12209 PR rtl-optimization/58079
12210 * combine.c (combine_simplify_rtx): Avoid using SUBST if
12211 simplify_comparison has widened a comparison with an integer.
12212
12213 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12214
12215 * config/arm/neon.md (movmisalign<mode>): Disable when we
12216 don't allow unaligned accesses.
12217 (*movmisalign<mode>_neon_store): Likewise.
12218 (*movmisalign<mode>_neon_load): Likewise.
12219 (*movmisalign<mode>_neon_store): Likewise.
12220 (*movmisalign<mode>_neon_load): Likewise.
12221
12222 2013-08-08 Jan Hubicka <jh@suse.cz>
12223
12224 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
12225 (make_pass_rebuild_cgraph_edges): Also clear references.
12226 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
12227 * ipa-inline-transform.c (inline_transform): Remove all references
12228 after inlining.
12229 * cgraphunit.c (expand_function): Remove all references after
12230 expansion.
12231 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
12232 (ipa_find_reference): Rewrite to iterator.
12233 (remove_stmt_references): Likewise.
12234 (ipa_clear_stmts_in_references): New function.
12235 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
12236 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
12237 clear references.
12238 * ipa-split.c (split_function): Remove references in split function.
12239
12240 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12241
12242 PR target/57431
12243 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
12244 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
12245
12246 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12247
12248 PR target/56979
12249 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
12250 suggested mode for the assignment isn't compatible with the
12251 registers required.
12252
12253 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12254
12255 PR target/58065
12256 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
12257
12258 2013-08-07 Xinliang David Li <davidxl@google.com>
12259
12260 * config/i386/i386.opt: New option -mtune-ctrl=.
12261 * config/i386/x86-tune.def: New file.
12262 * config/i386/i386.h: include x86-tune.def.
12263 * config/i386/i386.c (ix86_option_override_internal):
12264 Parsing -mtune-ctrl= option and set tune features.
12265
12266 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
12267
12268 PR other/12081
12269 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
12270 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
12271 to gen_2arg_fn_t.
12272
12273 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12274
12275 * rtl.h (update_alignments): Declare.
12276 * final.c (grow_label_align): New function extracted from...
12277 (shorten_branches): ...here. Call it.
12278 (update_alignments): New function.
12279 * reorg.c (sibling_labels): New variable.
12280 (get_label_before): Add SIBLING parameter. If it is non-zero, push
12281 the new label along with it onto the sibling_labels vector.
12282 (fill_simple_delay_slots): Adjust call to get_label_before.
12283 (fill_slots_from_thread): Likewise.
12284 (relax_delay_slots): Likewise.
12285 (make_return_insns): Likewise.
12286 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
12287
12288 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12289
12290 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
12291 document its semantics.
12292 (diagnostic_report_diagnostic): Adjust accordingly.
12293
12294 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12295
12296 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
12297 (sparc_option_override): ...and port to new C++ pass API.
12298 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
12299
12300 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
12301
12302 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
12303
12304 2013-08-06 Caroline Tice <cmtice@google.com>
12305
12306 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
12307 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
12308 * tree-pass.h: Add pass_vtable_verify.
12309 * varasm.c (assemble_variable): Add code to properly set the comdat
12310 section and name for the .vtable_map_vars section.
12311 (assemble_vtyv_preinit_initializer): New function.
12312 (default_sectin_type_flags): Make sure .vtable_map_vars section has
12313 LINK_ONCE flag.
12314 * output.h: Add function decl for assemble_vtv_preinit_initializer.
12315 * vtable-verify.c: New file.
12316 * vtable-verify.h: New file.
12317 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
12318 initialiation levels.
12319 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
12320 * passes.def: Insert pass_vtable_verify.
12321 * aclocal.m4: Reorder includes.
12322 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
12323 -fvtv-counts options.
12324 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
12325 as appropriate, if -fvtable-verify=... is used.
12326 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
12327 -fvtable-verify=... is used.
12328 * Makefile.in (OBJS): Add vtable-verify.o to list.
12329 (vtable-verify.o): Add new build rule.
12330 (GTFILES): Add vtable-verify.c to list.
12331 * common.opt (fvtable-verify=): New flag.
12332 (vtv_priority): Values for fvtable-verify= flag.
12333 (fvtv-counts): New flag.
12334 (fvtv-debug): New flag.
12335 * tree.h (save_vtable_map_decl): New extern function decl.
12336
12337 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12338
12339 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
12340 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
12341 (pass_data_rl78_devirt): ...new pass_data instance and...
12342 (make_pass_rl78_devirt): ...new function.
12343 (rl78_asm_file_start): Port pass registration to new C++ API.
12344
12345 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12346
12347 * coretypes.h (rtl_opt_pass): Add.
12348 (gcc::context): Add.
12349 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
12350 (epiphany_init): Port to new C++ pass API.
12351 (epiphany_optimize_mode_switching): Likewise.
12352 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
12353 (pass_manager::get_pass_mode_switching): New.
12354 (pass_manager::get_pass_peephole2): New.
12355 * mode-switching.c (pass_mode_switching): Add clone method.
12356 * recog.c (pass_peephole2): Add clone method.
12357 (pass_split_all_insns): Add clone method.
12358
12359 2013-08-06 David Malcolm <dmalcolm@redhat.com>
12360
12361 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
12362 (mips_option_override): ...here, porting to new C++ API for passes.
12363
12364 2013-08-06 Jan Hubicka <jh@suse.cz>
12365
12366 * cgraph.c (cgraph_get_body): New function based on lto.c
12367 implementation.
12368 * cgraph.h (cgraph_get_body): Declare.
12369 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
12370 LTO paths.
12371 * cgraphunit.c (expand_function): Get body prior expanding.
12372 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
12373 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
12374 really need.
12375 * passes.c (do_per_function_toporder): Get body.
12376 * tree-inline.c (expand_call_inline): Get body prior inlining it.
12377 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
12378
12379 2013-08-06 Martin Jambor <mjambor@suse.cz>
12380
12381 PR fortran/57987
12382 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
12383 is not re-finalized. Rename second parameter to no_collect.
12384
12385 2013-08-06 Martin Jambor <mjambor@suse.cz>
12386
12387 PR middle-end/58041
12388 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
12389 MEM_REF has proper alignment information.
12390
12391 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
12392
12393 PR other/12081
12394 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
12395 class insn_gen_fn.
12396 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
12397 rtx (*) (rtx, ...) with insn_gen_fn.
12398 * genoutput.c (output_insn_data): Cast gen_? function pointers to
12399 insn_gen_fn::stored_funcptr. Add initializer braces.
12400
12401 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12402
12403 Rewrite how instances of passes are cloned to remove assumptions
12404 about their sizes (thus allowing pass subclasses to have
12405 additional data fields, albeit non-GC-managed ones at this point).
12406
12407 * passes.c (make_pass_instance): Now that passes have clone
12408 methods, rewrite this function to eliminate XNEW and memcpy
12409 calls that used hardcoded sizes. Since this function no longer
12410 creates pass instances, rename it to...
12411 (add_pass_instance): ...this. Document the old way that passes were
12412 numbered and flagged, and rework this function to continue using it.
12413 (next_pass_1): Add an initial_pass argument for use by
12414 add_pass_instance.
12415 (position_pass): When adding multiple instances of a pass, use
12416 the pass's clone method, rather than relying on the XNEW/memcpy
12417 within the former make_pass_instance (now add_pass_instance).
12418 (pass_manager::pass_manager): When invoking next_pass_1, also supply
12419 the initial instance of the current pass within the pass manager.
12420
12421 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12422
12423 This is the automated part of the conversion of passes from C
12424 structs to C++ classes.
12425
12426 Patch autogenerated by refactor_passes.py from
12427 https://github.com/davidmalcolm/gcc-refactoring-scripts
12428 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
12429
12430 * asan.c (pass_asan): Convert from a global struct to a subclass of
12431 gimple_opt_pass along with...
12432 (pass_data_asan): ...new pass_data instance and...
12433 (make_pass_asan): ...new function.
12434 (pass_asan_O0): Convert from a global struct to a subclass of
12435 gimple_opt_pass along with...
12436 (pass_data_asan_O0): ...new pass_data instance and...
12437 (make_pass_asan_O0): ...new function.
12438 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
12439 subclass of rtl_opt_pass along with...
12440 (pass_data_inc_dec): ...new pass_data instance and...
12441 (make_pass_inc_dec): ...new function.
12442 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
12443 a subclass of rtl_opt_pass along with...
12444 (pass_data_reorder_blocks): ...new pass_data instance and...
12445 (make_pass_reorder_blocks): ...new function.
12446 (pass_duplicate_computed_gotos): Convert from a global struct to a
12447 subclass of rtl_opt_pass along with...
12448 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
12449 (make_pass_duplicate_computed_gotos): ...new function.
12450 (pass_partition_blocks): Convert from a global struct to a subclass of
12451 rtl_opt_pass along with...
12452 (pass_data_partition_blocks): ...new pass_data instance and...
12453 (make_pass_partition_blocks): ...new function.
12454 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
12455 struct to a subclass of rtl_opt_pass along with...
12456 (pass_data_branch_target_load_optimize1): ...new pass_data instance
12457 and...
12458 (make_pass_branch_target_load_optimize1): ...new function.
12459 (pass_branch_target_load_optimize2): Convert from a global struct to a
12460 subclass of rtl_opt_pass along with...
12461 (pass_data_branch_target_load_optimize2): ...new pass_data instance
12462 and...
12463 (make_pass_branch_target_load_optimize2): ...new function.
12464 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
12465 of rtl_opt_pass along with...
12466 (pass_data_jump): ...new pass_data instance and...
12467 (make_pass_jump): ...new function.
12468 (pass_jump2): Convert from a global struct to a subclass of
12469 rtl_opt_pass along with...
12470 (pass_data_jump2): ...new pass_data instance and...
12471 (make_pass_jump2): ...new function.
12472 * cfgexpand.c (pass_expand): Convert from a global struct to a
12473 subclass of rtl_opt_pass along with...
12474 (pass_data_expand): ...new pass_data instance and...
12475 (make_pass_expand): ...new function.
12476 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
12477 of rtl_opt_pass along with...
12478 (pass_data_free_cfg): ...new pass_data instance and...
12479 (make_pass_free_cfg): ...new function.
12480 (pass_into_cfg_layout_mode): Convert from a global struct to a
12481 subclass of rtl_opt_pass along with...
12482 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
12483 (make_pass_into_cfg_layout_mode): ...new function.
12484 (pass_outof_cfg_layout_mode): Convert from a global struct to a
12485 subclass of rtl_opt_pass along with...
12486 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
12487 (make_pass_outof_cfg_layout_mode): ...new function.
12488 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
12489 struct to a subclass of gimple_opt_pass along with...
12490 (pass_data_build_cgraph_edges): ...new pass_data instance and...
12491 (make_pass_build_cgraph_edges): ...new function.
12492 (pass_rebuild_cgraph_edges): Convert from a global struct to a
12493 subclass of gimple_opt_pass along with...
12494 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
12495 (make_pass_rebuild_cgraph_edges): ...new function.
12496 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
12497 subclass of gimple_opt_pass along with...
12498 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
12499 and...
12500 (make_pass_remove_cgraph_callee_edges): ...new function.
12501 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
12502 struct to a subclass of rtl_opt_pass along with...
12503 (pass_data_stack_adjustments): ...new pass_data instance and...
12504 (make_pass_stack_adjustments): ...new function.
12505 * combine.c (pass_combine): Convert from a global struct to a subclass
12506 of rtl_opt_pass along with...
12507 (pass_data_combine): ...new pass_data instance and...
12508 (make_pass_combine): ...new function.
12509 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
12510 global struct to a subclass of rtl_opt_pass along with...
12511 (pass_data_compare_elim_after_reload): ...new pass_data instance
12512 and...
12513 (make_pass_compare_elim_after_reload): ...new function.
12514 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
12515 of rtl_opt_pass along with...
12516 (pass_data_rtl_cprop): ...new pass_data instance and...
12517 (make_pass_rtl_cprop): ...new function.
12518 * cse.c (pass_cse): Convert from a global struct to a subclass of
12519 rtl_opt_pass along with...
12520 (pass_data_cse): ...new pass_data instance and...
12521 (make_pass_cse): ...new function.
12522 (pass_cse2): Convert from a global struct to a subclass of
12523 rtl_opt_pass along with...
12524 (pass_data_cse2): ...new pass_data instance and...
12525 (make_pass_cse2): ...new function.
12526 (pass_cse_after_global_opts): Convert from a global struct to a
12527 subclass of rtl_opt_pass along with...
12528 (pass_data_cse_after_global_opts): ...new pass_data instance and...
12529 (make_pass_cse_after_global_opts): ...new function.
12530 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
12531 of rtl_opt_pass along with...
12532 (pass_data_ud_rtl_dce): ...new pass_data instance and...
12533 (make_pass_ud_rtl_dce): ...new function.
12534 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
12535 rtl_opt_pass along with...
12536 (pass_data_fast_rtl_dce): ...new pass_data instance and...
12537 (make_pass_fast_rtl_dce): ...new function.
12538 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
12539 a subclass of rtl_opt_pass along with...
12540 (pass_data_df_initialize_opt): ...new pass_data instance and...
12541 (make_pass_df_initialize_opt): ...new function.
12542 (pass_df_initialize_no_opt): Convert from a global struct to a
12543 subclass of rtl_opt_pass along with...
12544 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
12545 (make_pass_df_initialize_no_opt): ...new function.
12546 (pass_df_finish): Convert from a global struct to a subclass of
12547 rtl_opt_pass along with...
12548 (pass_data_df_finish): ...new pass_data instance and...
12549 (make_pass_df_finish): ...new function.
12550 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
12551 rtl_opt_pass along with...
12552 (pass_data_rtl_dse1): ...new pass_data instance and...
12553 (make_pass_rtl_dse1): ...new function.
12554 (pass_rtl_dse2): Convert from a global struct to a subclass of
12555 rtl_opt_pass along with...
12556 (pass_data_rtl_dse2): ...new pass_data instance and...
12557 (make_pass_rtl_dse2): ...new function.
12558 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
12559 subclass of rtl_opt_pass along with...
12560 (pass_data_dwarf2_frame): ...new pass_data instance and...
12561 (make_pass_dwarf2_frame): ...new function.
12562 * except.c (pass_set_nothrow_function_flags): Convert from a global
12563 struct to a subclass of rtl_opt_pass along with...
12564 (pass_data_set_nothrow_function_flags): ...new pass_data instance
12565 and...
12566 (make_pass_set_nothrow_function_flags): ...new function.
12567 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
12568 subclass of rtl_opt_pass along with...
12569 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
12570 and...
12571 (make_pass_convert_to_eh_region_ranges): ...new function.
12572 * final.c (pass_compute_alignments): Convert from a global struct to a
12573 subclass of rtl_opt_pass along with...
12574 (pass_data_compute_alignments): ...new pass_data instance and...
12575 (make_pass_compute_alignments): ...new function.
12576 (pass_final): Convert from a global struct to a subclass of
12577 rtl_opt_pass along with...
12578 (pass_data_final): ...new pass_data instance and...
12579 (make_pass_final): ...new function.
12580 (pass_shorten_branches): Convert from a global struct to a subclass of
12581 rtl_opt_pass along with...
12582 (pass_data_shorten_branches): ...new pass_data instance and...
12583 (make_pass_shorten_branches): ...new function.
12584 (pass_clean_state): Convert from a global struct to a subclass of
12585 rtl_opt_pass along with...
12586 (pass_data_clean_state): ...new pass_data instance and...
12587 (make_pass_clean_state): ...new function.
12588 * function.c (pass_instantiate_virtual_regs): Convert from a global
12589 struct to a subclass of rtl_opt_pass along with...
12590 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
12591 (make_pass_instantiate_virtual_regs): ...new function.
12592 (pass_leaf_regs): Convert from a global struct to a subclass of
12593 rtl_opt_pass along with...
12594 (pass_data_leaf_regs): ...new pass_data instance and...
12595 (make_pass_leaf_regs): ...new function.
12596 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
12597 subclass of rtl_opt_pass along with...
12598 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
12599 and...
12600 (make_pass_thread_prologue_and_epilogue): ...new function.
12601 (pass_match_asm_constraints): Convert from a global struct to a
12602 subclass of rtl_opt_pass along with...
12603 (pass_data_match_asm_constraints): ...new pass_data instance and...
12604 (make_pass_match_asm_constraints): ...new function.
12605 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
12606 subclass of rtl_opt_pass along with...
12607 (pass_data_rtl_fwprop): ...new pass_data instance and...
12608 (make_pass_rtl_fwprop): ...new function.
12609 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
12610 rtl_opt_pass along with...
12611 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
12612 (make_pass_rtl_fwprop_addr): ...new function.
12613 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
12614 rtl_opt_pass along with...
12615 (pass_data_rtl_pre): ...new pass_data instance and...
12616 (make_pass_rtl_pre): ...new function.
12617 (pass_rtl_hoist): Convert from a global struct to a subclass of
12618 rtl_opt_pass along with...
12619 (pass_data_rtl_hoist): ...new pass_data instance and...
12620 (make_pass_rtl_hoist): ...new function.
12621 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
12622 subclass of gimple_opt_pass along with...
12623 (pass_data_lower_cf): ...new pass_data instance and...
12624 (make_pass_lower_cf): ...new function.
12625 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
12626 from a global struct to a subclass of gimple_opt_pass along with...
12627 (pass_data_strength_reduction): ...new pass_data instance and...
12628 (make_pass_strength_reduction): ...new function.
12629 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
12630 of rtl_opt_pass along with...
12631 (pass_data_rtl_ifcvt): ...new pass_data instance and...
12632 (make_pass_rtl_ifcvt): ...new function.
12633 (pass_if_after_combine): Convert from a global struct to a subclass of
12634 rtl_opt_pass along with...
12635 (pass_data_if_after_combine): ...new pass_data instance and...
12636 (make_pass_if_after_combine): ...new function.
12637 (pass_if_after_reload): Convert from a global struct to a subclass of
12638 rtl_opt_pass along with...
12639 (pass_data_if_after_reload): ...new pass_data instance and...
12640 (make_pass_if_after_reload): ...new function.
12641 * init-regs.c (pass_initialize_regs): Convert from a global struct to
12642 a subclass of rtl_opt_pass along with...
12643 (pass_data_initialize_regs): ...new pass_data instance and...
12644 (make_pass_initialize_regs): ...new function.
12645 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
12646 of ipa_opt_pass_d along with...
12647 (pass_data_ipa_cp): ...new pass_data instance and...
12648 (make_pass_ipa_cp): ...new function.
12649 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
12650 global struct to a subclass of gimple_opt_pass along with...
12651 (pass_data_inline_parameters): ...new pass_data instance and...
12652 (make_pass_inline_parameters): ...new function.
12653 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
12654 subclass of gimple_opt_pass along with...
12655 (pass_data_early_inline): ...new pass_data instance and...
12656 (make_pass_early_inline): ...new function.
12657 (pass_ipa_inline): Convert from a global struct to a subclass of
12658 ipa_opt_pass_d along with...
12659 (pass_data_ipa_inline): ...new pass_data instance and...
12660 (make_pass_ipa_inline): ...new function.
12661 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
12662 struct to a subclass of gimple_opt_pass along with...
12663 (pass_data_local_pure_const): ...new pass_data instance and...
12664 (make_pass_local_pure_const): ...new function.
12665 (pass_ipa_pure_const): Convert from a global struct to a subclass of
12666 ipa_opt_pass_d along with...
12667 (pass_data_ipa_pure_const): ...new pass_data instance and...
12668 (make_pass_ipa_pure_const): ...new function.
12669 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
12670 to a subclass of ipa_opt_pass_d along with...
12671 (pass_data_ipa_reference): ...new pass_data instance and...
12672 (make_pass_ipa_reference): ...new function.
12673 * ipa-split.c (pass_split_functions): Convert from a global struct to
12674 a subclass of gimple_opt_pass along with...
12675 (pass_data_split_functions): ...new pass_data instance and...
12676 (make_pass_split_functions): ...new function.
12677 (pass_feedback_split_functions): Convert from a global struct to a
12678 subclass of gimple_opt_pass along with...
12679 (pass_data_feedback_split_functions): ...new pass_data instance and...
12680 (make_pass_feedback_split_functions): ...new function.
12681 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
12682 global struct to a subclass of simple_ipa_opt_pass along with...
12683 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
12684 instance and...
12685 (make_pass_ipa_function_and_variable_visibility): ...new function.
12686 (pass_ipa_free_inline_summary): Convert from a global struct to a
12687 subclass of simple_ipa_opt_pass along with...
12688 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
12689 (make_pass_ipa_free_inline_summary): ...new function.
12690 (pass_ipa_whole_program_visibility): Convert from a global struct to a
12691 subclass of ipa_opt_pass_d along with...
12692 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
12693 and...
12694 (make_pass_ipa_whole_program_visibility): ...new function.
12695 (pass_ipa_profile): Convert from a global struct to a subclass of
12696 ipa_opt_pass_d along with...
12697 (pass_data_ipa_profile): ...new pass_data instance and...
12698 (make_pass_ipa_profile): ...new function.
12699 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
12700 ipa_opt_pass_d along with...
12701 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
12702 (make_pass_ipa_cdtor_merge): ...new function.
12703 * ira.c (pass_ira): Convert from a global struct to a subclass of
12704 rtl_opt_pass along with...
12705 (pass_data_ira): ...new pass_data instance and...
12706 (make_pass_ira): ...new function.
12707 (pass_reload): Convert from a global struct to a subclass of
12708 rtl_opt_pass along with...
12709 (pass_data_reload): ...new pass_data instance and...
12710 (make_pass_reload): ...new function.
12711 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
12712 subclass of rtl_opt_pass along with...
12713 (pass_data_cleanup_barriers): ...new pass_data instance and...
12714 (make_pass_cleanup_barriers): ...new function.
12715 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
12716 of rtl_opt_pass along with...
12717 (pass_data_loop2): ...new pass_data instance and...
12718 (make_pass_loop2): ...new function.
12719 (pass_rtl_loop_init): Convert from a global struct to a subclass of
12720 rtl_opt_pass along with...
12721 (pass_data_rtl_loop_init): ...new pass_data instance and...
12722 (make_pass_rtl_loop_init): ...new function.
12723 (pass_rtl_loop_done): Convert from a global struct to a subclass of
12724 rtl_opt_pass along with...
12725 (pass_data_rtl_loop_done): ...new pass_data instance and...
12726 (make_pass_rtl_loop_done): ...new function.
12727 (pass_rtl_move_loop_invariants): Convert from a global struct to a
12728 subclass of rtl_opt_pass along with...
12729 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
12730 (make_pass_rtl_move_loop_invariants): ...new function.
12731 (pass_rtl_unswitch): Convert from a global struct to a subclass of
12732 rtl_opt_pass along with...
12733 (pass_data_rtl_unswitch): ...new pass_data instance and...
12734 (make_pass_rtl_unswitch): ...new function.
12735 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
12736 subclass of rtl_opt_pass along with...
12737 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
12738 and...
12739 (make_pass_rtl_unroll_and_peel_loops): ...new function.
12740 (pass_rtl_doloop): Convert from a global struct to a subclass of
12741 rtl_opt_pass along with...
12742 (pass_data_rtl_doloop): ...new pass_data instance and...
12743 (make_pass_rtl_doloop): ...new function.
12744 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
12745 a subclass of rtl_opt_pass along with...
12746 (pass_data_lower_subreg): ...new pass_data instance and...
12747 (make_pass_lower_subreg): ...new function.
12748 (pass_lower_subreg2): Convert from a global struct to a subclass of
12749 rtl_opt_pass along with...
12750 (pass_data_lower_subreg2): ...new pass_data instance and...
12751 (make_pass_lower_subreg2): ...new function.
12752 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
12753 struct to a subclass of ipa_opt_pass_d along with...
12754 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
12755 (make_pass_ipa_lto_gimple_out): ...new function.
12756 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
12757 of ipa_opt_pass_d along with...
12758 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
12759 (make_pass_ipa_lto_finish_out): ...new function.
12760 * mode-switching.c (pass_mode_switching): Convert from a global struct
12761 to a subclass of rtl_opt_pass along with...
12762 (pass_data_mode_switching): ...new pass_data instance and...
12763 (make_pass_mode_switching): ...new function.
12764 * modulo-sched.c (pass_sms): Convert from a global struct to a
12765 subclass of rtl_opt_pass along with...
12766 (pass_data_sms): ...new pass_data instance and...
12767 (make_pass_sms): ...new function.
12768 * omp-low.c (pass_expand_omp): Convert from a global struct to a
12769 subclass of gimple_opt_pass along with...
12770 (pass_data_expand_omp): ...new pass_data instance and...
12771 (make_pass_expand_omp): ...new function.
12772 (pass_lower_omp): Convert from a global struct to a subclass of
12773 gimple_opt_pass along with...
12774 (pass_data_lower_omp): ...new pass_data instance and...
12775 (make_pass_lower_omp): ...new function.
12776 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
12777 of gimple_opt_pass along with...
12778 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
12779 (make_pass_diagnose_omp_blocks): ...new function.
12780 * passes.c (pass_early_local_passes): Convert from a global struct to
12781 a subclass of simple_ipa_opt_pass along with...
12782 (pass_data_early_local_passes): ...new pass_data instance and...
12783 (make_pass_early_local_passes): ...new function.
12784 (pass_all_early_optimizations): Convert from a global struct to a
12785 subclass of gimple_opt_pass along with...
12786 (pass_data_all_early_optimizations): ...new pass_data instance and...
12787 (make_pass_all_early_optimizations): ...new function.
12788 (pass_all_optimizations): Convert from a global struct to a subclass
12789 of gimple_opt_pass along with...
12790 (pass_data_all_optimizations): ...new pass_data instance and...
12791 (make_pass_all_optimizations): ...new function.
12792 (pass_all_optimizations_g): Convert from a global struct to a subclass
12793 of gimple_opt_pass along with...
12794 (pass_data_all_optimizations_g): ...new pass_data instance and...
12795 (make_pass_all_optimizations_g): ...new function.
12796 (pass_rest_of_compilation): Convert from a global struct to a subclass
12797 of rtl_opt_pass along with...
12798 (pass_data_rest_of_compilation): ...new pass_data instance and...
12799 (make_pass_rest_of_compilation): ...new function.
12800 (pass_postreload): Convert from a global struct to a subclass of
12801 rtl_opt_pass along with...
12802 (pass_data_postreload): ...new pass_data instance and...
12803 (make_pass_postreload): ...new function.
12804 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
12805 subclass of rtl_opt_pass along with...
12806 (pass_data_gcse2): ...new pass_data instance and...
12807 (make_pass_gcse2): ...new function.
12808 * postreload.c (pass_postreload_cse): Convert from a global struct to
12809 a subclass of rtl_opt_pass along with...
12810 (pass_data_postreload_cse): ...new pass_data instance and...
12811 (make_pass_postreload_cse): ...new function.
12812 * predict.c (pass_profile): Convert from a global struct to a subclass
12813 of gimple_opt_pass along with...
12814 (pass_data_profile): ...new pass_data instance and...
12815 (make_pass_profile): ...new function.
12816 (pass_strip_predict_hints): Convert from a global struct to a subclass
12817 of gimple_opt_pass along with...
12818 (pass_data_strip_predict_hints): ...new pass_data instance and...
12819 (make_pass_strip_predict_hints): ...new function.
12820 * recog.c (pass_peephole2): Convert from a global struct to a subclass
12821 of rtl_opt_pass along with...
12822 (pass_data_peephole2): ...new pass_data instance and...
12823 (make_pass_peephole2): ...new function.
12824 (pass_split_all_insns): Convert from a global struct to a subclass of
12825 rtl_opt_pass along with...
12826 (pass_data_split_all_insns): ...new pass_data instance and...
12827 (make_pass_split_all_insns): ...new function.
12828 (pass_split_after_reload): Convert from a global struct to a subclass
12829 of rtl_opt_pass along with...
12830 (pass_data_split_after_reload): ...new pass_data instance and...
12831 (make_pass_split_after_reload): ...new function.
12832 (pass_split_before_regstack): Convert from a global struct to a
12833 subclass of rtl_opt_pass along with...
12834 (pass_data_split_before_regstack): ...new pass_data instance and...
12835 (make_pass_split_before_regstack): ...new function.
12836 (pass_split_before_sched2): Convert from a global struct to a subclass
12837 of rtl_opt_pass along with...
12838 (pass_data_split_before_sched2): ...new pass_data instance and...
12839 (make_pass_split_before_sched2): ...new function.
12840 (pass_split_for_shorten_branches): Convert from a global struct to a
12841 subclass of rtl_opt_pass along with...
12842 (pass_data_split_for_shorten_branches): ...new pass_data instance
12843 and...
12844 (make_pass_split_for_shorten_branches): ...new function.
12845 * ree.c (pass_ree): Convert from a global struct to a subclass of
12846 rtl_opt_pass along with...
12847 (pass_data_ree): ...new pass_data instance and...
12848 (make_pass_ree): ...new function.
12849 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
12850 subclass of rtl_opt_pass along with...
12851 (pass_data_stack_regs): ...new pass_data instance and...
12852 (make_pass_stack_regs): ...new function.
12853 (pass_stack_regs_run): Convert from a global struct to a subclass of
12854 rtl_opt_pass along with...
12855 (pass_data_stack_regs_run): ...new pass_data instance and...
12856 (make_pass_stack_regs_run): ...new function.
12857 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
12858 subclass of rtl_opt_pass along with...
12859 (pass_data_cprop_hardreg): ...new pass_data instance and...
12860 (make_pass_cprop_hardreg): ...new function.
12861 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
12862 subclass of rtl_opt_pass along with...
12863 (pass_data_reginfo_init): ...new pass_data instance and...
12864 (make_pass_reginfo_init): ...new function.
12865 * regmove.c (pass_regmove): Convert from a global struct to a subclass
12866 of rtl_opt_pass along with...
12867 (pass_data_regmove): ...new pass_data instance and...
12868 (make_pass_regmove): ...new function.
12869 * regrename.c (pass_regrename): Convert from a global struct to a
12870 subclass of rtl_opt_pass along with...
12871 (pass_data_regrename): ...new pass_data instance and...
12872 (make_pass_regrename): ...new function.
12873 * reorg.c (pass_delay_slots): Convert from a global struct to a
12874 subclass of rtl_opt_pass along with...
12875 (pass_data_delay_slots): ...new pass_data instance and...
12876 (make_pass_delay_slots): ...new function.
12877 (pass_machine_reorg): Convert from a global struct to a subclass of
12878 rtl_opt_pass along with...
12879 (pass_data_machine_reorg): ...new pass_data instance and...
12880 (make_pass_machine_reorg): ...new function.
12881 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
12882 of rtl_opt_pass along with...
12883 (pass_data_sched): ...new pass_data instance and...
12884 (make_pass_sched): ...new function.
12885 (pass_sched2): Convert from a global struct to a subclass of
12886 rtl_opt_pass along with...
12887 (pass_data_sched2): ...new pass_data instance and...
12888 (make_pass_sched2): ...new function.
12889 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
12890 to a subclass of rtl_opt_pass along with...
12891 (pass_data_stack_ptr_mod): ...new pass_data instance and...
12892 (make_pass_stack_ptr_mod): ...new function.
12893 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
12894 to a subclass of rtl_opt_pass along with...
12895 (pass_data_rtl_store_motion): ...new pass_data instance and...
12896 (make_pass_rtl_store_motion): ...new function.
12897 * tracer.c (pass_tracer): Convert from a global struct to a subclass
12898 of gimple_opt_pass along with...
12899 (pass_data_tracer): ...new pass_data instance and...
12900 (make_pass_tracer): ...new function.
12901 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
12902 to a subclass of gimple_opt_pass along with...
12903 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
12904 (make_pass_diagnose_tm_blocks): ...new function.
12905 (pass_lower_tm): Convert from a global struct to a subclass of
12906 gimple_opt_pass along with...
12907 (pass_data_lower_tm): ...new pass_data instance and...
12908 (make_pass_lower_tm): ...new function.
12909 (pass_tm_init): Convert from a global struct to a subclass of
12910 gimple_opt_pass along with...
12911 (pass_data_tm_init): ...new pass_data instance and...
12912 (make_pass_tm_init): ...new function.
12913 (pass_tm_mark): Convert from a global struct to a subclass of
12914 gimple_opt_pass along with...
12915 (pass_data_tm_mark): ...new pass_data instance and...
12916 (make_pass_tm_mark): ...new function.
12917 (pass_tm_edges): Convert from a global struct to a subclass of
12918 gimple_opt_pass along with...
12919 (pass_data_tm_edges): ...new pass_data instance and...
12920 (make_pass_tm_edges): ...new function.
12921 (pass_tm_memopt): Convert from a global struct to a subclass of
12922 gimple_opt_pass along with...
12923 (pass_data_tm_memopt): ...new pass_data instance and...
12924 (make_pass_tm_memopt): ...new function.
12925 (pass_ipa_tm): Convert from a global struct to a subclass of
12926 simple_ipa_opt_pass along with...
12927 (pass_data_ipa_tm): ...new pass_data instance and...
12928 (make_pass_ipa_tm): ...new function.
12929 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
12930 subclass of gimple_opt_pass along with...
12931 (pass_data_call_cdce): ...new pass_data instance and...
12932 (make_pass_call_cdce): ...new function.
12933 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
12934 subclass of gimple_opt_pass along with...
12935 (pass_data_build_cfg): ...new pass_data instance and...
12936 (make_pass_build_cfg): ...new function.
12937 (pass_split_crit_edges): Convert from a global struct to a subclass of
12938 gimple_opt_pass along with...
12939 (pass_data_split_crit_edges): ...new pass_data instance and...
12940 (make_pass_split_crit_edges): ...new function.
12941 (pass_warn_function_return): Convert from a global struct to a
12942 subclass of gimple_opt_pass along with...
12943 (pass_data_warn_function_return): ...new pass_data instance and...
12944 (make_pass_warn_function_return): ...new function.
12945 (pass_warn_function_noreturn): Convert from a global struct to a
12946 subclass of gimple_opt_pass along with...
12947 (pass_data_warn_function_noreturn): ...new pass_data instance and...
12948 (make_pass_warn_function_noreturn): ...new function.
12949 (pass_warn_unused_result): Convert from a global struct to a subclass
12950 of gimple_opt_pass along with...
12951 (pass_data_warn_unused_result): ...new pass_data instance and...
12952 (make_pass_warn_unused_result): ...new function.
12953 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
12954 a subclass of gimple_opt_pass along with...
12955 (pass_data_merge_phi): ...new pass_data instance and...
12956 (make_pass_merge_phi): ...new function.
12957 * tree-complex.c (pass_lower_complex): Convert from a global struct to
12958 a subclass of gimple_opt_pass along with...
12959 (pass_data_lower_complex): ...new pass_data instance and...
12960 (make_pass_lower_complex): ...new function.
12961 (pass_lower_complex_O0): Convert from a global struct to a subclass of
12962 gimple_opt_pass along with...
12963 (pass_data_lower_complex_O0): ...new pass_data instance and...
12964 (make_pass_lower_complex_O0): ...new function.
12965 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
12966 subclass of gimple_opt_pass along with...
12967 (pass_data_lower_eh): ...new pass_data instance and...
12968 (make_pass_lower_eh): ...new function.
12969 (pass_refactor_eh): Convert from a global struct to a subclass of
12970 gimple_opt_pass along with...
12971 (pass_data_refactor_eh): ...new pass_data instance and...
12972 (make_pass_refactor_eh): ...new function.
12973 (pass_lower_resx): Convert from a global struct to a subclass of
12974 gimple_opt_pass along with...
12975 (pass_data_lower_resx): ...new pass_data instance and...
12976 (make_pass_lower_resx): ...new function.
12977 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
12978 of gimple_opt_pass along with...
12979 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
12980 (make_pass_lower_eh_dispatch): ...new function.
12981 (pass_cleanup_eh): Convert from a global struct to a subclass of
12982 gimple_opt_pass along with...
12983 (pass_data_cleanup_eh): ...new pass_data instance and...
12984 (make_pass_cleanup_eh): ...new function.
12985 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
12986 to a subclass of simple_ipa_opt_pass along with...
12987 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
12988 (make_pass_ipa_lower_emutls): ...new function.
12989 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
12990 a subclass of gimple_opt_pass along with...
12991 (pass_data_if_conversion): ...new pass_data instance and...
12992 (make_pass_if_conversion): ...new function.
12993 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
12994 subclass of gimple_opt_pass along with...
12995 (pass_data_build_ssa): ...new pass_data instance and...
12996 (make_pass_build_ssa): ...new function.
12997 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
12998 global struct to a subclass of gimple_opt_pass along with...
12999 (pass_data_loop_distribution): ...new pass_data instance and...
13000 (make_pass_loop_distribution): ...new function.
13001 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
13002 subclass of gimple_opt_pass along with...
13003 (pass_data_mudflap_1): ...new pass_data instance and...
13004 (make_pass_mudflap_1): ...new function.
13005 (pass_mudflap_2): Convert from a global struct to a subclass of
13006 gimple_opt_pass along with...
13007 (pass_data_mudflap_2): ...new pass_data instance and...
13008 (make_pass_mudflap_2): ...new function.
13009 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
13010 subclass of gimple_opt_pass along with...
13011 (pass_data_mudflap_1): ...new pass_data instance and...
13012 (make_pass_mudflap_1): ...new function.
13013 (pass_mudflap_2): Convert from a global struct to a subclass of
13014 gimple_opt_pass along with...
13015 (pass_data_mudflap_2): ...new pass_data instance and...
13016 (make_pass_mudflap_2): ...new function.
13017 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
13018 gimple_opt_pass along with...
13019 (pass_data_nrv): ...new pass_data instance and...
13020 (make_pass_nrv): ...new function.
13021 (pass_return_slot): Convert from a global struct to a subclass of
13022 gimple_opt_pass along with...
13023 (pass_data_return_slot): ...new pass_data instance and...
13024 (make_pass_return_slot): ...new function.
13025 * tree-object-size.c (pass_object_sizes): Convert from a global struct
13026 to a subclass of gimple_opt_pass along with...
13027 (pass_data_object_sizes): ...new pass_data instance and...
13028 (make_pass_object_sizes): ...new function.
13029 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
13030 global struct to a subclass of gimple_opt_pass along with...
13031 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
13032 and...
13033 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13034 (pass_fixup_cfg): Convert from a global struct to a subclass of
13035 gimple_opt_pass along with...
13036 (pass_data_fixup_cfg): ...new pass_data instance and...
13037 (make_pass_fixup_cfg): ...new function.
13038 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
13039 (make_pass_mudflap_1): ...new function.
13040 (pass_mudflap_2): Replace declaration with that of...
13041 (make_pass_mudflap_2): ...new function.
13042 (pass_asan): Replace declaration with that of...
13043 (make_pass_asan): ...new function.
13044 (pass_asan_O0): Replace declaration with that of...
13045 (make_pass_asan_O0): ...new function.
13046 (pass_tsan): Replace declaration with that of...
13047 (make_pass_tsan): ...new function.
13048 (pass_tsan_O0): Replace declaration with that of...
13049 (make_pass_tsan_O0): ...new function.
13050 (pass_lower_cf): Replace declaration with that of...
13051 (make_pass_lower_cf): ...new function.
13052 (pass_refactor_eh): Replace declaration with that of...
13053 (make_pass_refactor_eh): ...new function.
13054 (pass_lower_eh): Replace declaration with that of...
13055 (make_pass_lower_eh): ...new function.
13056 (pass_lower_eh_dispatch): Replace declaration with that of...
13057 (make_pass_lower_eh_dispatch): ...new function.
13058 (pass_lower_resx): Replace declaration with that of...
13059 (make_pass_lower_resx): ...new function.
13060 (pass_build_cfg): Replace declaration with that of...
13061 (make_pass_build_cfg): ...new function.
13062 (pass_early_tree_profile): Replace declaration with that of...
13063 (make_pass_early_tree_profile): ...new function.
13064 (pass_cleanup_eh): Replace declaration with that of...
13065 (make_pass_cleanup_eh): ...new function.
13066 (pass_sra): Replace declaration with that of...
13067 (make_pass_sra): ...new function.
13068 (pass_sra_early): Replace declaration with that of...
13069 (make_pass_sra_early): ...new function.
13070 (pass_early_ipa_sra): Replace declaration with that of...
13071 (make_pass_early_ipa_sra): ...new function.
13072 (pass_tail_recursion): Replace declaration with that of...
13073 (make_pass_tail_recursion): ...new function.
13074 (pass_tail_calls): Replace declaration with that of...
13075 (make_pass_tail_calls): ...new function.
13076 (pass_tree_loop): Replace declaration with that of...
13077 (make_pass_tree_loop): ...new function.
13078 (pass_tree_loop_init): Replace declaration with that of...
13079 (make_pass_tree_loop_init): ...new function.
13080 (pass_lim): Replace declaration with that of...
13081 (make_pass_lim): ...new function.
13082 (pass_tree_unswitch): Replace declaration with that of...
13083 (make_pass_tree_unswitch): ...new function.
13084 (pass_predcom): Replace declaration with that of...
13085 (make_pass_predcom): ...new function.
13086 (pass_iv_canon): Replace declaration with that of...
13087 (make_pass_iv_canon): ...new function.
13088 (pass_scev_cprop): Replace declaration with that of...
13089 (make_pass_scev_cprop): ...new function.
13090 (pass_empty_loop): Replace declaration with that of...
13091 (make_pass_empty_loop): ...new function.
13092 (pass_record_bounds): Replace declaration with that of...
13093 (make_pass_record_bounds): ...new function.
13094 (pass_graphite): Replace declaration with that of...
13095 (make_pass_graphite): ...new function.
13096 (pass_graphite_transforms): Replace declaration with that of...
13097 (make_pass_graphite_transforms): ...new function.
13098 (pass_if_conversion): Replace declaration with that of...
13099 (make_pass_if_conversion): ...new function.
13100 (pass_loop_distribution): Replace declaration with that of...
13101 (make_pass_loop_distribution): ...new function.
13102 (pass_vectorize): Replace declaration with that of...
13103 (make_pass_vectorize): ...new function.
13104 (pass_slp_vectorize): Replace declaration with that of...
13105 (make_pass_slp_vectorize): ...new function.
13106 (pass_complete_unroll): Replace declaration with that of...
13107 (make_pass_complete_unroll): ...new function.
13108 (pass_complete_unrolli): Replace declaration with that of...
13109 (make_pass_complete_unrolli): ...new function.
13110 (pass_parallelize_loops): Replace declaration with that of...
13111 (make_pass_parallelize_loops): ...new function.
13112 (pass_loop_prefetch): Replace declaration with that of...
13113 (make_pass_loop_prefetch): ...new function.
13114 (pass_iv_optimize): Replace declaration with that of...
13115 (make_pass_iv_optimize): ...new function.
13116 (pass_tree_loop_done): Replace declaration with that of...
13117 (make_pass_tree_loop_done): ...new function.
13118 (pass_ch): Replace declaration with that of...
13119 (make_pass_ch): ...new function.
13120 (pass_ccp): Replace declaration with that of...
13121 (make_pass_ccp): ...new function.
13122 (pass_phi_only_cprop): Replace declaration with that of...
13123 (make_pass_phi_only_cprop): ...new function.
13124 (pass_build_ssa): Replace declaration with that of...
13125 (make_pass_build_ssa): ...new function.
13126 (pass_build_alias): Replace declaration with that of...
13127 (make_pass_build_alias): ...new function.
13128 (pass_build_ealias): Replace declaration with that of...
13129 (make_pass_build_ealias): ...new function.
13130 (pass_dominator): Replace declaration with that of...
13131 (make_pass_dominator): ...new function.
13132 (pass_dce): Replace declaration with that of...
13133 (make_pass_dce): ...new function.
13134 (pass_dce_loop): Replace declaration with that of...
13135 (make_pass_dce_loop): ...new function.
13136 (pass_cd_dce): Replace declaration with that of...
13137 (make_pass_cd_dce): ...new function.
13138 (pass_call_cdce): Replace declaration with that of...
13139 (make_pass_call_cdce): ...new function.
13140 (pass_merge_phi): Replace declaration with that of...
13141 (make_pass_merge_phi): ...new function.
13142 (pass_split_crit_edges): Replace declaration with that of...
13143 (make_pass_split_crit_edges): ...new function.
13144 (pass_pre): Replace declaration with that of...
13145 (make_pass_pre): ...new function.
13146 (pass_profile): Replace declaration with that of...
13147 (make_pass_profile): ...new function.
13148 (pass_strip_predict_hints): Replace declaration with that of...
13149 (make_pass_strip_predict_hints): ...new function.
13150 (pass_lower_complex_O0): Replace declaration with that of...
13151 (make_pass_lower_complex_O0): ...new function.
13152 (pass_lower_complex): Replace declaration with that of...
13153 (make_pass_lower_complex): ...new function.
13154 (pass_lower_vector): Replace declaration with that of...
13155 (make_pass_lower_vector): ...new function.
13156 (pass_lower_vector_ssa): Replace declaration with that of...
13157 (make_pass_lower_vector_ssa): ...new function.
13158 (pass_lower_omp): Replace declaration with that of...
13159 (make_pass_lower_omp): ...new function.
13160 (pass_diagnose_omp_blocks): Replace declaration with that of...
13161 (make_pass_diagnose_omp_blocks): ...new function.
13162 (pass_expand_omp): Replace declaration with that of...
13163 (make_pass_expand_omp): ...new function.
13164 (pass_expand_omp_ssa): Replace declaration with that of...
13165 (make_pass_expand_omp_ssa): ...new function.
13166 (pass_object_sizes): Replace declaration with that of...
13167 (make_pass_object_sizes): ...new function.
13168 (pass_strlen): Replace declaration with that of...
13169 (make_pass_strlen): ...new function.
13170 (pass_fold_builtins): Replace declaration with that of...
13171 (make_pass_fold_builtins): ...new function.
13172 (pass_stdarg): Replace declaration with that of...
13173 (make_pass_stdarg): ...new function.
13174 (pass_early_warn_uninitialized): Replace declaration with that of...
13175 (make_pass_early_warn_uninitialized): ...new function.
13176 (pass_late_warn_uninitialized): Replace declaration with that of...
13177 (make_pass_late_warn_uninitialized): ...new function.
13178 (pass_cse_reciprocals): Replace declaration with that of...
13179 (make_pass_cse_reciprocals): ...new function.
13180 (pass_cse_sincos): Replace declaration with that of...
13181 (make_pass_cse_sincos): ...new function.
13182 (pass_optimize_bswap): Replace declaration with that of...
13183 (make_pass_optimize_bswap): ...new function.
13184 (pass_optimize_widening_mul): Replace declaration with that of...
13185 (make_pass_optimize_widening_mul): ...new function.
13186 (pass_warn_function_return): Replace declaration with that of...
13187 (make_pass_warn_function_return): ...new function.
13188 (pass_warn_function_noreturn): Replace declaration with that of...
13189 (make_pass_warn_function_noreturn): ...new function.
13190 (pass_cselim): Replace declaration with that of...
13191 (make_pass_cselim): ...new function.
13192 (pass_phiopt): Replace declaration with that of...
13193 (make_pass_phiopt): ...new function.
13194 (pass_forwprop): Replace declaration with that of...
13195 (make_pass_forwprop): ...new function.
13196 (pass_phiprop): Replace declaration with that of...
13197 (make_pass_phiprop): ...new function.
13198 (pass_tree_ifcombine): Replace declaration with that of...
13199 (make_pass_tree_ifcombine): ...new function.
13200 (pass_dse): Replace declaration with that of...
13201 (make_pass_dse): ...new function.
13202 (pass_nrv): Replace declaration with that of...
13203 (make_pass_nrv): ...new function.
13204 (pass_rename_ssa_copies): Replace declaration with that of...
13205 (make_pass_rename_ssa_copies): ...new function.
13206 (pass_sink_code): Replace declaration with that of...
13207 (make_pass_sink_code): ...new function.
13208 (pass_fre): Replace declaration with that of...
13209 (make_pass_fre): ...new function.
13210 (pass_check_data_deps): Replace declaration with that of...
13211 (make_pass_check_data_deps): ...new function.
13212 (pass_copy_prop): Replace declaration with that of...
13213 (make_pass_copy_prop): ...new function.
13214 (pass_vrp): Replace declaration with that of...
13215 (make_pass_vrp): ...new function.
13216 (pass_uncprop): Replace declaration with that of...
13217 (make_pass_uncprop): ...new function.
13218 (pass_return_slot): Replace declaration with that of...
13219 (make_pass_return_slot): ...new function.
13220 (pass_reassoc): Replace declaration with that of...
13221 (make_pass_reassoc): ...new function.
13222 (pass_rebuild_cgraph_edges): Replace declaration with that of...
13223 (make_pass_rebuild_cgraph_edges): ...new function.
13224 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
13225 (make_pass_remove_cgraph_callee_edges): ...new function.
13226 (pass_build_cgraph_edges): Replace declaration with that of...
13227 (make_pass_build_cgraph_edges): ...new function.
13228 (pass_local_pure_const): Replace declaration with that of...
13229 (make_pass_local_pure_const): ...new function.
13230 (pass_tracer): Replace declaration with that of...
13231 (make_pass_tracer): ...new function.
13232 (pass_warn_unused_result): Replace declaration with that of...
13233 (make_pass_warn_unused_result): ...new function.
13234 (pass_diagnose_tm_blocks): Replace declaration with that of...
13235 (make_pass_diagnose_tm_blocks): ...new function.
13236 (pass_lower_tm): Replace declaration with that of...
13237 (make_pass_lower_tm): ...new function.
13238 (pass_tm_init): Replace declaration with that of...
13239 (make_pass_tm_init): ...new function.
13240 (pass_tm_mark): Replace declaration with that of...
13241 (make_pass_tm_mark): ...new function.
13242 (pass_tm_memopt): Replace declaration with that of...
13243 (make_pass_tm_memopt): ...new function.
13244 (pass_tm_edges): Replace declaration with that of...
13245 (make_pass_tm_edges): ...new function.
13246 (pass_split_functions): Replace declaration with that of...
13247 (make_pass_split_functions): ...new function.
13248 (pass_feedback_split_functions): Replace declaration with that of...
13249 (make_pass_feedback_split_functions): ...new function.
13250 (pass_strength_reduction): Replace declaration with that of...
13251 (make_pass_strength_reduction): ...new function.
13252 (pass_ipa_lower_emutls): Replace declaration with that of...
13253 (make_pass_ipa_lower_emutls): ...new function.
13254 (pass_ipa_function_and_variable_visibility): Replace declaration with
13255 that of...
13256 (make_pass_ipa_function_and_variable_visibility): ...new function.
13257 (pass_ipa_tree_profile): Replace declaration with that of...
13258 (make_pass_ipa_tree_profile): ...new function.
13259 (pass_early_local_passes): Replace declaration with that of...
13260 (make_pass_early_local_passes): ...new function.
13261 (pass_ipa_whole_program_visibility): Replace declaration with that
13262 of...
13263 (make_pass_ipa_whole_program_visibility): ...new function.
13264 (pass_ipa_lto_gimple_out): Replace declaration with that of...
13265 (make_pass_ipa_lto_gimple_out): ...new function.
13266 (pass_ipa_increase_alignment): Replace declaration with that of...
13267 (make_pass_ipa_increase_alignment): ...new function.
13268 (pass_ipa_inline): Replace declaration with that of...
13269 (make_pass_ipa_inline): ...new function.
13270 (pass_ipa_free_lang_data): Replace declaration with that of...
13271 (make_pass_ipa_free_lang_data): ...new function.
13272 (pass_ipa_free_inline_summary): Replace declaration with that of...
13273 (make_pass_ipa_free_inline_summary): ...new function.
13274 (pass_ipa_cp): Replace declaration with that of...
13275 (make_pass_ipa_cp): ...new function.
13276 (pass_ipa_reference): Replace declaration with that of...
13277 (make_pass_ipa_reference): ...new function.
13278 (pass_ipa_pure_const): Replace declaration with that of...
13279 (make_pass_ipa_pure_const): ...new function.
13280 (pass_ipa_pta): Replace declaration with that of...
13281 (make_pass_ipa_pta): ...new function.
13282 (pass_ipa_lto_finish_out): Replace declaration with that of...
13283 (make_pass_ipa_lto_finish_out): ...new function.
13284 (pass_ipa_tm): Replace declaration with that of...
13285 (make_pass_ipa_tm): ...new function.
13286 (pass_ipa_profile): Replace declaration with that of...
13287 (make_pass_ipa_profile): ...new function.
13288 (pass_ipa_cdtor_merge): Replace declaration with that of...
13289 (make_pass_ipa_cdtor_merge): ...new function.
13290 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
13291 of...
13292 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13293 (pass_init_datastructures): Replace declaration with that of...
13294 (make_pass_init_datastructures): ...new function.
13295 (pass_fixup_cfg): Replace declaration with that of...
13296 (make_pass_fixup_cfg): ...new function.
13297 (pass_expand): Replace declaration with that of...
13298 (make_pass_expand): ...new function.
13299 (pass_instantiate_virtual_regs): Replace declaration with that of...
13300 (make_pass_instantiate_virtual_regs): ...new function.
13301 (pass_rtl_fwprop): Replace declaration with that of...
13302 (make_pass_rtl_fwprop): ...new function.
13303 (pass_rtl_fwprop_addr): Replace declaration with that of...
13304 (make_pass_rtl_fwprop_addr): ...new function.
13305 (pass_jump): Replace declaration with that of...
13306 (make_pass_jump): ...new function.
13307 (pass_jump2): Replace declaration with that of...
13308 (make_pass_jump2): ...new function.
13309 (pass_lower_subreg): Replace declaration with that of...
13310 (make_pass_lower_subreg): ...new function.
13311 (pass_cse): Replace declaration with that of...
13312 (make_pass_cse): ...new function.
13313 (pass_fast_rtl_dce): Replace declaration with that of...
13314 (make_pass_fast_rtl_dce): ...new function.
13315 (pass_ud_rtl_dce): Replace declaration with that of...
13316 (make_pass_ud_rtl_dce): ...new function.
13317 (pass_rtl_dce): Replace declaration with that of...
13318 (make_pass_rtl_dce): ...new function.
13319 (pass_rtl_dse1): Replace declaration with that of...
13320 (make_pass_rtl_dse1): ...new function.
13321 (pass_rtl_dse2): Replace declaration with that of...
13322 (make_pass_rtl_dse2): ...new function.
13323 (pass_rtl_dse3): Replace declaration with that of...
13324 (make_pass_rtl_dse3): ...new function.
13325 (pass_rtl_cprop): Replace declaration with that of...
13326 (make_pass_rtl_cprop): ...new function.
13327 (pass_rtl_pre): Replace declaration with that of...
13328 (make_pass_rtl_pre): ...new function.
13329 (pass_rtl_hoist): Replace declaration with that of...
13330 (make_pass_rtl_hoist): ...new function.
13331 (pass_rtl_store_motion): Replace declaration with that of...
13332 (make_pass_rtl_store_motion): ...new function.
13333 (pass_cse_after_global_opts): Replace declaration with that of...
13334 (make_pass_cse_after_global_opts): ...new function.
13335 (pass_rtl_ifcvt): Replace declaration with that of...
13336 (make_pass_rtl_ifcvt): ...new function.
13337 (pass_into_cfg_layout_mode): Replace declaration with that of...
13338 (make_pass_into_cfg_layout_mode): ...new function.
13339 (pass_outof_cfg_layout_mode): Replace declaration with that of...
13340 (make_pass_outof_cfg_layout_mode): ...new function.
13341 (pass_loop2): Replace declaration with that of...
13342 (make_pass_loop2): ...new function.
13343 (pass_rtl_loop_init): Replace declaration with that of...
13344 (make_pass_rtl_loop_init): ...new function.
13345 (pass_rtl_move_loop_invariants): Replace declaration with that of...
13346 (make_pass_rtl_move_loop_invariants): ...new function.
13347 (pass_rtl_unswitch): Replace declaration with that of...
13348 (make_pass_rtl_unswitch): ...new function.
13349 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
13350 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13351 (pass_rtl_doloop): Replace declaration with that of...
13352 (make_pass_rtl_doloop): ...new function.
13353 (pass_rtl_loop_done): Replace declaration with that of...
13354 (make_pass_rtl_loop_done): ...new function.
13355 (pass_web): Replace declaration with that of...
13356 (make_pass_web): ...new function.
13357 (pass_cse2): Replace declaration with that of...
13358 (make_pass_cse2): ...new function.
13359 (pass_df_initialize_opt): Replace declaration with that of...
13360 (make_pass_df_initialize_opt): ...new function.
13361 (pass_df_initialize_no_opt): Replace declaration with that of...
13362 (make_pass_df_initialize_no_opt): ...new function.
13363 (pass_reginfo_init): Replace declaration with that of...
13364 (make_pass_reginfo_init): ...new function.
13365 (pass_inc_dec): Replace declaration with that of...
13366 (make_pass_inc_dec): ...new function.
13367 (pass_stack_ptr_mod): Replace declaration with that of...
13368 (make_pass_stack_ptr_mod): ...new function.
13369 (pass_initialize_regs): Replace declaration with that of...
13370 (make_pass_initialize_regs): ...new function.
13371 (pass_combine): Replace declaration with that of...
13372 (make_pass_combine): ...new function.
13373 (pass_if_after_combine): Replace declaration with that of...
13374 (make_pass_if_after_combine): ...new function.
13375 (pass_ree): Replace declaration with that of...
13376 (make_pass_ree): ...new function.
13377 (pass_partition_blocks): Replace declaration with that of...
13378 (make_pass_partition_blocks): ...new function.
13379 (pass_match_asm_constraints): Replace declaration with that of...
13380 (make_pass_match_asm_constraints): ...new function.
13381 (pass_regmove): Replace declaration with that of...
13382 (make_pass_regmove): ...new function.
13383 (pass_split_all_insns): Replace declaration with that of...
13384 (make_pass_split_all_insns): ...new function.
13385 (pass_fast_rtl_byte_dce): Replace declaration with that of...
13386 (make_pass_fast_rtl_byte_dce): ...new function.
13387 (pass_lower_subreg2): Replace declaration with that of...
13388 (make_pass_lower_subreg2): ...new function.
13389 (pass_mode_switching): Replace declaration with that of...
13390 (make_pass_mode_switching): ...new function.
13391 (pass_sms): Replace declaration with that of...
13392 (make_pass_sms): ...new function.
13393 (pass_sched): Replace declaration with that of...
13394 (make_pass_sched): ...new function.
13395 (pass_ira): Replace declaration with that of...
13396 (make_pass_ira): ...new function.
13397 (pass_reload): Replace declaration with that of...
13398 (make_pass_reload): ...new function.
13399 (pass_clean_state): Replace declaration with that of...
13400 (make_pass_clean_state): ...new function.
13401 (pass_branch_prob): Replace declaration with that of...
13402 (make_pass_branch_prob): ...new function.
13403 (pass_value_profile_transformations): Replace declaration with that
13404 of...
13405 (make_pass_value_profile_transformations): ...new function.
13406 (pass_postreload_cse): Replace declaration with that of...
13407 (make_pass_postreload_cse): ...new function.
13408 (pass_gcse2): Replace declaration with that of...
13409 (make_pass_gcse2): ...new function.
13410 (pass_split_after_reload): Replace declaration with that of...
13411 (make_pass_split_after_reload): ...new function.
13412 (pass_branch_target_load_optimize1): Replace declaration with that
13413 of...
13414 (make_pass_branch_target_load_optimize1): ...new function.
13415 (pass_thread_prologue_and_epilogue): Replace declaration with that
13416 of...
13417 (make_pass_thread_prologue_and_epilogue): ...new function.
13418 (pass_stack_adjustments): Replace declaration with that of...
13419 (make_pass_stack_adjustments): ...new function.
13420 (pass_peephole2): Replace declaration with that of...
13421 (make_pass_peephole2): ...new function.
13422 (pass_if_after_reload): Replace declaration with that of...
13423 (make_pass_if_after_reload): ...new function.
13424 (pass_regrename): Replace declaration with that of...
13425 (make_pass_regrename): ...new function.
13426 (pass_cprop_hardreg): Replace declaration with that of...
13427 (make_pass_cprop_hardreg): ...new function.
13428 (pass_reorder_blocks): Replace declaration with that of...
13429 (make_pass_reorder_blocks): ...new function.
13430 (pass_branch_target_load_optimize2): Replace declaration with that
13431 of...
13432 (make_pass_branch_target_load_optimize2): ...new function.
13433 (pass_leaf_regs): Replace declaration with that of...
13434 (make_pass_leaf_regs): ...new function.
13435 (pass_split_before_sched2): Replace declaration with that of...
13436 (make_pass_split_before_sched2): ...new function.
13437 (pass_compare_elim_after_reload): Replace declaration with that of...
13438 (make_pass_compare_elim_after_reload): ...new function.
13439 (pass_sched2): Replace declaration with that of...
13440 (make_pass_sched2): ...new function.
13441 (pass_stack_regs): Replace declaration with that of...
13442 (make_pass_stack_regs): ...new function.
13443 (pass_stack_regs_run): Replace declaration with that of...
13444 (make_pass_stack_regs_run): ...new function.
13445 (pass_df_finish): Replace declaration with that of...
13446 (make_pass_df_finish): ...new function.
13447 (pass_compute_alignments): Replace declaration with that of...
13448 (make_pass_compute_alignments): ...new function.
13449 (pass_duplicate_computed_gotos): Replace declaration with that of...
13450 (make_pass_duplicate_computed_gotos): ...new function.
13451 (pass_variable_tracking): Replace declaration with that of...
13452 (make_pass_variable_tracking): ...new function.
13453 (pass_free_cfg): Replace declaration with that of...
13454 (make_pass_free_cfg): ...new function.
13455 (pass_machine_reorg): Replace declaration with that of...
13456 (make_pass_machine_reorg): ...new function.
13457 (pass_cleanup_barriers): Replace declaration with that of...
13458 (make_pass_cleanup_barriers): ...new function.
13459 (pass_delay_slots): Replace declaration with that of...
13460 (make_pass_delay_slots): ...new function.
13461 (pass_split_for_shorten_branches): Replace declaration with that of...
13462 (make_pass_split_for_shorten_branches): ...new function.
13463 (pass_split_before_regstack): Replace declaration with that of...
13464 (make_pass_split_before_regstack): ...new function.
13465 (pass_convert_to_eh_region_ranges): Replace declaration with that
13466 of...
13467 (make_pass_convert_to_eh_region_ranges): ...new function.
13468 (pass_shorten_branches): Replace declaration with that of...
13469 (make_pass_shorten_branches): ...new function.
13470 (pass_set_nothrow_function_flags): Replace declaration with that of...
13471 (make_pass_set_nothrow_function_flags): ...new function.
13472 (pass_dwarf2_frame): Replace declaration with that of...
13473 (make_pass_dwarf2_frame): ...new function.
13474 (pass_final): Replace declaration with that of...
13475 (make_pass_final): ...new function.
13476 (pass_rtl_seqabstr): Replace declaration with that of...
13477 (make_pass_rtl_seqabstr): ...new function.
13478 (pass_release_ssa_names): Replace declaration with that of...
13479 (make_pass_release_ssa_names): ...new function.
13480 (pass_early_inline): Replace declaration with that of...
13481 (make_pass_early_inline): ...new function.
13482 (pass_inline_parameters): Replace declaration with that of...
13483 (make_pass_inline_parameters): ...new function.
13484 (pass_update_address_taken): Replace declaration with that of...
13485 (make_pass_update_address_taken): ...new function.
13486 (pass_convert_switch): Replace declaration with that of...
13487 (make_pass_convert_switch): ...new function.
13488 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
13489 to a subclass of simple_ipa_opt_pass along with...
13490 (pass_data_ipa_tree_profile): ...new pass_data instance and...
13491 (make_pass_ipa_tree_profile): ...new function.
13492 * tree-sra.c (pass_sra_early): Convert from a global struct to a
13493 subclass of gimple_opt_pass along with...
13494 (pass_data_sra_early): ...new pass_data instance and...
13495 (make_pass_sra_early): ...new function.
13496 (pass_sra): Convert from a global struct to a subclass of
13497 gimple_opt_pass along with...
13498 (pass_data_sra): ...new pass_data instance and...
13499 (make_pass_sra): ...new function.
13500 (pass_early_ipa_sra): Convert from a global struct to a subclass of
13501 gimple_opt_pass along with...
13502 (pass_data_early_ipa_sra): ...new pass_data instance and...
13503 (make_pass_early_ipa_sra): ...new function.
13504 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
13505 subclass of gimple_opt_pass along with...
13506 (pass_data_ccp): ...new pass_data instance and...
13507 (make_pass_ccp): ...new function.
13508 (pass_fold_builtins): Convert from a global struct to a subclass of
13509 gimple_opt_pass along with...
13510 (pass_data_fold_builtins): ...new pass_data instance and...
13511 (make_pass_fold_builtins): ...new function.
13512 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
13513 subclass of gimple_opt_pass along with...
13514 (pass_data_copy_prop): ...new pass_data instance and...
13515 (make_pass_copy_prop): ...new function.
13516 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
13517 global struct to a subclass of gimple_opt_pass along with...
13518 (pass_data_rename_ssa_copies): ...new pass_data instance and...
13519 (make_pass_rename_ssa_copies): ...new function.
13520 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
13521 subclass of gimple_opt_pass along with...
13522 (pass_data_dce): ...new pass_data instance and...
13523 (make_pass_dce): ...new function.
13524 (pass_dce_loop): Convert from a global struct to a subclass of
13525 gimple_opt_pass along with...
13526 (pass_data_dce_loop): ...new pass_data instance and...
13527 (make_pass_dce_loop): ...new function.
13528 (pass_cd_dce): Convert from a global struct to a subclass of
13529 gimple_opt_pass along with...
13530 (pass_data_cd_dce): ...new pass_data instance and...
13531 (make_pass_cd_dce): ...new function.
13532 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
13533 subclass of gimple_opt_pass along with...
13534 (pass_data_dominator): ...new pass_data instance and...
13535 (make_pass_dominator): ...new function.
13536 (pass_phi_only_cprop): Convert from a global struct to a subclass of
13537 gimple_opt_pass along with...
13538 (pass_data_phi_only_cprop): ...new pass_data instance and...
13539 (make_pass_phi_only_cprop): ...new function.
13540 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
13541 subclass of gimple_opt_pass along with...
13542 (pass_data_dse): ...new pass_data instance and...
13543 (make_pass_dse): ...new function.
13544 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
13545 a subclass of gimple_opt_pass along with...
13546 (pass_data_forwprop): ...new pass_data instance and...
13547 (make_pass_forwprop): ...new function.
13548 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
13549 struct to a subclass of gimple_opt_pass along with...
13550 (pass_data_tree_ifcombine): ...new pass_data instance and...
13551 (make_pass_tree_ifcombine): ...new function.
13552 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
13553 subclass of gimple_opt_pass along with...
13554 (pass_data_ch): ...new pass_data instance and...
13555 (make_pass_ch): ...new function.
13556 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
13557 subclass of gimple_opt_pass along with...
13558 (pass_data_tree_loop): ...new pass_data instance and...
13559 (make_pass_tree_loop): ...new function.
13560 (pass_tree_loop_init): Convert from a global struct to a subclass of
13561 gimple_opt_pass along with...
13562 (pass_data_tree_loop_init): ...new pass_data instance and...
13563 (make_pass_tree_loop_init): ...new function.
13564 (pass_lim): Convert from a global struct to a subclass of
13565 gimple_opt_pass along with...
13566 (pass_data_lim): ...new pass_data instance and...
13567 (make_pass_lim): ...new function.
13568 (pass_tree_unswitch): Convert from a global struct to a subclass of
13569 gimple_opt_pass along with...
13570 (pass_data_tree_unswitch): ...new pass_data instance and...
13571 (make_pass_tree_unswitch): ...new function.
13572 (pass_predcom): Convert from a global struct to a subclass of
13573 gimple_opt_pass along with...
13574 (pass_data_predcom): ...new pass_data instance and...
13575 (make_pass_predcom): ...new function.
13576 (pass_vectorize): Convert from a global struct to a subclass of
13577 gimple_opt_pass along with...
13578 (pass_data_vectorize): ...new pass_data instance and...
13579 (make_pass_vectorize): ...new function.
13580 (pass_graphite): Convert from a global struct to a subclass of
13581 gimple_opt_pass along with...
13582 (pass_data_graphite): ...new pass_data instance and...
13583 (make_pass_graphite): ...new function.
13584 (pass_graphite_transforms): Convert from a global struct to a subclass
13585 of gimple_opt_pass along with...
13586 (pass_data_graphite_transforms): ...new pass_data instance and...
13587 (make_pass_graphite_transforms): ...new function.
13588 (pass_check_data_deps): Convert from a global struct to a subclass of
13589 gimple_opt_pass along with...
13590 (pass_data_check_data_deps): ...new pass_data instance and...
13591 (make_pass_check_data_deps): ...new function.
13592 (pass_iv_canon): Convert from a global struct to a subclass of
13593 gimple_opt_pass along with...
13594 (pass_data_iv_canon): ...new pass_data instance and...
13595 (make_pass_iv_canon): ...new function.
13596 (pass_scev_cprop): Convert from a global struct to a subclass of
13597 gimple_opt_pass along with...
13598 (pass_data_scev_cprop): ...new pass_data instance and...
13599 (make_pass_scev_cprop): ...new function.
13600 (pass_record_bounds): Convert from a global struct to a subclass of
13601 gimple_opt_pass along with...
13602 (pass_data_record_bounds): ...new pass_data instance and...
13603 (make_pass_record_bounds): ...new function.
13604 (pass_complete_unroll): Convert from a global struct to a subclass of
13605 gimple_opt_pass along with...
13606 (pass_data_complete_unroll): ...new pass_data instance and...
13607 (make_pass_complete_unroll): ...new function.
13608 (pass_complete_unrolli): Convert from a global struct to a subclass of
13609 gimple_opt_pass along with...
13610 (pass_data_complete_unrolli): ...new pass_data instance and...
13611 (make_pass_complete_unrolli): ...new function.
13612 (pass_parallelize_loops): Convert from a global struct to a subclass
13613 of gimple_opt_pass along with...
13614 (pass_data_parallelize_loops): ...new pass_data instance and...
13615 (make_pass_parallelize_loops): ...new function.
13616 (pass_loop_prefetch): Convert from a global struct to a subclass of
13617 gimple_opt_pass along with...
13618 (pass_data_loop_prefetch): ...new pass_data instance and...
13619 (make_pass_loop_prefetch): ...new function.
13620 (pass_iv_optimize): Convert from a global struct to a subclass of
13621 gimple_opt_pass along with...
13622 (pass_data_iv_optimize): ...new pass_data instance and...
13623 (make_pass_iv_optimize): ...new function.
13624 (pass_tree_loop_done): Convert from a global struct to a subclass of
13625 gimple_opt_pass along with...
13626 (pass_data_tree_loop_done): ...new pass_data instance and...
13627 (make_pass_tree_loop_done): ...new function.
13628 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
13629 struct to a subclass of gimple_opt_pass along with...
13630 (pass_data_cse_reciprocals): ...new pass_data instance and...
13631 (make_pass_cse_reciprocals): ...new function.
13632 (pass_cse_sincos): Convert from a global struct to a subclass of
13633 gimple_opt_pass along with...
13634 (pass_data_cse_sincos): ...new pass_data instance and...
13635 (make_pass_cse_sincos): ...new function.
13636 (pass_optimize_bswap): Convert from a global struct to a subclass of
13637 gimple_opt_pass along with...
13638 (pass_data_optimize_bswap): ...new pass_data instance and...
13639 (make_pass_optimize_bswap): ...new function.
13640 (pass_optimize_widening_mul): Convert from a global struct to a
13641 subclass of gimple_opt_pass along with...
13642 (pass_data_optimize_widening_mul): ...new pass_data instance and...
13643 (make_pass_optimize_widening_mul): ...new function.
13644 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
13645 subclass of gimple_opt_pass along with...
13646 (pass_data_phiopt): ...new pass_data instance and...
13647 (make_pass_phiopt): ...new function.
13648 (pass_cselim): Convert from a global struct to a subclass of
13649 gimple_opt_pass along with...
13650 (pass_data_cselim): ...new pass_data instance and...
13651 (make_pass_cselim): ...new function.
13652 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
13653 subclass of gimple_opt_pass along with...
13654 (pass_data_phiprop): ...new pass_data instance and...
13655 (make_pass_phiprop): ...new function.
13656 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
13657 subclass of gimple_opt_pass along with...
13658 (pass_data_pre): ...new pass_data instance and...
13659 (make_pass_pre): ...new function.
13660 (pass_fre): Convert from a global struct to a subclass of
13661 gimple_opt_pass along with...
13662 (pass_data_fre): ...new pass_data instance and...
13663 (make_pass_fre): ...new function.
13664 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
13665 subclass of gimple_opt_pass along with...
13666 (pass_data_reassoc): ...new pass_data instance and...
13667 (make_pass_reassoc): ...new function.
13668 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
13669 subclass of gimple_opt_pass along with...
13670 (pass_data_sink_code): ...new pass_data instance and...
13671 (make_pass_sink_code): ...new function.
13672 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
13673 subclass of gimple_opt_pass along with...
13674 (pass_data_strlen): ...new pass_data instance and...
13675 (make_pass_strlen): ...new function.
13676 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
13677 struct to a subclass of gimple_opt_pass along with...
13678 (pass_data_build_alias): ...new pass_data instance and...
13679 (make_pass_build_alias): ...new function.
13680 (pass_build_ealias): Convert from a global struct to a subclass of
13681 gimple_opt_pass along with...
13682 (pass_data_build_ealias): ...new pass_data instance and...
13683 (make_pass_build_ealias): ...new function.
13684 (pass_ipa_pta): Convert from a global struct to a subclass of
13685 simple_ipa_opt_pass along with...
13686 (pass_data_ipa_pta): ...new pass_data instance and...
13687 (make_pass_ipa_pta): ...new function.
13688 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
13689 subclass of gimple_opt_pass along with...
13690 (pass_data_uncprop): ...new pass_data instance and...
13691 (make_pass_uncprop): ...new function.
13692 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
13693 global struct to a subclass of gimple_opt_pass along with...
13694 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
13695 (make_pass_late_warn_uninitialized): ...new function.
13696 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
13697 to a subclass of gimple_opt_pass along with...
13698 (pass_data_init_datastructures): ...new pass_data instance and...
13699 (make_pass_init_datastructures): ...new function.
13700 (pass_early_warn_uninitialized): Convert from a global struct to a
13701 subclass of gimple_opt_pass along with...
13702 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
13703 (make_pass_early_warn_uninitialized): ...new function.
13704 (pass_update_address_taken): Convert from a global struct to a
13705 subclass of gimple_opt_pass along with...
13706 (pass_data_update_address_taken): ...new pass_data instance and...
13707 (make_pass_update_address_taken): ...new function.
13708 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
13709 struct to a subclass of gimple_opt_pass along with...
13710 (pass_data_release_ssa_names): ...new pass_data instance and...
13711 (make_pass_release_ssa_names): ...new function.
13712 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
13713 subclass of gimple_opt_pass along with...
13714 (pass_data_stdarg): ...new pass_data instance and...
13715 (make_pass_stdarg): ...new function.
13716 * tree-switch-conversion.c (pass_convert_switch): Convert from a
13717 global struct to a subclass of gimple_opt_pass along with...
13718 (pass_data_convert_switch): ...new pass_data instance and...
13719 (make_pass_convert_switch): ...new function.
13720 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
13721 to a subclass of gimple_opt_pass along with...
13722 (pass_data_tail_recursion): ...new pass_data instance and...
13723 (make_pass_tail_recursion): ...new function.
13724 (pass_tail_calls): Convert from a global struct to a subclass of
13725 gimple_opt_pass along with...
13726 (pass_data_tail_calls): ...new pass_data instance and...
13727 (make_pass_tail_calls): ...new function.
13728 * tree-vect-generic.c (pass_lower_vector): Convert from a global
13729 struct to a subclass of gimple_opt_pass along with...
13730 (pass_data_lower_vector): ...new pass_data instance and...
13731 (make_pass_lower_vector): ...new function.
13732 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
13733 gimple_opt_pass along with...
13734 (pass_data_lower_vector_ssa): ...new pass_data instance and...
13735 (make_pass_lower_vector_ssa): ...new function.
13736 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
13737 to a subclass of gimple_opt_pass along with...
13738 (pass_data_slp_vectorize): ...new pass_data instance and...
13739 (make_pass_slp_vectorize): ...new function.
13740 (pass_ipa_increase_alignment): Convert from a global struct to a
13741 subclass of simple_ipa_opt_pass along with...
13742 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
13743 (make_pass_ipa_increase_alignment): ...new function.
13744 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
13745 gimple_opt_pass along with...
13746 (pass_data_vrp): ...new pass_data instance and...
13747 (make_pass_vrp): ...new function.
13748 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
13749 subclass of simple_ipa_opt_pass along with...
13750 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
13751 (make_pass_ipa_free_lang_data): ...new function.
13752 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
13753 gimple_opt_pass along with...
13754 (pass_data_tsan): ...new pass_data instance and...
13755 (make_pass_tsan): ...new function.
13756 (pass_tsan_O0): Convert from a global struct to a subclass of
13757 gimple_opt_pass along with...
13758 (pass_data_tsan_O0): ...new pass_data instance and...
13759 (make_pass_tsan_O0): ...new function.
13760 * var-tracking.c (pass_variable_tracking): Convert from a global
13761 struct to a subclass of rtl_opt_pass along with...
13762 (pass_data_variable_tracking): ...new pass_data instance and...
13763 (make_pass_variable_tracking): ...new function.
13764 * web.c (pass_web): Convert from a global struct to a subclass of
13765 rtl_opt_pass along with...
13766 (pass_data_web): ...new pass_data instance and...
13767 (make_pass_web): ...new function.
13768 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
13769 declaration with that of...
13770 (make_pass_mode_switch_use): ...new function.
13771 (pass_resolve_sw_modes): Replace declaration with that of...
13772 (make_pass_resolve_sw_modes): ...new function.
13773 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
13774 from a global struct to a subclass of rtl_opt_pass along with...
13775 (pass_data_mode_switch_use): ...new pass_data instance and...
13776 (make_pass_mode_switch_use): ...new function.
13777 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
13778 from a global struct to a subclass of rtl_opt_pass along with...
13779 (pass_data_resolve_sw_modes): ...new pass_data instance and...
13780 (make_pass_resolve_sw_modes): ...new function.
13781 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
13782 struct to a subclass of rtl_opt_pass along with...
13783 (pass_data_insert_vzeroupper): ...new pass_data instance and...
13784 (make_pass_insert_vzeroupper): ...new function.
13785 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
13786 global struct to a subclass of rtl_opt_pass along with...
13787 (pass_data_work_around_errata): ...new pass_data instance and...
13788 (make_pass_work_around_errata): ...new function.
13789 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
13790 struct to a subclass of rtl_opt_pass along with...
13791 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
13792 (make_pass_mips_machine_reorg2): ...new function.
13793
13794 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13795
13796 * passes.c (pass_manager::operator new): New.
13797
13798 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13799
13800 Handwritten part of conversion of passes to C++ classes.
13801
13802 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
13803 (toplev.o): Add dep on PASS_MANAGER_H.
13804 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
13805 of early local pases to reflect this moving from a global to a
13806 member of gcc::pass_manager.
13807 (cgraph_add_new_function): Likewise.
13808 * lto-cgraph.c (lto_output_node): Update for conversion of
13809 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
13810 * passes.c (opt_pass::clone): New.
13811 (opt_pass::gate): New.
13812 (opt_pass::execute): New.
13813 (opt_pass::opt_pass): New.
13814 (pass_manager::execute_early_local_passes): New.
13815 (pass_manager::execute_pass_mode_switching): new.
13816 (finish_optimization_passes): Convert to...
13817 (pass_manager::finish_optimization_passes): ...this.
13818 (finish_optimization_passes): Update for conversion of passes to
13819 C++ classes.
13820 (register_dump_files_1): Use has_gate since we cannot portably
13821 check a vtable entry against NULL.
13822 (dump_one_pass): Likewise.
13823 (ipa_write_summaries_2): Likewise.
13824 (ipa_write_optimization_summaries_1): Likewise.
13825 (ipa_read_summaries_1): Likewise.
13826 (ipa_read_optimization_summaries_1): Likewise.
13827 (execute_ipa_stmt_fixups): Likewise.
13828 (pass_manager::pass_manager): Rewrite pass-creation, invoking
13829 pass-creation functions rather than wiring up globals, and
13830 storing the results in fields of pass_manager generated using
13831 pass-instances.def.
13832 (pass_manager::dump_profile_report): Update for conversion of
13833 passes to C++ classes.
13834 (pass_manager::execute_ipa_summary_passes): Likewise.
13835 (execute_one_ipa_transform_pass): Likewise.
13836 (execute_one_pass): Use has_gate and has_execute since we cannot
13837 portably check a vtable entry against NULL.
13838 * pass_manager.h (pass_manager::finish_optimization_passes): New.
13839 (pass_manager): Use pass-instances.def to add fields for the
13840 various pass instances.
13841 * toplev.c (finalize): Update for move of
13842 finish_optimization_passes to a method of gcc::pass_manager.
13843 * toplev.h (finish_optimization_passes): Move to method of class
13844 pass_manager.
13845 * tree-pass.h (struct pass_data): New.
13846 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
13847 (opt_pass::gate): Convert to virtual function.
13848 (opt_pass::~opt_pass): New.
13849 (opt_pass::clone): New.
13850 (opt_pass::execute): Convert to virtual function.
13851 (opt_pass::opt_pass): New.
13852 (opt_pass::ctxt_): new.
13853 (gimple_opt_pass): Convert to subclass of opt_pass.
13854 (gimple_opt_pass::gimple_opt_pass): New.
13855 (rtl_opt_pass): Convert to subclass of opt_pass.
13856 (rtl_opt_pass::rtl_opt_pass): New.
13857 (ipa_opt_pass_d): Convert to subclass of opt_pass.
13858 (ipa_opt_pass_d::ipa_opt_pass_d): New.
13859 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
13860 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
13861 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
13862 invocation of pass_mode_switching to reflect this moving from a
13863 global to a member of gcc::pass_manager.
13864 (ix86_option_override): Rework how pass_insert_vzeroupper is
13865 added to the pass_manager to reflect autogenerated changes.
13866 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
13867
13868 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
13869
13870 PR rtl-optimization/57708
13871 * recog.c (peep2_find_free_register): Validate all regs in a
13872 multi-reg mode.
13873
13874 2013-08-05 Jan Hubicka <jh@suse.cz>
13875
13876 PR lto/57602
13877 * cgraph.c (verify_cgraph_node): Accept local flags from other
13878 partitions.
13879 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
13880 (function_and_variable_visibility): Likewise.
13881 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
13882
13883 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13884
13885 * graph.c (init_graph_slim_pretty_print): Remove.
13886 (print_graph_cfg): Do not call it. Use local pretty printer.
13887 (start_graph_dump): Likewise.
13888
13889 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13890
13891 * gimple-pretty-print.c (buffer): Remove.
13892 (initialized): Likewise.
13893 (maybe_init_pretty_print): Likewise.
13894 (print_gimple_stmt): Do not call it. Use non-static local
13895 pretty_printer variable.
13896 (print_gimple_expr): Likewise.
13897 (print_gimple_seq): Likewise.
13898 (gimple_dump_bb): Likewise.
13899
13900 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13901
13902 * asan.c (asan_pp): Remove.
13903 (asan_pp_initialized): Likewise.
13904 (asan_pp_initialize): Likewise.
13905 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
13906 (asan_emit_stack_protection): Tidy. Use local pretty printer.
13907 (asan_add_global): Likewise.
13908
13909 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
13910
13911 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
13912 * diagnostic.h (diagnostic_flush_buffer): Adjust.
13913 * pretty-print.c (pp_formatted_text_data): Likewise.
13914 (pp_indent): Rename from pp_base_indent.
13915 (pp_format): Rename from pp_base_format.
13916 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
13917 (pp_format_verbatim): Rename from pp_base_format_verbatim.
13918 (pp_flush): Rename from pp_base_flush.
13919 (pp_set_line_maximum_length): Rename from
13920 pp_base_set_line_maximum_length.
13921 (pp_clear_output_area): Rename from pp_base_clear_output_area.
13922 (pp_set_prefix): Rename from pp_base_set_prefix.
13923 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
13924 (pp_emit_prefix): Rename from pp_base_emit_prefix.
13925 (pp_append_text): Rename from pp_base_append_text.
13926 (pp_formatted_text): Rename from pp_base_formatted_text.
13927 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
13928 (pp_remaining_character_count_for_line): Rename from
13929 pp_base_remaining_character_count_for_line.
13930 (pp_newline): Rename from pp_base_newline.
13931 (pp_character): Rename from pp_base_character.
13932 (pp_string): Rename from pp_base_string.
13933 (pp_maybe_space): Rename from pp_base_maybe_space.
13934 * asan.c (asan_pp_string): Adjust.
13935 (asan_emit_stack_protection): Likewise.
13936 (asan_add_global): Likewise.
13937 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
13938 * tree-mudflap.c (mf_varname_tree): Likewise.
13939 * tree-pretty-print.c (pp_tree_identifier): Rename from
13940 pp_base_tree_identifier.
13941 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
13942 Declare as function.
13943
13944 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13945
13946 * pretty-print.h (pp_bar_bar): New.
13947 (pp_ampersand_ampersand): Likewise.
13948 (pp_less_equal): Likewise.
13949 (pp_greater_equal): Likewise.
13950 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
13951 printer functions instead of pp_string or operators and punctuators.
13952 (dump_gimple_call): Likewise.
13953 (dump_gimple_omp_for): Likewise.
13954 (dump_gimple_transaction): Likewise.
13955 (dump_gimple_phi): Likewise.
13956 (pp_gimple_stmt_1): Likewise.
13957 * sched-vis.c (print_insn): Likewise.
13958 * tree-mudflap.c (mf_varname_tree): Likewise.
13959 * tree-pretty-print.c (dump_block_node): Likewise.
13960 (dump_generic_node): Likewise.
13961
13962 2013-08-02 Jan Hubicka <jh@suse.cz>
13963
13964 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
13965 boundaries.
13966 * lto-streamer-out.c (tree_is_indexable): Results decls and
13967 parm decls are not indexable.
13968 (DFS_write_tree_body): Do not follow args and results.
13969 (hash_tree): Likewise.
13970 (output_functions): Rearrange so struct function is needed
13971 only when real body is output; be able to also ouptut abstract
13972 functions; output DECL_ARGUMENTS and DECL_RESULT.
13973 (lto_output): When not in WPA, ale store abstract functions.
13974 (write_symbol): Do not care about RESULT_DECL.
13975 (output_symbol_p): Handle correctly sbtract decls.
13976 * lto-streamer-in.c (input_function): Rearrange so struct
13977 function can be NULL at entry; allow streaming of
13978 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
13979 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
13980 sanity check during LTO.
13981 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
13982 RESULT_DECl and DECL_ARGUMENTS.
13983 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
13984 Likewise.
13985
13986 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13987
13988 * pretty-print.h (pp_underscore): New.
13989 (pp_comma): Tidy.
13990 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
13991 printer functions instead of pp_character.
13992 (dump_binary_rhs): Likewise.
13993 (dump_ternary_rhs): Likewise.
13994 (dump_gimple_call_args): Likewise.
13995 (pp_points_to_solution): Likewise.
13996 (dump_gimple_call): Likewise.
13997 (dump_gimple_switch): Likewise.
13998 (dump_gimple_cond): Likewise.
13999 (dump_gimple_bind): Likewise.
14000 (dump_gimple_try): Likewise.
14001 (dump_gimple_omp_for): Likewise.
14002 (dump_gimple_omp_continue): Likewise.
14003 (dump_gimple_omp_single): Likewise.
14004 (dump_gimple_omp_sections): Likewise.
14005 (dump_gimple_omp_block): Likewise.
14006 (dump_gimple_omp_critical): Likewise.
14007 (dump_gimple_transaction): Likewise.
14008 (dump_gimple_asm): Likewise.
14009 (dump_gimple_phi): Likewise.
14010 (dump_gimple_omp_parallel): Likewise.
14011 (dump_gimple_omp_task): Likewise.
14012 (dump_gimple_omp_atomic_load): Likewise.
14013 (dump_gimple_omp_atomic_store): Likewise.
14014 (dump_gimple_mem_ops): Likewise.
14015 (pp_gimple_stmt_1): Likewise.
14016 (pp_cfg_jump): Likewise.
14017 (dump_implicit_edges): Likewise.
14018 (gimple_dump_bb_for_graph): Likewise.
14019 * graph.c (draw_cfg_node): Likewise.
14020 * langhooks.c (lhd_print_error_function): Likewise.
14021 * sched-vis.c (print_exp): Likewise.
14022 (print_value): Likewise.
14023 (print_pattern): Likewise.
14024 (print_insn): Likewise.
14025 (rtl_dump_bb_for_graph): Likewise.
14026 * tree-pretty-print.c (dump_function_declaration): Likewise.
14027 (dump_array_domain): Likewise.
14028 (dump_omp_clause): Likewise.
14029 (dump_location): Likewise.
14030 (dump_generic_node): Likewise.
14031 (print_struct_decl): Likewise.
14032 * diagnostic.c (diagnostic_show_locus): Use pp_space.
14033
14034 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14035
14036 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
14037 candidate table when replacing a candidate statement.
14038 (replace_rhs_if_not_dup): Likewise.
14039 (replace_one_candidate): Likewise.
14040
14041 2013-08-02 Jan Hubicka <jh@suse.cz>
14042 Martin Liska <marxin.liska@gmail.com>
14043
14044 * cgraphunit.c (add_new_function): Fix logic when adding from
14045 late IPA pass.
14046 (assemble_thunk): Rename to ...
14047 (expand_thunk); .. this one; export; get it working with
14048 general functions; make produced gimple valid.
14049 * cgraph.h (expand_thunk): Declare.
14050
14051 2013-08-02 Jan Hubicka <jh@suse.cz>
14052
14053 * ipa-cp.c (gather_context_independent_values): Use
14054 ipa_get_param_move_cost.
14055 (get_replacement_map): Remove PARAM; move parameter folding
14056 into tree-inline.c
14057 (create_specialized_node): Update.
14058 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
14059 assert that we have gimple body; update move_cost.
14060 (count_formal_params): Assert that we have gimple body.
14061 (ipa_dump_param): New function.
14062 (ipa_alloc_node_params): Break out from ...
14063 (ipa_initialize_node_params): ... here.
14064 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
14065 (ipa_write_node_info): Stream move costs.
14066 (ipa_read_node_info): Read move costs.
14067 (ipa_update_after_lto_read): Do not recompute node params.
14068 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
14069 (ipa_get_param): Check we are not in WPA.
14070 (ipa_get_param_move_cost): New.
14071 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
14072 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
14073 parm numbers to be present.
14074
14075 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14076
14077 PR rtl-optimization/58048
14078 * lra-constraints.c (process_alt_operands): Don't check asm
14079 operand on register.
14080
14081 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
14082
14083 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
14084 the implied StoreLoad barrier for atomic operations if before.
14085
14086 2013-08-02 Jan Hubicka <jh@suse.cz>
14087 Martin Liska <marxin.liska@gmail.com>
14088
14089 * cgraph.c (cgraph_function_body_availability): Do not check
14090 cgraph flags.
14091 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
14092 symtab_node_availability): Declare.
14093 * ipa.c (can_replace_by_local_alias): New.
14094 (function_and_variable_visibility): Use it.
14095 * symtab.c (symtab_for_node_and_aliases,
14096 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
14097
14098 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14099
14100 PR rtl-optimization/57963
14101 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
14102 (lra_constraints): Use them.
14103
14104 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
14105
14106 * config/arm/types.md (define_attr "type"): Add "load_acq"
14107 and "store_rel".
14108 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
14109 changes.
14110 (cortex_a53_store1): Likewise.
14111
14112 2013-08-01 Jan Hubicka <jh@suse.cz>
14113
14114 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
14115 partitions are not needed.
14116
14117 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
14118
14119 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
14120 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
14121 MAYBE_NON_Q_CLASS_P where appropriate.
14122
14123 2013-08-01 Jan Hubicka <jh@suse.cz>
14124
14125 * cgraph.h (release_function_body): Declare.
14126 * tree.c (free_lang_data_in_decl): Free, parameters and return values
14127 of unused delcarations.
14128
14129 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14130
14131 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
14132 RTL form when subtracting a constant.
14133
14134 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14135
14136 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
14137 Generate canonical plus rtx with negated immediate instead of minus
14138 where appropriate.
14139 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
14140
14141 2013-08-01 Jan Hubicka <jh@suse.cz>
14142
14143 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
14144 (cgraph_release_function_body): Likewise.
14145 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
14146 * cgraph.h (cgrpah_node): Rename abstract_and_needed
14147 to used_as_abstract_origin.
14148 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
14149 symbols used as abstract origins.
14150 * cgraphunit.c (analyze_functions): Update.
14151 * ipa.c (symtab_remove_unreachable_nodes): Recompute
14152 used_as_abstract_origin.
14153 * tree-inline.c (tree_function_versioning): Update
14154 used_as_abstract_origin; be ready for DECL_RESULT and
14155 DECL_ARGUMENTS to be NULL.
14156
14157 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
14158 for abstract functions.
14159 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
14160 real symbols.
14161
14162 2013-08-01 Jan Hubicka <jh@suse.cz>
14163
14164 * profile.c (compute_value_histograms): Fix thinko.
14165
14166 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14167
14168 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
14169 aarch-common-protos.h to extra_headers.
14170 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14171 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
14172 * config/aarch64/t-aarch64 (aarch-common.o): Define.
14173
14174 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14175
14176 * config/aarch64/aarch64.md (define_attr "type"): Delete.
14177 Include "../arm/types.md". Define "type" attribute for all patterns.
14178 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
14179 attribute changes.
14180
14181 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
14182
14183 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
14184 to support power8 load fusion.
14185 (fusion_gpr_mem_load): Likewise.
14186
14187 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
14188
14189 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
14190 declarations for power8 load fusion.
14191 (emit_fusion_gpr_load): Likewise.
14192
14193 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
14194 tuning for power8, turn on fusion mode by default. Turn on sign
14195 extending fusion mode if normal fusion mode is on, and we are at
14196 -O2 or -O3.
14197 (fusion_gpr_load_p): New function, return true if we can fuse an
14198 addis instruction with a dependent load to a GPR.
14199 (emit_fusion_gpr_load): Emit the instructions for power8 load
14200 fusion to GPRs.
14201
14202 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
14203 (VSX load fusion peepholes): New peepholes to fuse together an
14204 addi instruction with a VSX load instruction.
14205
14206 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
14207 peepholes to fuse an addis instruction with a load to a GPR base
14208 register. If we are supporting sign extending fusions, convert
14209 sign extending loads to zero extending loads and add an explicit
14210 sign extension.
14211
14212 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14213
14214 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
14215 aarch-common-protos.h to extra_headers.
14216 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14217 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
14218 (arm_early_store_addr_dep): Likewise.
14219 (arm_no_early_alu_shift_dep): Likewise.
14220 (arm_no_early_alu_shift_value_dep): Likewise.
14221 (arm_no_early_mul_dep): Likewise.
14222 (arm_no_early_store_addr_dep): Likewise.
14223 (arm_mac_accumulator_is_mul_result): Likewise.
14224 (arm_mac_accumulator_is_result): Likewise.
14225 * config/arm/aarch-common.c: ... here. New file.
14226 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
14227 here to ...
14228 (arm_early_store_addr_dep): Likewise.
14229 (arm_no_early_alu_shift_dep): Likewise.
14230 (arm_no_early_alu_shift_value_dep): Likewise.
14231 (arm_no_early_mul_dep): Likewise.
14232 (arm_no_early_store_addr_dep): Likewise.
14233 (arm_mac_accumulator_is_mul_result): Likewise.
14234 (arm_mac_accumulator_is_result): Likewise.
14235 * config/arm/aarch-common-protos.h: ... here. New file.
14236 * config/arm/t-arm (aarch-common.o): Define.
14237
14238 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14239
14240 * config/arm/arm.md: Include new file "types.md".
14241 (define_attr "type"): Move from here to ...
14242 (define_attr "mul32"): Likewise.
14243 (define_attr "mul64"): Likewise.
14244 * config/arm/types.md: ... here. New file.
14245
14246 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
14247
14248 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
14249 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
14250
14251 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
14252
14253 * gen-pass-instances.awk: Fix offset of substr().
14254
14255 2013-07-31 David Malcolm <dmalcolm@redhat.com>
14256
14257 * Makefile.in (pass-instances.def): New.
14258 (passes.o): Replace dependency on passes.def with one on
14259 pass-instances.def
14260
14261 * gen-pass-instances.awk: New.
14262
14263 * passes.c (pass_manager::pass_manager): Use pass-instances.def
14264 rather than passes.def, updating local definition of NEXT_PASS
14265 macro to add an extra NUM parameter (currently unused).
14266
14267 2013-07-30 David Malcolm <dmalcolm@redhat.com>
14268
14269 * Makefile.in (PASS_MANAGER_H): New.
14270 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
14271 (passes.o): Likewise.
14272 (statistics.o): Likewise.
14273 (cgraphunit.o): Likewise.
14274 (context.o): Depend on PASS_MANAGER_H.
14275
14276 * pass_manager.h: New.
14277
14278 * cgraphunit.c (cgraph_add_new_function): Update for moves
14279 of globals to fields of pass_manager.
14280 (analyze_function): Likewise.
14281 (expand_function): Likewise.
14282 (ipa_passes): Likewise.
14283 (compile): Likewise.
14284
14285 * context.c (context::context): New.
14286 * context.h (context::context): New.
14287 (context::get_passes): New.
14288 (context::passes_): New.
14289
14290 * lto-cgraph.c (input_node): Update for moves of globals to
14291 fields of pass_manager.
14292
14293 * passes.c (all_passes): Remove, in favor of a field of the
14294 same name within the new class pass_manager.
14295 (all_small_ipa_passes): Likewise.
14296 (all_lowering_passes): Likewise.
14297 (all_regular_ipa_passes): Likewise.
14298 (all_late_ipa_passes): Likewise.
14299 (all_lto_gen_passes): Likewise.
14300 (passes_by_id): Likewise.
14301 (passes_by_id_size): Likewise.
14302 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14303 the new class pass_manager.
14304 (set_pass_for_id): Convert to...
14305 (pass_manager::set_pass_for_id): ...method.
14306 (get_pass_for_id): Convert to...
14307 (pass_manager::get_pass_for_id): ...method.
14308 (register_one_dump_file): Move body of implementation into...
14309 (pass_manager::register_one_dump_file): ...here.
14310 (register_dump_files_1): Convert to...
14311 (pass_manager::register_dump_files_1): ...method.
14312 (register_dump_files): Convert to...
14313 (pass_manager::register_dump_files): ...method.
14314 (create_pass_tab): Update for moves of globals to fields of
14315 pass_manager.
14316 (dump_passes): Move body of implementation into...
14317 (pass_manager::dump_passes): ...here.
14318 (register_pass): Move body of implementation into...
14319 (pass_manager::register_pass): ...here.
14320 (init_optimization_passes): Convert into...
14321 (pass_manager::pass_manager): ...constructor for new
14322 pass_manager class, and initialize the pass_lists array.
14323 (check_profile_consistency): Update for moves of globals to
14324 fields of pass_manager.
14325 (dump_profile_report): Move body of implementation into...
14326 (pass_manager::dump_profile_report): ...here.
14327 (ipa_write_summaries_1): Update for moves of pass lists from
14328 being globals to fields of pass_manager.
14329 (ipa_write_optimization_summaries): Likewise.
14330 (ipa_read_summaries): Likewise.
14331 (ipa_read_optimization_summaries): Likewise.
14332 (execute_all_ipa_stmt_fixups): Likewise.
14333
14334 * statistics.c (statistics_fini): Update for moves of globals to
14335 fields of pass_manager.
14336
14337 * toplev.c (general_init): Replace call to
14338 init_optimization_passes with construction of the pass_manager
14339 instance.
14340
14341 * tree-pass.h (all_passes): Remove, in favor of a field of the
14342 same name within the new class pass_manager.
14343 (all_small_ipa_passes): Likewise.
14344 (all_lowering_passes): Likewise.
14345 (all_regular_ipa_passes): Likewise.
14346 (all_lto_gen_passes): Likewise.
14347 (all_late_ipa_passes): Likewise.
14348 (passes_by_id): Likewise.
14349 (passes_by_id_size): Likewise.
14350 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14351 the new class pass_manager.
14352 (get_pass_for_id): Remove.
14353
14354 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14355
14356 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
14357 configs.
14358
14359 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14360
14361 * arm.md (mulhi3): New expand pattern.
14362
14363 2013-07-30 Jan Hubicka <jh@suse.cz>
14364 Martin Liska <marxin.liska@gmail.com>
14365
14366 * profile.c (compute_value_histograms): Do not ICE when
14367 there is mismatch only on some counters.
14368
14369 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14370
14371 PR rtl-optimization/57637
14372 * function.c (move_insn_for_shrink_wrap): Also check the
14373 GEN set of the LIVE problem for the liveness analysis
14374 if it exists, otherwise give up.
14375
14376 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14377
14378 PR tree-optimization/57993
14379 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
14380 replaced statement in the candidate table.
14381 (phi_add_costs): Return infinite cost when the hidden basis does
14382 not dominate all phis on which the candidate is dependent.
14383 (replace_one_candidate): Record replaced statement in the
14384 candidate table.
14385
14386 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
14387
14388 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
14389 (ashlv2si3): New expander.
14390 (*ashlv2si3_i): New define_insn_and_split.
14391 * predicates.md (float_operation): Allow patterns with three
14392 basic sub-patterns.
14393
14394 PR rtl-optimization/58021
14395 * mode-switching.c (create_pre_exit): Always split off preceding
14396 insns if we are not at the basic block head.
14397
14398 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
14399
14400 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
14401 (UCLIBC_DYNAMIC_LINKER): New macro.
14402 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
14403 `-mnan=2008'.
14404 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
14405 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
14406 `-mnan=2008'.
14407 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
14408 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
14409 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
14410 for SF and DF modes. Use ieee_quad_format for TF mode.
14411 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
14412 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
14413 (mips_option_override): Handle `-mnan=legacy'.
14414 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
14415 `-mabs=2008' and `-mnan=2008'.
14416 (OPTION_DEFAULT_SPECS): Add "nan" default.
14417 (ASM_SPEC): Handle `-mnan='.
14418 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
14419 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
14420 comment accordingly.
14421 (neg<mode>2): Likewise.
14422 * config/mips/mips.opt (mabs, mnan): New options.
14423 * doc/install.texi (Configuration): Document `--with-nan=' option.
14424 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
14425 `-mnan=' options.
14426 (MIPS Options): Document them.
14427 * config.gcc <mips*-*-*>: Handle `--with-nan='.
14428 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
14429 * configure: Regenerate.
14430 * config.in: Regenerate.
14431
14432 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14433
14434 * config/i386/i386.md (float post-reload splitters): Do not check
14435 for subregs of SSE registers.
14436
14437 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14438 H.J. Lu <hongjiu.lu@intel.com>
14439
14440 PR target/57954
14441 PR target/57988
14442 * config/i386/i386.md (post-reload splitter
14443 to avoid partial SSE reg dependency stalls): New pattern.
14444
14445 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
14446
14447 * config/s390/s390.md ("movcc"): Swap load and store instructions.
14448
14449 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
14450
14451 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14452 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
14453
14454 2013-07-26 Cary Coutant <ccoutant@google.com>
14455
14456 * dwarf2out.c (die_checksum_ordered): Don't include template
14457 instantiations in signature.
14458 (is_template_parameter): New function.
14459 (is_template_instantiation): New function.
14460 (generate_skeleton_bottom_up): Don't include template instantiations
14461 in type unit DIE.
14462 (generate_skeleton): Likewise.
14463 (break_out_comdat_types): Move recursive call to break out nested
14464 types earlier.
14465 (prune_unused_types_mark_generic_parms_dies): Call
14466 is_template_parameter.
14467
14468 2013-07-26 Ian Bolton <ian.bolton@arm.com>
14469
14470 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
14471 uses vector registers.
14472 * config/aarch64/iterators.md: Add attributes rtn and vas.
14473
14474 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14475 Richard Earnshaw <richard.earnshaw@arm.com>
14476
14477 * combine.c (simplify_comparison): Re-canonicalize operands
14478 where appropriate.
14479 * config/arm/arm.md (movcond_addsi): New splitter.
14480
14481 2013-07-25 Sterling Augustine <saugustine@google.com>
14482
14483 * dwarf2out.c (size_of_pubnames): Move code to...
14484 (include_pubname_in_output): ...here. New.
14485 (want_pubnames): Rearrange.
14486 (output_pubnames): Call include_pubname_in_output. Move assertion.
14487
14488 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14489
14490 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
14491
14492 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14493
14494 PR target/38836
14495 * doc/extend.texi: Remove obsolete builtins. Fix
14496 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
14497
14498 2013-07-25 Jan Hubicka <jh@suse.cz>
14499
14500 * cgraph.c (release_function_body): Break out from ...
14501 (cgraph_release_function_body): ... this one; also release DECL_RESULT
14502 and DECL_ARGUMENTS.
14503 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
14504 old_tree in the map.
14505 (create_specialized_node): Update.
14506 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
14507 into index.
14508 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
14509 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
14510 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
14511 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
14512 DECL_RESULT.
14513
14514 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14515
14516 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
14517 addsi3_carryin_alt2_<optab>): Correct output template.
14518
14519 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14520
14521 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
14522 Adjust for arm_restrict_it.
14523 Remove trailing whitespace.
14524
14525 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
14526
14527 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
14528 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
14529
14530 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
14531
14532 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
14533
14534 PR rtl-optimization/57960
14535 * lra-constraints.c (process_alt_operands): Use the right mode
14536 when checking strict_low.
14537
14538 2013-07-25 Jan Hubicka <jh@suse.cz>
14539
14540 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
14541 * cgraph.c (cgraph_remove_node): Do not release function body
14542 when in cgraph streaming.
14543 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
14544 in other partitions are not considered reachable; fix handling of
14545 clones.
14546
14547 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14548
14549 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
14550
14551 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14552
14553 PR target/19599
14554 PR target/57731
14555 PR target/57837
14556 * config/arm/arm.md ("*sibcall_insn): Replace use of
14557 Ss with US. Adjust output for v5 and v4t.
14558 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
14559
14560 * config/arm/constraints.md ("Ss"): Rename to US.
14561
14562 2013-07-25 Terry Guo <terry.guo@arm.com>
14563
14564 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
14565 shift_add/shift_sub0/shift_sub1 RTXs.
14566
14567 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
14568 Anton Blanchard <anton@au1.ibm.com>
14569
14570 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
14571 (altivec_vpks<VI_char>ss): Likewise.
14572 (altivec_vpks<VI_char>us): Likewise.
14573 (altivec_vpku<VI_char>us): Likewise.
14574 (altivec_vpku<VI_char>um): Likewise.
14575
14576 2013-07-24 David Malcolm <dmalcolm@redhat.com>
14577
14578 Introduce context class.
14579
14580 * Makefile.in (CONTEXT_H): New.
14581 (OBJS): Add context.o.
14582 (toplev.o): Add CONTEXT_H to dependencies.
14583 (context.o): New.
14584
14585 * toplev.c (general_init): Create the singleton gcc::context instance.
14586
14587 * context.c: New.
14588
14589 * context.h: New.
14590
14591 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
14592
14593 PR rtl-optimization/57968
14594 * mode-switching.c (create_pre_exit): Allow instructions that
14595 don't set a return register to need a non-exit mode.
14596
14597 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14598 Anton Blanchard <anton@au1.ibm.com>
14599
14600 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
14601 operands to vperm for little endian.
14602 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
14603 of lvsl to create the control mask for a vperm for little endian.
14604
14605 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14606 Anton Blanchard <anton@au1.ibm.com>
14607
14608 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
14609 two operands for little-endian.
14610
14611 2013-07-23 Steve Ellcey <sellcey@mips.com>
14612
14613 * config/mips/mips.c (mips_case_values_threshold): New.
14614 (TARGET_CASE_VALUES_THRESHOLD): Define.
14615
14616 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14617 Anton Blanchard <anton@au1.ibm.com>
14618
14619 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
14620 selection of field for vector splat in little endian mode.
14621
14622 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
14623
14624 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
14625 expanders to rs6000.md.
14626 (ior<mode>3): Likewise.
14627 (and<mode>3): Likewise.
14628 (one_cmpl<mode>2): Likewise.
14629 (nor<mode>3): Likewise.
14630 (andc<mode>3): Likewise.
14631 (eqv<mode>3): Likewise.
14632 (nand<mode>3): Likewise.
14633 (orc<mode>3): Likewise.
14634
14635 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
14636 declaration.
14637
14638 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
14639 to split multi-word logical operations.
14640 (rs6000_split_logical_di): Likewise.
14641 (rs6000_split_logical): Likewise.
14642
14643 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
14644 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
14645 and allow TImode operations in 32-bit.
14646 (vsx_and<mode>3_64bit): Likewise.
14647 (vsx_ior<mode>3_32bit): Likewise.
14648 (vsx_ior<mode>3_64bit): Likewise.
14649 (vsx_xor<mode>3_32bit): Likewise.
14650 (vsx_xor<mode>3_64bit): Likewise.
14651 (vsx_one_cmpl<mode>2_32bit): Likewise.
14652 (vsx_one_cmpl<mode>2_64bit): Likewise.
14653 (vsx_nor<mode>3_32bit): Likewise.
14654 (vsx_nor<mode>3_64bit): Likewise.
14655 (vsx_andc<mode>3_32bit): Likewise.
14656 (vsx_andc<mode>3_64bit): Likewise.
14657 (vsx_eqv<mode>3_32bit): Likewise.
14658 (vsx_eqv<mode>3_64bit): Likewise.
14659 (vsx_nand<mode>3_32bit): Likewise.
14660 (vsx_nand<mode>3_64bit): Likewise.
14661 (vsx_orc<mode>3_32bit): Likewise.
14662 (vsx_orc<mode>3_64bit): Likewise.
14663
14664 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
14665 logical types in GPRs.
14666
14667 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
14668 logical insns to rs6000.md, and allow TImode operations in
14669 32-bit.
14670 (altivec_ior<mode>3): Likewise.
14671 (altivec_xor<mode>3): Likewise.
14672 (altivec_one_cmpl<mode>2): Likewise.
14673 (altivec_nor<mode>3): Likewise.
14674 (altivec_andc<mode>3): Likewise.
14675
14676 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
14677 attributes for moving the 128-bit logical operations into
14678 rs6000.md.
14679 (BOOL_REGS_OUTPUT): Likewise.
14680 (BOOL_REGS_OP1): Likewise.
14681 (BOOL_REGS_OP2): Likewise.
14682 (BOOL_REGS_UNARY): Likewise.
14683 (BOOL_REGS_AND_CR0): Likewise.
14684 (one_cmpl<mode>2): Add support for DI logical operations on
14685 32-bit, splitting the operations to 32-bit.
14686 (anddi3): Likewise.
14687 (iordi3): Likewise.
14688 (xordi3): Likewise.
14689 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
14690 changes to combine the 32/64-bit code, allow logical operations on
14691 TI mode in 32-bit, and to use similar match_operator patterns like
14692 scalar mode uses. Combine the Altivec and VSX code for logical
14693 operations, and move it here.
14694 (ior<mode>3, 128-bit types): Likewise.
14695 (xor<mode>3, 128-bit types): Likewise.
14696 (one_cmpl<mode>3, 128-bit types): Likewise.
14697 (nor<mode>3, 128-bit types): Likewise.
14698 (andc<mode>3, 128-bit types): Likewise.
14699 (eqv<mode>3, 128-bit types): Likewise.
14700 (nand<mode>3, 128-bit types): Likewise.
14701 (orc<mode>3, 128-bit types): Likewise.
14702 (and<mode>3_internal): Likewise.
14703 (bool<mode>3_internal): Likewise.
14704 (boolc<mode>3_internal1): Likewise.
14705 (boolc<mode>3_internal2): Likewise.
14706 (boolcc<mode>3_internal1): Likewise.
14707 (boolcc<mode>3_internal2): Likewise.
14708 (eqv<mode>3_internal1): Likewise.
14709 (eqv<mode>3_internal2): Likewise.
14710 (one_cmpl1<mode>3_internal): Likewise.
14711
14712 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14713
14714 * config/microblaze/microblaze.c (microblaze_expand_prologue):
14715 Rename flag_stack_usage to flag_stack_usage_info.
14716
14717 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14718
14719 * config/microblaze/sync.md: New file.
14720 * config/microblaze/microblaze.md: Include sync.md
14721 * config/microblaze/microblaze.c: Add print_operand 'y'.
14722 * config/microblaze/constraints.md: Add memory_contraint
14723 'Q' which is a single register.
14724
14725 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
14726
14727 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
14728
14729 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14730
14731 * reload.c (find_reloads): Exit loop once we find this operand
14732 cannot be reloaded somehow for this alternative.
14733
14734 * reload.c (find_reloads): Exit loop once we find a hard register.
14735
14736 * rtlanal.c (computed_jump_p): Exit loop once we find label
14737 reference is used.
14738
14739 * i386.c (ix86_pad_returns): Exit loop after setting replace.
14740
14741 * cfgloopmanip.c (remove_path): Exit loop after setting
14742 irred_invalidated.
14743
14744 * gensupport.c (subst_dup): Avoid loop if code is not
14745 MATCH_DUP nor MATCH_OP_DUP.
14746
14747 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
14748
14749 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
14750
14751 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14752
14753 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
14754 true for SP_REGNUM if mode == ptr_mode.
14755 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
14756 with value R0_REGNUM + 31.
14757
14758 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14759
14760 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
14761 pad pointer-typed argument downward.
14762
14763 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14764
14765 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
14766 and __ILP32__ when the ILP32 model is in use.
14767
14768 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14769
14770 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
14771 (aarch64_load_symref_appropriately): In the case of
14772 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
14773 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
14774 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
14775 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
14776 if 'mode' doesn't equal to 'ptr_mode'.
14777 (aarch64_output_mi_thunk): Add an assertion on the alignment of
14778 'vcall_offset'; change to call aarch64_emit_move differently depending
14779 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
14780 to calculate the upper bound of 'vcall_offset'.
14781 (aarch64_cannot_force_const_mem): Change to also return true if
14782 mode != ptr_mode.
14783 (aarch64_legitimize_reload_address): In the case of large
14784 displacements, add new local variable 'xmode' and an assertion
14785 based on it; change to use 'xmode' to generate the new rtx and
14786 reload.
14787 (aarch64_asm_trampoline_template): Change to generate the template
14788 differently depending on TARGET_ILP32 or not; change to use
14789 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
14790 (aarch64_trampoline_size): Removed.
14791 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
14792 and replace immediate literals with it. Change to use 'ptr_mode'
14793 instead of 'DImode' and call convert_memory_address if the mode
14794 of 'fnaddr' doesn't equal to 'ptr_mode'.
14795 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
14796 to output symbol.
14797 (aarch64_elf_asm_destructor): Likewise.
14798 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
14799 on TARGET_ILP32 instead of aarch64_trampoline_size.
14800 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
14801 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
14802 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
14803 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
14804 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
14805 (storewb_pair<GPI:mode>_<P:mode>): ... this.
14806 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
14807 depending on the value of 'mode'.
14808 (add_losym_<mode>): New.
14809 (ldr_got_small_<mode>): New, based on ldr_got_small.
14810 (ldr_got_small): Remove.
14811 (ldr_got_small_sidi): New.
14812 * config/aarch64/iterators.md (P): New.
14813 (PTR): Change to 'ptr_mode' in the condition.
14814
14815 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14816
14817 * config.gcc (aarch64*-*-*): Support --with-abi.
14818 (aarch64*-*-elf): Support --with-multilib-list.
14819 (aarch64*-*-linux*): Likewise.
14820 (supported_defaults): Add abi to aarch64*-*-*.
14821 * configure.ac: Mention AArch64 for --with-multilib-list.
14822 * configure: Re-generated.
14823 * config/aarch64/biarchilp32.h: New file.
14824 * config/aarch64/biarchlp64.h: New file.
14825 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
14826 (ABI_SPEC): Ditto.
14827 (MULTILIB_DEFAULTS): Ditto.
14828 (DRIVER_SELF_SPECS): Ditto.
14829 (ASM_SPEC): Update to also substitute -mabi.
14830 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
14831 file whose name depends on -mabi= and -mbig-endian.
14832 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
14833 TARGET_ILP32.
14834 (POINTER_SIZE): New define.
14835 (POINTERS_EXTEND_UNSIGNED): Ditto.
14836 (enum aarch64_abi_type): New enumeration tag.
14837 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
14838 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
14839 (TARGET_ILP32): New define.
14840 * config/aarch64/aarch64.opt (mabi): New.
14841 (aarch64_abi): New.
14842 (ilp32, lp64): New values for -mabi.
14843 * config/aarch64/t-aarch64 (comma): New define.
14844 (MULTILIB_OPTIONS): Ditto.
14845 (MULTILIB_DIRNAMES): Ditto.
14846 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
14847 * doc/invoke.texi: Document -mabi for AArch64.
14848
14849 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
14850
14851 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
14852
14853 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14854 Anton Blanchard <anton@au1.ibm.com>
14855
14856 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
14857 endianness when selecting field to splat.
14858
14859 2013-07-22 Eric Christopher <echristo@gmail.com>
14860
14861 * dwarf2out.c (die_odr_checksum): New function to use
14862 CHECKSUM_ macros and ULEB128 for DIE tag.
14863 (generate_type_signature): Use.
14864
14865 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
14866
14867 * config.gcc (sparc*-*-*): Accept leon3 processor.
14868 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
14869 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
14870 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
14871 * config/sparc/sparc.opt (enum processor_type): Add leon3.
14872 (mfix-ut699): Adjust comment.
14873 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
14874 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
14875 (CPP_CPU_SPEC): Likewise.
14876 (ASM_CPU_SPEC): Likewise.
14877 * config/sparc/sparc.c (leon3_cost): New constant.
14878 (sparc_option_override): Add leon3 support.
14879 (mem_ref): New function.
14880 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
14881 (sparc_do_work_around_errata): Look into the instruction in the delay
14882 slot and adjust accordingly. Add fix for the data cache nullify issues
14883 of the UT699. Change insertion position for the NOP.
14884 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
14885 (leon3_load): New reservation.
14886 (leon_store): Bump latency to 2.
14887 (grfpu): New automaton.
14888 (grfpu_alu): New unit.
14889 (grfpu_ds): Likewise.
14890 (leon_fp_alu): Adjust.
14891 (leon_fp_mult): Delete.
14892 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
14893 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
14894 * config/sparc/sparc.md (cpu): Add leon3.
14895 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
14896 (swapsi): Likewise.
14897 (atomic_test_and_set): Likewise.
14898 (ldstub): Likewise.
14899
14900 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
14901
14902 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
14903 default for R5900 targets.
14904 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
14905 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
14906 * config/mips/mips.c (mips_option_override): Report an error for
14907 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
14908 for -march=r5900 -mhard-float.
14909
14910 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14911
14912 * df-problems.c (can_move_insns_across): Exit loop once we
14913 find a non-fixed, non-global register.
14914
14915 * ipa-pure-const.c (propagate_nothrow): Exit loop after
14916 setting can_throw.
14917
14918 * omega.c (omega_eliminate_red): Break after setting red_found.
14919 (omega_problem_has_red_equations): Similarly after setting found.
14920 (omega_query_variable): Similarly after setting coupled.
14921
14922 2013-07-22 Marek Polacek <polacek@redhat.com>
14923
14924 * gimplify.c: Don't include gimple.h twice.
14925
14926 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14927
14928 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
14929 instead of TARGET_THUMB1.
14930 (Pz): New constraint.
14931 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
14932 encodings.
14933 (compare_negsi_si): Likewise.
14934 (compare_addsi2_op0): Likewise.
14935 (compare_addsi2_op1): Likewise.
14936 (addsi3_carryin_<optab>): Likewise.
14937 (addsi3_carryin_alt2_<optab>): Likewise.
14938 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
14939 for arm_restrict_it.
14940 (subsi3_carryin): Likewise.
14941 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
14942 (minmax_arithsi): Disable for arm_restrict_it.
14943 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
14944 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
14945 (satsi_<SAT:code>_shift): Likewise.
14946 (arm_shiftsi3): Add alternative for 16-bit encoding.
14947 (arm32_movhf): Disable for arm_restrict_it.
14948 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
14949 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
14950
14951 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
14952
14953 * config/arm/arm.md (attribute "insn"): Delete.
14954 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
14955 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
14956 (not_shiftsi): Update for attribute change.
14957 (not_shiftsi_compare0): Likewise.
14958 (not_shiftsi_compare0_scratch): Likewise.
14959 (arm_one_cmplsi2): Likewise.
14960 (thumb1_one_cmplsi2): Likewise.
14961 (notsi_compare0): Likewise.
14962 (notsi_compare0_scratch): Likewise.
14963 (thumb1_movdi_insn): Likewise.
14964 (arm_movsi_insn): Likewise.
14965 (movhi_insn_arch4): Likewise.
14966 (movhi_bytes): Likewise.
14967 (arm_movqi_insn): Likewise.
14968 (thumb1_movqi_insn): Likewise.
14969 (arm32_movhf): Likewise.
14970 (thumb1_movhf): Likewise.
14971 (arm_movsf_soft_insn): Likewise.
14972 (thumb1_movsf_insn): Likewise.
14973 (thumb_movdf_insn): Likewise.
14974 (movsicc_insn): Likewise.
14975 (movsfcc_soft_insn): Likewise.
14976 (and_scc): Likewise.
14977 (cond_move): Likewise.
14978 (if_move_not): Likewise.
14979 (if_not_move): Likewise.
14980 (if_shift_move): Likewise.
14981 (if_move_shift): Likewise.
14982 (if_shift_shift): Likewise.
14983 (if_not_arith): Likewise.
14984 (if_arith_not): Likewise.
14985 (cond_move_not): Likewise.
14986 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
14987 (neon_mov<mode>): Likewise.
14988 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
14989 (thumb2_movsi_vfp): Likewise.
14990 (movsf_vfp): Likewise.
14991 (thumb2_movsf_vfp): Likewise.
14992 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
14993 change.
14994 (cortexa7_older_only): Likewise.
14995 (cortexa7_younger): Likewise.
14996 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
14997 (1020alu_shift_op): Likewise.
14998 (1020alu_shift_reg_op): Likewise.
14999 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
15000 (alu_shift_op): Likewise.
15001 (alu_shift_reg_op): Likewise.
15002 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
15003 (11_alu_shift_op): Likewise.
15004 (11_alu_shift_reg_op): Likewise.
15005 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
15006 (9_alu_shift_reg_op): Likewise.
15007 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15008 change.
15009 (cortex_a15_alu_shift): Likewise.
15010 (cortex_a15_alu_shift_reg): Likewise.
15011 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
15012 (cortex_a5_alu_shift): Likewise.
15013 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
15014 change.
15015 (cortex_a53_alu_shift): Likewise.
15016 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15017 change.
15018 (cortex_a7_alu_reg): Likewise.
15019 (cortex_a7_alu_shift): Likewise.
15020 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15021 (cortex_a8_alu_shift): Likewise.
15022 (cortex_a8_alu_shift_reg): Likewise.
15023 (cortex_a8_mov): Likewise.
15024 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
15025 (cortex_a9_dp_shift): Likewise.
15026 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
15027 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
15028 (cortex_r4_mov): Likewise.
15029 (cortex_r4_alu_shift): Likewise.
15030 (cortex_r4_alu_shift_reg): Likewise.
15031 * config/arm/fa526.md (526_alu_op): Update for attribute change.
15032 (526_alu_shift_op): Likewise.
15033 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
15034 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
15035 (626te_alu_shift_op): Likewise.
15036 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
15037 (726te_alu_op): Likewise.
15038 (726te_alu_shift_op): Likewise.
15039 (726te_alu_shift_reg_op): Likewise.
15040 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
15041 (mp626_alu_shift_op): Likewise.
15042 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
15043 (pj4_alu_e1_conds): Likewise.
15044 (pj4_alu): Likewise.
15045 (pj4_alu_conds): Likewise.
15046 (pj4_shift): Likewise.
15047 (pj4_shift_conds): Likewise.
15048 (pj4_alu_shift): Likewise.
15049 (pj4_alu_shift_conds): Likewise.
15050
15051 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15052
15053 * config/arm/predicates.md (shiftable_operator_strict_it):
15054 New predicate.
15055 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
15056 Disable cond_exec version for arm_restrict_it.
15057 (thumb2_smaxsi3): Convert to generate cond_exec.
15058 (thumb2_sminsi3): Likewise.
15059 (thumb32_umaxsi3): Likewise.
15060 (thumb2_uminsi3): Likewise.
15061 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
15062 (thumb2_neg_abssi2): Likewise.
15063 (thumb2_mov_scc): Add alternative for 16-bit encoding.
15064 (thumb2_movsicc_insn): Adjust alternatives.
15065 (thumb2_mov_negscc): Disable for arm_restrict_it.
15066 (thumb2_mov_negscc_strict_it): New pattern.
15067 (thumb2_mov_notscc_strict_it): New pattern.
15068 (thumb2_mov_notscc): Disable for arm_restrict_it.
15069 (thumb2_ior_scc): Likewise.
15070 (thumb2_ior_scc_strict_it): New pattern.
15071 (thumb2_cond_move): Adjust for arm_restrict_it.
15072 (thumb2_cond_arith): Disable for arm_restrict_it.
15073 (thumb2_cond_arith_strict_it): New pattern.
15074 (thumb2_cond_sub): Adjust for arm_restrict_it.
15075 (thumb2_movcond): Likewise.
15076 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
15077 (thumb2_zero_extendhisi2_v6): Likewise.
15078 (thumb2_zero_extendqisi2_v6): Likewise.
15079 (orsi_notsi_si): Likewise.
15080 (orsi_not_shiftsi_si): Likewise.
15081
15082 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
15083
15084 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
15085 instruction sequence is 1 byte shorter.
15086
15087 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
15088
15089 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
15090 it is not needed after split.
15091
15092 2013-07-20 Iain Sandoe <iain@codesourcery.com>
15093
15094 PR target/51784
15095 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
15096 second label for nonlocal goto receivers. Don't output pic base labels
15097 unless we're producing PIC; mark that action unreachable().
15098 (ix86_save_reg): If the function contains a nonlocal label, save the
15099 PIC base reg.
15100 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
15101 * config/darwin.c (emitted_pic_label_num): New GTY.
15102 (update_pic_label_number_if_needed): New.
15103 (machopic_output_function_base_name): Adjust for nonlocal receiver
15104 case.
15105 (machopic_should_output_picbase_label): New.
15106 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
15107 (nonlocal_goto_receiver): New insn and split.
15108
15109 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
15110
15111 * config/aarch64/aarch64-builtins.c
15112 (aarch64_fold_builtin): Fold abs in all modes.
15113 * config/aarch64/aarch64-simd-builtins.def
15114 (abs): Enable for all modes.
15115 * config/aarch64/arm_neon.h
15116 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
15117 (vabs_f64): Add missing intrinsic.
15118
15119 2013-07-19 Ian Bolton <ian.bolton@arm.com>
15120
15121 * config/aarch64/arm_neon.h (vabs_s64): New function
15122
15123 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
15124
15125 PR target/57516
15126 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
15127 * config/avr/avr.md (adjust_len): Add `round'.
15128 * config/avr/avr-protos.h (avr_out_round): New prototype.
15129 (avr_out_plus): Add `out_label' argument.
15130 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
15131 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
15132 Handle the case where `insn' is just a pattern.
15133 (avr_out_bitop): Handle the case where `insn' is just a pattern.
15134 (avr_out_round): New function.
15135 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
15136
15137 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
15138
15139 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15140 Add check for flag_stack_usage to handle -fstack-usage support
15141
15142 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
15143
15144 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
15145 interaction for new Power8 flags and VSX.
15146
15147 2013-07-18 Sriraman Tallam <tmsriram@google.com>
15148
15149 PR middle-end/57698
15150 * tree-inline.c (expand_call_inline): Emit errors during
15151 early_inlining only if optimization is not turned on.
15152
15153 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15154
15155 * passes.def: New.
15156
15157 * passes.c (init_optimization_passes): Move the construction of
15158 the pass hierarchy into a new passes.def file.
15159
15160 * Makefile.in (passes.o): Add dependency on passes.def.
15161
15162 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15163
15164 * passes.c (init_optimization_passes): Introduce macros for
15165 constructing the tree of passes (INSERT_PASSES_AFTER,
15166 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
15167 TERMINATE_PASS_LIST).
15168
15169 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
15170 Wei Mi <wmi@google.com>
15171
15172 PR rtl-optimization/57878
15173 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
15174 top.
15175 (reload_pseudo_compare_func): Check nregs first for reload
15176 pseudos.
15177
15178 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15179
15180 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
15181
15182 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
15183
15184 * read-rtl.c (validate_const_int): Once an invalid character is
15185 seen, quit the loop.
15186
15187 * gengtype.c (write_roots): Similarly once we find the "deletable"
15188 or "if_marked" option.
15189
15190 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15191
15192 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
15193 "xtab" and "sat". Move value "clz" from here to ...
15194 (attriubte "type"): ... here.
15195 (satsi_<SAT:code>): Delete "insn" attribute.
15196 (satsi_<SAT:code>_shift): Likewise.
15197 (arm_zero_extendqisi2addsi): Likewise.
15198 (arm_extendqisi2addsi): Likewise.
15199 (clzsi2): Update for attribute changes.
15200 (rbitsi2): Likewise.
15201 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
15202 attribute.
15203 (arm_usatsihi): Likewise.
15204 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15205
15206 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15207
15208 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
15209 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
15210 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
15211 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
15212 in alphabetical order.
15213 (attribute "core_cycles"): Update for attribute changes.
15214 (arm_addsi3): Likewise.
15215 (addsi3_compare0): Likewise.
15216 (addsi3_compare0_scratch): Likewise.
15217 (addsi3_compare_op1): Likewise.
15218 (addsi3_compare_op2): Likewise.
15219 (compare_addsi2_op0): Likewise.
15220 (compare_addsi2_op1): Likewise.
15221 (addsi3_carryin_shift_<optab>): Likewise.
15222 (subsi3_carryin_shift): Likewise.
15223 (rsbsi3_carryin_shift): Likewise.
15224 (arm_subsi3_insn): Likewise.
15225 (subsi3_compare0): Likewise.
15226 (subsi3_compare): Likewise.
15227 (arm_andsi3_insn): Likewise.
15228 (thumb1_andsi3_insn): Likewise.
15229 (andsi3_compare0): Likewise.
15230 (andsi3_compare0_scratch): Likewise.
15231 (zeroextractsi_compare0_scratch
15232 (andsi_not_shiftsi_si): Likewise.
15233 (iorsi3_insn): Likewise.
15234 (iorsi3_compare0): Likewise.
15235 (iorsi3_compare0_scratch): Likewise.
15236 (arm_xorsi3): Likewise.
15237 (thumb1_xorsi3_insn): Likewise.
15238 (xorsi3_compare0): Likewise.
15239 (xorsi3_compare0_scratch): Likewise.
15240 (satsi_<SAT:code>_shift): Likewise.
15241 (rrx): Likewise.
15242 (arm_shiftsi3): Likewise.
15243 (shiftsi3_compare0): Likewise.
15244 (not_shiftsi): Likewise.
15245 (not_shiftsi_compare0): Likewise.
15246 (not_shiftsi_compare0_scratch): Likewise.
15247 (arm_one_cmplsi2): Likewise.
15248 (thumb_one_complsi2): Likewise.
15249 (notsi_compare0): Likewise.
15250 (notsi_compare0_scratch): Likewise.
15251 (thumb1_zero_extendhisi2): Likewise.
15252 (arm_zero_extendhisi2): Likewise.
15253 (arm_zero_extendhisi2_v6): Likewise.
15254 (arm_zero_extendhisi2addsi): Likewise.
15255 (thumb1_zero_extendqisi2): Likewise.
15256 (thumb1_zero_extendqisi2_v6): Likewise.
15257 (arm_zero_extendqisi2): Likewise.
15258 (arm_zero_extendqisi2_v6): Likewise.
15259 (arm_zero_extendqisi2addsi): Likewise.
15260 (thumb1_extendhisi2): Likewise.
15261 (arm_extendhisi2): Likewise.
15262 (arm_extendhisi2_v6): Likewise.
15263 (arm_extendqisi): Likewise.
15264 (arm_extendqisi_v6): Likewise.
15265 (arm_extendqisi2addsi): Likewise.
15266 (thumb1_extendqisi2): Likewise.
15267 (thumb1_movdi_insn): Likewise.
15268 (arm_movsi_insn): Likewise.
15269 (movsi_compare0): Likewise.
15270 (movhi_insn_arch4): Likewise.
15271 (movhi_bytes): Likewise.
15272 (arm_movqi_insn): Likewise.
15273 (thumb1_movqi_insn): Likewise.
15274 (arm32_movhf): Likewise.
15275 (thumb1_movhf): Likewise.
15276 (arm_movsf_soft_insn): Likewise.
15277 (thumb1_movsf_insn): Likewise.
15278 (movdf_soft_insn): Likewise.
15279 (thumb_movdf_insn): Likewise.
15280 (arm_cmpsi_insn): Likewise.
15281 (cmpsi_shiftsi): Likewise.
15282 (cmpsi_shiftsi_swp): Likewise.
15283 (arm_cmpsi_negshiftsi_si): Likewise.
15284 (movsicc_insn): Likewise.
15285 (movsfcc_soft_insn): Likewise.
15286 (arith_shiftsi): Likewise.
15287 (arith_shiftsi_compare0
15288 (arith_shiftsi_compare0_scratch
15289 (sub_shiftsi): Likewise.
15290 (sub_shiftsi_compare0
15291 (sub_shiftsi_compare0_scratch
15292 (and_scc): Likewise.
15293 (cond_move): Likewise.
15294 (if_plus_move): Likewise.
15295 (if_move_plus): Likewise.
15296 (if_move_not): Likewise.
15297 (if_not_move): Likewise.
15298 (if_shift_move): Likewise.
15299 (if_move_shift): Likewise.
15300 (if_shift_shift): Likewise.
15301 (if_not_arith): Likewise.
15302 (if_arith_not): Likewise.
15303 (cond_move_not): Likewise.
15304 (thumb1_ashlsi3): Set type attribute.
15305 (thumb1_ashrsi3): Likewise.
15306 (thumb1_lshrsi3): Likewise.
15307 (thumb1_rotrsi3): Likewise.
15308 (shiftsi3_compare0_scratch): Likewise.
15309 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
15310 (neon_mov<mode>): Likewise.
15311 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
15312 attribute changes.
15313 (thumb2_movsi_insn): Likewise.
15314 (thumb2_cmpsi_neg_shiftsi): Likewise.
15315 (thumb2_extendqisi_v6): Likewise.
15316 (thumb2_zero_extendhisi2_v6): Likewise.
15317 (thumb2_zero_extendqisi2_v6): Likewise.
15318 (thumb2_shiftsi3_short): Likewise.
15319 (thumb2_addsi3_compare0_scratch): Likewise.
15320 (orsi_not_shiftsi_si): Likewise.
15321 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
15322 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
15323 changes.
15324 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
15325 (1020alu_shift_op): Likewise.
15326 (1020alu_shift_reg_op): Likewise.
15327 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
15328 (alu_shift_op): Likewise.
15329 (alu_shift_reg_op): Likewise.
15330 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
15331 (11_alu_shift_op): Likewise.
15332 (11_alu_shift_reg_op): Likewise.
15333 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
15334 (9_alu_shift_reg_op): Likewise.
15335 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15336 changes.
15337 (cortex_a15_alu_shift): Likewise.
15338 (cortex_a15_alu_shift_reg): Likewise.
15339 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
15340 changes.
15341 (cortex_a5_alu_shift): Likewise.
15342 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
15343 changes.
15344 (cortex_a53_alu_shift): Likewise.
15345 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15346 changes.
15347 (cortex_a7_alu_reg): Likewise.
15348 (cortex_a7_alu_shift): Likewise.
15349 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
15350 changes.
15351 (cortex_a8_alu_shift): Likewise.
15352 (cortex_a8_alu_shift_reg): Likewise.
15353 (cortex_a8_mov): Likewise.
15354 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
15355 (cortex_a9_dp_shift): Likewise.
15356 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
15357 changes.
15358 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
15359 changes.
15360 (cortex_r4_mov): Likewise.
15361 (cortex_r4_alu_shift): Likewise.
15362 (cortex_r4_alu_shift_reg): Likewise.
15363 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
15364 (526_alu_shift_op): Likewise.
15365 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
15366 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
15367 (626te_alu_shift_op): Likewise.
15368 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
15369 (726te_alu_op): Likewise.
15370 (726te_alu_shift_op): Likewise.
15371 (726te_alu_shift_reg_op): Likewise.
15372 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
15373 (mp626_alu_shift_op): Likewise.
15374 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
15375 (pj4_alu_e1_conds): Likewise.
15376 (pj4_alu): Likewise.
15377 (pj4_alu_conds): Likewise.
15378 (pj4_shift): Likewise.
15379 (pj4_shift_conds): Likewise.
15380 (pj4_alu_shift): Likewise.
15381 (pj4_alu_shift_conds): Likewise.
15382 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15383 changes.
15384 (cortexa7_older_only): Likewise.
15385 (cortexa7_younger): Likewise.
15386
15387 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15388
15389 * ipa-pure-const.c (generate_summary): Rename to...
15390 (pure_const_generate_summary): ... this.
15391
15392 2013-07-17 Iain Sandoe <iain@codesourcery.com>
15393
15394 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
15395
15396 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
15397
15398 PR target/57909
15399 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
15400 usage in HI mode.
15401
15402 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15403
15404 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
15405 enabled without -march=zEC12.
15406 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
15407 flags to be set.
15408
15409 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15410
15411 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
15412 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
15413 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
15414 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
15415 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
15416 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
15417 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
15418 ISA_HAS_FP4.
15419 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
15420 and ISA_HAS_NMADD3_NMSUB3.
15421 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
15422 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
15423 (nmsub4<mode>, nmsub3<mode>): Likewise.
15424 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
15425
15426 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15427
15428 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
15429 TARGET_MIPS5400 checking.
15430
15431 2013-07-16 Jakub Jelinek <jakub@redhat.com>
15432 Peter Bergner <bergner@vnet.ibm.com>
15433
15434 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
15435 registers in the comment.
15436 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
15437 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
15438 rather than FIRST_PSEUDO_REGISTERS.
15439
15440 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
15441
15442 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
15443 enable extra ISA flags with TARGET_HTM.
15444
15445 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15446
15447 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
15448 Fix comment typos.
15449
15450 2013-07-15 Cong Hou <congh@google.com>
15451
15452 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
15453 in compare function for sorting.
15454
15455 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
15456
15457 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
15458 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
15459 * config/rs6000/rs6000.opt: Add -mhtm option.
15460 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
15461 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
15462 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
15463 __HTM__ if the HTM instructions are available.
15464 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
15465 htm_spr_reg_operand): New define_predicates.
15466 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
15467 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
15468 Include htm.md.
15469 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
15470 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
15471 HTM builtin functions.
15472 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
15473 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
15474 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
15475 (rs6000_builtin_mask_calculate): Likewise.
15476 (rs6000_option_override_internal): Likewise.
15477 (bdesc_htm): Add new HTM builtin support.
15478 (htm_spr_num): New function.
15479 (htm_spr_regno): Likewise.
15480 (rs6000_htm_spr_icode): Likewise.
15481 (htm_expand_builtin): Likewise.
15482 (htm_init_builtins): Likewise.
15483 (rs6000_expand_builtin): Add support for HTM builtin functions.
15484 (rs6000_init_builtins): Likewise.
15485 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
15486 option.
15487 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
15488 (TARGET_HTM, MASK_HTM): Define macros.
15489 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
15490 (FIXED_REGISTERS): Likewise.
15491 (CALL_USED_REGISTERS): Likewise.
15492 (CALL_REALLY_USED_REGISTERS): Likewise.
15493 (REG_ALLOC_ORDER): Likewise.
15494 (enum reg_class): Likewise.
15495 (REG_CLASS_NAMES): Likewise.
15496 (REG_CLASS_CONTENTS): Likewise.
15497 (REGISTER_NAMES): Likewise.
15498 (ADDITIONAL_REGISTER_NAMES): Likewise.
15499 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
15500 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
15501 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
15502 * config/rs6000/htm.md: New file.
15503 * config/rs6000/htmintrin.h: New file.
15504 * config/rs6000/htmxlintrin.h: New file.
15505
15506 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
15507
15508 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15509 Define SYMBOL_TINY_GOT, update comment.
15510 * config/aarch64/aarch64.c
15511 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
15512 (aarch64_expand_mov_immediate): Likewise.
15513 (aarch64_print_operand): Likewise.
15514 (aarch64_classify_symbol): Likewise.
15515 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
15516 (ldr_got_tiny): Define.
15517
15518 2013-07-13 Tobias Grosser <tobias@grosser.es>
15519
15520 PR tree-optimization/54094
15521 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
15522 scheduling dimension for the parallelism check from the polyhedral
15523 information in the AST.
15524 * graphite-dependences.c (carries_deps): Do not assume the schedule is
15525 in 2D + 1 form.
15526
15527 2013-07-13 Jason Merrill <jason@redhat.com>
15528
15529 * print-tree.c (debug_vec_tree): Use debug_raw.
15530 (debug_raw (vec<tree, va_gc> &)): New.
15531 (debug_raw (vec<tree, va_gc> *)): New.
15532 * tree.h: Declare them.
15533
15534 2013-07-13 Bin Cheng <bin.cheng@arm.com>
15535
15536 * ifcvt.c (ifcvt_after_combine): New static variable.
15537 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
15538 for size.
15539 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
15540 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
15541 rest_of_handle_if_after_reload): Pass new argument for if_convert.
15542
15543 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
15544
15545 * config/mips/mips.c (mips_expand_call): Remove empty statement.
15546
15547 2013-07-12 Michael Matz <matz@suse.de>
15548
15549 PR middle-end/55771
15550 * convert.c (convert_to_real): Reject non-float inner types.
15551
15552 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
15553
15554 * config/aarch64/aarch64-protos.h
15555 (aarch64_simd_immediate_valid_for_move): Remove.
15556 * config/aarch64/aarch64.c (simd_immediate_info): New member.
15557 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
15558 cases.
15559 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
15560
15561 2013-07-11 Steve Ellcey <sellcey@mips.com>
15562
15563 * config/mips/mips.c (mips_conditional_register_usage): Do not
15564 use t[0-7] registers in MIPS16 mode when optimizing for size.
15565
15566 2013-07-11 Sriraman Tallam <tmsriram@google.com>
15567
15568 * config/i386/i386.c (dispatch_function_versions): Fix array
15569 indexing of function_version_info to match actual_versions.
15570
15571 2013-07-11 Teresa Johnson <tejohnson@google.com>
15572
15573 * vec.h (struct va_gc): Move release out-of-line.
15574 (va_gc::release): Call ggc_free on released vec.
15575
15576 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15577
15578 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
15579 Require GOT register as additional operand in UNSPEC.
15580 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
15581 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15582 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15583 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
15584 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
15585 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
15586 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
15587
15588 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
15589
15590 PR target/57631
15591 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
15592 name seen by assembler/linker rather if available.
15593
15594 2013-07-11 Andreas Schwab <schwab@suse.de>
15595
15596 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
15597
15598 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
15599
15600 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
15601
15602 2013-07-10 Joseph Myers <joseph@codesourcery.com>
15603
15604 * doc/tm.texi.in: Move hook documentation to ....
15605 * target.def: ... here.
15606
15607 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
15608 text on @hook line.
15609 * doc/tm.texi: Regenerate.
15610
15611 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
15612
15613 PR c++/57869
15614 * doc/invoke.texi: Document -Wconditionally-supported.
15615
15616 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15617
15618 PR target/57844
15619 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
15620 of my_fp.
15621
15622 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15623
15624 PR target/57506
15625 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
15626 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
15627 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
15628 Remove duplicate devices.
15629 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
15630 * config/avr/t-multilib: Regenerate.
15631 * config/avr/avr-tables.opt: Regenerate.
15632 * doc/avr-mmcu.texi: Regenerate.
15633
15634 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15635
15636 PR target/56987
15637 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
15638
15639 2013-07-10 Graham Stott <graham.stott@btinternet.com>
15640
15641 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
15642 the cost of MULT when optimizing for size.
15643
15644 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
15645
15646 * config/cr16/cr16-protos.h: Don't include target.h.
15647
15648 2013-07-09 Joseph Myers <joseph@codesourcery.com>
15649
15650 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
15651 adjust register size for TDmode and TFmode for VSX registers.
15652
15653 2013-07-08 Kai Tietz <ktietz@redhat.com>
15654
15655 PR target/56892
15656 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
15657 hook_bool_const_tree_true.
15658
15659 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15660
15661 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
15662 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
15663 * config/s390/s390.md: Define FPR*_REGNUM constants.
15664 Fix FPR2_REGNUM constant (18 -> 17).
15665 ("*trunc<BFP:mode><DFP_ALL:mode>2")
15666 ("*trunc<DFP_ALL:mode><BFP:mode>2")
15667 ("trunc<BFP:mode><DFP_ALL:mode>2")
15668 ("trunc<DFP_ALL:mode><BFP:mode>2")
15669 ("*extend<BFP:mode><DFP_ALL:mode>2")
15670 ("*extend<DFP_ALL:mode><BFP:mode>2")
15671 ("extend<BFP:mode><DFP_ALL:mode>2")
15672 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
15673 FPR4_REGNUM.
15674
15675 2013-07-08 Graham Stott <graham.stott@btinternet.com>
15676
15677 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
15678
15679 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15680
15681 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
15682 and cfun_fpr_bit_p to cfun_fpr_save_p.
15683 (s390_frame_area, s390_register_info, s390_frame_info)
15684 (s390_emit_prologue, s390_emit_epilogue)
15685 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
15686 register numbers.
15687 * config/s390/s390.h: Define *_REGNUM macros for floating point
15688 register numbers.
15689
15690 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
15691
15692 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
15693
15694 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
15695
15696 PR rtl-optimization/57786
15697 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
15698 and break out of the loop when it is set to false.
15699
15700 2013-07-08 Jakub Jelinek <jakub@redhat.com>
15701
15702 PR target/57819
15703 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
15704 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
15705 (const_int 63)) 0)).
15706 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
15707 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
15708 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
15709
15710 PR rtl-optimization/57829
15711 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
15712 mask bits outside of mode are just sign-extension from mode to HWI.
15713
15714 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
15715
15716 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
15717 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
15718 adjust_address instead of change_address to keep info about alignment.
15719 (emit_strmov): Remove.
15720 (emit_memmov): New function.
15721 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
15722 (expand_movmem_epilogue): Likewise and return updated rtx for
15723 destination.
15724 (expand_constant_movmem_prologue): Likewise and return updated rtx for
15725 destination and source.
15726 (decide_alignment): Refactor, handle vector_loop.
15727 (ix86_expand_movmem): Likewise.
15728 (ix86_expand_setmem): Likewise.
15729 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
15730
15731 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
15732
15733 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
15734 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
15735
15736 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15737
15738 * config/i386/sse.md (sse_movlhps): Change alternative 3
15739 of operand 2 to "m".
15740
15741 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15742
15743 PR target/57807
15744 * config/i386/sse.md (iptr): New mode attribute.
15745 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
15746 (<sse>_vm<plusminus_insn><mode>3): Ditto.
15747 (<sse>_vmmul<mode>3): Ditto.
15748 (<sse>_vmdiv<mode>3): Ditto.
15749 (sse_vmrcpv4sf2): Ditto.
15750 (<sse>_vmsqrt<mode>2): Ditto.
15751 (sse_vmrsqrtv4sf2): Ditto.
15752 (<sse>_vm<code><mode>3): Ditto.
15753 (avx_vmcmp<mode>3): Ditto.
15754 (<sse>_vmmaskcmp<mode>3): Ditto.
15755 (<sse>_comi): Ditto.
15756 (<sse>_ucomi): Ditto.
15757 (*xop_vmfrcz_<mode>): Ditto.
15758 (*fmai_fmadd_<mode>): Ditto.
15759 (*fmai_fmsub_<mode>): Ditto.
15760 (*fmai_fnmadd_<mode>): Ditto.
15761 (*fmai_fnmsub_<mode>): Ditto.
15762 (*fma4i_vmfmadd_<mode>): Ditto.
15763 (*fma4i_vmfmsub_<mode>): Ditto.
15764 (*fma4i_vmfnmadd_<mode>): Ditto.
15765 (*fma4i_vmfnmsub_<mode>): Ditto.
15766 (*xop_vmfrcz_<mode>): Ditto.
15767 (sse_cvtps2pi): Ditto.
15768 (sse_cvttps2pi): Ditto.
15769 (sse_cvtss2si): Ditto.
15770 (sse_cvtss2si_2): Ditto.
15771 (sse_cvtss2siq_2): Ditto.
15772 (sse_cvttss2si): Ditto.
15773 (sse_cvttss2siq): Ditto.
15774 (sse_cvtsd2si): Ditto.
15775 (sse_cvtsd2si_2): Ditto.
15776 (sse_cvtsd2siq_2): Ditto.
15777 (sse_cvttsd2si): Ditto.
15778 (sse_cvttsd2siq): Ditto.
15779 (sse_cvtsd2ss): Ditto.
15780 (sse_cvtss2sd): Ditto.
15781 (avx2_pbroadcast<mode>): Ditto.
15782 (avx2_pbroadcast<mode>_1): Ditto.
15783 (*avx_vperm_broadcast_v4sf): Ditto.
15784
15785 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
15786 (sse_movlhps): Ditto.
15787 (sse_storehps): Ditto.
15788 (sse_loadhps): Ditto.
15789 (sse_storelps): Ditto.
15790 (sse_loadlps): Ditto.
15791 (*vec_concatv4sf): Ditto.
15792 (*vec_interleave_highv2df): Ditto.
15793 (*vec_interleave_lowv2df): Ditto.
15794 (*vec_extractv2df_1_sse): Ditto.
15795 (*vec_extractv2df_0_sse): Ditto.
15796 (sse2_storelpd): Ditto.
15797 (sse2_loadlpd): Ditto.
15798 (sse2_movsd): Ditto.
15799 (*vec_concatv4si): Ditto.
15800 (vec_concatv2di): Ditto.
15801
15802 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
15803 for Intel asm dialect.
15804 (mmx_punpcklwd): Ditto.
15805 (mmx_punpckldq): Ditto.
15806
15807 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
15808 for intel assembler dialect.
15809
15810 2013-07-06 Jakub Jelinek <jakub@redhat.com>
15811
15812 PR target/29776
15813 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
15814 for BUILT_IN_C{LZ,LRSB}*.
15815 * tree.h (CASE_INT_FN): Add FN##IMAX case.
15816 * tree-vrp.c (extract_range_basic): Handle
15817 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
15818 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
15819 fall thru to code calling set_value*.
15820 * builtins.c (expand_builtin): Remove *IMAX cases.
15821 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
15822 if width is bigger than 2*HWI.
15823
15824 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
15825
15826 PR rtl-optimization/55342
15827 * lra-int.h (lra_subreg_reload_pseudos): New.
15828 * lra.c: Add undoing optional reloads to the block diagram.
15829 (lra_subreg_reload_pseudos): New.
15830 (lra_optional_reload_pseudos): Change comments.
15831 (lra): Init and clear lra_subreg_reload_pseudos. Clear
15832 lra_optional_reload_pseudos after undo transformations.
15833 * lra-assigns.c (pseudo_prefix_title): New.
15834 (lra_setup_reg_renumber): Use it.
15835 (spill_for): Ditto. Check subreg reload pseudos too.
15836 (assign_by_spills): Consider subreg reload pseudos too.
15837 * lra-constraints.c (simplify_operand_subreg): Use
15838 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
15839 (curr_insn_transform): Recognize and do optional reloads.
15840 (undo_optional_reloads): New.
15841 (lra_undo_inheritance): Call undo_optional_reloads.
15842
15843 2013-07-05 Thomas Quinot <quinot@adacore.com>
15844
15845 * tree-complex.c (expand_complex_operations_1): Fix typo.
15846
15847 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
15848
15849 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
15850 (tune_params): New member 'const vec_costs'.
15851 * config/aarch64/aarch64.c (generic_vector_cost): New.
15852 (generic_tunings): New member 'generic_vector_cost'.
15853 (aarch64_builtin_vectorization_cost): New.
15854 (aarch64_add_stmt_cost): New.
15855 (TARGET_VECTORIZE_ADD_STMT_COST): New.
15856 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
15857
15858 2013-07-03 Jakub Jelinek <jakub@redhat.com>
15859
15860 PR target/57777
15861 * config/i386/predicates.md (vsib_address_operand): Disallow
15862 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
15863
15864 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
15865
15866 PR middle-end/55030
15867 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
15868 expand_builtin_setjmp_receiver.
15869 (expand_label): Adjust, call expand_builtin_setjmp_receiver
15870 with NULL for the label parameter.
15871 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
15872 the frame-pointer. Adjust comments.
15873 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
15874 only if LABEL is non-NULL.
15875
15876 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
15877
15878 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
15879 (ARM_ABI_AAPCS64): Ditto.
15880 (arm_abi): Ditto.
15881 (ARM_DEFAULT_ABI): Ditto.
15882
15883 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
15884
15885 * config/aarch64/aarch64-builtins.c
15886 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
15887 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
15888 (st1): Likewise.
15889 * config/aarch64/aarch64-simd.md
15890 (aarch64_ld1<VALL:mode>): New.
15891 (aarch64_st1<VALL:mode>): Likewise.
15892 * config/aarch64/arm_neon.h
15893 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
15894
15895 2013-07-02 Sriraman Tallam <tmsriram@google.com>
15896
15897 * config/i386/i386.c (gate_insert_vzeroupper): Check if
15898 target ISA is AVX.
15899 (ix86_option_override_internal):Turn on all -mavx target flags by
15900 default as they are dependent on AVX anyway.
15901
15902 2013-07-02 Cary Coutant <ccoutant@google.com>
15903
15904 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
15905 deterministic hash.
15906 (loc_checksum_ordered): Likewise.
15907 (hash_loc_operands): Remove inline keyword.
15908
15909 2013-07-02 Jakub Jelinek <jakub@redhat.com>
15910
15911 PR tree-optimization/57741
15912 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
15913 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
15914 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
15915 Allow REAL_CST step_exprs if flag_associative_math.
15916 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
15917
15918 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15919
15920 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
15921
15922 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15923
15924 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
15925
15926 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15927
15928 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
15929 encoding.
15930 (iorsi3_insn): Likewise.
15931 (arm_xorsi3): Likewise.
15932
15933 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
15934
15935 * arm.md (attribute "wtype"): Delete. Move attribute values from here
15936 to ...
15937 (attribute "type"): ... here, and prefix with "wmmx_".
15938 (attribute "core_cycles"): Update for attribute changes.
15939 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
15940 (tbcstv4hi): Likewise.
15941 (tbcstv2si): Likewise.
15942 (iwmmxt_iordi3): Likewise.
15943 (iwmmxt_xordi3): Likewise.
15944 (iwmmxt_anddi3): Likewise.
15945 (iwmmxt_nanddi3): Likewise.
15946 (iwmmxt_arm_movdi): Likewise.
15947 (iwmmxt_movsi_insn): Likewise.
15948 (mov<mode>_internal): Likewise.
15949 (and<mode>3_iwmmxt): Likewise.
15950 (ior<mode>3_iwmmxt): Likewise.
15951 (xor<mode>3_iwmmxt): Likewise.
15952 (add<mode>3_iwmmxt): Likewise.
15953 (ssaddv8qi3): Likewise.
15954 (ssaddv4hi3): Likewise.
15955 (ssaddv2si3): Likewise.
15956 (usaddv8qi3): Likewise.
15957 (usaddv4hi3): Likewise.
15958 (usaddv2si3): Likewise.
15959 (sub<mode>3_iwmmxt): Likewise.
15960 (sssubv8qi3): Likewise.
15961 (sssubv4hi3): Likewise.
15962 (sssubv2si3): Likewise.
15963 (ussubv8qi3): Likewise.
15964 (ussubv4hi3): Likewise.
15965 (ussubv2si3): Likewise.
15966 (mulv4hi3_iwmmxt): Likewise.
15967 (smulv4hi3_highpart): Likewise.
15968 (umulv4hi3_highpart): Likewise.
15969 (iwmmxt_wmacs): Likewise.
15970 (iwmmxt_wmacsz): Likewise.
15971 (iwmmxt_wmacu): Likewise.
15972 (iwmmxt_wmacuz): Likewise.
15973 (iwmmxt_clrdi): Likewise.
15974 (iwmmxt_clrv8qi): Likewise.
15975 (iwmmxt_clr4hi): Likewise.
15976 (iwmmxt_clr2si): Likewise.
15977 (iwmmxt_uavgrndv8qi3): Likewise.
15978 (iwmmxt_uavgrndv4hi3): Likewise.
15979 (iwmmxt_uavgv8qi3): Likewise.
15980 (iwmmxt_uavgv4hi3): Likewise.
15981 (iwmmxt_tinsrb): Likewise.
15982 (iwmmxt_tinsrh): Likewise.
15983 (iwmmxt_tinsrw): Likewise.
15984 (iwmmxt_textrmub): Likewise.
15985 (iwmmxt_textrmsb): Likewise.
15986 (iwmmxt_textrmuh): Likewise.
15987 (iwmmxt_textrmsh): Likewise.
15988 (iwmmxt_textrmw): Likewise.
15989 (iwmxxt_wshufh): Likewise.
15990 (eqv8qi3): Likewise.
15991 (eqv4hi3): Likewise.
15992 (eqv2si3): Likewise.
15993 (gtuv8qi3): Likewise.
15994 (gtuv4hi3): Likewise.
15995 (gtuv2si3): Likewise.
15996 (gtv8qi3): Likewise.
15997 (gtv4hi3): Likewise.
15998 (gtv2si3): Likewise.
15999 (smax<mode>3_iwmmxt): Likewise.
16000 (umax<mode>3_iwmmxt): Likewise.
16001 (smin<mode>3_iwmmxt): Likewise.
16002 (umin<mode>3_iwmmxt): Likewise.
16003 (iwmmxt_wpackhss): Likewise.
16004 (iwmmxt_wpackwss): Likewise.
16005 (iwmmxt_wpackdss): Likewise.
16006 (iwmmxt_wpackhus): Likewise.
16007 (iwmmxt_wpackwus): Likewise.
16008 (iwmmxt_wpackdus): Likewise.
16009 (iwmmxt_wunpckihb): Likewise.
16010 (iwmmxt_wunpckihh): Likewise.
16011 (iwmmxt_wunpckihw): Likewise.
16012 (iwmmxt_wunpckilb): Likewise.
16013 (iwmmxt_wunpckilh): Likewise.
16014 (iwmmxt_wunpckilw): Likewise.
16015 (iwmmxt_wunpckehub): Likewise.
16016 (iwmmxt_wunpckehuh): Likewise.
16017 (iwmmxt_wunpckehuw): Likewise.
16018 (iwmmxt_wunpckehsb): Likewise.
16019 (iwmmxt_wunpckehsh): Likewise.
16020 (iwmmxt_wunpckehsw): Likewise.
16021 (iwmmxt_wunpckelub): Likewise.
16022 (iwmmxt_wunpckeluh): Likewise.
16023 (iwmmxt_wunpckeluw): Likewise.
16024 (iwmmxt_wunpckelsb): Likewise.
16025 (iwmmxt_wunpckelsh): Likewise.
16026 (iwmmxt_wunpckelsw): Likewise.
16027 (ror<mode>3): Likewise.
16028 (ashr<mode>3_iwmmxt): Likewise.
16029 (lshr<mode>3_iwmmxt): Likewise.
16030 (ashl<mode>3_iwmmxt): Likewise.
16031 (ror<mode>3_di): Likewise.
16032 (ashr<mode>3_di): Likewise.
16033 (lshr<mode>3_di): Likewise.
16034 (ashl<mode>3_di): Likewise.
16035 (iwmmxt_wmadds): Likewise.
16036 (iwmmxt_wmaddu): Likewise.
16037 (iwmmxt_tmia): Likewise.
16038 (iwmmxt_tmiaph): Likewise.
16039 (iwmmxt_tmiabb): Likewise.
16040 (iwmmxt_tmiatb): Likewise.
16041 (iwmmxt_tmiabt): Likewise.
16042 (iwmmxt_tmiatt): Likewise.
16043 (iwmmxt_tmovmskb): Likewise.
16044 (iwmmxt_tmovmskh): Likewise.
16045 (iwmmxt_tmovmskw): Likewise.
16046 (iwmmxt_waccb): Likewise.
16047 (iwmmxt_wacch): Likewise.
16048 (iwmmxt_waccw): Likewise.
16049 (iwmmxt_waligni): Likewise.
16050 (iwmmxt_walignr): Likewise.
16051 (iwmmxt_walignr0): Likewise.
16052 (iwmmxt_walignr1): Likewise.
16053 (iwmmxt_walignr2): Likewise.
16054 (iwmmxt_walignr3): Likewise.
16055 (iwmmxt_wsadb): Likewise.
16056 (iwmmxt_wsadh): Likewise.
16057 (iwmmxt_wsadbz): Likewise.
16058 (iwmmxt_wsadhz): Likewise.
16059 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
16060 (iwmmxt_wabsdiffb): Likewise.
16061 (iwmmxt_wabsdiffh): Likewise.
16062 (iwmmxt_wabsdiffw): Likewise.
16063 (iwmmxt_waddsubhx): Likewise
16064 (iwmmxt_wsubaddhx): Likewise.
16065 (addc<mode>3): Likewise.
16066 (iwmmxt_avg4): Likewise.
16067 (iwmmxt_avg4r): Likewise.
16068 (iwmmxt_wmaddsx): Likewise.
16069 (iwmmxt_wmaddux): Likewise.
16070 (iwmmxt_wmaddsn): Likewise.
16071 (iwmmxt_wmaddun): Likewise.
16072 (iwmmxt_wmulwsm): Likewise.
16073 (iwmmxt_wmulwum): Likewise.
16074 (iwmmxt_wmulsmr): Likewise.
16075 (iwmmxt_wmulumr): Likewise.
16076 (iwmmxt_wmulwsmr): Likewise.
16077 (iwmmxt_wmulwumr): Likewise.
16078 (iwmmxt_wmulwl): Likewise.
16079 (iwmmxt_wqmulm): Likewise.
16080 (iwmmxt_wqmulwm): Likewise.
16081 (iwmmxt_wqmulmr): Likewise.
16082 (iwmmxt_wqmulwmr): Likewise.
16083 (iwmmxt_waddbhusm): Likewise.
16084 (iwmmxt_waddbhusl): Likewise.
16085 (iwmmxt_wqmiabb): Likewise.
16086 (iwmmxt_wqmiabt): Likewise.
16087 (iwmmxt_wqmiatb): Likewise.
16088 (iwmmxt_wqmiatt): Likewise.
16089 (iwmmxt_wqmiabbn): Likewise.
16090 (iwmmxt_wqmiabtn): Likewise.
16091 (iwmmxt_wqmiatbn): Likewise.
16092 (iwmmxt_wqmiattn): Likewise.
16093 (iwmmxt_wmiabb): Likewise.
16094 (iwmmxt_wmiabt): Likewise.
16095 (iwmmxt_wmiatb): Likewise.
16096 (iwmmxt_wmiatt): Likewise.
16097 (iwmmxt_wmiabbn): Likewise.
16098 (iwmmxt_wmiabtn): Likewise.
16099 (iwmmxt_wmiatbn): Likewise.
16100 (iwmmxt_wmiattn): Likewise.
16101 (iwmmxt_wmiawbb): Likewise.
16102 (iwmmxt_wmiawbt): Likewise.
16103 (iwmmxt_wmiawtb): Likewise.
16104 (iwmmxt_wmiawtt): Likewise.
16105 (iwmmxt_wmiawbbn): Likewise.
16106 (iwmmxt_wmiawbtn): Likewise.
16107 (iwmmxt_wmiawtbn): Likewise.
16108 (iwmmxt_wmiawttn): Likewise.
16109 (iwmmxt_wmerge): Likewise.
16110 (iwmmxt_tandc<mode>3): Likewise.
16111 (iwmmxt_torc<mode>3): Likewise.
16112 (iwmmxt_torvsc<mode>3): Likewise.
16113 (iwmmxt_textrc<mode>3): Likewise.
16114 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
16115 (wmmxt_pack): Likewise.
16116 (wmmxt_mult_c1): Likewise.
16117 (wmmxt_mult_c2): Likewise.
16118 (wmmxt_alu_c1): Likewise.
16119 (wmmxt_alu_c2): Likewise.
16120 (wmmxt_alu_c3): Likewise.
16121 (wmmxt_transfer_c1): Likewise.
16122 (wmmxt_transfer_c2): Likewise.
16123 (wmmxt_transfer_c3): Likewise.
16124 (marvell_f_iwmmxt_wstr): Likewise.
16125 (marvell_f_iwmmxt_wldr): Likewise.
16126
16127 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
16128
16129 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
16130
16131 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16132
16133 Revert:
16134 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16135 * lra-constraints.c (need_for_split_p): Check call used hard regs
16136 living through calls.
16137
16138 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
16139 call used regs for call insn.
16140
16141 2013-06-28 Jakub Jelinek <jakub@redhat.com>
16142
16143 PR target/57736
16144 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
16145 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
16146
16147 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16148
16149 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
16150
16151 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16152
16153 * lra-constraints.c (need_for_split_p): Check call used hard regs
16154 living through calls.
16155
16156 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
16157
16158 PR target/57744
16159 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
16160 to tie with any other modes. Eliminate Altivec vector mode tests,
16161 since these are a subset of ALTIVEC or VSX vector modes. Simplify
16162 code, to return 0 if testing MODE2 for a condition, if we've
16163 already tested MODE1 for the same condition.
16164
16165 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16166
16167 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
16168 layout.
16169
16170 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16171
16172 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
16173 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
16174
16175 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16176
16177 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
16178 Define.
16179 (aarch64_symbolic_constant_p): Remove.
16180 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
16181 static. Fix line length and white space.
16182 (aarch64_symbolic_constant_p): Remove.
16183 * config/aarch64/predicates.md (aarch64_valid_symref):
16184 Use aarch64_classify_symbol_expression.
16185
16186 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16187
16188 * config/arm/constraints.md (Ts): New constraint.
16189 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
16190 16-bit encodings.
16191 (compare_scc): Use "Ts" constraint for operand 0.
16192 (ior_scc_scc): Likewise.
16193 (and_scc_scc): Likewise.
16194 (and_scc_scc_nodom): Likewise.
16195 (ior_scc_scc_cmp): Likewise for operand 7.
16196 (and_scc_scc_cmp): Likewise.
16197 * config/arm/thumb2.md (thumb2_movsi_insn):
16198 Add alternatives for 16-bit encodings.
16199 (thumb2_movhi_insn): Likewise.
16200 (thumb2_movsicc_insn): Likewise.
16201 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
16202 (thumb2_negscc): Use "Ts" constraint.
16203 Move mvn instruction outside cond_exec block.
16204 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
16205 for 16-bit encodings.
16206
16207 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16208
16209 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
16210 encoding.
16211 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
16212 (mulsi3subsi): Likewise.
16213 (mulsidi3adddi): Likewise.
16214 (mulsidi3_v6): Likewise.
16215 (umulsidi3_v6): Likewise.
16216 (umulsidi3adddi_v6): Likewise.
16217 (smulsi3_highpart_v6): Likewise.
16218 (umulsi3_highpart_v6): Likewise.
16219 (mulhisi3tb): Likewise.
16220 (mulhisi3bt): Likewise.
16221 (mulhisi3tt): Likewise.
16222 (maddhisi4): Likewise.
16223 (maddhisi4tb): Likewise.
16224 (maddhisi4tt): Likewise.
16225 (maddhidi4): Likewise.
16226 (maddhidi4tb): Likewise.
16227 (maddhidi4tt): Likewise.
16228 (zeroextractsi_compare0_scratch): Likewise.
16229 (insv_zero): Likewise.
16230 (insv_t2): Likewise.
16231 (anddi_notzesidi_di): Likewise.
16232 (anddi_notsesidi_di): Likewise.
16233 (andsi_notsi_si): Likewise.
16234 (iordi_zesidi_di): Likewise.
16235 (xordi_zesidi_di): Likewise.
16236 (andsi_iorsi3_notsi): Likewise.
16237 (smax_0): Likewise.
16238 (smax_m1): Likewise.
16239 (smin_0): Likewise.
16240 (not_shiftsi): Likewise.
16241 (unaligned_loadsi): Likewise.
16242 (unaligned_loadhis): Likewise.
16243 (unaligned_loadhiu): Likewise.
16244 (unaligned_storesi): Likewise.
16245 (unaligned_storehi): Likewise.
16246 (extv_reg): Likewise.
16247 (extzv_t2): Likewise.
16248 (divsi3): Likewise.
16249 (udivsi3): Likewise.
16250 (arm_zero_extendhisi2addsi): Likewise.
16251 (arm_zero_extendqisi2addsi): Likewise.
16252 (compareqi_eq0): Likewise.
16253 (arm_extendhisi2_v6): Likewise.
16254 (arm_extendqisi2addsi): Likewise.
16255 (arm_movt): Likewise.
16256 (thumb2_ldrd): Likewise.
16257 (thumb2_ldrd_base): Likewise.
16258 (thumb2_ldrd_base_neg): Likewise.
16259 (thumb2_strd): Likewise.
16260 (thumb2_strd_base): Likewise.
16261 (thumb2_strd_base_neg): Likewise.
16262 (arm_negsi2): Add alternative for 16-bit encoding.
16263 (arm_one_cmplsi2): Likewise.
16264
16265 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16266
16267 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
16268 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
16269 (movdfcc): Likewise.
16270 * config/arm/vfp.md (*thumb2_movsf_vfp):
16271 Disable predication for arm_restrict_it.
16272 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
16273 (*thumb2_movdfcc_vfp): Likewise.
16274 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
16275 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
16276 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
16277 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
16278 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
16279 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
16280 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
16281 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
16282 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
16283 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
16284 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
16285 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
16286 Disable predication for arm_restrict_it.
16287
16288 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
16289
16290 * config/i386/bmiintrin.h (_bextr_u32): New.
16291 (_bextr_u64): Ditto.
16292
16293 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
16294
16295 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
16296 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
16297 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
16298 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
16299 * config/mips/n32-elf.h: ...this new file.
16300
16301 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
16302
16303 PR target/57224
16304 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
16305 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
16306
16307 2013-06-27 Catherine Moore <clm@codesourcery.com>
16308
16309 * config/mips/mips-tables.opt: Regenerate.
16310 * config/mips/mips-cpus.def: Add m14ke and m14kec.
16311 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
16312 * doc/invoke.texi: Add -m14kc.
16313
16314 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16315
16316 PR target/57623
16317 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
16318 constraints of operand 1 and 2.
16319
16320 PR target/57623
16321 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
16322 to match RTL canonicalization. Swap predicates and
16323 constraints of operand 1 and 2.
16324
16325 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
16326
16327 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
16328 Process OP_INOUT regs for splitting too.
16329
16330 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16331
16332 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
16333 decl before the loop, initialize to NULL.
16334 (vectorizable_load): Initialize ptr_incr to NULL.
16335
16336 2013-06-27 Martin Jambor <mjambor@suse.cz>
16337
16338 PR lto/57208
16339 * ipa-ref.h (ipa_maybe_record_reference): Declare.
16340 * ipa-ref.c (ipa_maybe_record_reference): New function.
16341 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
16342 * ipa-cp.c (create_specialized_node): Record potential references from
16343 aggvals.
16344 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
16345
16346 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16347
16348 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
16349 parameter 'mode' of type 'enum machine_mode mode'; change to pass
16350 'mode' to force_reg.
16351 (aarch64_add_offset): Update calls to aarch64_force_temporary.
16352 (aarch64_expand_mov_immediate): Likewise.
16353
16354 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16355
16356 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
16357 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
16358
16359 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16360
16361 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
16362 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16363 (struct machine_function): Add tbegin_p.
16364 (s390_canonicalize_comparison): Fold CC mode compares to
16365 conditional jump if possible.
16366 (s390_emit_jump): Return the emitted jump.
16367 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
16368 Handle CCRAWmode compares.
16369 (s390_option_override): Default to -mhtm if available.
16370 (s390_reg_clobbered_rtx): Handle floating point regs as well.
16371 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
16372 FPRs instead of df_regs_ever_live_p.
16373 (s390_optimize_nonescaping_tx): New function.
16374 (s390_init_frame_layout): Extend clobbered_regs array to cover
16375 FPRs as well.
16376 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
16377 (s390_expand_tbegin): New function.
16378 (enum s390_builtin): New enum definition.
16379 (code_for_builtin): New array definition.
16380 (s390_init_builtins): New function.
16381 (s390_expand_builtin): New function.
16382 (TARGET_INIT_BUILTINS): Define.
16383 (TARGET_EXPAND_BUILTIN): Define.
16384 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
16385 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
16386 (s390_alc_comparison): Likewise.
16387 * config/s390/s390-modes.def: Add CCRAWmode.
16388 * config/s390/s390.h (processor_flags): Add PF_TX.
16389 (TARGET_CPU_HTM): Define macro.
16390 (TARGET_HTM): Define macro.
16391 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
16392 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
16393 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16394 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
16395 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
16396 (TBEGIN_MASK, TBEGINC_MASK): New constants.
16397 ("*cc_to_int"): Move up.
16398 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
16399 constants other than 0.
16400 ("*ccraw_to_int"): New insn and splitter definition.
16401 ("tbegin", "tbegin_nofloat", "tbegin_retry")
16402 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
16403 ("tx_assist"): New expander.
16404 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
16405 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
16406 * config/s390/s390.opt: Add -mhtm option.
16407 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
16408 * config/s390/htmxlintrin.h: New file.
16409 * config/s390/htmintrin.h: New file.
16410 * config/s390/s390intrin.h: New file.
16411 * doc/extend.texi: Document htm builtins.
16412 * config.gcc: Add the new header files to extra_headers.
16413
16414 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
16415
16416 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
16417 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
16418
16419 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
16420 Pat Haugen <pthaugen@us.ibm.com>
16421 Peter Bergner <bergner@vnet.ibm.com>
16422
16423 * config/rs6000/power8.md: New.
16424 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
16425 setting for power8 entry.
16426 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
16427 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
16428 test for Power4/Power5 only.
16429 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
16430 support.
16431 (force_new_group): Adjust comment.
16432 * config/rs6000/rs6000.md: Include power8.md.
16433
16434 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
16435
16436 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
16437 * config/arm/arm-protos.h (arm_max_conditional_execute): New
16438 declaration.
16439 (tune_params): Update comment.
16440 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
16441 (arm_max_conditional_execute): New function.
16442 (thumb2_final_prescan_insn): Use max_insn_skipped and
16443 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
16444
16445 2013-06-25 Jakub Jelinek <jakub@redhat.com>
16446
16447 PR tree-optimization/57705
16448 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
16449 SSA_NAME step, provided that it is not defined inside the loop.
16450 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
16451 (get_initial_def_for_induction): Handle SSA_NAME IV step.
16452
16453 2013-06-25 Martin Jambor <mjambor@suse.cz>
16454
16455 PR middle-end/57670
16456 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
16457 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
16458 calls in the dump.
16459 (ipa_note_param_call): Initialize member_ptr flag.
16460 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
16461 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
16462 (ipa_write_indirect_edge_info): Stream member_ptr flag.
16463 (ipa_read_indirect_edge_info): Likewise.
16464
16465 2013-06-25 Richard Biener <rguenther@suse.de>
16466
16467 PR middle-end/56977
16468 * passes.c (init_optimization_passes): Move pass_fold_builtins
16469 and pass_dce earlier with -Og.
16470
16471 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
16472
16473 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
16474 <BIT_FIELD_REF>: Remove trailing TAB.
16475 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
16476 remove blank line.
16477
16478 2013-06-24 Martin Jambor <mjambor@suse.cz>
16479
16480 PR tree-optimization/57358
16481 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
16482 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
16483 (ipa_analyze_params_uses): Generate pessimistic info when true.
16484
16485 2013-06-24 Martin Jambor <mjambor@suse.cz>
16486
16487 PR tree-optimization/57539
16488 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
16489 global.inlined_to of the new node to it. All callers changed.
16490 * ipa-inline-transform.c (clone_inlined_nodes): New variable
16491 inlining_into, pass it to cgraph_clone_node.
16492 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
16493 ipa_free_edge_args_substructures.
16494 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
16495 rdesc linked list. Do not assert rdesc edges have inlined caller.
16496 Assert we have found an rdesc in the rdesc list.
16497
16498 2013-06-24 Richard Biener <rguenther@suse.de>
16499
16500 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
16501 (pointer_set_lookup): Declare.
16502 (class pointer_map): New template class implementing a
16503 generic pointer to T map.
16504 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
16505 pointer_map<T>::contains, pointer_map<T>::insert,
16506 pointer_map<T>::traverse): New functions.
16507 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
16508 (pointer_set_lookup): New function.
16509 (pointer_set_contains): Use pointer_set_lookup.
16510 (pointer_set_insert): Likewise.
16511 (insert_aux): Remove.
16512 (struct pointer_map_t): Embed a pointer_set_t.
16513 (pointer_map_create): Adjust.
16514 (pointer_map_destroy): Likewise.
16515 (pointer_map_contains): Likewise.
16516 (pointer_map_insert): Likewise.
16517 (pointer_map_traverse): Likewise.
16518 * tree-streamer.h (struct streamer_tree_cache_d): Use a
16519 pointer_map<unsigned> instead of a pointer_map_t.
16520 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
16521 (streamer_tree_cache_lookup): Likewise.
16522 (streamer_tree_cache_create): Likewise.
16523 (streamer_tree_cache_delete): Likewise.
16524 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
16525 pointer_map<unsigned> instead of a pointer_map_t.
16526 (lto_init_tree_ref_encoder): Adjust.
16527 (lto_destroy_tree_ref_encoder): Likewise.
16528 * lto-section-out.c (lto_output_decl_index): Likewise.
16529 (lto_record_function_out_decl_state): Likewise.
16530 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
16531
16532 2013-06-24 Richard Biener <rguenther@suse.de>
16533
16534 PR tree-optimization/57488
16535 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
16536
16537 2013-06-24 Alan Modra <amodra@gmail.com>
16538
16539 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
16540 (gen_easy_altivec_constant): Likewise.
16541 * config/rs6000/predicates.md (easy_vector_constant_add_self,
16542 easy_vector_constant_msb): Likewise.
16543
16544 2013-06-23 Jakub Jelinek <jakub@redhat.com>
16545
16546 PR target/57688
16547 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
16548 add missing return true.
16549
16550 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
16551
16552 PR target/52483
16553 * config/sh/predicates.md (general_extend_operand): Invoke
16554 general_movsrc_operand for memory operands.
16555 (general_movsrc_operand): Allow reg+reg addressing, do not use
16556 general_operand for memory operands.
16557
16558 2013-06-23 Sriraman Tallam <tmsriram@google.com>
16559
16560 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
16561 when current target options does not apply.
16562 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
16563 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
16564 * config/i386/bmiintrin.h: Pass appropriate target
16565 attributes to header.
16566 * config/i386/mmintrin.h: Ditto.
16567 * config/i386/nmmintrin.h: Ditto.
16568 * config/i386/avx2intrin.h: Ditto.
16569 * config/i386/fxsrintrin.h: Ditto.
16570 * config/i386/tbmintrin.h: Ditto.
16571 * config/i386/xsaveintrin.h: Ditto.
16572 * config/i386/f16cintrin.h: Ditto.
16573 * config/i386/xtestintrin.h: Ditto.
16574 * config/i386/xsaveoptintrin.h: Ditto.
16575 * config/i386/bmi2intrin.h: Ditto.
16576 * config/i386/lzcntintrin.h: Ditto.
16577 * config/i386/smmintrin.h: Ditto.
16578 * config/i386/wmmintrin.h: Ditto.
16579 * config/i386/x86intrin.h: Remove all header include guards.
16580 * config/i386/prfchwintrin.h: Ditto.
16581 * config/i386/pmmintrin.h: Ditto.
16582 * config/i386/tmmintrin.h: Ditto.
16583 * config/i386/xmmintrin.h: Ditto.
16584 * config/i386/popcntintrin.h: Ditto.
16585 * config/i386/rdseedintrin.h: Ditto.
16586 * config/i386/ammintrin.h: Ditto.
16587 * config/i386/emmintrin.h: Ditto.
16588 * config/i386/immintrin.h: Remove all header include guards.
16589 * config/i386/fma4intrin.h: Ditto.
16590 * config/i386/lwpintrin.h: Ditto.
16591 * config/i386/xopintrin.h: Ditto.
16592 * config/i386/ia32intrin.h: Ditto.
16593 * config/i386/avxintrin.h: Ditto.
16594 * config/i386/rtmintrin.h: Ditto.
16595 * config/i386/fmaintrin.h: Ditto.
16596 * config/i386/mm3dnow.h: Ditto.
16597
16598 2013-06-22 Sriraman Tallam <tmsriram@google.com>
16599
16600 * common/config/i386/i386-common.c: Handle LZCNT.
16601
16602 2013-06-22 Andi Kleen <ak@linux.intel.com>
16603
16604 * doc/extend.texi: Use __atomic_store_n instead of
16605 __atomic_store in HLE example.
16606
16607 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
16608
16609 * config/sh/sh.c: Remove <cstdlib> workaround.
16610
16611 2013-06-21 Andi Kleen <ak@linux.intel.com>
16612
16613 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
16614
16615 2013-06-21 Andi Kleen <ak@linux.intel.com>
16616
16617 * doc/extend.texi: Document that __atomic_clear and
16618 __atomic_test_and_set should only be used with bool.
16619
16620 2013-06-20 Jan Hubicka <jh@suse.cz>
16621
16622 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
16623 types_same_for_odr.
16624 * tree.c (decls_same_for_odr): New function.
16625 (same_for_edr): New function.
16626 (types_same_for_odr): New function.
16627 (get_binfo_at_offset): Use it.
16628 * tree.h (types_same_for_odr): Declare.
16629
16630 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
16631 Jason Merrill <jason@redhat.com>
16632
16633 * system.h: Include <cstdlib> as well as <stdlib.h>.
16634
16635 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
16636
16637 PR target/57655
16638 * config/i386/i386.c (construct_container): Report error if
16639 long double is used with disabled x87 float returns.
16640
16641 2013-06-20 Jan Hubicka <jh@suse.cz>
16642
16643 * lto-cgraph.c (input_symtab): Do not set cgraph state.
16644
16645 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
16646
16647 PR rtl-optimization/57425
16648 PR rtl-optimization/57569
16649 * alias.c (write_dependence_p): Remove parameters mem_mode and
16650 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
16651 Changed all callers.
16652 (canon_anti_dependence): Get comments and semantics in sync.
16653 Add parameter mem_canonicalized. Changed all callers.
16654 * rtl.h (canon_anti_dependence): Update prototype.
16655
16656 2013-06-20 Richard Biener <rguenther@suse.de>
16657
16658 * data-streamer-in.c (streamer_read_uhwi): Optimize single
16659 byte case, inline streamer_read_uchar and defer section
16660 overrun check.
16661
16662 2013-06-20 Richard Biener <rguenther@suse.de>
16663
16664 PR tree-optimization/57584
16665 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
16666 SSA names into the expanded expression that take part in
16667 abnormal coalescing.
16668
16669 2013-06-19 Sharad Singhai <singhai@google.com>
16670
16671 * gcov.c (print_usage): Handle new option.
16672 (process_args): Ditto.
16673 (get_gcov_intermediate_filename): New function.
16674 (output_intermediate_file): New function.
16675 (output_gcov_file): New function
16676 (generate_results): Handle new option.
16677 (release_function): Relase demangled name.
16678 (read_graph_file): Handle demangled name.
16679 (output_lines): Ditto.
16680 * doc/gcov.texi: Document gcov intermediate format.
16681
16682 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
16683
16684 PR bootstrap/57604
16685 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
16686 (lra_emit_add): Use the functions. Add comment about Y as an
16687 address segment.
16688
16689 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
16690
16691 PR driver/57652
16692 * collect2.c (collect_atexit): New.
16693 (collect_exit): Delete.
16694 (main): Register collect_atexit with atexit.
16695 (collect_wait): Change collect_exit to exit.
16696 (do_wait): Same.
16697 * collect2.h (collect_exit): Delete.
16698 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
16699
16700 2013-06-19 Wei Mi <wmi@google.com>
16701
16702 PR rtl-optimization/57518
16703 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
16704 if regno is used in paradoxical subreg.
16705 (update_equiv_regs): Check pdx_subregs[regno] before
16706 set a reg to be equivalent with a mem.
16707
16708 2013-06-19 Matthias Klose <doko@ubuntu.com>
16709
16710 PR driver/57651
16711 * file-find.h (find_a_file): Add a mode parameter.
16712 * file-find.c (find_a_file): Likewise.
16713 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
16714 with X_OK for the executables.
16715 * collect2.c (main): Call find_a_file with X_OK.
16716
16717 2013-06-19 Steve Ellcey <sellcey@mips.com>
16718
16719 PR target/56942
16720 * config/mips/mips.md (casesi_internal_mips16_<mode>):
16721 Use NEXT_INSN instead of next_real_insn.
16722
16723 2013-06-19 Jan Hubicka <jh@suse.cz>
16724
16725 * cgraph.h (const_value_known_p): Replace by ...
16726 (ctor_for_folding): .. this one.
16727 * cgraphunit.c (process_function_and_variable_attributes): Use it.
16728 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
16729 * expr.c (expand_expr_real_1): Likewise.
16730 (string_constant): Likewise.
16731 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
16732 * ipa.c (process_references): Likewise.
16733 (symtab_remove_unreachable_nodes): Likewise.
16734 * ipa-inline-analysis.c (param_change_prob): Likewise.
16735 * gimple-fold.c (canonicalize_constructor_val): Likewise.
16736 (get_base_constructor): Likwise.
16737 * varpool.c (varpool_remove_node): Likewise.
16738 (varpool_remove_initializer): LIkewise.
16739 (dump_varpool_node): LIkwise.
16740 (const_value_known_p): Rewrite to ...
16741 (ctor_for_folding): ... this one.
16742
16743 2013-06-19 Jakub Jelinek <jakub@redhat.com>
16744
16745 PR driver/57651
16746 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
16747 PERSONALITY in $PATH derived prefixes.
16748
16749 2013-06-19 Jeff Law <law@redhat.com>
16750
16751 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
16752 in comment.
16753
16754 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
16755 (simplify_bitwise_binary): Use it to simpify certain binary ops on
16756 booleans.
16757
16758 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
16759
16760 * config/arm/vfp.md: Move VFP instruction classification documentation
16761 to ...
16762 * config/arm/arm.md: ... here. Update instruction classification
16763 documentation.
16764
16765 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
16766
16767 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
16768 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
16769 pattern. Use more efficient sequences on ARMv5 and Thumb2.
16770
16771 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
16772
16773 PR target/57609
16774 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16775 with NEXT_INSN. Use tablejump_p to check for jump table data
16776 insns.
16777
16778 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
16779
16780 PR c++/56544
16781 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
16782 that now in C++ the value is correct per the C++ standards.
16783
16784 2013-06-19 Richard Biener <rguenther@suse.de>
16785
16786 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
16787 for global context.
16788
16789 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16790
16791 Revert:
16792 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16793
16794 PR target/57609
16795 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16796 with next_active_insn.
16797
16798 2013-06-18 Sriraman Tallam <tmsriram@google.com>
16799
16800 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
16801 functions are inlined during failures to flag an error.
16802 * tree-inline.c (expand_call_inline): Allow the error to be flagged
16803 in early inline pass.
16804
16805 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
16806
16807 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
16808 in comments.
16809
16810 2013-06-18 Julian Brown <julian@codesourcery.com>
16811
16812 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
16813 Permit virtual register pre-reload if !strict.
16814 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
16815 change.
16816 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
16817 prototype.
16818 * config/arm/neon.md (movmisalign<mode>): Use
16819 neon_perm_struct_or_reg_operand instead of
16820 neon_struct_or_register_operand.
16821 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
16822 neon_permissive_struct_operand instead of neon_struct_operand.
16823 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
16824 neon_vector_mem_operand.
16825 * config/arm/predicates.md (neon_struct_operand): Adjust call to
16826 neon_vector_mem_operand.
16827 (neon_permissive_struct_operand): New.
16828 (neon_struct_or_register_operand): Rename to...
16829 (neon_perm_struct_or_reg_operand): This. Adjust call to
16830 neon_vector_mem_operand.
16831
16832 2013-06-18 Richard Biener <rguenther@suse.de>
16833
16834 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
16835 * lto-streamer.h: Include pointer-set.h.
16836 (struct lto_decl_slot): Remove.
16837 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
16838 Remove next_index entry.
16839 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16840 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16841 (lto_init_tree_ref_encoder): Adjust.
16842 (lto_destroy_tree_ref_encoder): Likewise.
16843 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16844 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16845 (lto_output_decl_index): Adjust.
16846 (lto_new_out_decl_state): Likewise.
16847 (lto_record_function_out_decl_state): Likewise.
16848 * lto-streamer-out.c (copy_function): Likewise.
16849
16850 2013-06-18 Richard Biener <rguenther@suse.de>
16851
16852 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
16853 * cgraphunit.c: Include cfgloop.h.
16854 (init_lowered_empty_function): Initialize the loop tree.
16855 (assemble_thunk): Insert new BBs into loops.
16856
16857 2013-06-18 Richard Biener <rguenther@suse.de>
16858
16859 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
16860 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
16861 the map from cache entry to cache index optional.
16862 (streamer_tree_cache_replace_tree): Adjust accordingly.
16863 (streamer_tree_cache_append): Likewise.
16864 (streamer_tree_cache_delete): Likewise.
16865 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
16866 streamer cache map from cache entry to cache index.
16867 * lto-streamer-out.c (create_output_block): Adjust.
16868
16869 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
16870
16871 * config/arm/arm.md (attribute "insn"): Move multiplication and
16872 division attributes to...
16873 (attribute "type"): ... here. Remove mult.
16874 (attribute "mul32"): New attribute.
16875 (attribute "mul64"): Add umaal.
16876 (*arm_mulsi3): Update attributes.
16877 (*arm_mulsi3_v6): Likewise.
16878 (*thumb_mulsi3): Likewise.
16879 (*thumb_mulsi3_v6): Likewise.
16880 (*mulsi3_compare0): Likewise.
16881 (*mulsi3_compare0_v6): Likewise.
16882 (*mulsi_compare0_scratch): Likewise.
16883 (*mulsi_compare0_scratch_v6): Likewise.
16884 (*mulsi3addsi): Likewise.
16885 (*mulsi3addsi_v6): Likewise.
16886 (*mulsi3addsi_compare0): Likewise.
16887 (*mulsi3addsi_compare0_v6): Likewise.
16888 (*mulsi3addsi_compare0_scratch): Likewise.
16889 (*mulsi3addsi_compare0_scratch_v6): Likewise.
16890 (*mulsi3subsi): Likewise.
16891 (*mulsidi3adddi): Likewise.
16892 (*mulsi3addsi_v6): Likewise.
16893 (*mulsidi3adddi_v6): Likewise.
16894 (*mulsidi3_nov6): Likewise.
16895 (*mulsidi3_v6): Likewise.
16896 (*umulsidi3_nov6): Likewise.
16897 (*umulsidi3_v6): Likewise.
16898 (*umulsidi3adddi): Likewise.
16899 (*umulsidi3adddi_v6): Likewise.
16900 (*smulsi3_highpart_nov6): Likewise.
16901 (*smulsi3_highpart_v6): Likewise.
16902 (*umulsi3_highpart_nov6): Likewise.
16903 (*umulsi3_highpart_v6): Likewise.
16904 (mulhisi3): Likewise.
16905 (*mulhisi3tb): Likewise.
16906 (*mulhisi3bt): Likewise.
16907 (*mulhisi3tt): Likewise.
16908 (maddhisi4): Likewise.
16909 (*maddhisi4tb): Likewise.
16910 (*maddhisi4tt): Likewise.
16911 (maddhidi4): Likewise.
16912 (*maddhidi4tb): Likewise.
16913 (*maddhidi4tt): Likewise.
16914 (divsi3): Likewise.
16915 (udivsi3): Likewise.
16916 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
16917 (thumb2_mulsi_short_compare0): Likewise.
16918 (thumb2_mulsi_short_compare0_scratch): Likewise.
16919 * config/arm/arm1020e.md (1020mult1): Update attribute change.
16920 (1020mult2): Likewise.
16921 (1020mult3): Likewise.
16922 (1020mult4): Likewise.
16923 (1020mult5): Likewise.
16924 (1020mult6): Likewise.
16925 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
16926 change.
16927 (cortex_a15_mult64): Likewise.
16928 (cortex_a15_sdiv): Likewise.
16929 (cortex_a15_udiv): Likewise.
16930 * config/arm/arm1026ejs.md (mult1): Update attribute change.
16931 (mult2): Likewise.
16932 (mult3): Likewise.
16933 (mult4): Likewise.
16934 (mult5): Likewise.
16935 (mult6): Likewise.
16936 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
16937 (pj4_ir_div): Likewise.
16938 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
16939 (11_mult2): Likewise.
16940 (11_mult3): Likewise.
16941 (11_mult4): Likewise.
16942 (11_mult5): Likewise.
16943 (11_mult6): Likewise.
16944 (11_mult7): Likewise.
16945 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
16946 (cortex_a8_mla): Likewise.
16947 (cortex_a8_mull): Likewise.
16948 (cortex_a8_smulwy): Likewise.
16949 (cortex_a8_smlald): Likewise.
16950 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
16951 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
16952 (cortex_r4_mul_3): Likewise.
16953 (cortex_r4_mla_4): Likewise.
16954 (cortex_r4_mla_3): Likewise.
16955 (cortex_r4_smlald): Likewise.
16956 (cortex_r4_mull): Likewise.
16957 (cortex_r4_sdiv): Likewise.
16958 (cortex_r4_udiv): Likewise.
16959 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
16960 (cortex_a7_idiv): Likewise.
16961 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
16962 (9_mult2): Likewise.
16963 (9_mult3): Likewise.
16964 (9_mult4): Likewise.
16965 (9_mult5): Likewise.
16966 (9_mult6): Likewise.
16967 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
16968 (cortex_a53_sdiv): Likewise.
16969 (cortex_a53_udiv): Likewise.
16970 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
16971 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
16972 (mp626_mult2): Likewise.
16973 (mp626_mult3): Likewise.
16974 (mp626_mult4): Likewise.
16975 * config/arm/fa526.md (526_mult1): Update attribute change.
16976 (526_mult2): Likewise.
16977 * config/arm/arm-generic.md (mult): Update attribute change.
16978 (mult_ldsched_strongarm): Likewise.
16979 (mult_ldsched): Likewise.
16980 (multi_cycle): Likewise.
16981 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
16982 * config/arm/fa606te.md (606te_mult1): Update attribute change.
16983 (606te_mult2): Likewise.
16984 (606te_mult3): Likewise.
16985 (606te_mult4): Likewise.
16986 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
16987 (cortex_a9_mac16): Likewise.
16988 (cortex_a9_multiply): Likewise.
16989 (cortex_a9_mac): Likewise.
16990 (cortex_a9_multiply_long): Likewise.
16991 * config/arm/fa626te.md (626te_mult1): Update attribute change.
16992 (626te_mult2): Likewise.
16993 (626te_mult3): Likewise.
16994 (626te_mult4): Likewise.
16995
16996 2013-06-18 Richard Biener <rguenther@suse.de>
16997
16998 PR lto/57334
16999 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
17000
17001 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17002
17003 PR target/57609
17004 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17005 with next_active_insn.
17006
17007 2013-06-18 Alan Modra <amodra@gmail.com>
17008
17009 * config/rs6000/rs6000.h (enum data_align): New.
17010 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
17011 (DATA_ABI_ALIGNMENT): Define.
17012 (CONSTANT_ALIGNMENT): Correct comment.
17013 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
17014 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
17015
17016 2013-06-17 David Malcolm <dmalcolm@redhat.com>
17017
17018 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
17019 ATTRIBUTE_UNUSED marking.
17020
17021 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
17022
17023 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
17024 alternative and update.
17025 (aarch64_dup_lanedi): Delete.
17026 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
17027 * config/aarch64/aarch64-simd-builtins.def: Update.
17028
17029 2013-06-17 Richard Biener <rguenther@suse.de>
17030
17031 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
17032 (lto_input_scc): Declare.
17033 (lto_input_tree_1): Likewise.
17034 (struct lto_stats_d): Add num_tree_bodies_output and
17035 num_pickle_refs_output.
17036 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
17037 (lto_read_tree_1): Split out from ...
17038 (lto_read_tree): ... this.
17039 (lto_input_scc): New function.
17040 (lto_input_tree_1): Split out from ...
17041 (lto_input_tree): ... this. Handle LTO_tree_scc.
17042 (lto_data_in_create): Create the streamer cache without hashes.
17043 * lto-streamer-out.c (create_output_block): Create the streamer
17044 cache with hashes when not doing WPA.
17045 (lto_write_tree_1): Split out from ...
17046 (lto_write_tree): ... this.
17047 (get_symbol_initial_value): New function.
17048 (lto_output_tree_1): Split out from ...
17049 (lto_output_tree): ... this. Write trees as series of SCCs
17050 using a DFS walk via DFS_write_tree.
17051 (struct sccs, struct scc_entry): New types.
17052 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
17053 (DFS_write_tree_body): New function.
17054 (DFS_write_tree): Likewise.
17055 (hash_tree): Likewise.
17056 (scc_entry_compare): Likewise.
17057 (hash_scc): Likewise.
17058 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
17059 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
17060 TREE_CHAIN as regular reference.
17061 (streamer_read_integer_cst): Remove.
17062 (streamer_get_pickled_tree): Adjust.
17063 * tree-streamer-out.c (streamer_write_chain): Disable streaming
17064 of DECL_EXTERNALs in BLOCK_VARS for now.
17065 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
17066 reference.
17067 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
17068 Add hash value argument and record that if hashes are recorded
17069 in the cache.
17070 (streamer_tree_cache_insert_1): Adjust.
17071 (streamer_tree_cache_insert): Likewise.
17072 (streamer_tree_cache_insert_at): Rename to ...
17073 (streamer_tree_cache_replace_tree): ... this and adjust.
17074 (streamer_tree_cache_append): Adjust.
17075 (record_common_node): Likewise.
17076 (streamer_tree_cache_create): Add argument whether to
17077 record hash values together with trees.
17078 (streamer_tree_cache_delete): Adjust.
17079 * tree-streamer.h (struct streamer_tree_cache_d): Add
17080 vector of hashes.
17081 (streamer_read_integer_cst): Remove.
17082 (streamer_tree_cache_insert): Adjust.
17083 (streamer_tree_cache_append): Likewise.
17084 (streamer_tree_cache_insert_at): Rename to ...
17085 (streamer_tree_cache_replace_tree): ... this and adjust.
17086 (streamer_tree_cache_create): Add argument whether to record hashes.
17087 (streamer_tree_cache_get): Rename to ...
17088 (streamer_tree_cache_get_tree): ... this.
17089 (streamer_tree_cache_get_hash): New function.
17090 * tree.c (cache_integer_cst): New function.
17091 * tree.h (cache_integer_cst): Declare.
17092 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
17093 * lto-symtab.c (lto_varpool_replace_node): Only release
17094 DECL_INITIAL of non-prevailing decls.
17095 * varpool.c (varpool_remove_initializer): Do not release
17096 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
17097
17098 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17099
17100 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
17101 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
17102 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
17103 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
17104 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
17105 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
17106 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
17107 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
17108 instead of TARGET_64BIT.
17109 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
17110 Require ISA_HAS_<D>DIV.
17111
17112 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
17113
17114 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
17115 (mips*-*-linux*): Move default with_llsc setting to where other
17116 defaults are set.
17117 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
17118 with_arch block.
17119 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
17120 Likewise. Remove default with_tune setting. Move default float
17121 setting to its own block. Handle with_llsc in the same block as above.
17122
17123 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
17124
17125 PR rtl-optimization/57425
17126 PR rtl-optimization/57569
17127 * alias.c (write_dependence_p): Add new parameters mem_mode,
17128 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
17129 Changed all callers.
17130 (canon_anti_dependence): New function.
17131 * cse.c (check_dependence): Use canon_anti_dependence.
17132 * cselib.c (cselib_invalidate_mem): Likewise.
17133 * rtl.h (canon_anti_dependence): Declare.
17134
17135 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17136
17137 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
17138 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
17139 ".set mips3" for 64-bit targets.
17140
17141 2013-06-15 Dehao Chen <dehao@google.com>
17142
17143 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
17144 * gimple-low.c (gimple_check_call_matching_types): Likewise.
17145 (gimple_check_call_args): Likewise.
17146 * value-prof.c (check_ic_target): Likewise.
17147 * ipa-inline.c (early_inliner): Likewise.
17148 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
17149 * cgraph.c (cgraph_create_edge_1): Likewise.
17150 (cgraph_make_edge_direct): Likewise.
17151
17152 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
17153
17154 PR target/57615
17155 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
17156 rs6000_output_move_128bit to handle emitting quad memory
17157 operations. Set attribute length to 8 bytes.
17158
17159 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
17160
17161 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
17162 New pattern.
17163 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
17164 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
17165 (aarch64_<su>mlsl<mode>): Likewise.
17166
17167 2013-06-14 Mike Stump <mikestump@comcast.net>
17168
17169 * Makefile.in (TARGET_H): Add insn-codes.h.
17170
17171 2013-06-14 Alan Modra <amodra@gmail.com>
17172
17173 PR middle-end/57134
17174 PR middle-end/57586
17175 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
17176 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
17177 bitfield expansion when EXPAND_MEMORY.
17178 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
17179
17180 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
17181
17182 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
17183 test for clearing quad memory on 32-bit later.
17184
17185 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17186
17187 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
17188 (fold_negate_expr): Likewise.
17189 (fold_real_zero_addition_p): Handle vectors.
17190 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
17191
17192 2013-06-14 Alan Modra <amodra@gmail.com>
17193
17194 * varasm.c (force_const_mem): Revert 2013-06-07 change.
17195
17196 2013-06-13 Jan Hubicka <jh@suse.cz>
17197
17198 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17199 Local comdats are not externally visible.
17200 * symtab.c (dump_symtab_base): Dump externally visible.
17201 (verify_symtab_base): Verify back links in the symtab hash.
17202
17203 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17204
17205 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
17206 CONVERT_EXPR as equal nodes.
17207
17208 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17209
17210 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
17211
17212 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17213
17214 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
17215 Generalize to complex and vector.
17216 * tree.c (build_all_ones_cst): New function.
17217 * tree.h (build_all_ones_cst): Declare it.
17218
17219 2013-06-13 Alan Modra <amodra@gmail.com>
17220
17221 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
17222 * config/rs6000/rs6000.md (signbittf2): New insn.
17223 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
17224 (abstf2_internal, cmptf_internal2): Likewise.
17225 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
17226
17227 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
17228 Pat Haugen <pthaugen@us.ibm.com>
17229 Peter Bergner <bergner@vnet.ibm.com>
17230
17231 * config/rs6000/rs6000.c (emit_load_locked): Add support for
17232 power8 byte, half-word, and quad-word atomic instructions.
17233 (emit_store_conditional): Likewise.
17234 (rs6000_expand_atomic_compare_and_swap): Likewise.
17235 (rs6000_expand_atomic_op): Likewise.
17236
17237 * config/rs6000/sync.md (larx): Add new modes for power8.
17238 (stcx): Likewise.
17239 (AINT): New mode iterator to include TImode as well as normal
17240 integer modes on power8.
17241 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
17242 that VSX registers are not considered. Use AINT mode iterator
17243 instead of INT1 to allow inclusion of quad word atomic operations
17244 on power8.
17245 (load_locked<mode>): Likewise.
17246 (store_conditional<mode>): Likewise.
17247 (atomic_compare_and_swap<mode>): Likewise.
17248 (atomic_exchange<mode>): Likewise.
17249 (atomic_nand<mode>): Likewise.
17250 (atomic_fetch_<fetchop_name><mode>): Likewise.
17251 (atomic_nand_fetch<mode>): Likewise.
17252 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
17253 each type.
17254 (ATOMIC): On power8, add QImode, HImode modes.
17255 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
17256 modes that promote to SImode.
17257 (load_lockedti): Convert TImode arguments to PTImode, so that we
17258 get a guaranteed even/odd register pair.
17259 (load_lockedpti): Likewise.
17260 (store_conditionalti): Likewise.
17261 (store_conditionalpti): Likewise.
17262
17263 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
17264 atomic load/store instructions.
17265 (HSI): Likewise.
17266
17267 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
17268
17269 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
17270 loads.
17271 (insn_count): New attribute, with most cases extracted from...
17272 (length): ...here. Redefine most cases in terms of insn_count.
17273 (single_insn): Delete.
17274 (can_delay): Use insn_count to check for single instructions.
17275 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
17276 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
17277 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
17278 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
17279 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
17280 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
17281 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
17282 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
17283 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
17284 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
17285 rather than "length".
17286 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
17287 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
17288 Use "insn_count" rather than "length".
17289 * config/mips/mips-dsp.md
17290 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
17291 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
17292 length attributes.
17293
17294 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
17295
17296 PR tree-optimization/57361
17297 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
17298
17299 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
17300
17301 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
17302 to split.
17303 (aarch64_simd_combine<mode>): New instruction expansion.
17304 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
17305 function prototype.
17306 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
17307 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
17308
17309 2013-06-12 Jan Hubicka <jh@suse.cz>
17310
17311 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
17312 decl has when in streaming stage.
17313 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
17314 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
17315
17316 2013-06-12 Roland Stigge <stigge@antcom.de>
17317
17318 PR target/57578
17319 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
17320
17321 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17322
17323 PR tree-optimization/57537
17324 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
17325 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
17326
17327 2013-06-12 Richard Biener <rguenther@suse.de>
17328
17329 * data-streamer.h (streamer_write_char_stream): CSE
17330 obs->current_pointer.
17331 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
17332 streamer_write_char_stream manually and optimize the resulting loop.
17333 (streamer_write_hwi_stream): Likewise.
17334
17335 2013-06-12 Jan Hubicka <jh@suse.cz>
17336
17337 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
17338 * cgraph.h (varpool_create_empty_node): Declare.
17339 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
17340 duplicated nodes.
17341 * symtab.c (symtab_unregister_node): Be lax about missin entries
17342 in node hash.
17343 (symtab_get_node): Update comment.
17344 * varpool.c (varpool_create_empty_node): Break out from ...
17345 (varpool_node_for_decl): ... here.
17346 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
17347
17348 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
17349
17350 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
17351 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
17352 part. Use straight-line flow at the end.
17353 <COMPONENT_REF>: Remove superfluous else.
17354 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
17355
17356 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17357
17358 PR target/56564
17359 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
17360 target hook even for !TREE_PUBLIC decls. If no resolution info
17361 is available, return false for common and external decls.
17362
17363 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17364
17365 * config/rl78/constraints.md (U): New constraint.
17366 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
17367 valloc attribute.
17368
17369 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17370
17371 PR target/57589
17372 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
17373 to allow returning address to AT_PLATFORM name.
17374
17375 2013-06-11 Jan Hubicka <jh@suse.cz>
17376
17377 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
17378 * cgraph.h (symtab_node_base): Add weakref flag.
17379 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
17380 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
17381 (output_weakrefs): Use weakref flag.
17382 * fold-const.c (simple_operand_p): Handle WEAK.
17383 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
17384 * ipa.c (varpool_externally_visible_p): Drop weakref.
17385 (function_and_variable_visibility): Update comment; fix weakref
17386 sanity checks; do not clear DECL_WEAK on them.
17387 * lto-cgraph.c (lto_output_node): update.
17388 (lto_output_varpool_node): Update.
17389 (input_overwrite_node): Update.
17390 (input_node): Update.
17391 (input_varpool_node): Update.
17392 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
17393 (lto_symtab_merge_symbols): Add sanity check.
17394 (lto_symtab_prevailing_decl): Do not special case weakrefs.
17395 * passes.c (rest_of_decl_compilation): Set static flag, too.
17396 * symtab.c (dump_symtab_base): Dump weakref.
17397 (verify_symtab_base): Sanity check weakrefs.
17398 (symtab_make_decl_local): Remove duplicated code.
17399 (symtab_alias_ultimate_target): Simplify.
17400 * varpool.c (varpool_create_variable_alias): Set weakref flag.
17401
17402 2013-06-11 Tom de Vries <tom@codesourcery.com>
17403
17404 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
17405 sequence_vect == NULL.
17406
17407 2013-06-11 DJ Delorie <dj@redhat.com>
17408
17409 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
17410 (rl78_unwind_word_mode): New.
17411
17412 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17413
17414 * final.c (debug_prefix_maps): Make static.
17415
17416 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17417
17418 * function.c (initial_trampoline): Remove stray copy.
17419
17420 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
17421
17422 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
17423
17424 2013-06-11 Martin Jambor <mjambor@suse.cz>
17425
17426 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
17427 within bounds at the beginning of the function.
17428
17429 2013-06-11 Alan Modra <amodra@gmail.com>
17430
17431 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
17432 reporting.
17433 (get_named_section): Don't NULL !DECL_P decl.
17434
17435 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
17436
17437 * doc/invoke.texi (core-avx2): Document.
17438 (slm): Likewise.
17439 (atom): Updated with MOVBE.
17440
17441 2013-06-11 Richard Biener <rguenther@suse.de>
17442
17443 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
17444
17445 2013-06-11 Anton Blanchard <anton@samba.org>
17446
17447 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
17448 correct shift value in little-endian mode.
17449
17450 2013-06-11 Jakub Jelinek <jakub@redhat.com>
17451
17452 PR target/56564
17453 * varasm.c (get_variable_align): Move #endif to the right place.
17454
17455 2013-06-10 Cary Coutant <ccoutant@google.com>
17456
17457 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
17458 for hash so that hash table traversal order is deterministic.
17459
17460 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
17461 Pat Haugen <pthaugen@us.ibm.com>
17462 Peter Bergner <bergner@vnet.ibm.com>
17463
17464 * config/rs6000/vector.md (GPR move splitter): Do not split moves
17465 of vectors in GPRS if they are direct moves or quad word load or
17466 store moves.
17467
17468 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
17469 declaration.
17470 (direct_move_p): Likewise.
17471 (quad_load_store_p): Likewise.
17472
17473 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
17474 classes into bins based on the physical register type.
17475 (reg_class_to_reg_type): Likewise.
17476 (IS_STD_REG_TYPE): Likewise.
17477 (IS_FP_VECT_REG_TYPE): Likewise.
17478 (reload_fpr_gpr): Arrays to determine what insn to use if we can
17479 use direct move instructions.
17480 (reload_gpr_vsx): Likewise.
17481 (reload_vsx_gpr): Likewise.
17482 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
17483 information that is a simplification of register classes. Also
17484 precalculate direct move reload helpers.
17485 (direct_move_p): New function to return true if the operation can
17486 be done as a direct move instruciton.
17487 (quad_load_store_p): New function to return true if the operation
17488 is a quad memory operation.
17489 (rs6000_legitimize_address): If quad memory, only allow register
17490 indirect for TImode addresses.
17491 (rs6000_legitimate_address_p): Likewise.
17492 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
17493 (rs6000_reload_register_type): Likewise.
17494 (register_to_reg_type): Return register type.
17495 (rs6000_secondary_reload_simple_move): New helper function for
17496 secondary reload and secondary memory needed to identify anything
17497 that is a simple move, and does not need reloading.
17498 (rs6000_secondary_reload_direct_move): New helper function for
17499 secondary reload to identify cases that can be done with several
17500 instructions via the direct move instructions.
17501 (rs6000_secondary_reload_move): New helper function for secondary
17502 reload to identify moves between register types that can be done.
17503 (rs6000_secondary_reload): Add support for quad memory operations
17504 and for direct move.
17505 (rs6000_secondary_memory_needed): Likewise.
17506 (rs6000_debug_secondary_memory_needed): Change argument names.
17507 (rs6000_output_move_128bit): New function to return the move to
17508 use for 128-bit moves, including knowing about the various
17509 limitations of quad memory operations.
17510
17511 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
17512 memory operations. call rs6000_output_move_128bit for the actual
17513 instruciton(s) to generate.
17514 (vsx_movti_64bit): Likewise.
17515
17516 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
17517 (UNSPEC_P8V_MTVSRWZ): Likewise.
17518 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
17519 (UNSPEC_P8V_MTVSRD): Likewise.
17520 (UNSPEC_P8V_XXPERMDI): Likewise.
17521 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
17522 (UNSPEC_FUSION_GPR): Likewise.
17523 (FMOVE128_GPR): New iterator for direct move.
17524 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
17525 (f32_sv): Likewise.
17526 (f32_dm): Likewise.
17527 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
17528 loads and direct move instructions.
17529 (zero_extendsidi2_lfiwzx): Likewise.
17530 (extendsidi2_lfiwax): Likewise.
17531 (extendsidi2_nocell): Likewise.
17532 (floatsi<mode>2_lfiwax): Likewise.
17533 (lfiwax): Likewise.
17534 (floatunssi<mode>2_lfiwzx): Likewise.
17535 (lfiwzx): Likewise.
17536 (fix_trunc<mode>_stfiwx): Likewise.
17537 (fixuns_trunc<mode>_stfiwx): Likewise.
17538 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
17539 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
17540 (parity<mode>2_cmpb): Set length/type attr.
17541 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
17542 for 'mr.' to fast_compare.
17543 (bpermd_<mode>): Change type attr to popcnt.
17544 (p8_fmrgow_<mode>): New insns for power8 direct move support.
17545 (p8_mtvsrwz_1): Likewise.
17546 (p8_mtvsrwz_2): Likewise.
17547 (reload_fpr_from_gpr<mode>): Likewise.
17548 (p8_mtvsrd_1): Likewise.
17549 (p8_mtvsrd_2): Likewise.
17550 (p8_xxpermdi_<mode>): Likewise.
17551 (reload_vsx_from_gpr<mode>): Likewise.
17552 (reload_vsx_from_gprsf): Likewise.
17553 (p8_mfvsrd_3_<mode>): LIkewise.
17554 (reload_gpr_from_vsx<mode>): Likewise.
17555 (reload_gpr_from_vsxsf): Likewise.
17556 (p8_mfvsrd_4_disf): Likewise.
17557 (multi-word GPR splits): Do not split direct moves or quad memory
17558 operations.
17559
17560 2013-06-10 David Malcolm <dmalcolm@redhat.com>
17561
17562 * tree-into-ssa.c (interesting_blocks): Make static.
17563
17564 2013-06-10 Jakub Jelinek <jakub@redhat.com>
17565
17566 PR target/56564
17567 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
17568 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
17569 Use DATA_ABI_ALIGNMENT for that case instead if defined.
17570 (get_variable_align): New function.
17571 (get_variable_section, emit_bss, emit_common,
17572 assemble_variable_contents, place_block_symbol): Use
17573 get_variable_align instead of DECL_ALIGN.
17574 (assemble_noswitch_variable): Add align argument, use it
17575 instead of DECL_ALIGN.
17576 (assemble_variable): Adjust caller. Use get_variable_align
17577 instead of DECL_ALIGN.
17578 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
17579 caller.
17580 (DATA_ABI_ALIGNMENT): Define.
17581 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
17582 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
17583 opt is false, only return the psABI mandated alignment increase.
17584 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
17585 (DATA_ABI_ALIGNMENT): ... this.
17586 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
17587 (DATA_ABI_ALIGNMENT): ... this.
17588 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
17589 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
17590 (DATA_ABI_ALIGNMENT): ... this.
17591 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
17592 * doc/tm.texi: Regenerated.
17593
17594 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
17595
17596 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
17597 cmp_code to construct REG_EQUAL note.
17598
17599 2013-06-09 Jakub Jelinek <jakub@redhat.com>
17600
17601 PR target/57568
17602 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
17603 that operands[2] doesn't overlap with operands[0].
17604
17605 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
17606 Jan Hubicka <jh@suse.cz>
17607
17608 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
17609 hack to mark symbols as used.
17610
17611 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
17612
17613 PR rtl-optimization/57559
17614 * lra-constraints.c (process_alt_operands): Don't discourage
17615 memory with known offset for offsetable memory constraint.
17616 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
17617
17618 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
17619
17620 * varasm.c (struct oc_local_state): Reorder fields.
17621 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
17622 and adjust accordingly.
17623 (output_constructor): Reorder initialization code and adjust call to
17624 output_constructor_bitfield.
17625
17626 2013-06-07 Jan Hubicka <jh@suse.cz>
17627
17628 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
17629
17630 2013-06-07 David Malcolm <dmalcolm@redhat.com>
17631
17632 * tree-object-size.c (unknown): Make const.
17633
17634 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17635
17636 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
17637 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
17638 for last alternative in the cpu_facility attribute.
17639
17640 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17641
17642 PR target/56315
17643 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
17644 (xordi3): Change operand 2 constraint to arm_xordi_operand.
17645 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
17646 * config/arm/constraints.md (Dg): New constraint.
17647 * config/arm/neon.md (xordi3_neon): Remove.
17648 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
17649 * config/arm/predicates.md (arm_xordi_operand): New predicate.
17650
17651 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17652
17653 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
17654 Clean up alternatives.
17655
17656 2013-06-07 Alan Modra <amodra@gmail.com>
17657
17658 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
17659 va_list_gpr_size.
17660
17661 2013-06-07 Alan Modra <amodra@gmail.com>
17662
17663 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
17664
17665 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17666
17667 * config/arm/constraints.md (Df): New constraint.
17668 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
17669 Correct length attribute for last two alternatives.
17670
17671 2013-06-07 Alan Modra <amodra@gmail.com>
17672
17673 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
17674 override user -mfp-in-toc.
17675 (offsettable_ok_by_alignment): Consider just the current access
17676 rather than the whole object, unless BLKmode. Handle
17677 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
17678 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
17679 for -mcmodel=medium.
17680 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
17681 override user -mfp-in-toc or -msum-in-toc. Default to
17682 -mno-fp-in-toc for -mcmodel=medium.
17683
17684 2013-06-06 DJ Delorie <dj@redhat.com>
17685
17686 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
17687 TARGET_VALID_POINTER_MODE.
17688
17689 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
17690 Pat Haugen <pthaugen@us.ibm.com>
17691 Peter Bergner <bergner@vnet.ibm.com>
17692
17693 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
17694 Document new power8 builtins.
17695
17696 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
17697 condition code register, to allow 128-bit logical operations to be
17698 done in the VSX or GPR registers.
17699 (nor<mode>3): Use the canonical form for nor.
17700 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
17701 vclz*, and vpopcnt* vector instructions.
17702 (nand<mode>3): Likewise.
17703 (orc<mode>3): Likewise.
17704 (clz<mode>2): LIkewise.
17705 (popcount<mode>2): Likewise.
17706
17707 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
17708 that only the GPRs are recognized.
17709
17710 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
17711 support for new power8 builtins.
17712
17713 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
17714 builtin functions.
17715 (xscvdpspn): Likewise.
17716 (vclz): Likewise.
17717 (vclzb): Likewise.
17718 (vclzh): Likewise.
17719 (vclzw): Likewise.
17720 (vclzd): Likewise.
17721 (vpopcnt): Likewise.
17722 (vpopcntb): Likewise.
17723 (vpopcnth): Likewise.
17724 (vpopcntw): Likewise.
17725 (vpopcntd): Likewise.
17726 (vgbbd): Likewise.
17727 (vmrgew): Likewise.
17728 (vmrgow): Likewise.
17729 (eqv): Likewise.
17730 (eqv_v16qi3): Likewise.
17731 (eqv_v8hi3): Likewise.
17732 (eqv_v4si3): Likewise.
17733 (eqv_v2di3): Likewise.
17734 (eqv_v4sf3): Likewise.
17735 (eqv_v2df3): Likewise.
17736 (nand): Likewise.
17737 (nand_v16qi3): Likewise.
17738 (nand_v8hi3): Likewise.
17739 (nand_v4si3): Likewise.
17740 (nand_v2di3): Likewise.
17741 (nand_v4sf3): Likewise.
17742 (nand_v2df3): Likewise.
17743 (orc): Likewise.
17744 (orc_v16qi3): Likewise.
17745 (orc_v8hi3): Likewise.
17746 (orc_v4si3): Likewise.
17747 (orc_v2di3): Likewise.
17748 (orc_v4sf3): Likewise.
17749 (orc_v2df3): Likewise.
17750
17751 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
17752 allow power8 quad mode in 64-bit.
17753 (rs6000_builtin_vectorized_function): Add support to vectorize
17754 ISA 2.07 count leading zeros, population count builtins.
17755 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
17756 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
17757 (builtin_function_type): Add vgbbd builtin function which takes an
17758 unsigned argument.
17759 (altivec_expand_vec_perm_const): Add support for new power8 merge
17760 instructions.
17761
17762 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
17763 that does not include TImdoe for use with 32-bit.
17764 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
17765 instructions.
17766 (UNSPEC_VSX_CVDPSPN): Likewise.
17767 (vsx_xscvdpspn): Likewise.
17768 (vsx_xscvspdpn): Likewise.
17769 (vsx_xscvdpspn_scalar): Likewise.
17770 (vsx_xscvspdpn_directmove): Likewise.
17771 (vsx_and<mode>3): Split logical operations into 32-bit and
17772 64-bit. Add support to do logical operations on TImode as well as
17773 VSX vector types. Allow logical operations to be done in either
17774 VSX registers or in general purpose registers in 64-bit mode. Add
17775 splitters if GPRs were used. For AND, add clobber of CCmode to
17776 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
17777 encoding.
17778 (vsx_and<mode>3_32bit): Likewise.
17779 (vsx_and<mode>3_64bit): Likewise.
17780 (vsx_ior<mode>3): Likewise.
17781 (vsx_ior<mode>3_32bit): Likewise.
17782 (vsx_ior<mode>3_64bit): Likewise.
17783 (vsx_xor<mode>3): Likewise.
17784 (vsx_xor<mode>3_32bit): Likewise.
17785 (vsx_xor<mode>3_64bit): Likewise.
17786 (vsx_one_cmpl<mode>2): Likewise.
17787 (vsx_one_cmpl<mode>2_32bit): Likewise.
17788 (vsx_one_cmpl<mode>2_64bit): Likewise.
17789 (vsx_nor<mode>3): Likewise.
17790 (vsx_nor<mode>3_32bit): Likewise.
17791 (vsx_nor<mode>3_64bit): Likewise.
17792 (vsx_andc<mode>3): Likewise.
17793 (vsx_andc<mode>3_32bit): Likewise.
17794 (vsx_andc<mode>3_64bit): Likewise.
17795 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
17796 and xxlorc instructions.
17797 (vsx_eqv<mode>3_64bit): Likewise.
17798 (vsx_nand<mode>3_32bit): Likewise.
17799 (vsx_nand<mode>3_64bit): Likewise.
17800 (vsx_orc<mode>3_32bit): Likewise.
17801 (vsx_orc<mode>3_64bit): Likewise.
17802
17803 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
17804
17805 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
17806 instruction.
17807 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
17808 (p8_vmrgow): Likewise.
17809 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
17810 GPRs to be split under VSX.
17811 (p8v_clz<mode>2): Add power8 count leading zero support.
17812 (p8v_popcount<mode>2): Add power8 population count support.
17813 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
17814 support.
17815
17816 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
17817 instruction.
17818
17819 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
17820 builtin functions.
17821 (vec_nand): Likewise.
17822 (vec_vclz): Likewise.
17823 (vec_vclzb): Likewise.
17824 (vec_vclzd): Likewise.
17825 (vec_vclzh): Likewise.
17826 (vec_vclzw): Likewise.
17827 (vec_vgbbd): Likewise.
17828 (vec_vmrgew): Likewise.
17829 (vec_vmrgow): Likewise.
17830 (vec_vpopcnt): Likewise.
17831 (vec_vpopcntb): Likewise.
17832 (vec_vpopcntd): Likewise.
17833 (vec_vpopcnth): Likewise.
17834 (vec_vpopcntw): Likewise.
17835
17836 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17837
17838 PR rtl-optimization/57468
17839 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
17840 spilled pseudos.
17841
17842 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17843
17844 PR rtl-optimization/57459
17845 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
17846 type when setting live regs.
17847
17848 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17849
17850 * config/s390/s390.opt (mlra): New option.
17851 * config/s390/s390.c (s390_decompose_address): Check displacement
17852 for all registers for LRA.
17853 (s390_secondary_reload): Don't used secondary reloads for LRA.
17854 (s390_lra_p): New function.
17855 (TARGET_LRA_P): Define.
17856 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
17857 of attribute cpu_facility to zarch for the last alternative.
17858 (*cmpmem_short): Ditto.
17859
17860 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
17861
17862 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
17863 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
17864 (arm_expand_prologue): Likewise.
17865
17866 2013-06-06 Teresa Johnson <tejohnson@google.com>
17867
17868 PR c++/53743
17869 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
17870 as this is now done by redirect_edge_and_branch_force.
17871 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
17872 barriers, and fix interaction with splitting.
17873 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
17874 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
17875 reflect changes made in the routine.
17876 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
17877 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
17878 since this is called in cfglayout mode, and replace partition fixup
17879 with assert as that is now done by force_nonfallthru_and_redirect.
17880 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
17881 already be marked with region crossing note.
17882 (insert_section_boundary_note): Make non-static, gate on flag
17883 has_bb_partition, rewrite to also check for multiple partitions.
17884 (rest_of_handle_reorder_blocks): Remove call to
17885 insert_section_boundary_note, now done later during free_cfg.
17886 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
17887 * bb-reorder.h (insert_section_boundary_note): Declare.
17888 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
17889 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
17890 invoke insert_section_boundary_note.
17891 (try_redirect_by_replacing_jump): Remove unnecessary
17892 check for region crossing note.
17893 (fixup_partition_crossing): New function.
17894 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
17895 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
17896 in non-cfglayout mode.
17897 (force_nonfallthru_and_redirect): Fixup partition boundaries,
17898 remove old code that tried to do this. Emit barrier correctly
17899 when we are in cfglayout mode.
17900 (last_bb_in_partition): New function.
17901 (rtl_split_edge): Correctly fixup partition boundaries.
17902 (commit_one_edge_insertion): Remove old code that tried to
17903 fixup region crossing edge since this is now handled in
17904 split_block, and set up insertion point correctly since
17905 block may now end in a jump.
17906 (verify_hot_cold_block_grouping): Guard against checking when not in
17907 linearized RTL mode.
17908 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
17909 notes.
17910 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
17911 rtl_verify_flow_info, so not called in cfglayout mode.
17912 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
17913 (fixup_reorder_chain): Remove old code that attempted to fixup region
17914 crossing note as this is now handled in force_nonfallthru_and_redirect.
17915 (duplicate_insn_chain): Don't duplicate switch section notes.
17916 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
17917 note.
17918 * basic-block.h (emit_barrier_after_bb): Declare.
17919
17920 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17921
17922 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
17923 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
17924 arm_usatsihi): Adjust alternatives for arm_restrict_it.
17925
17926 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17927
17928 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
17929 where appropriate.
17930 * config/arm/ldmstm.md: Regenerate.
17931
17932 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17933
17934 * config/arm/sync.md (atomic_loaddi_1):
17935 Disable predication for arm_restrict_it.
17936 (arm_load_exclusive<mode>): Likewise.
17937 (arm_load_exclusivesi): Likewise.
17938 (arm_load_exclusivedi): Likewise.
17939 (arm_load_acquire_exclusive<mode>): Likewise.
17940 (arm_load_acquire_exclusivesi): Likewise.
17941 (arm_load_acquire_exclusivedi): Likewise.
17942 (arm_store_exclusive<mode>): Likewise.
17943 (arm_store_exclusive<mode>): Likewise.
17944 (arm_store_release_exclusivedi): Likewise.
17945 (arm_store_release_exclusive<mode>): Likewise.
17946
17947 2013-06-06 Richard Biener <rguenther@suse.de>
17948
17949 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
17950 after LTO_null.
17951 (lto_tag_is_tree_code_p): Adjust.
17952 (lto_tag_is_gimple_code_p): Likewise.
17953 (lto_gimple_code_to_tag): Likewise.
17954 (lto_tag_to_gimple_code): Likewise.
17955 (lto_tree_code_to_tag): Likewise.
17956 (lto_tag_to_tree_code): Likewise.
17957 * data-streamer.h (streamer_write_hwi_in_range): Use
17958 uhwi streaming to stream the normalized range.
17959 (streamer_read_hwi_in_range): Likewise.
17960
17961 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17962
17963 * config/arm/arm.md (enabled_for_depr_it): New attribute.
17964 (predicable_short_it): Likewise.
17965 (predicated): Likewise.
17966 (enabled): Handle above.
17967 (define_cond_exec): Set predicated attribute to yes.
17968
17969 2013-06-05 Mike Stump <mikestump@comcast.net>
17970
17971 * gdbinit.in (__FUNCTION__): Add.
17972
17973 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
17974
17975 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
17976 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
17977
17978 2013-06-05 Jan Hubicka <jh@suse.cz>
17979
17980 * varasm.c (mark_decl_referenced): Revert the removal until targets
17981 are fixed.
17982
17983 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
17984
17985 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
17986 instead of mark_decl_referenced.
17987
17988 2013-06-05 Jan Hubicka <jh@suse.cz>
17989
17990 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
17991 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
17992 and symtab_used_from_object_file_p.
17993 (cgraph_make_node_local_1): Clear forced_by_abi.
17994 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
17995 * cgraph.h (symtab_node_base): Add forced_by_abi.
17996 (decide_is_variable_needed): Remove.
17997 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
17998 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
17999 (decide_is_symbol_needed): ... this one; handle symbols in general;
18000 always analyze virtuals; honnor forced_by_abi.
18001 (cgraph_finalize_function): Update.
18002 (varpool_finalize_decl): Update.
18003 (symbol_defined_and_needed): Remove.
18004 (analyze_functions): Update.
18005 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18006 output_refs, input_overwrite_node): Handle forced_by_abi.
18007 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
18008 (address_taken_from_non_vtable_p): ... this one.
18009 (comdat_can_be_unshared_p_1): New function.
18010 (cgraph_comdat_can_be_unshared_p): Rename to ...
18011 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
18012 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
18013 (function_and_variable_visibility): Clear forced_by_abi as needed.
18014 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
18015 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
18016 * symtab.c (dump_symtab_base): Dump forced_by_abi.
18017 * varpool.c (decide_is_variable_needed): Remove.
18018
18019 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18020
18021 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
18022 (arm_option_override): Override arm_restrict_it where appropriate.
18023 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
18024 * config/arm/arm.opt (mrestrict-it): New command-line option.
18025 * doc/invoke.texi: Document -mrestrict-it.
18026
18027 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18028
18029 * tsan.c (tsan_atomic_table): Make const.
18030
18031 2013-06-05 Richard Biener <rguenther@suse.de>
18032
18033 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
18034 index associated with the tree we are supposed to replace.
18035 * tree-streamer-out.c (pack_ts_base_value_fields): Output
18036 TREE_ASM_WRITTEN as zero for everything but SSA names.
18037
18038 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18039
18040 * tree-ssa-structalias.c (call_stmt_vars): Make static.
18041
18042 2013-06-04 Jan Hubicka <jh@suse.cz>
18043
18044 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
18045 (input_node, input_varpool_node): Handle correctly external same
18046 body aliases.
18047 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
18048 nodes at ltrans stage.
18049
18050 2013-06-04 Jan Hubicka <jh@suse.cz>
18051
18052 * ipa-inline.c (update_caller_keys): Fix availability test.
18053 (update_callee_keys): Likewise.
18054 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
18055 to follow ELF standard.
18056
18057 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
18058
18059 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
18060 (mips64r5900el-*-elf*): New configurations.
18061 * config/mips/mips-cpus.def (r5900): New processor.
18062 * config/mips/mips-tables.opt: Regenerate.
18063 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
18064 (mips_issue_rate): Handle PROCESSOR_R5900.
18065 (mips_reorg_process_insns): Force reorder mode for the R5900.
18066 * config/mips/mips.h (TARGET_MIPS5900): Define.
18067 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
18068 TARGET_MIPS5900.
18069 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
18070 TARGET_MIPS5900.
18071 * config/mips/mips.md (processor): Add r5900.
18072 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
18073
18074 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18075
18076 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
18077 into function to generate MOVI instruction.
18078 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
18079 (aarch64_preferred_simd_mode): Turn into wrapper.
18080 (aarch64_output_scalar_simd_mov_immediate): New function.
18081 * config/aarch64/aarch64-protos.h: Add prototype for above.
18082
18083 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18084
18085 * config/aarch64/aarch64.c (simd_immediate_info): Remove
18086 element_char member.
18087 (sizetochar): Return signed char.
18088 (aarch64_simd_valid_immediate): Remove elchar and other
18089 unnecessary variables.
18090 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
18091 Calculate element_char as required.
18092 * config/aarch64/aarch64-protos.h: Update and move prototype
18093 for aarch64_output_simd_mov_immediate.
18094 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
18095 Update arguments.
18096
18097 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18098
18099 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
18100 information completed by aarch64_simd_valid_immediate.
18101 (aarch64_legitimate_constant_p): Update arguments.
18102 (aarch64_simd_valid_immediate): Work with struct rather than many
18103 pointers.
18104 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
18105 (aarch64_simd_make_constant): Update arguments.
18106 (aarch64_output_simd_mov_immediate): Work with struct rather than
18107 many pointers. Output immediate directly rather than as operand.
18108 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18109 Update prototype.
18110 * config/aarch64/constraints.md (Dn): Update arguments.
18111
18112 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18113
18114 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
18115 longer static.
18116 (aarch64_simd_immediate_valid_for_move): Remove.
18117 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
18118 (aarch64_simd_make_constant): Update call.
18119 (aarch64_output_simd_mov_immediate): Update call.
18120 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18121 Add prototype.
18122 * config/aarch64/constraints.md (Dn): Update call.
18123
18124 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18125
18126 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
18127 return type to bool for prototype.
18128 (aarch64_legitimate_constant_p): Check for true instead of not -1.
18129 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
18130 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
18131
18132 2013-06-04 Catherine Moore <clm@codesourcery.com>
18133
18134 * config/mips/mips.opt (meva): New.
18135 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
18136 (ASM_SPEC): Handle -meva.
18137 * doc/invoke.texi (meva): Document.
18138
18139 2013-06-04 Alan Modra <amodra@gmail.com>
18140
18141 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
18142 constant output.
18143
18144 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18145
18146 * rtl.def: Add extra fourth optional field to define_cond_exec.
18147 * gensupport.c (process_one_cond_exec): Process attributes from
18148 define_cond_exec.
18149 * doc/md.texi: Document fourth field in define_cond_exec.
18150
18151 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
18152
18153 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
18154 out the processing order as in store_bit_field_1.
18155
18156 2013-06-04 Jan Hubicka <jh@suse.cz>
18157
18158 PR middle-end/57500
18159 * cgraphunit.c (cgraph_process_same_body_aliases): Create
18160 non-VAR_DECL node if it does not exist yet.
18161
18162 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
18163
18164 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
18165 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
18166 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
18167 target_cpu_default setting.
18168
18169 2013-06-03 Teresa Johnson <tejohnson@google.com>
18170
18171 * dumpfile.c (opt_info_switch_p): Change -fopt-info
18172 default to -fopt-info=optimized instead of all.
18173 * doc/invoke.texi: Ditto.
18174 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
18175 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
18176 (execute_vect_slp): Emit BB vectorization success under
18177 MSG_OPTIMIZED_LOCATIONS.
18178 * tree-vect-slp.c (vect_slp_transform_bb): Change
18179 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
18180 * tree-vect-loop.c (vect_transform_loop): Ditto.
18181
18182 2013-06-03 Jason Merrill <jason@redhat.com>
18183
18184 PR c++/57415
18185 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
18186 Use TARGET_EXPR for C++.
18187
18188 2013-06-03 Jakub Jelinek <jakub@redhat.com>
18189
18190 PR rtl-optimization/57268
18191 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
18192 if DEBUG_INSN_P (insn).
18193
18194 Reapply
18195 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18196
18197 PR rtl-optimization/57268
18198 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18199 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18200
18201 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18202
18203 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
18204 (ix86_avoid_lea_for_addr): Likewise.
18205 (exact_dependency_1): Likewise.
18206 (ix86_adjust_cost): Likewise.
18207 (swap_top_of_ready_list): Fix formatting and !reload_completed check
18208 removed.
18209 (do_reorder_for_imul): Fix typo, formatting and
18210 !reload_completed check removed.
18211 (ix86_sched_reorder): Fix typo and formatting.
18212 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
18213 list.
18214
18215 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
18216
18217 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
18218
18219 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
18220
18221 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
18222 <STRING_CST>: Likewise.
18223 <VECTOR_CST>: Likewise.
18224
18225 2013-06-01 Janus Weil <janus@gcc.gnu.org>
18226 Mikael Morin <mikael@gcc.gnu.org>
18227
18228 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
18229 * config.in: Regenerated.
18230 * configure: Regenerated.
18231
18232 2013-06-01 Jan Hubicka <jh@suse.cz>
18233
18234 PR middle-end/57366
18235 * cgraphunit.c (compile): When weakref is not supported,
18236 set up transparent aliases before final output pass.
18237 * varasm.c (assemble_alias): Do not try to do it here.
18238
18239 2013-06-01 Jan Hubicka <jh@suse.cz>
18240
18241 PR middle-end/57467
18242 * passes.c (for_per_function): Skip unanalyzed functions.
18243
18244 2013-06-01 Jan Hubicka <jh@suse.cz>
18245
18246 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
18247 (lto_symtab_merge_symbols_1): ... this one.
18248 (lto_symtab_merge_cgraph_nodes): Rename to ...
18249 (lto_symtab_merge_symbols): ... this one; simplify.
18250 * cgraph.c (same_body_aliases_done): Rename to ...
18251 (cpp_implicit_aliases_done): ... this one.
18252 (cgraph_create_function_alias): Update.
18253 (cgraph_same_body_alias): Update.
18254 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
18255 (verify_edge_corresponds_to_fndecl): Simplify.
18256 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
18257 (cgraph_node): Remove same_body_alias.
18258 (varpool_node): Remove alias_of and extra_name_alias.
18259 (same_body_aliases_done): Rename to ..
18260 (cpp_implicit_aliases_done): ... this one.
18261 (symtab_alias_ultimate_target): Add default parameter.
18262 (symtab_resolve_alias): New function.
18263 (fixup_same_cpp_alias_visibility): Declare.
18264 (cgraph_function_node): Add default parameter.
18265 (cgraph_node_asm_name): Likewise.
18266 (cgraph_function_or_thunk_node): Add default parameter; do
18267 not ICE when it is NULL.
18268 (varpool_variable_node): Likewise.
18269 * tree-emutls.c (create_emultls_var): Update.
18270 (ipa_lower_emutls): Update.
18271 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
18272 (cgraph_reset_node): Reset alias info.
18273 (cgraph_finalize_function): Update.
18274 (fixup_same_cpp_alias_visibility): Move to symtab.c.
18275 (analyze_function): Simplify.
18276 (cgraph_process_same_body_aliases): Simplify.
18277 (analyze_functions): Fixup same body aliases.
18278 (handle_alias_pairs): Simplify.
18279 (assemble_thunk): Update.
18280 (assemble_thunks_and_aliases): Update.
18281 (output_weakrefs): Rewrite.
18282 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
18283 (lto_output_varpool_node): Likewise.
18284 (compute_ltrans_boundary): Remve assert.
18285 (get_alias_symbol): New functoin.
18286 (input_node): Rewrite alias handling.
18287 (input_varpool_node): Likewise.
18288 * ipa-pure-const.c (propagate_pure_const): Fix formating.
18289 * ipa.c (process_references): Handle weakrefs correctly.
18290 (symtab_remove_unreachable_nodes): Likewise.
18291 * trans-mem.c (get_cg_data): Update.
18292 (ipa_tm_create_version_alias): Update.
18293 (ipa_tm_execute): Update.
18294 * symtab.c (dump_symtab_base): Dump aliases.
18295 (verify_symtab_base): Verify aliases.
18296 (symtab_node_availability): New function.
18297 (symtab_alias_ultimate_target): Simplify.
18298 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
18299 handle all the fixup cases.
18300 (symtab_resolve_alias): New function.
18301 * passes.c (ipa_write_summaries): Handle weakrefs.
18302 * varpool.c (varpool_analyze_node): Simplify.
18303 (assemble_aliases): Update.
18304 (varpool_create_variable_alias): Simplify.
18305 (varpool_extra_name_alias): Simplify.
18306 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
18307 (lto_symtab_merge_symbols): ... this one.
18308
18309 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
18310
18311 Revert
18312 PR rtl-optimization/57268
18313 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18314 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18315
18316 2013-06-01 Tobias Burnus <burnus@net-b.de>
18317
18318 Partially reverted:
18319 2013-05-31 Tobias Burnus <burnus@net-b.de>
18320
18321 PR middle-end/57073
18322 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
18323 further up.
18324
18325 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18326
18327 PR rtl-optimization/57268
18328 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18329 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18330
18331 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18332
18333 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
18334 unordered comparison operators when -fno-trapping-math is in effect
18335 on the e500.
18336 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
18337 and implement unordered comparison operators properly on the e500.
18338
18339 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18340
18341 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
18342 for constant scalar integers.
18343 (simplify_relational_operation_1): Likewise.
18344
18345 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
18346
18347 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
18348 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
18349 Fix comment.
18350
18351 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18352 Igor Zamyatin <igor.zamyatin@intel.com>
18353
18354 Silvermont (SLM) architecture performance tuning.
18355 * config/i386/i386.h (enum ix86_tune_indices): Add
18356 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
18357 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
18358
18359 * config/i386/i386.c (initial_ix86_tune_features)
18360 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
18361 (ix86_lea_outperforms): Handle Silvermont tuning.
18362 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
18363 call.
18364 (ix86_use_lea_for_mov): Likewise.
18365 (ix86_avoid_lea_for_addr): Likewise.
18366 (ix86_lea_for_add_ok): Likewise.
18367 (exact_dependency_1): New function.
18368 (exact_store_load_dependency): Likewise.
18369 (ix86_adjust_cost): Handle Silvermont tuning.
18370 (do_reoder_for_imul): Likewise.
18371 (swap_top_of_ready_list): New function.
18372 (ix86_sched_reorder): Changed to handle Silvermont tuning.
18373
18374 * config/i386/i386.md (peepholes that split memory operand in fp
18375 converts): New.
18376
18377 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18378
18379 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
18380 Remove un-necessary braces.
18381
18382 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18383
18384 * config/aarch64/aarch64.c (aarch64_classify_symbol):
18385 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
18386
18387 2013-05-31 Tobias Burnus <burnus@net-b.de>
18388
18389 PR middle-end/57073
18390 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
18391
18392 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18393
18394 PR target/56315
18395 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
18396 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
18397 * config/arm/neon.md (iordi3_neon): Remove.
18398 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
18399 * config/arm/predicates.md (imm_for_neon_logic_operand):
18400 Move to earlier in the file.
18401 (neon_logic_op2): Likewise.
18402 (arm_iordi_operand_neon): New predicate.
18403
18404 2013-05-31 Richard Biener <rguenther@suse.de>
18405
18406 PR tree-optimization/57478
18407 PR tree-optimization/57453
18408 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
18409 are life as well.
18410
18411 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18412
18413 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
18414 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
18415
18416 2013-05-30 Tobias Burnus <burnus@net-b.de>
18417 Thomas Koenig <tkoenig@gcc.gnu.org>
18418
18419 PR middle-end/57073
18420 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
18421 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
18422
18423 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
18424
18425 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
18426
18427 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
18428
18429 * target.def (register_usage_leveling_p): New hook.
18430 * targhooks.c (default_register_usage_leveling_p): New.
18431 * targhooks.h (default_register_usage_leveling_p): New prototype.
18432 * lra-assigns.c (register_usage_leveling_p): Use the hook.
18433 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
18434 * doc/tm.texi: Update.
18435 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
18436
18437 2013-05-30 Ian Bolton <ian.bolton@arm.com>
18438
18439 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
18440 (*insv_reg<mode>): New define_insn.
18441
18442 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
18443
18444 PR rtl-optimization/57439
18445 * postreload.c (move2add_valid_value_p): Check that we have
18446 a zero subreg_regno_offset when accessing the register in
18447 the requested mode.
18448
18449 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18450 Igor Zamyatin <igor.zamyatin@intel.com>
18451
18452 Silvermont (SLM) architecture pipeline model, tuning and
18453 insn selection.
18454 * config.gcc: Add slm config options and target.
18455
18456 * config/i386/slm.md: New.
18457
18458 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
18459
18460 * config/i386/i386-c.c (ix86_target_macros_internal): New case
18461 PROCESSOR_SLM.
18462 (ix86_target_macros_internal): Likewise.
18463
18464 * config/i386/i386.c (slm_cost): New cost.
18465 (m_SLM): New macro flag.
18466 (initial_ix86_tune_features): Set m_SLM.
18467 (x86_accumulate_outgoing_args): Likewise.
18468 (x86_arch_always_fancy_math_387): Likewise.
18469 (processor_target_table): Add slm cost.
18470 (cpu_names): Add slm cpu name.
18471 (x86_option_override_internal): Set SLM ISA.
18472 (ix86_issue_rate): New case PROCESSOR_SLM.
18473 (ia32_multipass_dfa_lookahead): Likewise.
18474 (fold_builtin_cpu): Add slm.
18475
18476 * config/i386/i386.h (TARGET_SLM): New target macro.
18477 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
18478 (processor_type): Add PROCESSOR_SLM.
18479
18480 * config/i386/i386.md (cpu): Add new value "slm".
18481 (slm.md): Include slm.md.
18482
18483 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
18484 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18485
18486 * config/arm/arm-protos.h: Add and update function protos.
18487 * config/arm/arm.c (use_simple_return_p): New added.
18488 (thumb2_expand_return): Check simple_return flag.
18489 * config/arm/arm.md: Add simple_return and conditional simple_return.
18490 * config/arm/iterators.md: Add iterator for return and simple_return.
18491
18492 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18493
18494 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
18495 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
18496 (arm_emit_vfp_multi_reg_pop): Likewise.
18497 (thumb2_emit_ldrd_pop): Likewise.
18498 (arm_expand_epilogue): Add misc REG_CFA notes.
18499 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
18500
18501 2013-05-29 Lawrence Crowl <crowl@google.com>
18502
18503 * config/arm/t-arm: Update for below.
18504
18505 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
18506 Change type to hash_table. Update dependent calls and types.
18507
18508 * config/i386/t-cygming: Update for below.
18509
18510 * config/i386/t-interix: Update for below.
18511
18512 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
18513 Change type to hash_table. Update dependent calls and types.
18514 (i386_find_on_wrapper_list::wrappers): Likewise.
18515
18516 * config/ia64/t-ia64: Update for below.
18517
18518 * config/ia64/ia64.c (bundle_state_table):
18519 Change type to hash_table. Update dependent calls and types.
18520
18521 * config/mips/mips.c (mips_reorg_process_insns::htab):
18522 Change type to hash_table. Update dependent calls and types.
18523
18524 * config/sol2.c (solaris_comdat_htab):
18525 Change type to hash_table. Update dependent calls and types.
18526
18527 * config/t-sol2: Update for above.
18528
18529 2013-05-29 Teresa Johnson <tejohnson@google.com>
18530
18531 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
18532 functions are not yet marked as defined.
18533
18534 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
18535 Pat Haugen <pthaugen@us.ibm.com>
18536 Peter Bergner <bergner@vnet.ibm.com>
18537
18538 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
18539 instructions.
18540 (VEC_A): Likewise.
18541 (VEC_C): Likewise.
18542 (vrotl<mode>3): Likewise.
18543 (vashl<mode>3): Likewise.
18544 (vlshr<mode>3): Likewise.
18545 (vashr<mode>3): Likewise.
18546
18547 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18548 support for power8 V2DI builtins.
18549
18550 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
18551 power8 V2DI builtins.
18552 (vupkhsw): Likewise.
18553 (vupklsw): Likewise.
18554 (vaddudm): Likewise.
18555 (vminsd): Likewise.
18556 (vmaxsd): Likewise.
18557 (vminud): Likewise.
18558 (vmaxud): Likewise.
18559 (vpkudum): Likewise.
18560 (vpksdss): Likewise.
18561 (vpkudus): Likewise.
18562 (vpksdus): Likewise.
18563 (vrld): Likewise.
18564 (vsld): Likewise.
18565 (vsrd): Likewise.
18566 (vsrad): Likewise.
18567 (vsubudm): Likewise.
18568 (vcmpequd): Likewise.
18569 (vcmpgtsd): Likewise.
18570 (vcmpgtud): Likewise.
18571 (vcmpequd_p): Likewise.
18572 (vcmpgtsd_p): Likewise.
18573 (vcmpgtud_p): Likewise.
18574 (vupkhsw): Likewise.
18575 (vupklsw): Likewise.
18576 (vaddudm): Likewise.
18577 (vmaxsd): Likewise.
18578 (vmaxud): Likewise.
18579 (vminsd): Likewise.
18580 (vminud): Likewise.
18581 (vpksdss): Likewise.
18582 (vpksdus): Likewise.
18583 (vpkudum): Likewise.
18584 (vpkudus): Likewise.
18585 (vrld): Likewise.
18586 (vsld): Likewise.
18587 (vsrad): Likewise.
18588 (vsrd): Likewise.
18589 (vsubudm): Likewise.
18590
18591 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
18592 support for power8 V2DI instructions.
18593
18594 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
18595 power8 V2DI instructions. Combine pack and unpack insns to use an
18596 iterator for each mode. Check whether a particular mode supports
18597 Altivec instructions instead of just checking TARGET_ALTIVEC.
18598 (UNSPEC_VPKUWUM): Likewise.
18599 (UNSPEC_VPKSHSS): Likewise.
18600 (UNSPEC_VPKSWSS): Likewise.
18601 (UNSPEC_VPKUHUS): Likewise.
18602 (UNSPEC_VPKSHUS): Likewise.
18603 (UNSPEC_VPKUWUS): Likewise.
18604 (UNSPEC_VPKSWUS): Likewise.
18605 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
18606 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
18607 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
18608 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
18609 (UNSPEC_VUPKHSB): Likewise.
18610 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
18611 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
18612 (UNSPEC_VUPKHSH): Likewise.
18613 (UNSPEC_VUPKLSB): Likewise.
18614 (UNSPEC_VUPKLSH): Likewise.
18615 (VI2): Likewise.
18616 (VI_char): Likewise.
18617 (VI_scalar): Likewise.
18618 (VI_unit): Likewise.
18619 (VP): Likewise.
18620 (VP_small): Likewise.
18621 (VP_small_lc): Likewise.
18622 (VU_char): Likewise.
18623 (add<mode>3): Likewise.
18624 (altivec_vaddcuw): Likewise.
18625 (altivec_vaddu<VI_char>s): Likewise.
18626 (altivec_vadds<VI_char>s): Likewise.
18627 (sub<mode>3): Likewise.
18628 (altivec_vsubcuw): Likewise.
18629 (altivec_vsubu<VI_char>s): Likewise.
18630 (altivec_vsubs<VI_char>s): Likewise.
18631 (altivec_vavgs<VI_char>): Likewise.
18632 (altivec_vcmpbfp): Likewise.
18633 (altivec_eq<mode>): Likewise.
18634 (altivec_gt<mode>): Likewise.
18635 (altivec_gtu<mode>): Likewise.
18636 (umax<mode>3): Likewise.
18637 (smax<mode>3): Likewise.
18638 (umin<mode>3): Likewise.
18639 (smin<mode>3): Likewise.
18640 (altivec_vpkuhum): Likewise.
18641 (altivec_vpkuwum): Likewise.
18642 (altivec_vpkshss): Likewise.
18643 (altivec_vpkswss): Likewise.
18644 (altivec_vpkuhus): Likewise.
18645 (altivec_vpkshus): Likewise.
18646 (altivec_vpkuwus): Likewise.
18647 (altivec_vpkswus): Likewise.
18648 (altivec_vpks<VI_char>ss): Likewise.
18649 (altivec_vpks<VI_char>us): Likewise.
18650 (altivec_vpku<VI_char>us): Likewise.
18651 (altivec_vpku<VI_char>um): Likewise.
18652 (altivec_vrl<VI_char>): Likewise.
18653 (altivec_vsl<VI_char>): Likewise.
18654 (altivec_vsr<VI_char>): Likewise.
18655 (altivec_vsra<VI_char>): Likewise.
18656 (altivec_vsldoi_<mode>): Likewise.
18657 (altivec_vupkhsb): Likewise.
18658 (altivec_vupkhs<VU_char>): Likewise.
18659 (altivec_vupkls<VU_char>): Likewise.
18660 (altivec_vupkhsh): Likewise.
18661 (altivec_vupklsb): Likewise.
18662 (altivec_vupklsh): Likewise.
18663 (altivec_vcmpequ<VI_char>_p): Likewise.
18664 (altivec_vcmpgts<VI_char>_p): Likewise.
18665 (altivec_vcmpgtu<VI_char>_p): Likewise.
18666 (abs<mode>2): Likewise.
18667 (vec_unpacks_hi_v16qi): Likewise.
18668 (vec_unpacks_hi_v8hi): Likewise.
18669 (vec_unpacks_lo_v16qi): Likewise.
18670 (vec_unpacks_hi_<VP_small_lc>): Likewise.
18671 (vec_unpacks_lo_v8hi): Likewise.
18672 (vec_unpacks_lo_<VP_small_lc>): Likewise.
18673 (vec_pack_trunc_v8h): Likewise.
18674 (vec_pack_trunc_v4si): Likewise.
18675 (vec_pack_trunc_<mode>): Likewise.
18676
18677 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
18678 V2DI builtins.
18679 (vec_vmaxsd): Likewise.
18680 (vec_vmaxud): Likewise.
18681 (vec_vminsd): Likewise.
18682 (vec_vminud): Likewise.
18683 (vec_vpksdss): Likewise.
18684 (vec_vpksdus): Likewise.
18685 (vec_vpkudum): Likewise.
18686 (vec_vpkudus): Likewise.
18687 (vec_vrld): Likewise.
18688 (vec_vsld): Likewise.
18689 (vec_vsrad): Likewise.
18690 (vec_vsrd): Likewise.
18691 (vec_vsubudm): Likewise.
18692 (vec_vupkhsw): Likewise.
18693 (vec_vupklsw): Likewise.
18694
18695 2013-05-29 Jan Hubicka <jh@suse.cz>
18696
18697 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
18698 flags; reorder rest of fields in more consistent way.
18699 (varpool_node): Remove analyzed, finalized and alias.
18700 (cgraph_ndoe): Likewise.
18701 (symtab_alias_ultimate_target): New function.
18702 (cgraph_function_node): Move offline.
18703 (cgraph_reset_node): Declare.
18704 (cgraph_comdat_can_be_unshared_p): Remove.
18705 (varpool_remove_initializer): Declare.
18706 (varpool_first_defined_variable, varpool_next_defined_variable
18707 cgraph_first_defined_function, cgraph_next_defined_function): Update.
18708 (cgraph_function_with_gimple_body_p): Update.
18709 (varpool_all_refs_explicit_p): Update.
18710 (symtab_alias_target): New function.
18711 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
18712 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
18713 (cgraph_function_or_thunk_node): Simplify using
18714 symtab_alias_ultimate_target.
18715 (varpool_variable_node): Likewise.
18716 * cgraph.c (cgraph_create_function_alias): Update.
18717 (cgraph_add_thunk): Update.
18718 (cgraph_remove_node): Update.
18719 (dump_cgraph_node): Do not dump removed flags.
18720 (cgraph_function_body_availability): Update.
18721 (cgraph_propagate_frequency): Update.
18722 (verify_cgraph_node): Check sanity of local flag.
18723 (cgraph_function_node): Move here from cgraph.h; revamp for
18724 cgraph_function_or_thunk_node.
18725 * lto-symtab.c (lto_varpool_replace_node): Update.
18726 (lto_symtab_resolve_can_prevail_p): Update.
18727 (lto_symtab_merge_cgraph_nodes): Update.
18728 * ipa-cp.c (determine_versionability, initialize_node_lattices,
18729 propagate_constants_accross_call, devirtualization_time_bonus,
18730 ipcp_propagate_stage): Update.
18731 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
18732 * ipa-inline-transform.c (clone_inlined_nodes,
18733 preserve_function_body_p): Update.
18734 * ipa-reference.c (propagate): Update.
18735 (write_node_summary_p): Update.
18736 * toplev.c (wrapup_global_declaration_2): Update.
18737 * cgraphunit.c (cgraph_analyze_function): Rename to ...
18738 (analyze_function) ... this one.
18739 (cgraph_process_new_functions): Update.
18740 (cgraph_reset_node): Export.
18741 (cgraph_finalize_function): Update.
18742 (cgraph_add_new_function): Update.
18743 (process_function_and_variable_attributes): Update.
18744 (varpool_finalize_decl): Update.
18745 (symbol_finalized): Remove.
18746 (symbol_finalized_and_needed): Rename to ...
18747 (symbol_defined_and_needed): ... update.
18748 (cgraph_analyze_functions): Update.
18749 (handle_alias_pairs): Update.
18750 (mark_functions_to_output): Update.
18751 (assemble_thunk): Update.
18752 (output_in_order): Update.
18753 (output_weakrefs): Update.
18754 (finalize_compilation_unit): Update.
18755 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
18756 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
18757 input_node, input_varpool_node): Update.
18758 * dbxout.c (dbxout_expand_expr): Update.
18759 * cgraphclones.c (cgraph_clone_node): Update.
18760 (cgraph_copy_node_for_versioning): Update.
18761 (cgraph_materialize_clone): Update.
18762 (cgraph_materialize_all_clones): Update.
18763 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
18764 propagate_pure_const, propagate_nothrow): Update.
18765 * lto-streamer-out.c (lto_output, write_symbol): Update.
18766 * ipa-utils.c (ipa_reverse_postorder): Update.
18767 * ipa-inline.c (can_inline_edge_p): Update.
18768 (update_caller_keys, ipa_inline): Update.
18769 * dwarf2out.c (reference_to_unused,
18770 premark_types_used_by_global_vars_helper): Update.
18771 * tree-eh.c (tree_could_trap_p): Update.
18772 * ipa-split.c (consider_split, execute_split_functions): Update.
18773 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
18774 has_addr_references_p): Update; move ahead in file for better
18775 readability.
18776 (process_references): Simplify.
18777 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
18778 bodies are removed.
18779 (cgraph_comdat_can_be_unshared_p): Make static.
18780 (cgraph_externally_visible_p): Update.
18781 (varpool_externally_visible_p): Update.
18782 (function_and_variable_visibility): Update.
18783 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
18784 ipa_tm_mark_force_output_node): Update.
18785 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
18786 estimate_edge_devirt_benefit, inline_generate_summary,
18787 inline_write_summary): Update.
18788 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
18789 * ipa-prop.c (ipa_compute_jump_functions): Update.
18790 (ipa_print_node_params, ipa_prop_read_section,
18791 ipa_update_after_lto_read, read_replacements_section): Update.
18792 * varasm.c (mark_decl_referenced): Update.
18793 (assemble_alias, dump_tm_clone_pairs): Update.
18794 * tree-inline.c (copy_bb): Update.
18795 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
18796 Update.
18797 * symtab.c (dump_symtab_base): Print new flags.
18798 (verify_symtab_base): Verify new flags.
18799 (symtab_alias_ultimate_target): New function.
18800 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
18801 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
18802 Update.
18803 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
18804 Update.
18805 * i386.c (ix86_get_function_versions_dispatcher,
18806 ix86_generate_version_dispatcher_body): Update.
18807 (fold_builtin_cpu): Use varpool_add_new_variable.
18808 * varpool.c (varpool_remove_initializer): Break out from ...
18809 (varpool_remove_node): ... this one.
18810 (dump_varpool_node, varpool_node_for_asm,
18811 cgraph_variable_initializer_availability, varpool_analyze_node,
18812 varpool_assemble_decl, varpool_remove_unreferenced_decls,
18813 varpool_finalize_named_section_flags, varpool_create_variable_alias):
18814 Update.
18815
18816 2013-05-29 Jan Hubicka <jh@suse.cz>
18817
18818 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
18819
18820 2013-05-29 Easwaran Raman <eraman@google.com>
18821
18822 PR tree-optimization/57442
18823 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
18824 when control exits the main loop.
18825
18826 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
18827
18828 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
18829 and RX600.
18830 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
18831 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
18832 * rx/t-rx: Add rx100 under multi library matches option for nofpu
18833 option.
18834
18835 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18836
18837 PR tree-optimization/57441
18838 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
18839 Don't limit size of incr_vec to number of candidates.
18840
18841 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
18842
18843 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
18844 and mips16 directories.
18845 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
18846 (MULTILIB_DIRNAMES): Ditto.
18847 (MULTILIB_EXCEPTIONS): Add new exceptions.
18848 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
18849 (MULTILIB_DIRNAMES): Ditto.
18850 (MULTILIB_EXCEPTIONS): Add new exceptions.
18851
18852 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18853 Marcus Shawcroft <marcus.shawcroft@arm.com>
18854
18855 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
18856 SYMBOL_TINY_ABSOLUTE.
18857 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
18858 SYMBOL_TINY_ABSOLUTE.
18859 (aarch64_expand_mov_immediate): Likewise.
18860 (aarch64_classify_symbol): Likewise.
18861 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
18862 Permit SYMBOL_TINY_ABSOLUTE.
18863 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
18864
18865 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18866 Marcus Shawcroft <marcus.shawcroft@arm.com>
18867
18868 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
18869 Refactor if/switch. Replace gcc_assert with if.
18870
18871 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
18872
18873 * config/i386/i386.c (initial_ix86_tune_features): Enable
18874 FP Reassociation for AMD bdver1 and bdver2.
18875
18876 2013-05-29 Martin Jambor <mjambor@suse.cz>
18877
18878 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
18879 and IMAGPART_EXPR do not occur within other handled_components.
18880
18881 2013-05-29 Richard Biener <rguenther@suse.de>
18882
18883 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
18884 access on whether the use is in the BB we currently try to
18885 vectorize.
18886 (vect_bb_vectorization_profitable_p): Pass the BB we currently
18887 vectorize to vect_bb_slp_scalar_cost.
18888
18889 2013-05-29 Richard Biener <rguenther@suse.de>
18890
18891 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
18892 computing scalar cost offsetted by stmts that are kept live
18893 by scalar uses.
18894 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
18895 for computation of scalar cost.
18896
18897 2013-05-28 Steve Ellcey <sellcey@mips.com>
18898
18899 * config/mips/mips-cpus.def (mips32r2): Change processor type.
18900
18901 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
18902
18903 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
18904 array notation built-in reduction functions.
18905 * doc/passes.texi (Passes): Added documentation about changes done
18906 for Cilk Plus.
18907 * doc/invoke.texi (C Dialect Options): Added documentation about
18908 the -fcilkplus flag.
18909 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
18910 (BUILTINS_DEF): Depend on cilkplus.def.
18911 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
18912 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
18913 * cilkplus.def: New file.
18914
18915 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
18916
18917 PR rtl-optimization/57439
18918 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
18919
18920 2013-05-28 Easwaran Raman <eraman@google.com>
18921
18922 PR tree-optimization/57337
18923 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
18924 (find_insert_point): Correctly identify the insertion point
18925 when two statements with the same UID is compared.
18926
18927 2013-05-28 Richard Biener <rguenther@suse.de>
18928
18929 PR tree-optimization/56787
18930 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
18931 from the list of data references.
18932 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
18933 clobbers.
18934 (vect_analyze_loop_operations): Likewise.
18935 (vect_transform_loop): Remove clobbers.
18936
18937 2013-05-28 Martin Jambor <mjambor@suse.cz>
18938
18939 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
18940 and REALPART_EXPRs have scalar type.
18941
18942 2013-05-28 Richard Biener <rguenther@suse.de>
18943
18944 PR tree-optimization/57411
18945 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
18946 virtual operands.
18947 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
18948 virtual operand propagation.
18949
18950 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18951
18952 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
18953 destination register for bmasksi_vis.
18954 (vector_init_bshuffle): Likewise.
18955 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
18956
18957 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18958
18959 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
18960 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
18961 mode if the instruction isn't available in the original mode.
18962 * config/sparc/sparc.opt (mfix-ut699): New option.
18963 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
18964 (divdf3): Turn into expander.
18965 (divdf3_nofix): New insn.
18966 (divdf3_fix): Likewise.
18967 (divsf3): Disable if -mfix-ut699.
18968 (sqrtdf2): Turn into expander.
18969 (sqrtdf2_nofix): New insn.
18970 (sqrtdf2_fix): Likewise.
18971 (sqrtsf2): Disable if -mfix-ut699.
18972
18973 2013-05-27 Richard Biener <rguenther@suse.de>
18974
18975 PR middle-end/57412
18976 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
18977 block for the new loop.
18978
18979 2013-05-27 Richard Biener <rguenther@suse.de>
18980
18981 PR tree-optimization/57343
18982 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
18983 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
18984 (number_of_iterations_cond): Do not build the folded tree.
18985
18986 2013-05-27 Richard Biener <rguenther@suse.de>
18987
18988 Revert
18989 PR middle-end/57381
18990 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
18991 OEP_CONSTANT_ADDRESS_OF retained.
18992
18993 PR tree-optimization/57417
18994 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
18995 for unchanged base.
18996 (set_ssa_val_to): Compare addresses using
18997 get_addr_base_and_unit_offset.
18998
18999 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
19000
19001 PR rtl-optimization/56833
19002 * postreload.c (move2add_record_mode): New function.
19003 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
19004 (move2add_use_add2_insn): Use move2add_record_sym_value.
19005 (move2add_use_add3_insn): Likewise.
19006 (reload_cse_move2add): Use move2add_valid_value_p and
19007 move2add_record_mode. Invalidate call-clobbered and REG_INC
19008 affected regs by setting reg_mode to VOIDmode.
19009 (move2add_note_store): Don't pretend the inside of a SUBREG is
19010 the actual destination. Invalidate single/leading registers by
19011 setting reg_mode to VOIDmode.
19012 Use move2add_record_sym_value, move2add_valid_value_p and
19013 move2add_record_mode.
19014
19015 2013-05-27 Richard Biener <rguenther@suse.de>
19016
19017 PR tree-optimization/57396
19018 * tree-affine.c (double_int_constant_multiple_p): Properly
19019 return false for val == 0 and div != 0.
19020
19021 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19022
19023 * config/mips/mips.h: Use #elif in preprocessor conditions.
19024
19025 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19026
19027 PR target/53916
19028 * config/mips/constraints.md (kl): New constraint.
19029 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
19030 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
19031 constraint for operand 0. Split after CSE for MIPS16. Emit a move
19032 from LO for MIPS16.
19033 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
19034
19035 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19036
19037 PR target/55777
19038 * config/mips/mips.c (mips_can_inline_p): New function.
19039 (TARGET_CAN_INLINE_P): Define.
19040
19041 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
19042
19043 * sched-int.h (ds_t, dw_t): Make unsigned int.
19044 Fix documentation that describes how all the ds_t bits are used.
19045 Reserve the last bit for delayed-branch scheduling.
19046 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
19047 (BITS_PER_DEP_WEAK): Fix definition and documentation.
19048 (gen_dep_weak_1): Remove prototype.
19049 * sched-deps.c (get_dep_weak_1): Make static.
19050 * target.def (speculate_insn, needs_block_p, gen_spec_check,
19051 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
19052 * doc/tm.texi: Regenerate.
19053 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
19054
19055 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
19056
19057 PR debug/56950
19058 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
19059
19060 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
19061 Sandra Loosemore <sandra@codesourcery.com>
19062
19063 * config.gcc (powerpc-*): Allow native for with-cpu.
19064
19065 2013-05-24 Jeff Law <law@redhat.com>
19066
19067 PR tree-optimization/57124
19068 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
19069 conversion feeding a condition if the range has an overflow
19070 if -fstrict-overflow. Add warnings for when we do make the
19071 transformation.
19072
19073 2013-05-24 Dehao Chen <dehao@google.com>
19074
19075 * tree-cfg.c (locus_discrim_map): Fix the typo.
19076 (locus_discrim_hasher): Likewise.
19077 (locus_discrim_hasher::hash): Likewise.
19078 (locus_discrim_hasher::equal): Likewise.
19079
19080 2013-05-24 Martin Jambor <mjambor@suse.cz>
19081
19082 PR tree-optimization/57294
19083 * cgraph.h (ipa_record_stmt_references): Declare.
19084 * cgraphbuild.c (ipa_record_stmt_references): New function.
19085 (build_cgraph_edges): Use ipa_record_stmt_references.
19086 (rebuild_cgraph_edges): Likewise.
19087 (cgraph_rebuild_references): Likewise.
19088 * ipa-prop.c (ipa_modify_call_arguments): Discard references
19089 associated with the old statement and build references from the
19090 newly built statements.
19091 * ipa-ref.c (ipa_remove_stmt_references): New function.
19092 * ipa-ref.h (ipa_remove_stmt_references): Declare.
19093
19094 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
19095
19096 * lra-constraints.c (emit_spill_move): Use smaller mode for
19097 mem-mem moves.
19098 (check_and_process_move): Consider mem-reg moves for secondary
19099 too.
19100 (curr_insn_transform): Don't lose insns emitted before for
19101 secondary memory moves.
19102 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
19103 reg set up in the current insn.
19104
19105 2013-05-24 Dehao Chen <dehao@google.com>
19106
19107 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
19108 hash function.
19109 (locus_descrim_hasher::equal): Likewise.
19110 (build_gimple_cfg): New discriminator assignment algorithm.
19111 (make_edges): Likewise.
19112 (next_discriminator_for_locus): Likewise.
19113 (same_line_p): Likewise.
19114 (assign_discriminators): Likewise.
19115 (make_cond_expr_edges): Likewise.
19116 (make_gimple_switch_edges): Likewise.
19117 (make_goto_expr_edges): Likewise.
19118 (make_gimple_asm_edges): Likewise.
19119
19120 2013-05-24 Ian Bolton <ian.bolton@arm.com>
19121
19122 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
19123 X format specifier to only display bottom 16 bits.
19124 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
19125 immediate to match for operand 2, since it will be masked.
19126
19127 2013-05-24 Richard Biener <rguenther@suse.de>
19128
19129 PR tree-optimization/57287
19130 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
19131 all SSA names that occur in abnormal PHIs.
19132
19133 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
19134
19135 PR tree-ssa/57385
19136 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
19137 that index is not negative.
19138
19139 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
19140
19141 PR rtl-optimization/55177
19142 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
19143 (simplify_byte_swapping_operation): New.
19144 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
19145 (simplify_relational_operation_1): Deal with BSWAP.
19146
19147 2013-05-23 Richard Henderson <rth@redhat.com>
19148
19149 PR target/56742
19150 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
19151 (ix86_reorg): Call it.
19152
19153 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
19154
19155 PR target/57379
19156 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
19157 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
19158 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
19159
19160 2013-05-23 Christian Bruel <christian.bruel@st.com>
19161
19162 PR debug/57351
19163 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
19164
19165 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19166 Marcus Shawcroft <marcus.shawcroft@arm.com>
19167
19168 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
19169 * config/aarch64/constraints.md (Usa): Remove.
19170 * doc/md.texi (AArch64 Usa): Remove.
19171
19172 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19173 Marcus Shawcroft <marcus.shawcroft@arm.com>
19174
19175 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
19176 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
19177 * config/aarch64/predicates.md (aarch64_const_address): Remove.
19178 (aarch64_mov_operand): Use aarch64_mov_operand_p.
19179
19180 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
19181
19182 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
19183 instruction (AdvSIMD).
19184 * config/aarch64/aarch64-builtins.c
19185 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
19186 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
19187
19188 2013-05-23 Martin Jambor <mjambor@suse.cz>
19189
19190 PR middle-end/57347
19191 * tree.h (contains_bitfld_component_ref_p): Declare.
19192 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
19193 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
19194 caller.
19195 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
19196 not access a bit-field. Assert all final offsets are byte-aligned.
19197
19198 2013-05-23 Richard Biener <rguenther@suse.de>
19199
19200 PR tree-optimization/57380
19201 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
19202 least one invariant or re-used load.
19203 * passes.c (init_optimization_passes): Move pass_phiprop before
19204 pass_forwprop.
19205
19206 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
19207
19208 * config/aarch64/aarch64-simd.md
19209 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
19210
19211 2013-05-23 Richard Biener <rguenther@suse.de>
19212
19213 PR middle-end/57381
19214 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19215 OEP_CONSTANT_ADDRESS_OF retained.
19216
19217 2013-05-23 Jakub Jelinek <jakub@redhat.com>
19218
19219 PR middle-end/57344
19220 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
19221 don't lower unit. Handle unit not being always BITS_PER_WORD.
19222
19223 2013-05-23 Richard Biener <rguenther@suse.de>
19224
19225 PR rtl-optimization/57341
19226 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
19227 instead of true_dependence.
19228
19229 2013-05-22 David Malcolm <dmalcolm@redhat.com>
19230
19231 * bb-reorder.c (branch_threshold): Make const.
19232 (exec_threshold): Ditto.
19233
19234 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19235 Pat Haugen <pthaugen@us.ibm.com>
19236 Peter Bergner <bergner@vnet.ibm.com>
19237
19238 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
19239 documentation for the power8 crypto builtins.
19240
19241 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
19242
19243 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
19244 macros for defining power8 builtin functions.
19245 (BU_P8V_AV_2): Likewise.
19246 (BU_P8V_AV_P): Likewise.
19247 (BU_P8V_VSX_1): Likewise.
19248 (BU_P8V_OVERLOAD_1): Likewise.
19249 (BU_P8V_OVERLOAD_2): Likewise.
19250 (BU_CRYPTO_1): Likewise.
19251 (BU_CRYPTO_2): Likewise.
19252 (BU_CRYPTO_3): Likewise.
19253 (BU_CRYPTO_OVERLOAD_1): Likewise.
19254 (BU_CRYPTO_OVERLOAD_2): Likewise.
19255 (XSCVSPDP): Fix typo, point to the correct instruction.
19256 (VCIPHER): Add power8 crypto builtins.
19257 (VCIPHERLAST): Likewise.
19258 (VNCIPHER): Likewise.
19259 (VNCIPHERLAST): Likewise.
19260 (VPMSUMB): Likewise.
19261 (VPMSUMH): Likewise.
19262 (VPMSUMW): Likewise.
19263 (VPERMXOR_V2DI): Likewise.
19264 (VPERMXOR_V4SI: Likewise.
19265 (VPERMXOR_V8HI: Likewise.
19266 (VPERMXOR_V16QI: Likewise.
19267 (VSHASIGMAW): Likewise.
19268 (VSHASIGMAD): Likewise.
19269 (VPMSUM): Likewise.
19270 (VPERMXOR): Likewise.
19271 (VSHASIGMA): Likewise.
19272
19273 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
19274 __CRYPTO__ if the crypto instructions are available.
19275 (altivec_overloaded_builtins): Add support for overloaded power8
19276 builtins.
19277
19278 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
19279 support for power8 crypto builtins.
19280 (builtin_function_type): Likewise.
19281 (altivec_init_builtins): Add support for builtins that take vector
19282 long long (V2DI) arguments.
19283
19284 * config/rs6000/crypto.md: New file, define power8 crypto
19285 instructions.
19286
19287 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19288 Pat Haugen <pthaugen@us.ibm.com>
19289 Peter Bergner <bergner@vnet.ibm.com>
19290
19291 * doc/invoke.texi (Option Summary): Add power8 options.
19292 (RS/6000 and PowerPC Options): Likewise.
19293
19294 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
19295 constraints.md instead of rs6000.h. Reorder w* constraints. Add
19296 wm, wn, wr documentation.
19297
19298 * config/rs6000/constraints.md (wm): New constraint for VSX
19299 registers if direct move instructions are enabled.
19300 (wn): New constraint for no registers.
19301 (wq): New constraint for quad word even GPR registers.
19302 (wr): New constraint if 64-bit instructions are enabled.
19303 (wv): New constraint if power8 vector instructions are enabled.
19304 (wQ): New constraint for quad word memory locations.
19305
19306 * config/rs6000/predicates.md (const_0_to_15_operand): New
19307 constraint for 0..15 for crypto instructions.
19308 (gpc_reg_operand): If VSX allow registers in VSX registers as well
19309 as GPR and floating point registers.
19310 (int_reg_operand): New predicate to match only GPR registers.
19311 (base_reg_operand): New predicate to match base registers.
19312 (quad_int_reg_operand): New predicate to match even GPR registers
19313 for quad memory operations.
19314 (vsx_reg_or_cint_operand): New predicate to allow vector logical
19315 operations in both GPR and VSX registers.
19316 (quad_memory_operand): New predicate for quad memory operations.
19317 (reg_or_indexed_operand): New predicate for direct move support.
19318
19319 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
19320 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
19321 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
19322 (POWERPC_MASKS): Add power8 options.
19323 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
19324 various options.
19325
19326 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
19327 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
19328
19329 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
19330 (-mpower8-fusion): New power8 options.
19331 (-mpower8-fusion-sign): Likewise.
19332 (-mpower8-vector): Likewise.
19333 (-mcrypto): Likewise.
19334 (-mdirect-move): Likewise.
19335 (-mquad-memory): Likewise.
19336
19337 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
19338 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
19339 registers.
19340 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
19341 (rs6000_debug_vector_unit): Add p8_vector.
19342 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
19343 definitions. Also print fusion state.
19344 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
19345 (rs6000_builtin_mask_calculate): Add power8 builtin support.
19346 (rs6000_option_override_internal): Add support for power8.
19347 (rs6000_common_init_builtins): Add debugging for skipped builtins
19348 if -mdebug=builtin.
19349 (rs6000_adjust_cost): Add power8 support.
19350 (rs6000_issue_rate): Likewise.
19351 (insn_must_be_first_in_group): Likewise.
19352 (insn_must_be_last_in_group): Likewise.
19353 (force_new_group): Likewise.
19354 (rs6000_register_move_cost): Likewise.
19355 (rs6000_opt_masks): Likewise.
19356
19357 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
19358 power8 capable assembler, default to power7 options.
19359 (TARGET_DIRECT_MOVE): Likewise.
19360 (TARGET_CRYPTO): Likewise.
19361 (TARGET_P8_VECTOR): Likewise.
19362 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
19363 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
19364 (VECTOR_MEM_P8_VECTOR_P): Likewise.
19365 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
19366 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
19367 (TARGET_XSCVDPSPN): Likewise.
19368 (TARGET_XSCVSPDPN): Likewsie.
19369 (TARGET_SYNC_HI_QI): Likewise.
19370 (TARGET_SYNC_TI): Likewise.
19371 (MASK_CRYPTO): Likewise.
19372 (MASK_DIRECT_MOVE): Likewise.
19373 (MASK_P8_FUSION): Likewise.
19374 (MASK_P8_VECTOR): Likewise.
19375 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
19376 temporary used by some of the direct move instructions to get two FP
19377 temporary registers does not force creation of a stack frame.
19378 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
19379 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
19380 that any VSX registers are tieable, even if they are also an
19381 Altivec vector mode.
19382 (r6000_reg_class_enum): Add wm, wr, wv constraints.
19383 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
19384 (RS6000_BTM_CRYPTO): Likewise.
19385 (RS6000_BTM_COMMON): Likewise.
19386
19387 * config/rs6000/rs6000.md (cpu attribute): Add power8.
19388 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
19389 (enum rs6000_vector): Add power8 vector support.
19390
19391 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19392
19393 PR target/19599
19394 PR target/57340
19395 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
19396 (any_sibcall_could_use_r3): this and handle indirect calls.
19397 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
19398
19399 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19400
19401 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
19402
19403 2013-05-22 Richard Biener <rguenther@suse.de>
19404
19405 PR middle-end/57349
19406 * profile.c (branch_prob): Do not split blocks that are
19407 abnormally receiving from ECF_RETURNS_TWICE functions.
19408
19409 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19410
19411 * recog.c (offsettable_address_addr_space_p): Fix calculation of
19412 address mode. Move pointer mode initialization to the same place.
19413
19414 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
19415
19416 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
19417 while it has any effect.
19418
19419 2013-05-21 Easwaran Raman <eraman@google.com>
19420
19421 PR tree-optimization/57322
19422 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
19423 UID of the statement added to the BB to be 1.
19424
19425 2013-05-21 Jakub Jelinek <jakub@redhat.com>
19426
19427 PR tree-optimization/57331
19428 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
19429 of conversion from pointer type to integral type with integer.
19430
19431 2013-05-21 Martin Jambor <mjambor@suse.cz>
19432
19433 PR lto/57289
19434 * ipa-prop.c (ipa_read_node_info): Process param_used and
19435 controlled_uses in the same order as when writing.
19436
19437 2013-05-21 Magnus Granberg <baldrick@free.fr>
19438
19439 PR plugins/56754
19440 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
19441
19442 2013-05-21 Richard Biener <rguenther@suse.de>
19443
19444 PR tree-optimization/57318
19445 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
19446 estimate stmts with side-effects as likely eliminated.
19447
19448 2013-05-21 Richard Biener <rguenther@suse.de>
19449
19450 PR tree-optimization/57330
19451 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
19452 preserve the call stmts fntype.
19453
19454 2013-05-21 Richard Biener <rguenther@suse.de>
19455
19456 PR tree-optimization/57303
19457 * tree-ssa-sink.c (statement_sink_location): Improve killing
19458 stmt detection and properly handle self-assignments.
19459
19460 2013-05-21 Christian Bruel <christian.bruel@st.com>
19461
19462 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
19463 spanning registers. LEAF_REG_REMAP is supported only for contiguous
19464 registers. Set register size out of the PARALLEL loop.
19465
19466 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
19467
19468 PR target/56547
19469 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
19470 (*fmasf4, *fmasf4_media): New insns.
19471
19472 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19473
19474 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
19475 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
19476 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
19477 (mips_idiv_insns): Update the comments to say that the returned
19478 instruction counts are in units of BASE_INSN_LENGTH.
19479 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
19480 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
19481 using 2 rather than 4 as the length of indirect MIPS16 and
19482 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
19483 length of a NOP. Don't divide MIPS16 lengths by 2.
19484 (mips16_split_long_branches): Assume a branch is long if the
19485 length is greater than 4 rather than 8.
19486 * config/mips/mips.md (length): Give MIPS16 lengths directly,
19487 rather than multiplying them by 2. Multiply instruction counts
19488 by BASE_INSN_LENGTH rather than 4.
19489 (*jump_mips16, tls_get_tp_mips16_<mode>)
19490 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
19491
19492 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19493
19494 * config/mips/mips.md (extended_mips16): Remove branch case.
19495 (length): Remove duplicated extended_mips16 test.
19496
19497 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19498
19499 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
19500
19501 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
19502
19503 * recog.h (Recog_data): Rename to...
19504 (recog_data_d): ...this.
19505 (recog_data): Update accordingly.
19506 * recog.c (recog_data): Likewise.
19507 * reload.c (save_recog_data): Likewise.
19508 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
19509 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
19510
19511 2013-05-17 Julian Brown <julian@codesourcery.com>
19512
19513 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
19514 found in a REG_EQUAL note, invalidate it.
19515
19516 2013-05-17 Easwaran Raman <eraman@google.com>
19517
19518 * tree-ssa-reassoc.c (find_insert_point): New function.
19519 (insert_stmt_after): Likewise.
19520 (get_def_stmt): Likewise.
19521 (ensure_ops_are_available): Likewise.
19522 (not_dominated_by): Likewise.
19523 (rewrite_expr_tree): Do not move statements beyond what is
19524 necessary. Remove call to swap_ops_for_binary_stmt...
19525 (reassociate_bb): ... and move it here.
19526 (build_and_add_sum): Assign UIDs for new statements.
19527 (linearize_expr): Likewise.
19528 (do_reassoc): Renumber gimple statement UIDs.
19529
19530 2013-05-17 Jan Hubicka <jh@suse.cz>
19531
19532 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
19533 weakrefs.
19534 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
19535 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
19536 weakrefs.
19537 (output_weakrefs): Update.
19538
19539 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
19540 Martin Jambor <mjambor@suse.cz>
19541
19542 PR middle-end/57276
19543 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
19544 value that corresponds to the given aggval is found in values vector.
19545
19546 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
19547
19548 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
19549 sse, sse2, sse3, ssse3 and sse4a flags to options.
19550
19551 2013-05-17 David Malcolm <dmalcolm@redhat.com>
19552
19553 * gengtype-state.c: (s_expr_writer): New class, to handle
19554 prettifying of output layout of s-expressions.
19555 (state_writer): New class, to write out gtype.state.
19556 (state_written_type_count): Move this variable into member data of
19557 state_writer.
19558 (s_expr_writer::s_expr_writer): New code: constructor for new class
19559 (state_writer::state_writer(): ditto
19560 (s_expr_writer::write_new_line): New function
19561 (s_expr_writer::write_any_indent): ditto
19562 (s_expr_writer::begin_s_expr): ditto
19563 (s_expr_writer::end_s_expr): ditto
19564 (write_state_fileloc): convert to method of state_writer...
19565 (state_writer:: write_state_fileloc): ...and use methods of
19566 s_expr_writer to write indentation into the gtype.state output file
19567 to visually represent the hierarchical structure of the list
19568 structures
19569 (write_state_fields): ditto, renaming to...
19570 (state_writer::write_state_fields)
19571 (write_state_a_string): ditto, renaming to...
19572 (state_writer::write_state_a_string)
19573 (write_state_string_option): ditto, renaming to...
19574 (state_writer::write_state_string_option)
19575 (write_state_type_option): ditto, renaming to...
19576 (state_writer::write_state_type_option)
19577 (write_state_nested_option): ditto, renaming to...
19578 (state_writer::write_state_nested_option)
19579 (write_state_option): ditto, renaming to...
19580 (state_writer::write_state_option)
19581 (write_state_options): ditto, renaming to...
19582 (state_writer::write_state_options)
19583 (write_state_lang_bitmap): ditto, renaming to...
19584 (state_writer::write_state_lang_bitmap)
19585 (write_state_version): ditto, renaming to...
19586 (state_writer::write_state_version)
19587 (write_state_scalar_type): ditto, renaming to...
19588 (state_writer::write_state_scalar_type)
19589 (write_state_string_type): ditto, renaming to...
19590 (state_writer::write_state_string_type)
19591 (write_state_undefined_type): ditto, renaming to...
19592 (state_writer::write_state_undefined_type)
19593 (write_state_struct_union_type): ditto, renaming to...
19594 (state_writer::write_state_struct_union_type)
19595 (write_state_struct_type): ditto, renaming to...
19596 (state_writer::write_state_struct_type)
19597 (write_state_user_struct_type): ditto, renaming to...
19598 (state_writer::write_state_user_struct_type)
19599 (write_state_lang_struct_type): ditto, renaming to...
19600 (state_writer::write_state_lang_struct_type)
19601 (write_state_param_struct_type): ditto, renaming to...
19602 (state_writer::write_state_param_struct_type)
19603 (write_state_pointer_type): ditto, renaming to...
19604 (state_writer::write_state_pointer_type)
19605 (write_state_array_type): ditto, renaming to...
19606 (state_writer::write_state_array_type)
19607 (write_state_gc_used): ditto, renaming to...
19608 (state_writer::write_state_gc_used)
19609 (write_state_common_type_content): ditto, renaming to...
19610 (state_writer::write_state_common_type_content)
19611 (write_state_type): ditto, renaming to...
19612 (state_writer::write_state_type)
19613 (write_state_pair_list): ditto, renaming to...
19614 (state_writer::write_state_pair_list)
19615 (write_state_pair): ditto, renaming to...
19616 (state_writer::write_state_pair)
19617 (write_state_typedefs): ditto, renaming to...
19618 (state_writer::write_state_typedefs)
19619 (write_state_structures): ditto, renaming to...
19620 (state_writer::write_state_structures)
19621 (write_state_param_structs): ditto, renaming to...
19622 (state_writer::write_state_param_structs)
19623 (write_state_variables): ditto, renaming to...
19624 (state_writer::write_state_variables)
19625 (write_state_srcdir): ditto, renaming to...
19626 (state_writer::write_state_srcdir)
19627 (write_state_files_list): ditto, renaming to...
19628 (state_writer::write_state_files_list)
19629 (write_state_languages): ditto, renaming to...
19630 (state_writer::write_state_languages)
19631 (write_state): create a state_writer instance and use it when
19632 writing out the state file
19633
19634 2013-05-17 Mike Stump <mikestump@comcast.net>
19635
19636 PR rtl-optimization/57304
19637 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
19638 accessing DF_REF_REAL_LOC.
19639
19640 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19641
19642 PR rtl-optimization/57281
19643 PR rtl-optimization/57300
19644 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
19645 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
19646 what the other splitter did if the registers are dead.
19647
19648 2013-05-17 Richard Biener <rguenther@suse.de>
19649
19650 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
19651 MEM_REF offsets.
19652
19653 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19654
19655 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
19656 linking.
19657
19658 2013-05-17 Marek Polacek <polacek@redhat.com>
19659
19660 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
19661 length when doing non-zero store of storing '\0' to '\0'.
19662
19663 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19664
19665 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
19666 vect_external_def oprnd1 with loop_vinfo, try to emit
19667 optional cast, negation and and stmts on the loop preheader
19668 edge instead of into the pattern def seq.
19669
19670 PR tree-optimization/57051
19671 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
19672 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
19673
19674 2013-05-16 Nick Clifton <nickc@redhat.com>
19675
19676 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
19677 (rl78_is_naked_func): New function.
19678 (rl78_expand_prologue): Skip prologue generation for naked functions.
19679 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
19680 * doc/extend.texi (naked): Add RL78 to the list of processors
19681 that supports this attribute.
19682
19683 2013-05-16 Jeff Law <law@redhat.com>
19684
19685 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
19686
19687 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19688
19689 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
19690 cache parameters using detect_caches_amd also for CYRIX,
19691 NSC and TM2 signatures.
19692
19693 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19694 Dzianis Kahanovich <mahatma@eu.by>
19695
19696 PR target/45359
19697 PR target/46396
19698 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
19699 VIA/Centaur processors and determine their cache parameters
19700 using detect_caches_amd.
19701
19702 2013-05-16 Teresa Johnson <tejohnson@google.com>
19703
19704 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
19705 (rtl_verify_edges): New function.
19706 (rtl_verify_bb_insns): Ditto.
19707 (rtl_verify_bb_pointers): Ditto.
19708 (rtl_verify_bb_insn_chain): Ditto.
19709 (rtl_verify_fallthru): Ditto.
19710 (rtl_verify_bb_layout): Ditto.
19711 (rtl_verify_flow_info_1): Outline checks into new functions.
19712 (rtl_verify_flow_info): Ditto.
19713
19714 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
19715
19716 * cfghooks.c (copy_bbs): Add update_dominance argument.
19717 * cfghooks.h (copy_bbs): Update prototype.
19718 * tree-cfg.c (gimple_duplicate_sese_region):
19719 Add update_dominance argument.
19720 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
19721 * tree-ssa-loop-ch.c (copy_loop_headers): Update
19722 gimple_duplicate_sese_region call.
19723 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
19724 Update copy_bbs call.
19725 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
19726 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
19727
19728 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19729
19730 * tree-vectorizer.h (NUM_PATTERNS): Increment.
19731 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
19732 vect_recog_rotate_pattern.
19733 (vect_recog_rotate_pattern): New function.
19734
19735 2013-05-16 Jason Merrill <jason@redhat.com>
19736
19737 * Makefile.in (LLINKER): New variable.
19738 (mostlyclean): Remove link mutex.
19739 * configure.ac: Handle --enable-link-mutex.
19740 * lock-and-run.sh: New script.
19741
19742 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19743
19744 PR target/19599
19745 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
19746 for NULL decl.
19747
19748 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19749
19750 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
19751
19752 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
19753
19754 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
19755 * config/arm/arm.c (next_consecutive_mem): New function.
19756 (gen_movmem_ldrd_strd): Likewise.
19757 * config/arm/arm.md (movmemqi): Update condition and code.
19758 (unaligned_loaddi, unaligned_storedi): New patterns.
19759
19760 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19761
19762 * config.gcc: Obsolete *-*-solaris2.9*.
19763 * doc/install.texi (Specific, *-*-solaris2*): Document it.
19764
19765 2013-05-16 Richard Biener <rguenther@suse.de>
19766
19767 * passes.c (init_optimization_passes): Move pass_parallelize_loops
19768 earlier, after GRAPHITE transforms and IV canonicalization.
19769
19770 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19771
19772 * omp-low.c (extract_omp_for_data): For collapsed loops,
19773 if at least one of the loops is known at compile time to
19774 iterate zero times, set count to 0.
19775 (expand_omp_regimplify_p): New function.
19776 (expand_omp_for_generic): For collapsed loops, if at least
19777 one of the loops isn't known to iterate at least once,
19778 add runtime check with setting count to 0.
19779 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
19780 For unsigned types if it isn't known at compile time that
19781 the loop will iterate at least once, add runtime check to bypass
19782 the whole loop if initial condition isn't true.
19783
19784 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
19785
19786 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
19787
19788 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
19789
19790 PR middle-end/57286
19791 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
19792 transformations to avoid an infinite loop.
19793
19794 2013-05-16 Marek Polacek <polacek@redhat.com>
19795
19796 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
19797
19798 2013-05-15 Leif Ekblad <leif@rdos.net>
19799
19800 * config/i386/i386.c (ix86_decompose_address): Use
19801 DEFAULT_TLS_SEG_REG to access TLS segment register.
19802 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
19803 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
19804 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
19805
19806 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
19807
19808 PR target/57260
19809 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
19810 sibling calls to functions that would normally be lazily bound,
19811 unless $gp is call-clobbered.
19812
19813 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
19814
19815 * config/i386/i386.c (ix86_option_override_internal): Update
19816 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
19817 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
19818 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
19819 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
19820 of TARGET_3DNOW.
19821 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
19822
19823 2013-05-15 Andreas Schwab <schwab@suse.de>
19824
19825 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
19826 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
19827 third operand.
19828
19829 2013-05-15 Teresa Johnson <tejohnson@google.com>
19830
19831 * loop-unroll.c (report_unroll_peel): Check decision before
19832 emitting unroll/peel message.
19833
19834 2013-05-15 Teresa Johnson <tejohnson@google.com>
19835
19836 * function.h (has_bb_partition): New rtl_data flag.
19837 (bb_reorder_complete): Ditto.
19838 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
19839 instead of flag_reorder_blocks_and_partition.
19840 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
19841 with some enhancements.
19842 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
19843 * bb-reorder.c (connect_traces): Check for has_bb_partition
19844 instead of flag_reorder_blocks_and_partition.
19845 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
19846 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
19847 verify_hot_cold_block_grouping.
19848 (partition_hot_cold_basic_blocks): Set has_bb_partition.
19849
19850 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19851
19852 PR target/19599
19853 * config/arm/predicates.md (call_insn_operand): New predicate.
19854 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
19855 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
19856 if insn is not a tail call.
19857 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
19858 registers.
19859 * config/arm/arm.h (enum reg_class): New caller save register class.
19860 (REG_CLASS_NAMES): Likewise.
19861 (REG_CLASS_CONTENTS): Likewise.
19862 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
19863 without decls.
19864
19865 2013-05-15 Richard Biener <rguenther@suse.de>
19866
19867 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
19868 of MSG_OPTIMIZED_LOCATIONS.
19869 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
19870 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
19871 message.
19872 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
19873 of MSG_OPTIMIZED_LOCATIONS.
19874 (execute_vect_slp): Likewise.
19875 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
19876 (vect_create_cond_for_alias_checks): Likewise.
19877 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
19878 (vect_recog_widen_mult_pattern): Likewise.
19879 (vect_recog_widen_sum_pattern): Likewise.
19880 (vect_recog_over_widening_pattern): Likewise.
19881 (vect_recog_widen_shift_pattern): Likewise.
19882 (vect_recog_vector_vector_shift_pattern): Likewise.
19883 (vect_recog_divmod_pattern): Likewise.
19884 (vect_recog_mixed_size_cond_pattern): Likewise.
19885 (vect_recog_bool_pattern): Likewise.
19886 (vect_pattern_recog_1): Likewise.
19887
19888 2013-05-15 Martin Jambor <mjambor@suse.cz>
19889
19890 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
19891 non-functions to builtin_unreachable.
19892 * ipa-inline-transform.c (inline_call): Do not assert estimates were
19893 correct when new direct edges were discovered.
19894
19895 2013-05-15 Martin Jambor <mjambor@suse.cz>
19896
19897 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
19898 header, print symbol order instead of node uid, print more information
19899 about indirect edge targets.
19900 (ipa_make_edge_direct_to_target): Print symbol order instead of node
19901 uids.
19902 (ipa_make_edge_direct_to_target): Likewise.
19903 (remove_described_reference): Likewise.
19904 (propagate_controlled_uses): Likewise.
19905 (ipa_print_node_params): Also print symbol order.
19906 (ipcp_transform_function): Print symbol order instead of node uids.
19907 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
19908 (cgraph_get_create_real_symbol_node): Likewise.
19909 * ipa-cp.c (print_lattice): Likewise.
19910 (print_all_lattices): Likewise.
19911 (determine_versionability): Likewise.
19912 (initialize_node_lattices): Likewise.
19913 (estimate_local_effects): Likewise.
19914 (update_profiling_info): Likewise.
19915 (create_specialized_node): Likewise.
19916 (perhaps_add_new_callers): Likewise.
19917 (decide_about_value): Likewise.
19918 (decide_whether_version_node): Likewise.
19919 (identify_dead_nodes): Likewise.
19920 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
19921 (dump_inline_summary): Likewise.
19922 (estimate_node_size_and_time): Likewise.
19923 (inline_analyze_function): Likewise.
19924 * ipa-inline.c (report_inline_failed_reason): Likewise.
19925 (want_early_inline_function_p): Likewise.
19926 (edge_badness): Likewise.
19927 (update_edge_key): Likewise.
19928 (inline_small_functions): Likewise. Add dumping of order to two other
19929 dumps.
19930 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
19931 instead of node uids.
19932 (propagate_pure_const): Likewise.
19933 (propagate_pure_const): Likewise.
19934 * ipa-utils.c (dump_cgraph_node_set): Likewise.
19935 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
19936 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
19937 of node uids.
19938 * tree-pretty-print.c (dump_function_header): Likewise.
19939 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
19940 Print symbol order instead of node uids.
19941
19942 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19943
19944 * config/s390/s390.c (s390_register_move_cost): Don't impose the
19945 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
19946
19947 2013-05-15 Richard Biener <rguenther@suse.de>
19948
19949 PR tree-optimization/57275
19950 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
19951 return value for fail to do runtime alias checks for gather loads.
19952
19953 2013-05-15 Jan Hubicka <jh@suse.cz>
19954
19955 PR lto/57038
19956 PR lto/47375
19957 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
19958 weakrefs are not external.
19959 (lto_symtab_merge_decls): Fix thinko when dealing with
19960 non-lto_symtab decls.
19961 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
19962 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
19963 * varpool.c (dump_varpool_node): Dump more flags.
19964
19965 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19966
19967 * config/i386/i386.c (processor_alias_table): Add instruction
19968 FSGSBASE for AMD bdver3 architecture.
19969
19970 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19971
19972 * tree.c (warn_deprecated_use): Print file:line using locus color.
19973 * diagnostic.c (diagnostic_report_current_module): Print file:line
19974 and file:line:column using locus color.
19975
19976 2013-05-14 Mike Stump <mikestump@comcast.net>
19977
19978 * gdbinit.in: Add __null.
19979
19980 2013-05-14 Mike Stump <mikestump@comcast.net>
19981
19982 * recog.h: Rename struct recog_data to Recog_data.
19983 * recog.c: Likewise.
19984 * reload.c (can_reload_into): Likewise.
19985 * config/picochip/picochip.c: Likewise.
19986
19987 2013-05-14 Mike Stump <mikestump@comcast.net>
19988
19989 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
19990
19991 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
19992
19993 * resource.h (struct resources): Remove unch_memory member.
19994 (CLEAR_RESOURCE): Don't clear unch_memory.
19995 * resource.c (mark_referenced_resources): Don't set it.
19996 (mark_set_resources): Likewise.
19997 (mark_target_live_regs): Don't clear it.
19998 (init_resource_info): Likewise.
19999 * reorg.c (resource_conflicts_p): Don't compare it.
20000 (redundant_insn): Don't set it.
20001
20002 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
20003 Remove prototypes.
20004 * emit-rtl.c (next_label): Remove unused function.
20005 (skip_consecutive_labels, link_cc0_insns): Move to ...
20006 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
20007 only place where these functions are used, and make them static.
20008
20009 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20010
20011 * fold-const.c (fold_negate_expr): Handle vectors.
20012 (fold_truth_not_expr): Make it static.
20013 (fold_invert_truthvalue): New static function.
20014 (invert_truthvalue_loc): Handle vectors. Do not call
20015 fold_truth_not_expr directly.
20016 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
20017 <TRUTH_NOT_EXPR>: Do not cast to boolean.
20018 (fold_comparison): Handle vector constants.
20019 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
20020 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
20021 * tree.h (fold_truth_not_expr): Remove declaration.
20022
20023 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
20024
20025 * config/aarch64/aarch64-simd.md
20026 (aarch64_vcond_internal<mode>): Rename to...
20027 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
20028 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
20029 float modes. Clarify all iterator modes.
20030 (vcond<mode><mode>): Use new name for vcond expanders.
20031 (vcond<v_cmp_result><mode>): Likewise.
20032 (vcondu<mode><mode>: Likewise.
20033 * config/aarch64/iterators.md (VDQF_COND): New.
20034
20035 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20036
20037 PR bootstrap/57266
20038 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
20039 variable for the shift amount. Check that we shift by non-negative
20040 amounts.
20041
20042 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
20043
20044 PR target/42017
20045 * config/arm/arm.h (EPILOGUE_USES): Only return true
20046 for LR_REGNUM after epilogue_completed.
20047
20048 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
20049
20050 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
20051 is error_mark_node.
20052
20053 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20054
20055 PR target/57261
20056 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
20057 and Solaris 11+/x86 with gld.
20058 * configure: Regenerate.
20059
20060 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20061
20062 * expmed.c (expand_shift_1): Canonicalize rotates by
20063 constant bitsize / 2 to bitsize - 1.
20064 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
20065 case ROTATERT>: Likewise.
20066
20067 Revert:
20068 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20069
20070 * config/i386/i386.md (rotateinv): New code attr.
20071 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20072 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20073 roll $31, %eax, etc.
20074
20075 2013-05-14 Richard Biener <rguenther@suse.de>
20076
20077 PR middle-end/57235
20078 * tree-eh.c (sink_clobbers): Give up for successors with
20079 multiple predecessors and no virtual uses.
20080
20081 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
20082
20083 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
20084 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
20085
20086 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20087
20088 PR middle-end/57251
20089 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
20090 the case when both op0 and op1 have VOIDmode.
20091
20092 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20093
20094 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
20095 in multiply-accumulate mode.
20096
20097 2013-05-13 Guozhi Wei <carrot@google.com>
20098
20099 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
20100
20101 2013-05-13 Kai Tietz <ktietz@redhat.com>
20102
20103 PR target/56975
20104 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
20105 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
20106 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
20107 * config/i386/i386.c (ix86_option_override_internal): Likewise.
20108 (ix86_expand_prologue): Likewise.
20109 (ix86_expand_split_stack_prologue): Likewise.
20110 (legitimate_pic_address_disp_p): Likewise.
20111 (legitimize_pic_address): Likewise.
20112 (legitimize_tls_address): Likewise.
20113 (legitimize_pe_coff_symbol): Likewise.
20114 (output_pic_addr_const): Likewise.
20115 (construct_plt_address): Likewise.
20116 (ix86_expand_call): Likewise.
20117 (x86_output_mi_thunk): Likewise.
20118 (x86_function_profiler): Likewise.
20119
20120 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
20121
20122 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
20123 similar switch cases.
20124 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
20125 (aarch64_simd_mov_to_<mode>low): Delete.
20126 (aarch64_simd_mov_to_<mode>high): Delete.
20127 (move_lo_quad_<mode>): Add w<-r alternative.
20128 (aarch64_simd_move_hi_quad_<mode>): Likewise.
20129 (aarch64_simd_mov_from_*): Update type attribute.
20130 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
20131 statement.
20132
20133 2013-05-13 Jan Hubicka <jh@suse.cz>
20134
20135 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
20136 * config/i386/i386.c (ix86_compute_frame_layout,
20137 ix86_expand_epilogue, emit_i387_cw_initialization,
20138 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
20139 ix86_local_alignment): Fix use of size/speed predicates.
20140
20141 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20142
20143 PR tree-optimization/45216
20144 PR tree-optimization/57157
20145 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
20146 the (-Y) & (B - 1) variant if OP is |.
20147 * expmed.c (expand_shift_1): For rotations by const0_rtx just
20148 return shifted. Use (-op1) & (prec - 1) as other_amount
20149 instead of prec - op1.
20150
20151 2013-05-13 Martin Jambor <mjambor@suse.cz>
20152
20153 PR middle-end/42371
20154 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
20155 (ipa_constant_data): New type.
20156 (ipa_jump_func): Use ipa_constant_data to hold information about
20157 constant jump functions.
20158 (ipa_get_jf_constant): Adjust to jump function type changes.
20159 (ipa_get_jf_constant_rdesc): New function.
20160 (ipa_param_descriptor): New field controlled_uses.
20161 (ipa_get_controlled_uses): New function.
20162 (ipa_set_controlled_uses): Likewise.
20163 * ipa-ref.h (ipa_find_reference): Declare.
20164 * ipa-prop.c (ipa_cst_ref_desc): New type.
20165 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
20166 changes.
20167 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
20168 New parameter cs. Adjust all callers.
20169 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
20170 (remove_described_reference): New function.
20171 (jfunc_rdesc_usable): Likewise.
20172 (try_make_edge_direct_simple_call): Decrement controlled use count,
20173 attempt to remove reference if it hits zero.
20174 (combine_controlled_uses_counters): New function.
20175 (propagate_controlled_uses): Likewise.
20176 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
20177 (ipa_edge_duplication_hook): Duplicate reference descriptions.
20178 (ipa_print_node_params): Print described use counter.
20179 (ipa_write_jump_function): Adjust to jump function type changes.
20180 (ipa_read_jump_function): New parameter CS, pass it to
20181 ipa_set_jf_constant. Adjust caller.
20182 (ipa_write_node_info): Stream controlled use count
20183 (ipa_read_node_info): Likewise.
20184 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
20185 asserting.
20186 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
20187 count. Remove cloning-added reference if it reaches zero.
20188 * ipa-ref.c (ipa_find_reference): New function.
20189
20190 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20191
20192 * config/i386/i386.c (processor_target_table): Modified default
20193 alignment values for AMD BD and BT architectures.
20194
20195 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
20196
20197 * tree-vect-generic.c (uniform_vector_p): Move ...
20198 * tree.c (uniform_vector_p): ... here.
20199 * tree.h (uniform_vector_p): Declare it.
20200 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
20201 into a scalar.
20202
20203 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20204
20205 PR tree-optimization/57230
20206 * tree-ssa-strlen.c (handle_char_store): Record length for
20207 array store from STRING_CST.
20208
20209 PR tree-optimization/57230
20210 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
20211 check.
20212
20213 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
20214
20215 * config/epiphany/epiphany.c (epiphany_init): Check size of
20216 NUM_MODES_FOR_MODE_SWITCHING.
20217 (epiphany_expand_prologue):
20218 Remove CONFIG_REGNUM initial value handling code.
20219 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
20220 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
20221 (emit_set_fp_mode, epiphany_mode_after): Likewise.
20222 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
20223 Don't return 1 for FP_MODE_NONE.
20224 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
20225 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
20226 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
20227 * config/epiphany/epiphany.md (save_config): New pattern.
20228
20229 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
20230
20231 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
20232
20233 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
20234
20235 * config/i386/i386.md (memory): Handle sseishft1.
20236 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
20237 (*vec_extractv2di_1): Ditto.
20238
20239 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
20240
20241 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
20242 saved registers.
20243
20244 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
20245
20246 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
20247 Add mthumb/march=armv7-a multilib.
20248 Add mthumb/march=armv7-r multilib.
20249 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
20250
20251 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
20252
20253 * config/v850/t-rtems: Add more multilibs.
20254
20255 2013-05-10 Richard Biener <rguenther@suse.de>
20256
20257 PR tree-optimization/57214
20258 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
20259 not propagate from SSA names that occur in abnormal PHI nodes.
20260
20261 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
20262
20263 * stor-layout.c (element_precision): New function.
20264 * machmode.h (element_precision): Declare it.
20265 * tree.c (build_minus_one_cst): New function.
20266 (element_precision): Likewise.
20267 * tree.h (build_minus_one_cst): Declare new function.
20268 (element_precision): Likewise.
20269 * fold-const.c (operand_equal_p): Use element_precision.
20270 (fold_binary_loc): Handle vector types.
20271 * convert.c (convert_to_integer): Use element_precision.
20272 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
20273 separately.
20274
20275 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
20276
20277 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
20278 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
20279 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
20280 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20281 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
20282 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
20283 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
20284 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20285 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
20286 (Uuw8): New constraints.
20287 (Usb4): Move into alphabetical order.
20288 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
20289 (sd8_operand, ub8_operand, uw8_operand): New predicates.
20290 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
20291 previously unnamed patterns.
20292 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
20293 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
20294 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
20295 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
20296 of set_attr_alternative/if_then_else. Use extended_mips16 instead
20297 of specific lengths.
20298
20299 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20300
20301 * config/i386/i386.md (rotateinv): New code attr.
20302 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20303 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20304 roll $31, %eax, etc.
20305
20306 PR tree-optimization/45216
20307 PR tree-optimization/57157
20308 * tree-ssa-forwprop.c (simplify_rotate): New function.
20309 (ssa_forward_propagate_and_combine): Call it.
20310
20311 2013-05-10 Richard Biener <rguenther@suse.de>
20312
20313 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
20314 disable peeling when we version for aliasing.
20315 (vector_alignment_reachable_p): Honor explicit user alignment.
20316 (vect_supportable_dr_alignment): Likewise.
20317 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
20318 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
20319 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
20320 then peeling to arrange for the cost-model check to come first.
20321
20322 2013-05-10 Alan Modra <amodra@gmail.com>
20323
20324 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
20325 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
20326 * configure: Regenerate.
20327
20328 2013-05-10 Alan Modra <amodra@gmail.com>
20329
20330 PR target/55033
20331 * varasm.c (default_elf_select_section): Move !DECL_P check..
20332 (get_named_section): ..to here before calling get_section_name.
20333 Adjust assertion.
20334 (default_section_type_flags): Add DECL_P check.
20335 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
20336 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
20337
20338 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
20339
20340 * config/epiphany/epiphany.c (epiphany_expand_prologue):
20341 When using gen_stack_adjust_str with a register offset, add a
20342 REG_FRAME_RELATED_EXPR note.
20343
20344 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
20345
20346 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
20347 (*vec_extractv4si_zext_mem): Ditto.
20348 (*vec_extractv2di): Add 0->x and x->x alternatives.
20349 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
20350 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
20351
20352 2013-05-09 Jason Merrill <jason@redhat.com>
20353
20354 N3639 C++1y VLA support
20355 * gimplify.c (gimplify_vla_decl): Don't touch an existing
20356 DECL_VALUE_EXPR.
20357
20358 * tree.c (build_constructor_va): New.
20359 * tree.h: Declare it.
20360
20361 2013-05-09 Martin Jambor <mjambor@suse.cz>
20362
20363 PR lto/57084
20364 * gimple-fold.c (canonicalize_constructor_val): Call
20365 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
20366
20367 2013-05-09 Jan Hubicka <jh@suse.cz>
20368 Richard Biener <rguenther@suse.de>
20369
20370 PR lto/54095
20371 * symtab.c (symtab_make_decl_local): Do not add private names.
20372
20373 2013-05-09 Jan Hubicka <jh@suse.cz>
20374
20375 PR lto/54095
20376 * symtab.c (insert_to_assembler_name_hash): Handle clones.
20377 (unlink_from_assembler_name_hash): Likewise.
20378 (symtab_prevail_in_asm_name_hash, symtab_register_node,
20379 symtab_unregister_node, symtab_initialize_asm_name_hash,
20380 change_decl_assembler_name): Update.
20381
20382 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
20383
20384 * config/aarch64/aarch64.md: New movtf split.
20385 (*movtf_aarch64): Update.
20386 (aarch64_movdi_tilow): Handle TF modes and rename to
20387 aarch64_movdi_<mode>low.
20388 (aarch64_movdi_tihigh): Handle TF modes and rename to
20389 aarch64_movdi_<mode>high
20390 (aarch64_movtihigh_di): Handle TF modes and rename to
20391 aarch64_mov<mode>high_di
20392 (aarch64_movtilow_di): Handle TF modes and rename to
20393 aarch64_mov<mode>low_di
20394 (aarch64_movtilow_tilow): Remove spurious whitespace.
20395 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
20396 splits.
20397 (aarch64_print_operand): Update.
20398
20399 2013-05-09 Alan Modra <amodra@gmail.com>
20400
20401 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
20402 powerpc64le.
20403 * configure: Regenerate.
20404
20405 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20406
20407 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
20408 splitter preparation statements.
20409 * config/i386/sse.md (*vec_extract* splitters): Ditto.
20410 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
20411 adjust_address_nv.
20412
20413 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20414
20415 * gimple-ssa-strength-reduction.c (count_candidates): Change
20416 return value to int.
20417 (analyze_candidates_and_replace): Change type of length to int.
20418
20419 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20420
20421 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
20422 (*vec_extract<mode>): Use VI12_128 mode iterator.
20423 (*vec_extract<mode>_mem): Ditto.
20424 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
20425 attribute.
20426
20427 2013-05-08 Diego Novillo <dnovillo@google.com>
20428
20429 PR bootstrap/54659
20430
20431 Revert:
20432 2012-08-17 Diego Novillo <dnovillo@google.com>
20433
20434 PR bootstrap/54281
20435 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
20436 * config.in: Regenerate.
20437 * configure: Regenerate.
20438 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
20439
20440 2013-05-08 Jan Hubicka <jh@suse.cz>
20441
20442 PR lto/54095
20443 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
20444 * cgraph.h (symtab_node_base): Add unique_name.
20445 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20446 input_overwrite_node, input_varpool_node): Stream unique_name.
20447 * cgraphclones.c (cgraph_create_virtual_clone,
20448 cgraph_function_versioning): Set unique_name.
20449 * ipa.c (function_and_variable_visibility): Set unique_name.
20450
20451 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20452
20453 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
20454 (alloc_cand_and_find_basis): Restrict conditional candidate
20455 processing to CAND_MULTs.
20456
20457 2013-05-08 Jan Hubicka <jh@suse.cz>
20458
20459 PR lto/54095
20460 lto-symtab.c (lto_symtab_symbol_p): New function.
20461 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
20462 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
20463 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
20464 Skip static symbols.
20465
20466 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
20467
20468 PR tree-optimization/57200
20469 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
20470 Only call inform if the preceding warning_at returns true.
20471
20472 2013-05-07 Han Shen <shenhan@google.com>
20473
20474 * cfgexpand.c (record_or_union_type_has_array_p): New function.
20475 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
20476 * common.opt (fstack-protector-strong): New option.
20477 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
20478 * doc/invoke.texi (Optimization Options): Document
20479 "-fstack-protector-strong".
20480 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
20481
20482 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20483
20484 * config/mips/mips.c (mips_machine_reorg2): Return 0.
20485
20486 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
20487
20488 * ira.c (update_equiv_regs): Add insn having equiv memory even if
20489 it is not lhs of the insn.
20490 (setup_reg_equiv): Remove insn having equiv memory which it is not
20491 lhs of the insn.
20492 * lra-constraints.c (process_address): Try to improve generation
20493 code for address base + disp.
20494 (lra_constraints): Make correct the code for checking insn setting
20495 up backward equivalence. Remove insn only if it is in the init
20496 insn list.
20497 * lra-eliminations.c (update_reg_eliminate): Change return value.
20498 (lra_eliminate): Use the result.
20499
20500 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
20501
20502 * config/i386/sse.md (ssescalarnummask): New mode attribute.
20503 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
20504 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
20505 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
20506 register target operands.
20507 (*vec_extractv8hi_sse2): New pattern.
20508 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
20509 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
20510 (*vec_extract<mode>_mem): New insn and split pattern.
20511
20512 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
20513
20514 * config/arm/arm.c (arm_asan_shadow_offset): New function.
20515 (TARGET_ASAN_SHADOW_OFFSET): Define.
20516 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
20517 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
20518
20519 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20520
20521 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
20522 (incr_vec_index): Return -1 if increment not found.
20523 (create_add_on_incoming_edge): Assert if increment not found.
20524 (record_increment): Limit number of increments recorded.
20525 (all_phi_incrs_profitable): Return false if an increment not found.
20526 (replace_profitable_candidates): Don't process increments that were
20527 not recorded.
20528 (analyze_candidates_and_replace): Limit size of incr_vec.
20529
20530 2013-05-07 Richard Biener <rguenther@suse.de>
20531
20532 * calls.c (special_function_p): setjmp-like functions are leaf.
20533 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
20534 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
20535
20536 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
20537
20538 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
20539 (aarch64_simd_mov<mode>): New expander.
20540 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
20541 (aarch64_simd_mov_to_<mode>high): Likewise.
20542 (aarch64_simd_mov_from_<mode>low): Likewise.
20543 (aarch64_simd_mov_from_<mode>high): Likewise.
20544 (aarch64_dup_lane<mode>): Update.
20545 (aarch64_dup_lanedi): New instruction pattern.
20546 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
20547 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
20548
20549 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20550
20551 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
20552 (replace_mult_candidate): Remove unnecessary argument; remove
20553 unnecessary parameter from call to introduce_cast_before_cand.
20554 (replace_unconditional_candidate): Remove unnecessary parameter
20555 from call to replace_mult_candidate.
20556 (replace_conditional_candidate): Likewise.
20557 (insert_initializers): Use make_temp_ssa_name.
20558 (introduce_cast_before_cand): Remove unnecessary argument; use
20559 make_temp_ssa_name.
20560 (replace_one_candidate): Remove unnecessary argument; remove
20561 unnecessary parameter from calls to introduce_cast_before_cand.
20562 (replace_profitable_candidates): Remove unnecessary parameters
20563 from calls to replace_one_candidate.
20564
20565 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20566
20567 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
20568 phi def as possibly hiding a basis for a CAND_ADD whose operands
20569 have been commuted in the analysis.
20570 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
20571
20572 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20573
20574 * config/aarch64/aarch64.md
20575 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
20576 shift value between 0-4.
20577
20578 2013-05-07 Richard Biener <rguenther@suse.de>
20579
20580 * double-int.h (rshift): New overload.
20581 * double-int.c (rshift): New function.
20582 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
20583 (create_reference_ops_from_ref): Remove.
20584 (vn_reference_insert): Use shared ops for constructing the
20585 reference and copy it.
20586
20587 2013-05-07 Richard Biener <rguenther@suse.de>
20588
20589 PR middle-end/57190
20590 * tree-eh.c (sink_clobbers): Properly propagate
20591 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
20592
20593 2013-05-07 Jakub Jelinek <jakub@redhat.com>
20594
20595 PR tree-optimization/57149
20596 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
20597 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
20598 collect_phi_def_edges, execute_late_warn_uninitialized): Use
20599 uninit_undefined_value_p instead of ssa_undefined_value_p.
20600
20601 PR debug/57184
20602 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
20603 for modifier == EXPAND_INITIALIZER.
20604
20605 2013-05-07 Anton Blanchard <anton@samba.org>
20606
20607 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
20608 for powerpc64 little endian.
20609 * configure: Regenerate.
20610
20611 2013-05-06 Graham Stott <grahams@btinternet.com>
20612
20613 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
20614 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
20615 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
20616 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
20617
20618 2013-05-06 Graham Stott <grahams@btinternet.com>
20619
20620 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
20621 codes which allow non-lvalues.
20622
20623 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
20624
20625 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
20626 components are all 1s.
20627 (integer_minus_onep): New function.
20628 * tree.h (integer_minus_onep): Declare it.
20629 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
20630 integer_minus_onep instead of integer_all_onesp.
20631
20632 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20633
20634 PR target/52933
20635 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
20636 variations of these patterns.
20637
20638 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20639
20640 * config/i386/i386.md (isa): Add x64_sse4 member.
20641 (enabled): Handle x64_sse4.
20642 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
20643 instruction for 64bit SSE4_1 targets. Update insn attributes.
20644 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
20645 instruction for SSE4_1 targets. Update insn attributes.
20646 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20647 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
20648 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
20649 const_1 selector.
20650 (*vec_extractv4si): Rename from *sse4_1_pextrd.
20651 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
20652 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
20653
20654 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20655
20656 PR target/57108
20657 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
20658
20659 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
20660
20661 * final.c (do_assembler_dialects): Don't handle curly braces and
20662 vertical bar escaped by % as dialect delimiters.
20663 (output_asm_insn): Print curly braces and vertical bar if escaped
20664 by % and ASSEMBLER_DIALECT defined.
20665 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
20666 * doc/tm.texi: Regenerated.
20667
20668 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20669
20670 * config/mips/mips.c: Include tree-pass.h.
20671 (mips_reorg): Split in pre- and post-dbr_schedule parts.
20672 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
20673 (pass_mips_machine_reorg2): New machine specific pass.
20674 (insert_pass_mips_machine_reorg2): New pass plugin definition.
20675 (mips_option_override): Register the new pass.
20676 * rtl.h (cleanup_barriers): Remove prototype.
20677 (dbr_schedule): Likewise.
20678 * jump.c (cleanup_barriers): Make static.
20679 * reorg.c (dbr_schedule): Likewise.
20680
20681 2013-05-06 Richard Biener <rguenther@suse.de>
20682
20683 PR tree-optimization/57185
20684 * tree-parloops.c (add_field_for_reduction): Handle anonymous
20685 SSA names properly.
20686
20687 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20688
20689 PR target/57106
20690 * config/i386/i386.c (add_parameter_dependencies): Add dependence
20691 between "first_arg" and "insn", not "last" and "insn".
20692
20693 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20694
20695 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
20696 (find_candidates_in_block): Re-enable slsr_process_phi.
20697 (create_phi_basis): Fix double counting of candidate adjustment.
20698
20699 2013-05-06 Richard Biener <rguenther@suse.de>
20700
20701 PR middle-end/57147
20702 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
20703 the edge is also fallthru, preserve it and just clear the
20704 abnormal flag.
20705 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
20706 also complex, preserve that and just clear the fallthru flag.
20707 * tree-inline.c (update_ssa_across_abnormal_edges): Also
20708 update virtual operands.
20709
20710 2013-05-06 Alan Modra <amodra@gmail.com>
20711
20712 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
20713 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
20714 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
20715 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
20716 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20717 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
20718
20719 2013-05-06 Alan Modra <amodra@gmail.com>
20720
20721 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
20722 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
20723 (DEFAULT_ASM_ENDIAN): Define.
20724 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20725 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
20726 Update -K PIC clause from sysv4.h.
20727 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
20728 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
20729
20730 2013-05-06 Alan Modra <amodra@gmail.com>
20731
20732 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
20733 twice for little-endian.
20734 (ashrdi3_no_power, ashrdi3): Support little-endian.
20735
20736 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20737
20738 PR target/55303
20739 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
20740 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
20741 related expanders.
20742 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
20743 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
20744 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
20745 New predicates.
20746
20747 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
20748 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20749
20750 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
20751 * config/pa/pa.opt: Make mbig-switch a no-op.
20752 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
20753 (CASE_VECTOR_MODE): Always return SImode.
20754 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
20755 for the !TARGET_BIG_SWITCH case.
20756 * config/pa/pa-linux.h: Likewise.
20757 * config/pa/pa-openbsd.h: Likewise.
20758 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
20759 * config/pa/pa.md (short_jump): Remove define_insn.
20760 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
20761 (casesi0): Remove define_insn.
20762 (type): Remove btable_branch.
20763 (pa_combine_type): Likewise.
20764 (in_nullified_branch_delay): Likewise.
20765 (in_call_delay): Likewise.
20766 (define_delay): Likewise.
20767 (define_insn_reservation "Z3"): Likewise.
20768 (define_insn_reservation "Z4"): Likewise.
20769 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
20770 (pa_adjust_insn_length): Remove adjustment for btable branches.
20771 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
20772 and mno-big-switch
20773
20774 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
20775
20776 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20777 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
20778 Add m->r,x alternatives.
20779 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
20780 splitters using SWI48x mode iterator.
20781 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
20782 TARGET_64BIT. Add m->x alternative.
20783 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
20784 Add o->x alternative. Enable for TARGET_SSE.
20785 (sse_storeq): Remove expander.
20786 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
20787 with memory input operand.
20788 (*vec_extractv2di_1 splitter): New.
20789 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
20790 * config/i386/i386.md (ssevecmodelower): New mode attribute.
20791
20792 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20793
20794 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
20795 (INT_LOWPART): Delete.
20796 (extract_MB): Adjust.
20797 (extract_ME): Adjust.
20798 (print_operand): Adjust.
20799
20800 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20801
20802 * config/rs6000/predicates.md (reg_or_add_cint_operand,
20803 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20804 (reg_or_logical_cint_operand, easy_fp_constant,
20805 logical_const_operand): Delete "CONST_DOUBLE" case.
20806 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
20807 "HOST_BITS_PER_WIDE_INT == 64" test.
20808 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
20809 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
20810 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
20811 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
20812 test.
20813 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
20814 CONST_DOUBLE DImode/VOIDmode case.
20815 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
20816 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20817 CONST_DOUBLE VOIDmode case.
20818 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20819 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
20820 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
20821 Delete CONST_DOUBLE case.
20822 (splitters for mov FMOVE64 const_double): Delete
20823 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20824 "HOST_BITS_PER_WIDE_INT >= 64" test.
20825 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
20826 case.
20827 (mov DI const_double): Delete.
20828
20829 2013-05-04 Jakub Jelinek <jakub@redhat.com>
20830
20831 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
20832 on op shows all bits zero in mode of a lowpart subreg, return zero.
20833
20834 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
20835
20836 PR target/57150
20837 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
20838 to save TFmode registers and DImode to save TImode registers for
20839 caller save operations.
20840 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
20841 mark being partially clobbered since they only use the first
20842 double word.
20843
20844 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
20845 and TDmode only use the upper 64-bits of each VSX register.
20846
20847 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20848
20849 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
20850 (find_candidates_in_block): Disable slsr_process_phi.
20851
20852 2013-05-03 Guozhi Wei <carrot@google.com>
20853
20854 * coverage.c (coverage_obj_init): Move the construction of gcov
20855 constructor to ...
20856 (build_init_ctor): ... here.
20857
20858 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20859
20860 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
20861 (slsr_cand_d): Redefine def_phi.
20862 (stride_status, phi_adjust_status, count_phis_status): New enums.
20863 (find_phi_def): New.
20864 (find_basis_for_base_expr): New.
20865 (find_basis_for_candidate): Handle hidden bases.
20866 (alloc_cand_and_find_basis): Handle phi candidates.
20867 (slsr_process_phi): New.
20868 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
20869 (create_mul_imm_cand): Likewise.
20870 (create_add_ssa_cand): Exclude phi base candidates.
20871 (create_add_imm_cand): Likewise.
20872 (slsr_process_cast): Likewise.
20873 (slsr_process_copy): Likewise.
20874 (find_candidates_in_block): Handle phi candidates.
20875 (dump_candidate): Likewise.
20876 (unconditional_cands): Delete.
20877 (unconditional_cands_with_known_stride_p): Delete.
20878 (phi_dependent_cand_p): New.
20879 (cand_increment): Handle phi-dependent candidates.
20880 (replace_dependent): Delete.
20881 (replace_mult_candidate): New.
20882 (replace_unconditional_candidate): New.
20883 (incr_vec_index): Move to avoid forward reference.
20884 (create_add_on_incoming_edge): New.
20885 (create_phi_basis): New.
20886 (replace_dependents): Delete.
20887 (replace_conditional_candidate): New.
20888 (phi_add_costs): New.
20889 (replace_uncond_cands_and_profitable_phis): New.
20890 (record_increment): Handle phi adjustments.
20891 (record_phi_increments): New.
20892 (record_increments): Handle phi adjustments.
20893 (phi_incr_cost): New.
20894 (lowest_cost_path): Handle phis.
20895 (total_savings): Likewise.
20896 (analyze_increments): Likewise.
20897 (ncd_with_phi): New.
20898 (ncd_of_cand_and_phis): New.
20899 (nearest_common_dominator_for_cands): Handle phi increments.
20900 (all_phi_incrs_profitable): New.
20901 (replace_profitable_candidates): Handle phi-dependent candidates.
20902 (analyze_candidates_and_replace): Likewise.
20903
20904 2013-05-03 Teresa Johnson <tejohnson@google.com>
20905
20906 PR bootstrap/57154
20907 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
20908 do not exceed REG_BR_PROB_BASE.
20909
20910 2013-05-03 Jeff Law <law@redhat.com>
20911
20912 PR tree-optimization/57144
20913 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
20914 operand of the condition will bit into the new type when eliminating
20915 a cast feeding a condition.
20916
20917 2013-05-03 Jakub Jelinek <jakub@redhat.com>
20918
20919 PR rtl-optimization/57130
20920 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
20921 of COMPARE as in_code to the recursive call if needed.
20922
20923 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
20924
20925 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
20926 (enabled): Handle new members.
20927 * config/i386/sse.md (*vec_concatv2si): Merge from
20928 *vec_concatv2si_sse2 and vec_concatv2si_sse.
20929 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
20930
20931 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
20932
20933 PR tree-optimization/57027
20934 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
20935 for fnms opportunity, check we got the prerequisite kind
20936 of tree / gimple before using accessor functions.
20937
20938 2013-05-03 Richard Biener <rguenther@suse.de>
20939
20940 * double-int.h (lshift): New overload without precision
20941 and arith argument.
20942 (operator *=, operator +=, operator -=): Move ...
20943 * double-int.c (operator *=, operator +=, operator -=): ... here
20944 and implement more efficiently.
20945 (mul_double_with_sign): Remove.
20946 (lshift_double): Adjust to take unsinged shift argument, push
20947 dispatching code to callers.
20948 (mul_double_wide_with_sign): Add early out for callers that
20949 are not interested in high parts or overflow.
20950 (lshift): New function.
20951 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
20952 dispatch code here.
20953 (lrotate, rrotate): Use logical shifts.
20954 * expr.c (get_inner_reference): Use lshift.
20955 * fixed-value.c (do_fixed_divide): Likewise.
20956 * tree-dfa.c (get_ref_base_and_extent): Likewise.
20957 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
20958 (indirect_refs_may_alias_p): Likewise.
20959 (stmt_kills_ref_p_1): Likewise.
20960
20961 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20962
20963 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
20964
20965 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20966
20967 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
20968 scalar form of FABD instruction.
20969
20970 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
20971
20972 * lra-constraints.c (process_alt_operands): Add checking alt
20973 number to choose the best alternative.
20974
20975 2013-05-02 Richard Biener <rguenther@suse.de>
20976
20977 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
20978 bitmap and its handling.
20979 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
20980
20981 2013-05-02 Richard Biener <rguenther@suse.de>
20982
20983 PR middle-end/57140
20984 * tree-inline.c (copy_loops): Properly handle removed loops.
20985 (copy_cfg_body): Mark destination loops for fixup if source
20986 loops needed fixup.
20987
20988 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
20989
20990 PR target/56732
20991 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
20992 generating simple_return for naked functions.
20993
20994 2013-05-02 Martin Jambor <mjambor@suse.cz>
20995
20996 PR middle-end/56988
20997 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
20998 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
20999 flags match.
21000 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
21001 ipa_agg_replacement_value structures.
21002 (known_aggs_to_agg_replacement_list): Likewise.
21003 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
21004 (read_agg_replacement_chain): Likewise.
21005 (ipcp_transform_function): Also check that by_ref flags match.
21006
21007 2013-05-02 Richard Biener <rguenther@suse.de>
21008
21009 * graphds.h (struct graph): Add obstack member.
21010 * graphds.c (new_graph): Initialize obstack and allocate
21011 vertices from it.
21012 (add_edge): Allocate edge from the obstack.
21013 (free_graph): Free the obstack instead of all edges and vertices.
21014
21015 2013-05-02 Teresa Johnson <tejohnson@google.com>
21016
21017 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
21018 divides.
21019 * cfg.c (update_bb_profile_for_threading): Ditto.
21020 * tree-inline.c (copy_bb): Ditto.
21021 (copy_edges_for_bb): Ditto.
21022 (initialize_cfun): Ditto.
21023 (copy_cfg_body): Ditto.
21024 (expand_call_inline): Ditto.
21025 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
21026 (estimate_node_size_and_time): Ditto.
21027 (inline_merge_summary): Ditto.
21028 * cgraphclones.c (cgraph_clone_edge): Ditto.
21029 (cgraph_clone_node): Ditto.
21030 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21031 (compute_trg_info): Ditto.
21032
21033 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21034
21035 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
21036 S reg when fp attribute set.
21037 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
21038
21039 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21040
21041 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
21042 New pattern.
21043 (*and_one_cmplsi3_compare0_uxtw): Likewise.
21044 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
21045 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
21046
21047 2013-05-02 Richard Biener <rguenther@suse.de>
21048
21049 * tree-scalar-evolution.c (scev_info_hasher): Remove.
21050 (struct instantiate_cache_entry): New type.
21051 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
21052 (struct instantiate_cache_type): New type.
21053 (set_instantiated_value, get_instantiated_value): Remove.
21054 (get_instantiated_value_entry): New function.
21055 (instantiate_scev_name): Use the new cache and adjust.
21056 (instantiate_scev_poly): Adjust.
21057 (instantiate_scev_binary): Likewise.
21058 (instantiate_array_ref): Likewise.
21059 (instantiate_scev_convert): Likewise.
21060 (instantiate_scev_not): Likewise.
21061 (instantiate_scev_3): Likewise.
21062 (instantiate_scev_2): Likewise.
21063 (instantiate_scev_r): Likewise.
21064 (instantiate_scev): Likewise.
21065 (resolve_mixers): Likewise.
21066
21067 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
21068
21069 PR target/57091
21070 * lra-constraints.c (best_small_class_operands_num): Remove.
21071 (process_alt_operands): Remove small_class_operands_num. Take
21072 small classes operands into losers and only if the operand is not
21073 matched. Modify debugging output.
21074 (curr_insn_transform): Remove best_small_class_operands_num.
21075 Print insn name.
21076
21077 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21078
21079 * config/aarch64/aarch64-builtins.c
21080 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
21081 * config/aarch64/aarch64-simd-builtins.def
21082 (reduc_splus_): Add new modes.
21083 (reduc_uplus_): New.
21084 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
21085 (reduc_uplus_v4sf): Likewise.
21086 (reduc_splus_v4sf): Likewise.
21087 (aarch64_addv<mode>): Likewise.
21088 (reduc_uplus_<mode>): Likewise.
21089 (reduc_splus_<mode>): Likewise.
21090 (aarch64_addvv2di): Likewise.
21091 (reduc_uplus_v2di): Likewise.
21092 (reduc_splus_v2di): Likewise.
21093 (aarch64_addvv2si): Likewise.
21094 (reduc_uplus_v2si): Likewise.
21095 (reduc_splus_v2si): Likewise.
21096 (reduc_<sur>plus_<mode>): New.
21097 (reduc_<sur>plus_v2di): Likewise.
21098 (reduc_<sur>plus_v2si): Likewise.
21099 (reduc_<sur>plus_v4sf): Likewise.
21100 (aarch64_addpv4sf): Likewise.
21101 * config/aarch64/arm_neon.h
21102 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
21103 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
21104 add UNSPEC_SADDV, UNSPEC_UADDV.
21105 (SUADDV): New.
21106 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
21107
21108 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21109
21110 * config/aarch64/arm_neon.h
21111 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
21112
21113 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21114
21115 * config/aarch64/aarch64-builtins
21116 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
21117
21118 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21119
21120 * config/aarch64/aarch64-simd-builtins.def
21121 (reduc_smax_): New.
21122 (reduc_smin_): Likewise.
21123 (reduc_umax_): Likewise.
21124 (reduc_umin_): Likewise.
21125 (reduc_smax_nan_): Likewise.
21126 (reduc_smin_nan_): Likewise.
21127 (fmax): Remove.
21128 (fmin): Likewise.
21129 (smax): Update for V2SF, V4SF and V2DF modes.
21130 (smin): Likewise.
21131 (smax_nan): New.
21132 (smin_nan): Likewise.
21133 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
21134 (<su><maxmin><mode>3): ...This, refactor.
21135 (s<maxmin><mode>3): New.
21136 (<maxmin_uns><mode>3): Likewise.
21137 (reduc_<maxmin_uns>_<mode>): Refactor.
21138 (reduc_<maxmin_uns>_v4sf): Likewise.
21139 (reduc_<maxmin_uns>_v2si): Likewise.
21140 (aarch64_<fmaxmin><mode>: Remove.
21141 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
21142 new builtin names.
21143 (vmin<q>_f<32,64>): Likewise.
21144 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21145 (FMAXMIN): New.
21146 (su): Add mappings for smax, smin, umax, umin.
21147 (maxmin): New.
21148 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21149 (FMAXMIN): Rename as...
21150 (FMAXMIN_UNS): ...This.
21151 (maxminv): Remove.
21152 (fmaxminv): Likewise.
21153 (fmaxmin): Likewise.
21154 (maxmin_uns): New.
21155 (maxmin_uns_op): Likewise.
21156
21157 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21158
21159 * config/aarch64/arm_neon.h
21160 (vac<ge, gt><sd>_f<32, 64>): Rename to...
21161 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
21162 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
21163
21164 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21165
21166 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
21167 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
21168
21169 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21170
21171 * config/aarch64/aarch64-simd.md
21172 (vcond<mode>_internal): Handle special cases for constant masks.
21173 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
21174 (vcondu<mode><mode>): Likewise.
21175 (vcond<v_cmp_result><mode>): New.
21176
21177 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21178
21179 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
21180 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
21181 * config/aarch64/aarch64-simd-builtins.def
21182 (cmeq): Update to BUILTIN_VALLDI.
21183 (cmgt): Likewise.
21184 (cmge): Likewise.
21185 (cmle): Likewise.
21186 (cmlt): Likewise.
21187 * config/aarch64/arm_neon.h
21188 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
21189 to builtins or C as appropriate.
21190
21191 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21192
21193 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
21194 (cmgeu): ...This.
21195 (cmhi): Rename to...
21196 (cmgtu): ...This.
21197 * config/aarch64/aarch64-simd.md
21198 (simd_mode): Add SF.
21199 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
21200 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
21201 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
21202 (cstore<mode>_neg): ...This.
21203 * config/aarch64/iterators.md
21204 (VALLF): new.
21205 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
21206 (COMPARISONS): New.
21207 (UCOMPARISONS): Likewise.
21208 (optab): Add missing comparisons.
21209 (n_optab): New.
21210 (cmp_1): Likewise.
21211 (cmp_2): Likewise.
21212 (CMP): Likewise.
21213 (cmp): Remove.
21214 (VCMP_S): Likewise.
21215 (VCMP_U): Likewise.
21216 (V_cmp_result): Add DF, SF modes.
21217 (v_cmp_result): Likewise.
21218 (v): Likewise.
21219 (vmtype): Likewise.
21220 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
21221
21222 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
21223
21224 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
21225 define_insn to define_insn_and_split.
21226 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
21227 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
21228 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
21229 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
21230 (thumb2_negscc): Likewise.
21231
21232 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21233
21234 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
21235
21236 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21237
21238 * config/arm/thumb2.md: Remove trailing whitespaces.
21239
21240 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21241
21242 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
21243 Use gen_int_mode rather than GEN_INT.
21244
21245 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
21246
21247 * value-prof.c (stream_in_histogram_value): Remove the strayed
21248 debug_gimple_stmt.
21249
21250 2013-04-30 Richard Biener <rguenther@suse.de>
21251
21252 PR middle-end/57122
21253 * cfghooks.c (split_edge): Properly check for the loop latch edge.
21254
21255 2013-04-30 Richard Biener <rguenther@suse.de>
21256
21257 PR middle-end/57107
21258 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
21259
21260 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
21261
21262 PR rtl-optimization/56957
21263 PR rtl-optimization/57105
21264 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
21265 variable. Use just INSN_UID for determining whether an insn
21266 should be only disconnected from the insn stream.
21267 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
21268
21269 2013-04-30 Jakub Jelinek <jakub@redhat.com>
21270
21271 PR tree-optimization/57104
21272 * tsan.c (instrument_expr): Don't instrument accesses to
21273 DECL_HARD_REGISTER VAR_DECLs.
21274
21275 2013-04-30 Richard Biener <rguenther@suse.de>
21276
21277 * function.h (loops_for_fn): New inline function.
21278 (set_loops_for_fn): Likewise.
21279 * cfgloop.h (place_new_loop): Add struct function parameter.
21280 (get_loop): Likewise.
21281 (get_loops): Likewise.
21282 (number_of_loops): Likewise.
21283 (fel_next): Adjust.
21284 (fel_init): Likewise.
21285 * cfg.c (get_loop_copy): Adjust.
21286 * cfgloop.c (flow_loops_dump): Likewise.
21287 (record_loop_exits): Likewise.
21288 (verify_loop_structure): Likewise.
21289 * cfgloopanal.c (mark_irreducible_loops): Likewise.
21290 (estimate_reg_pressure_cost): Likewise.
21291 (mark_loop_exit_edges): Likewise.
21292 * cfgloopmanip.c (place_new_loop): Likewise.
21293 (add_loop): Likewise.
21294 (duplicate_loop): Likewise.
21295 * graph.c (draw_cfg_nodes): Likewise.
21296 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
21297 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
21298 (extract_affine_chrec): Likewise.
21299 (build_scop_iteration_domain): Likewise.
21300 * graphite.c (graphite_initialize): Likewise.
21301 * ira-build.c (create_loop_tree_nodes): Likewise.
21302 (more_one_region_p): Likewise.
21303 (rebuild_regno_allocno_maps): Likewise.
21304 (mark_loops_for_removal): Likewise.
21305 (mark_all_loops_for_removal): Likewise.
21306 (remove_unnecessary_regions): Likewise.
21307 (ira_build): Likewise.
21308 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21309 * loop-init.c (fix_loop_structure): Likewise.
21310 (gate_rtl_move_loop_invariants): Likewise.
21311 (gate_rtl_unswitch): Likewise.
21312 (gate_rtl_unroll_and_peel_loops): Likewise.
21313 (rtl_doloop): Likewise.
21314 * lto-streamer-in.c (input_cfg): Likewise.
21315 * lto-streamer-out.c (output_cfg): Likewise.
21316 * modulo-sched.c (sms_schedule): Likewise.
21317 * predict.c (tree_estimate_probability): Likewise.
21318 (tree_estimate_probability_driver): Likewise.
21319 (estimate_loops): Likewise.
21320 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
21321 (move_sese_region_to_fn): Likewise.
21322 (debug_loop_num): Likewise.
21323 * tree-chrec.c (chrec_evaluate): Likewise.
21324 (hide_evolution_in_other_loops_than_loop): Likewise.
21325 (chrec_component_in_loop_num): Likewise.
21326 (reset_evolution_in_loop): Likewise.
21327 (evolution_function_is_invariant_rec_p): Likewise.
21328 * tree-if-conv.c (main_tree_if_conversion): Likewise.
21329 * tree-inline.c (copy_loops): Likewise.
21330 (copy_cfg_body): Likewise.
21331 (tree_function_versioning): Likewise.
21332 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
21333 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
21334 Likewise.
21335 (add_to_evolution_1): Likewise.
21336 (scev_const_prop): Likewise.
21337 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
21338 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
21339 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
21340 (tree_ssa_lim_initialize): Likewise.
21341 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
21342 (verify_loop_closed_ssa): Likewise.
21343 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
21344 (tree_ssa_loop_im): Likewise.
21345 (tree_ssa_loop_unswitch): Likewise.
21346 (tree_vectorize): Likewise.
21347 (check_data_deps): Likewise.
21348 (tree_ssa_loop_ivcanon): Likewise.
21349 (tree_ssa_loop_bounds): Likewise.
21350 (tree_complete_unroll): Likewise.
21351 (tree_complete_unroll_inner): Likewise.
21352 (tree_parallelize_loops): Likewise.
21353 (tree_ssa_loop_prefetch): Likewise.
21354 (tree_ssa_loop_ivopts): Likewise.
21355 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
21356 * tree-vectorizer.c (vectorize_loops): Likewise.
21357
21358 2013-04-29 Mike Frysinger <vapier@gentoo.org>
21359
21360 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
21361 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
21362 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
21363 with EABI_LINK_SPEC.
21364
21365 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21366
21367 PR target/44578
21368 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
21369 alternative.
21370
21371 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
21372
21373 PR target/57097
21374 * lra-constraints.c (process_alt_operands): Discourage a bit more
21375 using memory for pseudos. Print cost dump for alternatives.
21376 Modify cost values for conflicts with early clobbers.
21377 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
21378
21379 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21380
21381 PR target/57098
21382 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
21383
21384 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21385
21386 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
21387 from/to S register.
21388 (movdi_aarch64): Support LDR/STR from/to D register.
21389
21390 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21391
21392 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
21393 or higher by default.
21394
21395 2013-04-29 Richard Biener <rguenther@suse.de>
21396
21397 PR middle-end/57075
21398 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
21399 even if not adding abnormal edges for calls that can make
21400 abnormal gotos.
21401
21402 2013-04-29 Richard Biener <rguenther@suse.de>
21403
21404 PR middle-end/57103
21405 * tree-cfg.c (move_stmt_op): Fix condition under which to update
21406 TREE_BLOCK.
21407 (move_stmt_r): Remove redundant checking.
21408
21409 2013-04-29 Teresa Johnson <tejohnson@google.com>
21410
21411 PR bootstrap/57077
21412 * basic-block.h (apply_scale): New function.
21413 (apply_probability): Use apply_scale.
21414 * gimple-streamer-in.c (input_bb): Ditto.
21415 * lto-streamer-in.c (input_cfg): Ditto.
21416 * lto-cgraph.c (merge_profile_summaries): Ditto.
21417 * tree-optimize.c (execute_fixup_cfg): Ditto.
21418 * tree-inline.c (copy_bb): Update comment to use apply_scale.
21419 (copy_edges_for_bb): Ditto.
21420 (copy_cfg_body): Ditto.
21421
21422 2013-04-29 Tom de Vries <tom@codesourcery.com>
21423
21424 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
21425 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
21426 (tail_merge_optimize): Handle current_loops == NULL.
21427
21428 2013-04-26 Jeff Law <law@redhat.com>
21429
21430 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
21431 (simplify_cond_using_ranges): Generalize code to simplify
21432 COND_EXPRs where one argument is a constant and the other
21433 is an SSA_NAME created by an integral type conversion.
21434
21435 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21436
21437 * config/arm/arm.md (store_minmaxsi): Use only when
21438 optimize_insn_for_size_p.
21439
21440 2013-04-29 Christian Bruel <christian.bruel@st.com>
21441
21442 PR target/57108
21443 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
21444
21445 2013-04-29 Richard Biener <rguenther@suse.de>
21446
21447 PR middle-end/57089
21448 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
21449 loop tree make sure to schedule a fixup for the child as well.
21450 (expand_omp_for_generic): Properly add loops.
21451 (expand_omp_for_static_nochunk): Likewise.
21452 (expand_omp_for_static_chunk): Likewise.
21453 (expand_omp_for): For the degenerate case fixup loops.
21454 (expand_omp_sections): Fix default bb placement in loops.
21455 (expand_omp_atomic_pipeline): Properly add loops.
21456
21457 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21458
21459 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
21460
21461 2013-04-29 Tom de Vries <tom@codesourcery.com>
21462
21463 * tree-ssa-tail-merge.c: Update header comment.
21464
21465 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21466
21467 * config/aarch64/arm_neon.h
21468 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
21469 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
21470 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
21471 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
21472 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
21473 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
21474 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
21475 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
21476
21477 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21478
21479 * config/aarch64/aarch64-simd.md
21480 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
21481 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
21482 fix_trunc, fixuns_trunc.
21483 (ftrunc<VDQF:mode>2): New.
21484 * config/aarch64/iterators.md (optab): Add fix, fixuns.
21485 (fix_trunc_optab): New.
21486
21487 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21488
21489 * config/aarch64/aarch64-builtins.c
21490 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
21491 iceilf, lround, iroundf.
21492
21493 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21494
21495 PR target/54349
21496 * config/i386/i386.h (enum ix86_tune_indices)
21497 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
21498 New, split from X86_TUNE_INTER_UNIT_MOVES.
21499 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
21500 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
21501 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
21502 (TARGET_INTER_UNIT_MOVES): Remove.
21503 * config/i386/i386.c (initial_ix86_tune_features): Update.
21504 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
21505 (ix86_expand_convert_uns_didf_sse): Use
21506 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21507 (ix86_expand_vector_init_one_nonzero): Ditto.
21508 (ix86_expand_vector_init_interleave): Ditto.
21509 (inline_secondary_memory_needed): Return true for moves from SSE class
21510 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
21511 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
21512 * config/i386/constraints.md (Yi, Ym): Depend on
21513 TARGET_INTER_UNIT_MOVES_TO_VEC.
21514 (Yj, Yn): New constraints.
21515 * config/i386/i386.md (*movdi_internal): Change constraints of
21516 operand 1 from Yi to Yj and from Ym to Yn.
21517 (*movsi_internal): Ditto.
21518 (*movdf_internal): Ditto.
21519 (*movsf_internal): Ditto.
21520 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
21521 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21522 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
21523 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
21524 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
21525 * config/i386/sse.md (movdi_to_sse): Ditto.
21526 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
21527 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
21528 TARGET_INTER_UNIT_MOVES.
21529 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
21530 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
21531 instead of TARGET_INTER_UNIT_MOVES.
21532 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
21533 operand 1 from Yi to Yj and from Ym to Yn.
21534
21535 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21536
21537 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
21538 (float_truncate_hi_): Likewise.
21539 (float_extend_lo_): Likewise.
21540 (float_truncate_lo_): Likewise.
21541 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
21542 (aarch64_float_extend_lo_v2df): Likewise.
21543 (vec_unpacks_hi_v4sf): Likewise.
21544 (aarch64_float_truncate_lo_v2sf): Likewise.
21545 (aarch64_float_truncate_hi_v4sf): Likewise.
21546 (vec_pack_trunc_v2df): Likewise.
21547 (vec_pack_trunc_df): Likewise.
21548
21549 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21550
21551 * config/aarch64/aarch64-builtins.c
21552 (aarch64_fold_builtin): Fold float conversions.
21553 * config/aarch64/aarch64-simd-builtins.def
21554 (floatv2si, floatv4si, floatv2di): New.
21555 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
21556 * config/aarch64/aarch64-simd.md
21557 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
21558 * config/aarch64/iterators.md (FLOATUORS): New.
21559 (optab): Add float, floatuns.
21560 (su_optab): Likewise.
21561
21562 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21563
21564 * config/aarch64/aarch64-builtins.c
21565 (aarch64_builtin_vectorized_function): Use new names for
21566 fcvt builtins.
21567 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
21568 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
21569 (fcvtzu): Split as...
21570 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
21571 (fcvtas): Split as...
21572 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
21573 (fcvtau): Split as...
21574 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
21575 (fcvtps): Split as...
21576 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
21577 (fcvtpu): Split as...
21578 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
21579 (fcvtms): Split as...
21580 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
21581 (fcvtmu): Split as...
21582 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
21583 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
21584 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
21585 (lfrintnusf, lfrintnudf): Likewise.
21586 * config/aarch64/aarch64-simd.md
21587 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
21588 define_insn.
21589 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
21590 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
21591 (fcvt_pattern): Likewise.
21592
21593 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21594
21595 * config/aarch64/aarch64-simd.md
21596 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
21597 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
21598
21599 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21600
21601 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
21602 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
21603 (vrnd<a,m,n,p>_f32): Implement using builtins.
21604 (vrnd<i,x><q>_f<32, 64>): New.
21605
21606 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21607
21608 * config/aarch64/aarch64-builtins.c
21609 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
21610 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
21611 (frintz): Rename to...
21612 (btrunc): ...this.
21613 (frintp): Rename to...
21614 (ceil): ...this.
21615 (frintm): Rename to...
21616 (floor): ...this.
21617 (frinti): Rename to...
21618 (nearbyint): ...this.
21619 (frintx): Rename to...
21620 (rint): ...this.
21621 (frinta): Rename to...
21622 (round): ...this.
21623 * config/aarch64/aarch64-simd.md
21624 (aarch64_frint<frint_suffix><mode>): Delete.
21625 (<frint_pattern><mode>2): Convert to insn.
21626 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
21627 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
21628 (frint_pattern): Likewise.
21629 (frint_suffix): Likewise.
21630
21631 2013-04-29 Richard Biener <rguenther@suse.de>
21632
21633 PR tree-optimization/57081
21634 * loop-init.c: Include tree-flow.h.
21635 (loop_optimizer_finalize): Free number of iteration estimates.
21636 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
21637
21638 2013-04-29 Jakub Jelinek <jakub@redhat.com>
21639
21640 PR tree-optimization/57083
21641 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
21642 non-singleton shift count range, zero extend low_bound for uns case.
21643
21644 * config/i386/predicates.md (general_vector_operand): New predicate.
21645 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
21646 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
21647 if they aren't nonimmediate operands. If their original values
21648 satisfy const_vector_equal_evenodd_p, don't shift them.
21649 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
21650 predicates. For the SSE4.1 case force operands[{1,2}] into registers
21651 if not nonimmediate_operand.
21652 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
21653 instead of register_operand.
21654 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
21655
21656 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
21657
21658 * stor-layout.c (finalize_size_functions): Allocate a structure and
21659 reset cfun before dumping the functions.
21660
21661 2013-04-27 Jakub Jelinek <jakub@redhat.com>
21662
21663 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
21664
21665 PR target/56866
21666 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
21667 use xop_pmacsdqh if uns_p.
21668 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
21669 the immediate rotate count.
21670
21671 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
21672
21673 * rtl.h (struct rtx_def): Add comment for field jump.
21674 (LRA_SUBREG_P): New macro.
21675 * recog.c (register_operand): Check LRA_SUBREG_P.
21676 * lra.c (lra): Add note at the end of RTL code. Align non-empty
21677 stack frame.
21678 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
21679 (lra_final_code_change): Skip subreg change for operators.
21680 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
21681 if there are no operand changes.
21682 * lra-constraints.c (curr_insn_set): New.
21683 (match_reload): Set LRA_SUBREG_P.
21684 (emit_spill_move): Ditto.
21685 (check_and_process_move): Use curr_insn_set. Process only single
21686 set insns. Don't initialize sec_mem_p and change_p.
21687 (simplify_operand_subreg): Use LRA_SUBREG_P.
21688 (reg_in_class_p): New function.
21689 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
21690 of #ifdef. Add code to remove cycling.
21691 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
21692 non-null disp. Reload inner instead of disp when base and index
21693 are null. Try to put lo_sum into register.
21694 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
21695 (check_and_process_move): Move code for move cost check to
21696 simple_move_p. Remove equiv_substitution.
21697 (simple_move_p): New function.
21698 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
21699 curr_insn_set. Call check_and_process_move only for single set
21700 insns. Use the new function. Move call of check_and_process_move
21701 after operand equiv substitution and address process.
21702
21703 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21704
21705 PR go/57045
21706 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
21707 with nonlocal goto receivers or returns twice calls, ignore
21708 unininitialized values from abnormal edges to nl goto receiver
21709 or returns twice call.
21710
21711 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21712
21713 PR tree-optimization/57051
21714 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
21715 and VEC_RSHIFT_EXPR if shift count is a multiple of element
21716 bitsize.
21717
21718 2013-04-26 Richard Biener <rguenther@suse.de>
21719
21720 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
21721 (expand_omp_taskreg): Likewise. Mark loops for fixup.
21722 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
21723 (fixup_loop_arrays_after_move): New function.
21724 (move_sese_region_to_fn): Properly outline the loop tree parts
21725 of the SESE region.
21726
21727 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
21728
21729 * config/i386/i386.md (type, unit): Fix long lines.
21730
21731 2013-04-26 Richard Biener <rguenther@suse.de>
21732
21733 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
21734 (lto-streamer-out.o): Likewise.
21735 * cfgloop.c (init_loops_structure): Export, add struct function
21736 argument and adjust.
21737 (flow_loops_find): Adjust.
21738 * cfgloop.h (enum loop_estimation): Add EST_LAST.
21739 (init_loops_structure): Declare.
21740 * lto-streamer-in.c: Include cfgloop.h.
21741 (input_cfg): Input the loop tree.
21742 * lto-streamer-out.c: Include cfgloop.h.
21743 (output_cfg): Output the loop tree.
21744 (output_struct_function_base): Do not drop PROP_loops.
21745
21746 2013-03-26 Richard Biener <rguenther@suse.de>
21747
21748 * tree-cfg.c (execute_build_cfg): Build the loop tree.
21749 (pass_build_cfg): Provide PROP_loops.
21750 (move_sese_region_to_fn): Remove loops that are outlined into fn
21751 for now.
21752 * tree-inline.c: Include cfgloop.h.
21753 (initialize_cfun): Do not drop PROP_loops.
21754 (copy_loops): New function.
21755 (copy_cfg_body): Copy loop structure.
21756 (tree_function_versioning): Initialize destination loop tree.
21757 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
21758 (pass_parallelize_loops): Do IL verification.
21759 * loop-init.c (loop_optimizer_init): Fixup loops if required.
21760 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
21761 the CFG make sure we fixup loops as well.
21762 * tree-ssa-tail-merge.c: Include cfgloop.h.
21763 (replace_block_by): When merging loop latches mark loops for fixup.
21764 * lto-streamer-out.c (output_struct_function_base): Drop
21765 PROP_loops for now.
21766 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
21767 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
21768 * ipa-split.c: Include cfgloop.h.
21769 (split_function): Add the new return block to the loop tree root.
21770 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
21771 whether we have removed the forwarder block.
21772 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
21773 * cfgloop.h (place_new_loop): Declare.
21774 * cfgloopmanip.c (place_new_loop): Export.
21775 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
21776 (tree-switch-conversion.o): Likewise.
21777 (tree-complex.o): Likewise.
21778 (tree-inline.o): Likewise.
21779 (tree-ssa-tailmerge.o): Likewise.
21780 (ipa-split.o): Likewise.
21781 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
21782 (tree-ssa-copy.o): Likewise.
21783 * tree-switch-conversion.c: Include cfgloop.h
21784 (process_switch): If we emit a bit-test cascade, schedule loops
21785 for fixup.
21786 * tree-complex.c: Include cfgloop.h.
21787 (expand_complex_div_wide): Properly add new basic-blocks to loops.
21788 * asan.c: Include cfgloop.h.
21789 (create_cond_insert_point): Properly add new basic-blocks to
21790 loops, schedule loop fixup.
21791 * cfgloop.c (verify_loop_structure): Check that looks are not
21792 marked for fixup.
21793 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
21794 to loops.
21795 (expand_omp_for_generic): Likewise.
21796 (expand_omp_sections): Likewise.
21797 (expand_omp_atomic_pipeline): Schedule loops for fixup.
21798 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
21799 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
21800 is initialized, not when loops are present.
21801 * tree-parloops.c (parallelize_loops): Remove checking here.
21802 * passes.c (init_optimization_passes): Schedule a copy-propagation
21803 pass before complete unrolling of inner loops.
21804
21805 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21806
21807 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
21808 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
21809 (colorize_init): Add argument to _WIN32 version.
21810 * toplev.c: Include diagnostic-color.h.
21811 (process_options): Default to -fdiagnostics-color=auto if
21812 GCC_COLORS env var is in the environment.
21813 * common.opt (fdiagnostics-color=): Add Var and Init.
21814 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
21815 env var is in the environment, the default is auto rather than never.
21816
21817 * diagnostic.h (file_name_as_prefix): Add context argument.
21818 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
21819 the string as locus.
21820 * langhooks.c (lhd_print_error_function): Adjust caller.
21821
21822 2013-04-25 Lawrence Crowl <crowl@google.com>
21823
21824 * var-tracking.c (shared_hash_def::htab):
21825 Change type to hash_table. Update dependent calls and types.
21826
21827 2013-04-25 Lawrence Crowl <crowl@google.com>
21828
21829 * Makefile.in: Update as needed below.
21830
21831 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
21832 Move declaration to after the type's method definitons.
21833
21834 * attribs.c (htab_t scoped_attributes::attribute_hash):
21835 Change type to hash_table. Update dependent calls and types.
21836
21837 * bitmap.c (htab_t bitmap_desc_hash):
21838 Change type to hash_table. Update dependent calls and types.
21839
21840 * cselib.c (htab_t cselib_hash_table):
21841 Change type to hash_table. Update dependent calls and types.
21842
21843 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
21844 (hash_string_slot_node): Move implementation into lto-streamer.h
21845 struct string_slot_hasher.
21846 (eq_string_slot_node): Likewise.
21847
21848 * data-streamer-out.c: Update output_block::string_hash_table
21849 dependent calls and types.
21850
21851 * dwarf2cfi.c (htab_t trace_index):
21852 Change type to hash_table. Update dependent calls and types.
21853
21854 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
21855 Change type to hash_table. Update dependent calls and types.
21856 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
21857 (htab_t optimize_external_refs::map): Likewise.
21858 (htab_t output_comp_unit::extern_map): Likewise.
21859 (htab_t output_comdat_type_unit::extern_map): Likewise.
21860 (htab_t output_macinfo::macinfo_htab): Likewise.
21861 (htab_t optimize_location_lists::htab): Likewise.
21862 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
21863
21864 * except.c (htab_t ehspec_hash_type):
21865 Change type to hash_table. Update dependent calls and types.
21866 (assign_filter_values::ttypes): Likewise.
21867 (assign_filter_values::ehspec): Likewise.
21868 (sjlj_assign_call_site_values::ar_hash): Likewise.
21869 (convert_to_eh_region_ranges::ar_hash): Likewise.
21870
21871 * gcse.c (htab_t pre_ldst_table):
21872 Change type to hash_table. Update dependent calls and types.
21873
21874 * ggc-common.c (htab_t saving_htab):
21875 Change type to hash_table. Update dependent calls and types.
21876 (htab_t loc_hash): Likewise.
21877 (htab_t ptr_hash): Likewise.
21878 (call_count): Rename ggc_call_count.
21879 (call_alloc): Rename ggc_call_alloc.
21880 (loc_descriptor): Rename make_loc_descriptor.
21881 (add_statistics): Rename ggc_add_statistics.
21882
21883 * ggc-common.c (saving_htab):
21884 Change type to hash_table. Update dependent calls and types.
21885
21886 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
21887 (push_gimplify_context): Likewise.
21888 (pop_gimplify_context): Likewise.
21889 (struct gimple_temp_hash_elt): Added.
21890 (struct gimplify_hasher): Likewise.
21891 (struct gimplify_ctx.temp_htab):
21892 Change type to hash_table. Update dependent calls and types.
21893
21894 * gimple-fold.c: Include gimplify-ctx.h.
21895
21896 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
21897 Change type to hash_table. Update dependent calls and types.
21898 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
21899 avoid potential global name collision.
21900
21901 * gimplify.c: Include gimplify-ctx.h.
21902 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
21903 (htab_t gimplify_ctx::temp_htab):
21904 Update dependent calls and types for new type hash_table.
21905 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
21906 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
21907
21908 * gimplify-ctx.h: New.
21909 (struct gimple_temp_hash_elt): Move from gimplify.c.
21910 (class gimplify_hasher): New.
21911 (struct gimplify_ctx): Move from gimple.h.
21912 (htab_t gimplify_ctx::temp_htab):
21913 Change type to hash_table. Update dependent calls and types.
21914
21915 * graphite-clast-to-gimple.c: Include graphite-htab.h.
21916 (htab_t ivs_params::newivs_index):
21917 Change type to hash_table. Update dependent calls and types.
21918 (htab_t ivs_params::params_index): Likewise.
21919 (htab_t print_generated_program::params_index): Likewise.
21920 (htab_t gloog::newivs_index): Likewise.
21921 (htab_t gloog::params_index): Likewise.
21922
21923 * graphite.c: Include graphite-htab.h.
21924 4htab_t graphite_transform_loops::bb_pbb_mapping):
21925 Change type to hash_table. Update dependent calls and types.
21926
21927 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
21928 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
21929 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
21930
21931 * graphite-dependences.c: Include graphite-htab.h.
21932 (loop_is_parallel_p): Change hash table type of parameter.
21933
21934 * graphite-htab.h: New.
21935 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
21936 (extern find_pbb_via_hash): Move from graphite-poly.h.
21937 (extern loop_is_parallel_p): Move from graphite-poly.h.
21938 (extern get_loop_body_pbbs): Move from graphite-poly.h.
21939
21940 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
21941 (extern loop_is_parallel_p): Move to graphite-htab.h.
21942 (extern get_loop_body_pbbs): Move to graphite-htab.h.
21943
21944 * haifa-sched.c (htab_t delay_htab):
21945 Change type to hash_table. Update dependent calls and types.
21946 (htab_t delay_htab_i2): Likewise.
21947
21948 * ira-color.c (htab_t allocno_hard_regs_htab):
21949 Change type to hash_table. Update dependent calls and types.
21950
21951 * ira-costs.c (htab_t cost_classes_htab):
21952 Change type to hash_table. Update dependent calls and types.
21953
21954 * loop-invariant.c (htab_t merge_identical_invariants::eq):
21955 Change type to hash_table. Update dependent calls and types.
21956
21957 * loop-iv.c (htab_t bivs):
21958 Change type to hash_table. Update dependent calls and types.
21959
21960 * loop-unroll.c (htab_t opt_info::insns_to_split):
21961 Change type to hash_table. Update dependent calls and types.
21962 (htab_t opt_info::insns_with_var_to_expand): Likewise.
21963
21964 * lto-streamer.h (struct string_slot): Move from data-streamer.h
21965 (struct string_slot_hasher): New.
21966 (htab_t output_block::string_hash_table):
21967 Change type to hash_table. Update dependent calls and types.
21968
21969 * lto-streamer-in.c (freeing_string_slot_hasher): New.
21970 (htab_t file_name_hash_table):
21971 Change type to hash_table. Update dependent calls and types.
21972
21973 * lto-streamer-out.c: Update output_block::string_hash_table dependent
21974 calls and types.
21975
21976 * lto-streamer.c (htab_t tree_htab):
21977 Change type to hash_table. Update dependent calls and types.
21978
21979 * omp-low.c: Include gimplify-ctx.h.
21980
21981 * passes.c (htab_t name_to_pass_map):
21982 Change type to hash_table. Update dependent calls and types.
21983 (pass_traverse): Rename to passes_pass_traverse.
21984
21985 * plugin.c (htab_t event_tab):
21986 Change type to hash_table. Update dependent calls and types.
21987
21988 * postreload-gcse.c (htab_t expr_table):
21989 Change type to hash_table. Update dependent calls and types.
21990 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
21991
21992 * sese.c (debug_rename_map_1): Make extern.
21993 (htab_t copy_bb_and_scalar_dependences::rename_map):
21994 Change type to hash_table. Update dependent calls and types.
21995
21996 * sese.h (extern debug_rename_map): Move to .c file.
21997
21998 * store-motion.c (htab_t store_motion_mems_table):
21999 Change type to hash_table. Update dependent calls and types.
22000
22001 * trans-mem.c (htab_t tm_new_mem_hash):
22002 Change type to hash_table. Update dependent calls and types.
22003
22004 * tree-browser.c (htab_t TB_up_ht):
22005 Change type to hash_table. Update dependent calls and types.
22006
22007 * tree-cfg.c (htab_t discriminator_per_locus):
22008 Change type to hash_table. Update dependent calls and types.
22009
22010 * tree-complex.c: Include tree-hasher.h
22011 (htab_t complex_variable_components):
22012 Change type to hash_table. Update dependent calls and types.
22013
22014 * tree-eh.c (htab_t finally_tree):
22015 Change type to hash_table. Update dependent calls and types.
22016
22017 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
22018 struct int_tree_hasher.
22019 (extern int_tree_map_eq): Likewise.
22020 (uid_decl_map_hash): Removed.
22021 (extern decl_tree_map_eq): Likewise.
22022
22023 * tree-hasher.h: New.
22024 (struct int_tree_hasher): New.
22025 (typedef int_tree_htab_type): New.
22026
22027 * tree-inline.c: Include gimplify-ctx.h.
22028
22029 * tree-mudflap.c: Include gimplify-ctx.h.
22030
22031 * tree-parloops.c: Include tree-hasher.h.
22032 (htab_t eliminate_local_variables_stmt::decl_address):
22033 Change type to hash_table. Update dependent calls and types.
22034 (htab_t separate_decls_in_region::decl_copies): Likewise.
22035
22036 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
22037 Change type to hash_table. Update dependent calls and types.
22038
22039 * tree-sra.c (candidates):
22040 Change type to hash_table. Update dependent calls and types.
22041
22042 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
22043 in tree-flow.h.
22044 (int_tree_map_hash): Likewise.
22045
22046 * tree-ssa-dom.c (htab_t avail_exprs):
22047 Change type to hash_table. Update dependent calls and types.
22048
22049 * tree-ssa-live.c (var_map_base_init::tree_to_index):
22050 Change type to hash_table. Update dependent calls and types.
22051
22052 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
22053 Change type to hash_table. Update dependent calls and types.
22054
22055 * tree-ssa-phiopt.c (seen_ssa_names):
22056 Change type to hash_table. Update dependent calls and types.
22057
22058 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
22059 Change type to hash_table. Update dependent calls and types.
22060
22061 * tree-ssa-uncprop.c (equiv):
22062 Change type to hash_table. Update dependent calls and types.
22063
22064 2013-04-25 Jakub Jelinek <jakub@redhat.com>
22065
22066 PR rtl-optimization/57003
22067 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
22068 call note_stores with kill_clobbered_value callback again after
22069 killing regs_invalidated_by_call.
22070
22071 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22072
22073 * config/aarch64/aarch64-simd.md
22074 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
22075 (aarch64_simd_bsl<mode>): Likewise.
22076 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
22077
22078 2013-04-25 Marek Polacek <polacek@redhat.com>
22079
22080 PR tree-optimization/57066
22081 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
22082
22083 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22084
22085 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
22086
22087 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22088
22089 * config/aarch64/aarch64-builtins.c
22090 (aarch64_fold_builtin): New.
22091 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
22092 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
22093 * config/aarch64/aarch64-simd-builtins.def (abs): New.
22094 * config/aarch64/arm_neon.h
22095 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
22096
22097 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22098 Tejas Belagod <tejas.belagod@arm.com>
22099
22100 * config/aarch64/aarch64-builtins.c
22101 (aarch64_gimple_fold_builtin): New.
22102 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
22103 * config/aarch64/aarch64-simd-builtins.def (addv): New.
22104 * config/aarch64/aarch64-simd.md (addpv4sf): New.
22105 (addvv4sf): Update.
22106 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
22107
22108 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22109
22110 * config/aarch64/aarch64.md
22111 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
22112
22113 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22114
22115 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
22116 (*ngcsi_uxtw): New pattern.
22117
22118 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22119 Julian Brown <julian@codesourcery.com>
22120
22121 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
22122 (TB_DREG): Add T_V4HF.
22123 (v4hf_UP): New macro.
22124 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22125 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22126 Handle initialisation of V4HF. Adjust initialisation of reinterpret
22127 built-ins.
22128 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22129 (arm_vector_mode_supported_p): Handle V4HF.
22130 (arm_mangle_map): Handle V4HFmode.
22131 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
22132 * config/arm/arm_neon_builtins.def: Add entries for
22133 vcvtv4hfv4sf, vcvtv4sfv4hf.
22134 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
22135 (neon_vcvtv4hfv4sf): Likewise.
22136 * config/arm/neon-gen.ml: Handle half-precision floating point
22137 features.
22138 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
22139 * config/arm/arm_neon.h: Regenerate.
22140 * config/arm/neon.ml (type elts): Add F16.
22141 (type vectype): Add T_float16x4, T_floatHF.
22142 (type vecmode): Add V4HF.
22143 (type features): Add Requires_FP_bit feature.
22144 (elt_width): Handle F16.
22145 (elt_class): Likewise.
22146 (elt_of_class_width): Likewise.
22147 (mode_of_elt): Refactor.
22148 (type_for_elt): Handle F16, fix error messages.
22149 (vectype_size): Handle T_float16x4.
22150 (vcvt_sh): New function.
22151 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
22152 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
22153 (string_of_mode): Handle V4HF.
22154 * doc/arm-neon-intrinsics.texi: Regenerate.
22155
22156 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22157
22158 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
22159 format specifier in 'X' case.
22160
22161 2013-04-25 Alan Modra <amodra@gmail.com>
22162
22163 PR target/57052
22164 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
22165 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
22166 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
22167 Repeat for many other rotate/shift and mask patterns using subregs.
22168 Name lshiftrt insns.
22169 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
22170 on WORDS_BIG_ENDIAN.
22171
22172 2013-04-25 Alan Modra <amodra@gmail.com>
22173
22174 * config.gcc: Support little-endian powerpc-linux targets.
22175 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
22176 (LINK_OS_LINUX_SPEC): Define.
22177 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
22178 Preserve MASK_LITTLE_ENDIAN.
22179 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
22180 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
22181 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
22182 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
22183 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
22184 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
22185 Correct fp word order for little-endian. Don't shift toc entries
22186 smaller than a word for little-endian.
22187 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
22188 (bswapdi2 splits): Correct low-part subreg for little-endian.
22189 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
22190 low/high where such is correct only for be.
22191 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
22192 little-endian for -mcall-aixdesc.
22193
22194 2013-04-25 Alan Modra <amodra@gmail.com>
22195
22196 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
22197 replace_equiv_address_nv.
22198
22199 2013-04-25 Alan Modra <amodra@gmail.com>
22200
22201 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
22202
22203 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22204
22205 Revert:
22206 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22207 * rtl.h (struct rtx_def): ...
22208
22209 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22210
22211 PR rtl-optimizations/57046
22212 * lra-constraints (split_reg): Set up lra_risky_transformations_p
22213 for multi-reg splits.
22214
22215 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
22216
22217 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
22218
22219 2013-04-24 Sterling Augustine <saugustine@google.com>
22220
22221 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
22222 (comp_dir_string, debug_str_dwo_section): New.
22223 (DEBUG_STR_DWO_SECTION): Rename to ...
22224 (DEBUG_DWO_STR_SECTION): ... this.
22225 (DEBUG_NORM_STR_SECTION): Delete.
22226 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
22227 (DEBUG_STR_DWO_SECTION_FLAGS): New.
22228 (find_AT_string): Move most logic to ...
22229 (find_AT_string_in_table): ... here. New.
22230 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
22231 add_skeleton_AT_string. Delete logic.
22232 (output_skeleton_debug_sections): Remove call to
22233 add_top_level_skeleton_die_attrs.
22234 (add_comp_dir_attribute): Move logic to comp_dir_string.
22235 (dwarf2out_init): Initialize debug_str_dwo_section.
22236 (output_indirect_string): Call find_string_form.
22237 (output_indirect_strings): Rewrite.
22238 (prune_unused_types): Empty skeleton_debug_str_hash.
22239 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
22240 (dwarf2out_finish): Call output_indirect_strings.
22241
22242 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22243
22244 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
22245
22246 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22247
22248 * rtl.h (struct rtx_def): Add comment for field jump.
22249 (LRA_SUBREG_P): New macro.
22250 * recog.c (register_operand): Check LRA_SUBREG_P.
22251 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22252 stack frame.
22253 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22254 (lra_final_code_change): Skip subreg change for operators.
22255 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22256 if there are no operand changes.
22257 * lra-constraints.c (curr_insn_set): New.
22258 (match_reload): Set LRA_SUBREG_P.
22259 (emit_spill_move): Ditto.
22260 (check_and_process_move): Use curr_insn_set. Process only single
22261 set insns. Don't initialize sec_mem_p and change_p.
22262 (simplify_operand_subreg): Use LRA_SUBREG_P.
22263 (reg_in_class_p): New function.
22264 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22265 of #ifdef. Add code to remove cycling.
22266 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22267 non-null disp. Reload inner instead of disp when base and index
22268 are null. Try to put lo_sum into register.
22269 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22270 (check_and_process_move): Move code for move cost check to
22271 simple_move_p. Remove equiv_substitution.
22272 (simple_move_p): New function.
22273 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22274 curr_insn_set. Call check_and_process_move only for single set
22275 insns. Use the new function. Move call of check_and_process_move
22276 after operand equiv substitution and address process.
22277
22278 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
22279
22280 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
22281 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
22282 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
22283
22284 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22285
22286 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
22287
22288 2013-04-24 Marek Polacek <polacek@redhat.com>
22289
22290 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
22291 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
22292 (select_loops_exit_conditions): Likewise.
22293 (number_of_iterations_for_all_loops): Likewise.
22294 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
22295 (scev_analysis): Likewise.
22296
22297 2013-04-02 Catherine Moore <clm@codesourcery.com>
22298 Chao-ying Fu <fu@mips.com>
22299
22300 * config/mips/micromips.md (jraddiusp): New pattern.
22301 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
22302 instruction if possible.
22303
22304 2013-04-24 Alan Modra <amodra@gmail.com>
22305
22306 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
22307
22308 2013-04-24 Julian Brown <julian@codesourcery.com>
22309 Chung-Lin Tang <cltang@codesourcery.com>
22310
22311 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
22312 dependency behavior in enumeration type DIE generation. Add TODO note
22313 to comments about future DW_FORM_sdata/udata re-work of related code.
22314
22315 2013-04-23 Lawrence Crowl <crowl@google.com>
22316
22317 * Makefile.in: Update as needed below.
22318
22319 * hash-table.h (class hash_table):
22320 Correct many methods with parameter types compare_type to the correct
22321 value_type. (Correct code was unlikely to notice the change.)
22322 (hash_table::elements_with_deleted) New.
22323 (class hashtable::iterator): New.
22324 (hashtable::begin()): New.
22325 (hashtable::end()): New.
22326 (FOR_EACH_HASH_TABLE_ELEMENT): New.
22327
22328 * statistics.c (statistics_hashes):
22329 Change type to hash_table. Update dependent calls and types.
22330
22331 * tree-into-ssa.c (var_infos):
22332 Change type to hash_table. Update dependent calls and types.
22333
22334 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
22335 Change type to hash_table. Update dependent calls and types.
22336
22337 * tree-ssa-loop-im.c (struct mem_ref.refs):
22338 Change type to hash_table. Update dependent calls and types.
22339
22340 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
22341 Change type to hash_table. Update dependent calls and types.
22342
22343 * tree-ssa-sccvn.c (vn_tables_s::nary):
22344 Change type to hash_table. Update dependent calls and types.
22345 (vn_tables_s::phis): Likewise.
22346 (vn_tables_s::references): Likewise.
22347
22348 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
22349 (vn_reference_eq): Update parameter and return types.
22350
22351 * tree-ssa-structalias.c (pointer_equiv_class_table):
22352 Change type to hash_table. Update dependent calls and types.
22353 (location_equiv_class_table): Likewise.
22354
22355 * tree-vect-data-refs.c: Consequential changes for making
22356 peeling a hash_table.
22357
22358 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
22359 (destroy_loop_vec_info): Dependent hash_table update.
22360
22361 * tree-vectorizer.h (peeling_htab):
22362 Change type to hash_table. Update dependent calls and types.
22363
22364 2013-04-23 Shiva Chen <shiva0217@gmail.com>
22365
22366 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
22367 to check the register content is equal or not.
22368 * lra-constraints.c (match_reload): Use lra_assign_reg_val
22369 to assign register content record.
22370 * lra-eliminations.c (update_reg_eliminate): Use
22371 lra_update_reg_val_offset to update register content offset.
22372 * lra-int.h (struct lra_reg): Add offset member.
22373 (lra_reg_val_equal_p): New static inline function.
22374 (lra_update_reg_val_offset): New static inline function.
22375 (lra_assign_reg_val): New static inline function.
22376 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
22377 to assign register content record.
22378 (initialize_lra_reg_info_element): Initial offset to zero.
22379
22380 2013-04-23 Catherine Moore <clm@codesourcery.com>
22381
22382 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
22383 operands. Record compression.
22384
22385 2013-04-23 Xinliang David Li <davidxl@google.com>
22386
22387 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
22388
22389 2013-04-23 Richard Biener <rguenther@suse.de>
22390
22391 PR middle-end/57036
22392 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
22393 parameter, only add abnormal goto edges from the copied body
22394 if the call could perform abnormal gotos.
22395 (copy_cfg_body): Adjust.
22396
22397 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
22398
22399 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
22400
22401 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
22402
22403 * coretypes.h (gimple_stmt_iterator): Add struct to make
22404 compatible with C.
22405
22406 2013-04-23 Richard Biener <rguenther@suse.de>
22407
22408 PR tree-optimization/57026
22409 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
22410 from SSA names occuring in abnormal PHI nodes.
22411
22412 2013-04-22 Andi Kleen <ak@linux.intel.com>
22413
22414 * lto/lto.c (print_lto_report_1): Fix LTO report names.
22415
22416 2013-04-22 Andi Kleen <ak@linux.intel.com>
22417
22418 * lto/lto.c (print_lto_report_1): Declare early.
22419 (read_cgraph_and_symbols): Call print_lto_report_1 early.
22420
22421 2013-04-22 Andi Kleen <ak@linux.intel.com>
22422
22423 * common.opt (-flto-report-wpa): Add.
22424 * doc/invoke.texi (-flto-report-wpa): Add.
22425 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
22426 (lto_main): dito.
22427
22428 2013-04-22 Xinliang David Li <davidxl@google.com>
22429
22430 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
22431 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
22432 * Makefile.in: New dependency
22433
22434 David Daney <ddaney.cavm@gmail.com>
22435
22436 * configure.ac (gcc_cv_as_micromips_support): Use the
22437 --fatal-warnings option.
22438 * configure: Regenerate.
22439
22440 2013-04-22 Marek Polacek <polacek@redhat.com>
22441
22442 PR sanitizer/56990
22443 * tsan.c (instrument_expr): Don't instrument expression
22444 in case its size is zero.
22445
22446 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
22447
22448 PR target/57032
22449 Revert:
22450 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22451
22452 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22453
22454 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22455
22456 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
22457 (gimple_stmt_iterator): New typedef.
22458 * gimple.h (gimple_stmt_iterator): Rename to...
22459 (gimple_stmt_iterator_d): ... This.
22460 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
22461 trees be valid for GIMPLE and GENERIC.
22462 (TARGET_GIMPLE_FOLD_BUILTIN): New.
22463 * gimple-fold.c (gimple_fold_call): Call target hook
22464 gimple_fold_builtin.
22465 * hooks.c (hook_bool_gsiptr_false): New.
22466 * hooks.h (hook_bool_gsiptr_false): New.
22467 * target.def (fold_stmt): New.
22468 * doc/tm.texi: Regenerate.
22469
22470 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
22471
22472 PR target/57018
22473 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
22474 a set sp if no stack realignment.
22475
22476 2013-04-22 Nick Clifton <nickc@redhat.com>
22477
22478 * config.gcc (tilegx-linux): Extend extra_objs rather than
22479 overwriting it.
22480 (tilepro-linux): Likewise.
22481
22482 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22483
22484 * config/aarch64/aarch64-builtins.c
22485 (CF): Remove.
22486 (CF0, CF1, CF2, CF3, CF4, CF10): New.
22487 (VAR<1-12>): Add MAP parameter.
22488 (BUILTIN_*): Likewise.
22489 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
22490 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
22491 (aarch64_ushl_n<mode>): Likewise.
22492 (aarch64_sshr_n<mode>): Likewise.
22493 (aarch64_ushr_n<mode>): Likewise.
22494 (aarch64_<maxmin><mode>): Likewise.
22495 (aarch64_sqrt<mode>): Likewise.
22496 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
22497 (vshr<q>_n_*): Likewise.
22498
22499 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22500
22501 * config/aarch64/aarch64-builtins.c
22502 (aarch64_simd_builtin_type_mode): Handle SF types.
22503 (sf_UP): Define.
22504 (BUILTIN_GPF): Define.
22505 (aarch64_init_simd_builtins): Handle SF types.
22506 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
22507 (frecps): Likewise.
22508 (frecpx): Likewise.
22509 * config/aarch64/aarch64-simd.md
22510 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
22511 (aarch64_frecpe<mode>): New.
22512 (aarch64_frecps<mode>): Likewise.
22513 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
22514 (v8type): Add frecp<esx>.
22515 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
22516 (aarch64_frecps<mode>): Likewise.
22517 * config/aarch64/iterators.md (FRECP): New.
22518 (frecp_suffix): Likewise.
22519 * config/aarch64/arm_neon.h
22520 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
22521
22522 2013-04-22 Christian Bruel <christian.bruel@st.com>
22523
22524 PR target/56995
22525 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
22526 (REG_CLASS_NAMES): Idem.
22527 (REG_CLASS_CONTENTS): Idem.
22528 (REGCLASS_HAS_FP_REG): Idem.
22529 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
22530 (sh_conditional_register_usage): Idem.
22531
22532 2013-04-21 Jeff Law <law@redhat.com>
22533
22534 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
22535 (ssa_forward_propagate_and_combine): Use it.
22536
22537 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22538
22539 * lra.c: Update the flow chart diagram.
22540
22541 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22542
22543 PR rtl-optimization/56847
22544 * lra-constraints.c (process_alt_operands): Discourage alternative
22545 with non-matche doffsettable memory constraint fro memory with
22546 known offset.
22547
22548 2013-04-19 Richard Biener <rguenther@suse.de>
22549
22550 PR tree-optimization/56982
22551 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
22552 function.
22553 * gimplify.c (gimplify_call_expr): Notice special calls.
22554 (gimplify_modify_expr): Likewise.
22555 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
22556 abnormal control flow receivers.
22557 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
22558 in the same way as cfun->has_nonlocal_labels.
22559 (gimple_purge_dead_abnormal_call_edges): Likewise.
22560 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
22561 receivers start a basic-block.
22562
22563 2013-04-19 Richard Biener <rguenther@suse.de>
22564
22565 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
22566 member ...
22567 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
22568 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
22569 (SLP_TREE_LOAD_PERMUTATION): Add.
22570 (vect_transform_slp_perm_load): Adjust prototype.
22571 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
22572 (vect_free_slp_instance): Likewise.
22573 (vect_create_new_slp_node): Likewise.
22574 (vect_supported_slp_permutation_p): Remove.
22575 (vect_slp_rearrange_stmts): Adjust.
22576 (vect_supported_load_permutation_p): Likewise. Inline
22577 vect_supported_slp_permutation_p here.
22578 (vect_analyze_slp_instance): Compute load permutations per
22579 slp node instead of per instance.
22580 (vect_get_slp_defs): Adjust.
22581 (vect_transform_slp_perm_load): Likewise.
22582 (vect_schedule_slp_instance): Remove redundant code.
22583 (vect_schedule_slp): Remove hack for PR56270, add it ...
22584 * tree-vect-stmts.c (vectorizable_load): ... here, do not
22585 CSE loads for SLP. Adjust.
22586
22587 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22588
22589 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
22590 spelling in two comments.
22591
22592 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22593
22594 PR target/56797
22595 * config/arm/arm.c (load_multiple_sequence): Require SP
22596 as base register for loads if SP is in the register list.
22597
22598 2013-04-19 Martin Jambor <mjambor@suse.cz>
22599
22600 PR tree-optimization/56718
22601 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
22602 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
22603 and made public. Adjusted all callers.
22604 (ipa_intraprocedural_devirtualization): New function.
22605 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
22606 (ipa_intraprocedural_devirtualization): Likewise.
22607 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
22608
22609 2013-04-19 Richard Biener <rguenther@suse.de>
22610
22611 PR tree-optimization/57000
22612 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
22613
22614 2013-04-19 Terry Guo <terry.guo@arm.com>
22615
22616 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
22617 Replace with ...
22618 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
22619 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
22620 (cortex_m4_fmacs): Use new reservations.
22621 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
22622
22623 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
22624
22625 PR rtl-optimization/56999
22626 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
22627 related code.
22628 (lra_coalesce): Remove split_origin_bitmap and related code.
22629 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
22630 ranges if necessary.
22631
22632 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
22633
22634 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
22635 New array.
22636 (ix86_expand_call): Remove clobbered_registers array and use
22637 x86_64_ms_sysv_extra_clobbered_registers instead.
22638 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
22639 Declare here.
22640 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
22641 predicate.
22642 * config/i386/i386.md (*call_rex64_ms_sysv): Use
22643 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
22644 (*call_value_rex64_ms_sysv): Ditto.
22645
22646 2013-04-18 Cary Coutant <ccoutant@google.com>
22647
22648 * dwarf2out.c (output_pubnames): Check die_perennial_p of
22649 parent instead of die_mark.
22650
22651 2013-04-18 Diego Novillo <dnovillo@google.com>
22652
22653 * gimple.c (create_gimple_tmp): New.
22654 (get_expr_type): New.
22655 (build_assign): New.
22656 (build_type_cast): New.
22657 * gimple.h (enum ssa_mode): Define.
22658 (gimple_seq_set_location): New.
22659 * asan.c (build_check_stmt): Change some gimple_build_* calls
22660 to use build_assign and build_type_cast.
22661
22662 2013-04-18 Richard Biener <rguenther@suse.de>
22663
22664 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
22665 handle negative step. Remove redundant checks.
22666 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
22667 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
22668 for negative step and grouped loads fail to vectorize.
22669
22670 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22671
22672 * emit-rtl.c (reset_insn_used_flags): New function.
22673 (reset_all_used_flags): Use it.
22674 (verify_insn_sharing): New function.
22675 (verify_rtl_sharing): Fix verification for SEQUENCEs.
22676
22677 2013-04-18 Jakub Jelinek <jakub@redhat.com>
22678
22679 PR tree-optimization/56984
22680 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
22681 and (x >> M) >= N don't register any assertion if N << M is the
22682 minimum value.
22683
22684 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22685
22686 * lower-subreg.c (resolve_simple_move): If called self-recursive,
22687 do not delete_insn insns that have not yet been emitted, only
22688 unlink them with remove_insn.
22689 * df-scan.c (df_insn_delete): Revert r197492.
22690
22691 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
22692
22693 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
22694 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
22695
22696 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22697
22698 * config/arm/arm.md (movsicc_insn): Convert define_insn into
22699 define_insn_and_split.
22700 (and_scc,ior_scc,negscc): Likewise.
22701 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
22702
22703 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22704
22705 * config/arm/arm.c (use_return_insn): Return 0 for targets that
22706 can benefit from using a sequence of LDRD instructions in epilogue
22707 instead of a single LDM instruction.
22708
22709 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
22710
22711 PR 45688
22712 * doc/extend.texi: Fix typo.
22713
22714 2013-04-17 Richard Biener <rguenther@suse.de>
22715
22716 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
22717 (vect_build_slp_tree): ... here.
22718 (vect_build_slp_tree_1): Compute which stmts of the SLP group
22719 match. Remove special-casing of mismatched complex loads.
22720 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
22721 re-try the match with swapped commutative operands.
22722 (vect_supported_load_permutation_p): Remove special-casing of
22723 mismatched complex loads.
22724 (vect_analyze_slp_instance): Adjust.
22725
22726 2013-04-17 Richard Biener <rguenther@suse.de>
22727
22728 PR rtl-optimization/56921
22729 * cfgloop.h (struct loop): Add simple_loop_desc member.
22730 (struct niter_desc): Mark with GTY(()).
22731 (simple_loop_desc): Do not use aux field but simple_loop_desc.
22732 * loop-iv.c (get_simple_loop_desc): Likewise.
22733 (free_simple_loop_desc): Likewise.
22734
22735 Revert
22736 2013-04-16 Richard Biener <rguenther@suse.de>
22737
22738 PR rtl-optimization/56921
22739 * loop-init.c (pass_rtl_move_loop_invariants): Add
22740 TODO_do_not_ggc_collect to todo_flags_finish.
22741 (pass_rtl_unswitch): Same.
22742 (pass_rtl_unroll_and_peel_loops): Same.
22743 (pass_rtl_doloop): Same.
22744
22745 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
22746
22747 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
22748 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
22749 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
22750 references.
22751 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
22752 * tree-streamer.c (record_common_node): Adjust reference in comment.
22753
22754 2013-04-17 Terry Guo <terry.guo@arm.com>
22755
22756 * config/arm/cortex-m4.md: Add a new bypass.
22757
22758 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22759
22760 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
22761 New pattern.
22762 (*subs_<optab><mode>_multp2): New pattern.
22763 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22764 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22765
22766 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22767
22768 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
22769 (*subs_mul_imm_<mode>): New pattern.
22770
22771 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
22772
22773 PR target/56948
22774 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
22775 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
22776 (vsx_movti_32bit): Change j->wa to O->wa.
22777
22778 2013-04-16 Richard Biener <rguenther@suse.de>
22779
22780 PR rtl-optimization/56921
22781 * loop-init.c (pass_rtl_move_loop_invariants): Add
22782 TODO_do_not_ggc_collect to todo_flags_finish.
22783 (pass_rtl_unswitch): Same.
22784 (pass_rtl_unroll_and_peel_loops): Same.
22785 (pass_rtl_doloop): Same.
22786
22787 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
22788
22789 * config/arm/arm.c (emit_multi_reg_push): New declaration
22790 for an existing function.
22791 (arm_emit_strd_push): New function.
22792 (arm_expand_prologue): Used here.
22793 (arm_emit_ldrd_pop): New function.
22794 (arm_expand_epilogue): Used here.
22795 (arm_get_frame_offsets): Update condition.
22796 (arm_emit_multi_reg_pop): Add a special case for load of a single
22797 register with writeback.
22798
22799 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
22800
22801 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
22802 description.
22803
22804 2013-04-16 Richard Biener <rguenther@suse.de>
22805
22806 PR tree-optimization/56756
22807 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
22808 (first_mem_ref_loc): New.
22809 (execute_sm): Place the load temporarily before a previous
22810 access instead of in the latch edge to ensure its SSA dependencies
22811 are defined at points dominating the load.
22812
22813 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
22814
22815 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
22816 correct fix by moving header and footer insn to the footer of
22817 the merged basic block. Clear BB_END of the merged-away block.
22818
22819 PR middle-end/43631
22820 * emit-rtl.c (make_note_raw): New function.
22821 (link_insn_into_chain): New static inline function.
22822 (add_insn): Use it.
22823 (add_insn_before, add_insn_after): Factor insn chain linking code...
22824 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
22825 using link_insn_into_chain.
22826 (note_outside_basic_block_p): New helper function for emit_note_after
22827 and emit_note_before.
22828 (emit_note_after): Use nobb variant of add_insn_after if the note
22829 should not be contained in a basic block.
22830 (emit_note_before): Use nobb variant of add_insn_before if the note
22831 should not be contained in a basic block.
22832 (emit_note_copy): Use make_note_raw.
22833 (emit_note): Likewise.
22834 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
22835 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
22836 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
22837 the moved barrier the tail of the basic block it follows.
22838 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
22839
22840 2013-04-15 Jakub Jelinek <jakub@redhat.com>
22841
22842 PR tree-optimization/56962
22843 * gimple-ssa-strength-reduction.c (record_increment): Only set
22844 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
22845 either rhs1 or rhs2 is equal to c->base_expr.
22846
22847 2013-04-15 Richard Biener <rguenther@suse.de>
22848
22849 PR tree-optimization/56933
22850 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
22851 member.
22852 (GROUP_READ_WRITE_DEPENDENCE): Remove.
22853 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
22854 * tree-vect-data-refs.c (vect_analyze_group_access): Move
22855 dependence check ...
22856 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
22857 ... here.
22858 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
22859 GROUP_READ_WRITE_DEPENDENCE.
22860
22861 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22862
22863 * emit-rtl.c (reset_all_used_flags): New function.
22864 (verify_rtl_sharing): Call reset_all_used_flags before and after
22865 performing the checks.
22866
22867 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22868
22869 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
22870 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
22871 * config/arm/constraints.md (De): New constraint.
22872 * config/arm/neon.md (anddi3_neon): Delete.
22873 (neon_vand<mode>): Expand to standard anddi3 pattern.
22874 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
22875 Move earlier in the file.
22876 (neon_inv_logic_op2): Likewise.
22877 (arm_anddi_operand_neon): New predicate.
22878
22879 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22880
22881 * configure.ac (gcc_cv_ld_as_needed): Set
22882 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
22883 Use -z ignore, -z record on *-*-solaris2*.
22884 (HAVE_LD_AS_NEEDED): Update comment.
22885 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
22886 * configure: Regenerate.
22887 * config.in: Regenerate.
22888 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
22889 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
22890 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
22891 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
22892 equivalents. Fix markup.
22893 * doc/tm.texi: Regenerate.
22894
22895 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
22896
22897 * config/i386/i386.opt: New option mstack-protector-guard=.
22898 * config/i386/i386-opts.h: Add enum stack_protector_guard.
22899 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
22900 TARGET_SSP_TLS_GUARD.
22901 * config/i386/i386.c (ix86_option_override_internal): Set
22902 ix86_stack_protector_guard.
22903 * config/i386/i386.md (stack_protect_set): Enable for
22904 TARGET_SSP_TLS_GUARD only.
22905 (stack_protect_set_<mode>): Ditto.
22906 (stack_protect_test): Ditto.
22907 (stack_protect_test_<mode>): Ditto.
22908 * doc/invoke.texi (i386 Option): Document.
22909
22910 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
22911
22912 PR target/56890
22913 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
22914 (S_MODES): Set H_MODE bit.
22915 (SF_MODES): Set only S_MODE and SF_MODE bits.
22916 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
22917 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
22918 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
22919 <MODE_FLOAT>: Likewise.
22920
22921 2013-04-15 Joey Ye <joey.ye@arm.com>
22922
22923 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
22924
22925 2013-04-15 Joey Ye <joey.ye@arm.com>
22926
22927 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
22928 for real far jump.
22929 (thumb_far_jump_used_p): Count instruction size and set
22930 far_jump_used.
22931
22932 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
22933
22934 * reorg.c (fill_simple_delay_slots): Reindent block of code.
22935 * resource.c (mark_target_live_regs): Reformat conditional block.
22936
22937 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22938
22939 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
22940 notes, they are emitted only just before final.
22941 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
22942
22943 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22944
22945 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
22946 * cfgrtl.c (delete_insn): Call it here instead.
22947 * lra-spills.c (lra_final_code_change): Use delete_insn.
22948 * haifa-sched.c (sched_remove_insn): Likewise.
22949 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
22950 returning to the nop pool.
22951 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
22952 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
22953
22954 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22955
22956 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
22957 * doc/tm.texi: Regenerated.
22958
22959 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
22960
22961 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
22962 QImode checks.
22963
22964 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22965
22966 * df-core.c (df_find_def): Compare register numbers.
22967 (df_find_use): Likewise.
22968
22969 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
22970
22971 PR target/56903
22972 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
22973 lra_in_progress for return.
22974
22975 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22976
22977 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
22978 define_insn into define_insn_and_split and emit movsicc patterns.
22979
22980 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22981
22982 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
22983
22984 2013-04-12 Richard Biener <rguenther@suse.de>
22985
22986 * tree-pass.h (TODO_do_not_ggc_collect): New.
22987 * passes.c (execute_one_ipa_transform_pass): Honor
22988 TODO_do_not_ggc_collect.
22989 (execute_one_pass): Likewise.
22990
22991 Revert
22992 2013-04-10 Richard Biener <rguenther@suse.de>
22993
22994 * passes.c (init_optimization_passes): Remove reload pass.
22995 * ira.c (do_reload): Merge into ...
22996 (ira): ... this.
22997 (rest_of_handle_reload): Remove.
22998 (pass_reload): Likewise.
22999 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23000 of reload for vzeroupper pass placement.
23001
23002 2013-04-12 Jakub Jelinek <jakub@redhat.com>
23003
23004 PR tree-optimization/56918
23005 PR tree-optimization/56920
23006 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
23007 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
23008 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
23009 use wide_mul_with_sign method.
23010
23011 2013-04-12 Richard Biener <rguenther@suse.de>
23012
23013 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
23014 not be considered a gimple constant.
23015
23016 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
23017
23018 * fold-const.c (const_binop): Handle vector shifts by a scalar.
23019 (fold_binary_loc): Call const_binop also for mixed vector-scalar
23020 operations.
23021
23022 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
23023 Jakub Jelinek <jakub@redhat.com>
23024
23025 * opts.c: Include diagnostic-color.h.
23026 (common_handle_option): Handle OPT_fdiagnostics_color_.
23027 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
23028 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
23029 (diagnostic-color.o): New.
23030 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
23031 (diagnostic_color_rule): New enum.
23032 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
23033 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
23034 the location string.
23035 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
23036 either NULL, or color kind.
23037 * diagnostic-color.c: New file.
23038 * diagnostic-color.h: New file.
23039 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
23040 arguments.
23041 * doc/invoke.texi (-fdiagnostics-color): Document.
23042 * pretty-print.h (pp_show_color): Define.
23043 (struct pretty_print_info): Add show_color field.
23044 * diagnostic.c: Include diagnostic-color.h.
23045 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
23046 macros. Colorize error:, warning: etc. strings and also the location
23047 string.
23048 (diagnostic_show_locus): Colorize the caret line.
23049 * pretty-print.c: Include diagnostic-color.h.
23050 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
23051 inside of %< %> quotes or quoted through q format modifier.
23052
23053 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23054
23055 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
23056
23057 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23058
23059 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
23060 code in CC_NZ mode.
23061 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
23062 pattern.
23063
23064 2013-04-11 Marek Polacek <polacek@redhat.com>
23065
23066 PR tree-optimization/48184
23067 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
23068
23069 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
23070
23071 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
23072 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
23073 (skip_simple_arithmetic): Tidy up.
23074 * tree.h (skip_simple_constant_arithmetic): Declare.
23075
23076 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23077
23078 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
23079
23080 2013-04-11 Richard Biener <rguenther@suse.de>
23081
23082 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23083 generate vector constants.
23084
23085 2013-04-11 Richard Biener <rguenther@suse.de>
23086
23087 PR tree-optimization/56878
23088 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
23089 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
23090 New function.
23091 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
23092 Prefer to align the DR with the most invariant base address.
23093
23094 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23095
23096 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
23097 comment.
23098
23099 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
23100
23101 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
23102 floating-point vector comparisons against 0.
23103
23104 2013-04-11 Jakub Jelinek <jakub@redhat.com>
23105
23106 PR tree-optimization/56899
23107 * fold-const.c (extract_muldiv_1): Apply distributive law
23108 only if TYPE_OVERFLOW_WRAPS (ctype).
23109
23110 2013-04-11 Bin Cheng <bin.cheng@arm.com>
23111
23112 PR target/56124
23113 * ira-costs.c (scan_one_insn): Check whether the source rtx of
23114 loading has side effect.
23115
23116 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
23117
23118 * config/sparc/sparc.c: Include tree-pass.h.
23119 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
23120 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
23121 head of file. Change return type. Split off gate function.
23122 (sparc_gate_work_around_errata): New function.
23123 (pass_work_around_errata): New pass definition.
23124 (insert_pass_work_around_errata) New pass insert definition to
23125 insert pass_work_around_errata just after delayed-branch scheduling.
23126 (sparc_option_override): Insert the pass.
23127 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
23128
23129 2013-04-10 David S. Miller <davem@davemloft.net>
23130
23131 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
23132 or -mcpu=hypersparc.
23133
23134 * target.def (cstore_mode): New hook.
23135 * target.h: Include insn-codes.h
23136 * targhooks.c: Likewise.
23137 (default_cstore_mode): New function.
23138 * targhooks.h: Declare it.
23139 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
23140 * doc/tm.texi: Rebuild.
23141 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
23142 target hook, rather than inspecting the insn_data.
23143 * config/sparc/sparc.c (sparc_cstore_mode): New function.
23144 (TARGET_CSTORE_MODE): Redefine.
23145 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
23146 result patterns.
23147 * config/sparc/predicates.md (cstore_result_operand): New special
23148 predicate.
23149 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
23150 Use it for operand 0.
23151 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
23152 (*snesi_special): Likewise.
23153 (*snesi_zero): Likewise.
23154 (*seqsi_zero): Likewise.
23155 (*sltu_insn): Likewise.
23156 (*sgeu_insn): Likewise.
23157 (*seqdi_special): Make operand 0 and comparison operation be of
23158 DImode.
23159 (*snedi_special): Likewise.
23160 (*snedi_special_vis3): Likewise.
23161 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
23162 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
23163 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
23164 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
23165 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
23166 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
23167 (*sltu_extend_sp64): Likewise.
23168 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
23169 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
23170 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
23171 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
23172 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
23173
23174 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
23175
23176 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
23177 (aarch64_start_file): Use the new function.
23178
23179 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23180 Jason Merrill <jason@redhat.com>
23181
23182 * common.opt: Add -gdwarf.
23183 * opts.c (common_handle_option): Handle it.
23184 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
23185
23186 2013-04-10 Richard Biener <rguenther@suse.de>
23187
23188 * passes.c (execute_todo): Do not call ggc_collect conditional here.
23189 (execute_one_ipa_transform_pass): But unconditionally here.
23190 (execute_one_pass): And here.
23191 (init_optimization_passes): Remove reload pass.
23192 * tree-pass.h (TODO_ggc_collect): Remove.
23193 (pass_reload): Likewise.
23194 * ira.c (do_reload): Merge into ...
23195 (ira): ... this.
23196 (rest_of_handle_reload): Remove.
23197 (pass_reload): Likewise.
23198 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23199 of reload for vzeroupper pass placement.
23200 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
23201 and todo_flags_finish of all passes.
23202
23203 2013-04-10 Richard Biener <rguenther@suse.de>
23204
23205 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
23206 first_const_oprnd field, rename first_def_type to first_op_type.
23207 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
23208 (vect_get_and_check_slp_defs): Always use the type of the
23209 operand. Allow mixed vect_external_def, vect_constant_def types.
23210 (vect_get_constant_vectors): Handle mixed vect_external_def,
23211 vect_constant_def types.
23212
23213 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
23214
23215 PR tree-optimization/55524
23216 * tree-ssa-math-opts.c
23217 (convert_mult_to_fma): Don't use an fms construct
23218 when we don't have an fms operation, but fnma, and it looks
23219 likely that we'll be able to use the latter.
23220
23221 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
23222
23223 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
23224 function.
23225 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
23226 inline fail caused by overwritable functions.
23227
23228 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
23229
23230 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
23231 unnecessary bits in the constant power of two case.
23232
23233 2013-04-10 Richard Biener <rguenther@suse.de>
23234
23235 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
23236 broken code swapping operands.
23237 (vect_build_slp_tree): Do not compute load permutations here.
23238 (vect_analyze_slp_instance): Compute load permutations here,
23239 after building the SLP tree.
23240
23241 2013-04-09 Christian Bruel <christian.bruel@st.com>
23242
23243 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
23244 of next/prev_real_insn.
23245
23246 2013-04-09 Jan Hubicka <jh@suse.cz>
23247
23248 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
23249 Drop aliased parameter.
23250 (function_and_variable_visibility): Do not handle alias pairs.
23251 * cgraph.c (varpool_externally_visible_p): Update prototype.
23252 * varpool.c (varpool_add_new_variable): Update.
23253
23254 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23255
23256 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
23257
23258 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
23259
23260 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
23261
23262 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
23263
23264 2013-04-09 Marek Polacek <polacek@redhat.com>
23265
23266 PR tree-optimization/48762
23267 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
23268
23269 2013-04-09 Richard Biener <rguenther@suse.de>
23270
23271 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
23272 dealing with cost.
23273 (vect_build_slp_tree): Likewise.
23274 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
23275 calculating the cost of a SLP instance.
23276 (vect_analyze_slp_instance): Use it from here, after building
23277 the SLP tree.
23278
23279 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23280
23281 PR middle-end/56883
23282 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
23283 expand_omp_for_static_chunk): Use simple_p = true in
23284 force_gimple_operand_gsi calls when assigning to addressable decls.
23285
23286 2013-04-09 Jeff Law <law@redhat.com>
23287
23288 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
23289 when the boolean was created by converting a wider object which
23290 had a boolean range.
23291
23292 2013-04-09 Richard Biener <rguenther@suse.de>
23293
23294 * tree-vectorizer.h (slp_void_p): Remove.
23295 (slp_tree): Typedef before _slp_tree declaration.
23296 (struct _slp_tree): Use a vector of slp_tree as children.
23297 (vect_get_place_in_interleaving_chain): Remove.
23298 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
23299 Move ...
23300 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
23301 and make static.
23302 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
23303 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
23304 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
23305 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
23306 Use slp_node instead of slp_void_p and adjust.
23307
23308 2013-04-09 Richard Biener <rguenther@suse.de>
23309
23310 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
23311 work that is not necessary.
23312
23313 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23314
23315 PR tree-optimization/56854
23316 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
23317 forward into clobber stmts if it would change MEM_REF lhs into
23318 non-MEM_REF.
23319
23320 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
23321
23322 * tree.c (type_hash_lookup, type_hash_add): Make static.
23323 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
23324
23325 2013-04-09 Richard Biener <rguenther@suse.de>
23326
23327 * tree.h (unsave_expr_now): Remove.
23328 * tree-inline.c (mark_local_for_remap_r): Remove.
23329 (unsave_expr_1): Likewise.
23330 (unsave_r): Likewise.
23331 (unsave_expr_now): Likewise.
23332 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
23333 (propagate_tree_value): Likewise.
23334
23335 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
23336
23337 * doc/rtl.texi (sequence): Rewrite documentation to match the
23338 current use of SEQUENCE rtl objects.
23339 * rtl.def (SEQUENCE): Likewise.
23340
23341 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
23342 Update documentation.
23343 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
23344 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
23345
23346 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
23347
23348 2013-04-08 Teresa Johnson <tejohnson@google.com>
23349
23350 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
23351 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
23352 methods.
23353 (estimate_edge_size_and_time): Add comment to suggest using rounding
23354 methods.
23355 (estimate_node_size_and_time): Ditto.
23356 (remap_edge_change_prob): Use helper rounding divide methods.
23357 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
23358 (gimple_mod_pow2_value_transform): Ditto.
23359 (gimple_mod_subtract_transform): Ditto.
23360 (gimple_ic_transform): Ditto.
23361 (gimple_stringops_transform): Ditto.
23362 * stmt.c (conditional_probability): Ditto.
23363 (emit_case_dispatch_table): Ditto.
23364 * lto-cgraph.c (merge_profile_summaries): Ditto.
23365 * tree-optimize.c (execute_fixup_cfg): Ditto.
23366 * cfgcleanup.c (try_forward_edges): Ditto.
23367 * cfgloopmanip.c (scale_loop_profile): Ditto.
23368 (loopify): Ditto.
23369 (duplicate_loop_to_header_edge): Ditto.
23370 (lv_adjust_loop_entry_edge): Ditto.
23371 * tree-vect-loop.c (vect_transform_loop): Ditto.
23372 * profile.c (compute_branch_probabilities): Ditto.
23373 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
23374 * lto-streamer-in.c (input_cfg): Ditto.
23375 * gimple-streamer-in.c (input_bb): Ditto.
23376 * ipa-cp.c (update_profiling_info): Ditto.
23377 (update_specialized_profile): Ditto.
23378 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
23379 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
23380 rounding methods.
23381 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23382 (compute_trg_info): Ditto.
23383 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
23384 (purge_dead_edges): Ditto.
23385 * loop-unswitch.c (unswitch_loop): Ditto.
23386 * cgraphclones.c (cgraph_clone_edge): Ditto.
23387 (cgraph_clone_node): Ditto.
23388 * tree-inline.c (copy_bb): Ditto.
23389 (copy_edges_for_bb): Ditto.
23390 (initialize_cfun): Ditto.
23391 (copy_cfg_body): Ditto.
23392 (expand_call_inline): Ditto.
23393
23394 2013-04-08 Kai Tietz <ktietz@redhat.com>
23395
23396 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
23397 TARGET_CYGWIN64 by TARGET_64BIT.
23398
23399 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23400
23401 * config/epiphany/epiphany.md (GPR_1): New constant.
23402 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
23403 * config/epiphany/epiphany.c (gen_compare_reg):
23404 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
23405 is already in place.
23406 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
23407 Don't require being called during rtl expansion; If y operlaps r0,
23408 return 0.
23409 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
23410 (epiphany_expand_epilogue): Likewise.
23411
23412 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
23413 Don't use CC_FPmode for ORDERED / UNORDERED.
23414 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
23415
23416 * config/epiphany/constraints.md (CnL): New constraint.
23417 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
23418 * config/epiphany/predicates.md (add_operand): Allow 1024.
23419
23420 * config/epiphany/epiphany.md (logical_op): New code iterator.
23421 (op_mnc): New code attribute.
23422 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
23423 (mov_f+1, mov_f+2): New peephole2 patterns.
23424
23425 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
23426 (cstoresi4): Also allow re-use of zero result when doing a NE
23427 comparison to a non-zero operand.
23428 Use (clobber (scratch)) for first insn if the gpr output is not needed.
23429
23430 * config/epiphany/epiphany.md (<insn_opname>v2si3):
23431 Use gen_addsi3_i / gen_subsi3_i.
23432
23433 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23434
23435 PR c++/34949
23436 PR c++/50243
23437 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
23438 contain anything but clobbers, at most one __builtin_stack_restore,
23439 optionally debug stmts and final resx, and if it has at least one
23440 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
23441 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
23442 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
23443 which isn't defaut definition, remove them.
23444 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
23445 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
23446 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
23447 with MEM_REF LHS with SSA_NAME address.
23448
23449 2013-04-08 Jeff Law <law@redhat.com>
23450
23451 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
23452
23453 2013-04-08 Richard Biener <rguenther@suse.de>
23454
23455 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
23456 extra newline.
23457 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
23458 determined vector type.
23459 (vect_analyze_data_refs): Likewise.
23460 (vect_get_new_vect_var): Adjust.
23461 (vect_create_destination_var): Preserve SSA name versions.
23462 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
23463 not dump anything here.
23464
23465 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23466
23467 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
23468 Add member lr_slot_known.
23469 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
23470 if necessary.
23471 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
23472 Remove code that sets lr_slot_offset according to what a previous
23473 version of epiphany_emit_save_restore used to do.
23474 (epiphany_emit_save_restore): When doing an lr save or restore,
23475 set/verify lr_slot_known and lr_slot_offset.
23476
23477 2013-04-08 Xinyu Qi <xyqi@marvell.com>
23478
23479 PR target/54338
23480 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
23481 in ALL_REGS.
23482
23483 2013-04-08 Richard Biener <rguenther@suse.de>
23484
23485 * alias.c (find_base_term): Fix thinko in previous change.
23486
23487 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23488
23489 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
23490 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
23491 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
23492 if possible to compute val.
23493 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
23494 For QImode integers don't require anything about precision. Use
23495 const_with_all_bytes_same to find out if the constant doesn't have
23496 repeated bytes in it.
23497
23498 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23499
23500 * config/s390/s390.c (s390_expand_insv): Only accept insertions
23501 within mode size.
23502
23503 2013-04-08 Marek Polacek <polacek@redhat.com>
23504
23505 PR rtl-optimization/48182
23506 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
23507 value to 1.
23508
23509 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23510
23511 PR target/55487
23512 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
23513 nuses, make sure we have a label.
23514
23515 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23516
23517 PR target/56843
23518 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
23519 (rs6000_emit_swdiv_low_precision): Remove.
23520 (rs6000_emit_swdiv): Rewrite to handle between one and four
23521 iterations of Newton-Raphson generally; modify required number of
23522 iterations for some cases.
23523 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
23524
23525 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
23526
23527 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
23528 set-but-unused variable.
23529
23530 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
23531 basic blocks of released function bodies garbage-collectable.
23532
23533 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
23534 (struct rtl_opt_pass): Add TODO_df_finish.
23535
23536 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
23537
23538 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23539
23540 * config/arm/constraints.md (q): New constraint.
23541 * config/arm/ldrdstrd.md: New file.
23542 * config/arm/arm.md (ldrdstrd.md) New include.
23543 (arm_movdi): Use "q" instead of "r" constraint
23544 for double-word memory access.
23545 (movdf_soft_insn): Likewise.
23546 * config/arm/vfp.md (movdi_vfp): Likewise.
23547 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
23548 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
23549 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
23550 (mem_ok_for_ldrd_strd): Likewise.
23551 (output_move_double): Update assertion.
23552
23553 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23554
23555 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
23556
23557 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23558
23559 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
23560 define_insn_and_split.
23561 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
23562
23563 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23564
23565 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
23566 define_insn_and_split.
23567 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
23568 (shiftsi3_compare): New pattern.
23569 (rrx): New pattern.
23570 * config/arm/unspecs.md (UNSPEC_RRX): New.
23571
23572 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23573
23574 * config/arm/arm.md (negdi_extendsidi): New pattern.
23575 (negdi_zero_extendsidi): Likewise.
23576
23577 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23578
23579 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
23580 define_insn_and_split.
23581 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
23582 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
23583
23584 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23585
23586 * config/arm/arm.md (arm_subdi3): Convert define_insn into
23587 define_insn_and_split.
23588 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
23589 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
23590
23591 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23592
23593 * config/arm/arm.md (subsi3_carryin): New pattern.
23594 (subsi3_carryin_const): Likewise.
23595 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
23596 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
23597
23598 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23599
23600 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
23601
23602 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23603
23604 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
23605 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
23606
23607 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23608
23609 * config/arm/arm.c (arm_expand_builtin): Change fcode
23610 type to unsigned int.
23611
23612 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23613
23614 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
23615
23616 2013-04-04 Ian Lance Taylor <iant@google.com>
23617
23618 * doc/standards.texi (Standards): The Go frontend supports the Go 1
23619 language standard.
23620
23621 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
23622
23623 PR middle-end/56729
23624 * df-scan.c (df_insn_delete): Disable failing assert.
23625
23626 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23627
23628 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
23629 New function prototype.
23630 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
23631 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
23632 (arm_builtin_vectorized_function): New function.
23633
23634 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23635
23636 * config/arm/arm_neon_builtins.def: New file.
23637 * config/arm/arm.c (neon_builtin_data): Move contents to
23638 arm_neon_builtins.def.
23639 (enum arm_builtins): Include neon builtin definitions.
23640 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
23641 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
23642
23643 2013-04-04 Marek Polacek <polacek@redhat.com>
23644
23645 PR tree-optimization/48186
23646 * predict.c (maybe_hot_frequency_p): Return false if
23647 HOT_BB_FREQUENCY_FRACTION is 0.
23648 (cgraph_maybe_hot_edge_p): Likewise.
23649
23650 2013-04-04 Richard Biener <rguenther@suse.de>
23651
23652 PR tree-optimization/56826
23653 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
23654 more accurately.
23655
23656 2013-04-04 Richard Biener <rguenther@suse.de>
23657
23658 PR tree-optimization/56213
23659 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
23660 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
23661
23662 2013-04-04 Richard Biener <rguenther@suse.de>
23663
23664 PR tree-optimization/56837
23665 * tree-loop-distribution.c (classify_partition): For non-zero
23666 values require that the value has the same precision as its
23667 mode to be useful as memset value.
23668
23669 2013-04-03 Nick Clifton <nickc@redhat.com>
23670
23671 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
23672 (fmssf4): Use fmsf.s on E3V5 architectures.
23673 (fnmasf4): Use fnmaf.s on E3V5 architectures.
23674 (fnmssf4): Use fnmsf.s on E3V5 architectures.
23675
23676 2013-04-03 Jeff Law <law@redhat.com>
23677
23678 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
23679 (lra-eliminations.o): Likewise.
23680
23681 2013-04-03 Teresa Johnson <tejohnson@google.com>
23682
23683 * gcov-io.c (compute_working_sets): Moved most of body of old
23684 compute_working_sets here from profile.c.
23685 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
23686 (gcov_working_set_t): Moved typedef here from basic-block.h
23687 (compute_working_set): Declare.
23688 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
23689 (get_working_sets): Renamed from compute_working_set,
23690 replace most of body with call to new compute_working_sets.
23691 (get_exec_counts): Replace call to compute_working_sets
23692 to get_working_sets.
23693 * profile.h (get_working_sets): Renamed from compute_working_set.
23694 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
23695 to get_working_sets.
23696 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
23697 * gcov-dump.c (dump_working_sets): New function.
23698
23699 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
23700
23701 * hwint.c (sext_hwi, zext_hwi): New functions.
23702 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
23703 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
23704 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
23705 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
23706 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
23707 (sext_hwi, zext_hwi): New functions.
23708
23709 2013-04-03 Jeff Law <law@redhat.com>
23710
23711 PR tree-optimization/56799
23712 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
23713 back test for widening conversion erroneously dropped in prior change.
23714
23715 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23716
23717 PR target/56809
23718 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
23719 instead of next_real_insn.
23720
23721 2013-04-03 Marek Polacek <polacek@redhat.com>
23722
23723 PR sanitizer/55702
23724 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
23725
23726 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23727
23728 PR target/56809
23729 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
23730 next_real_insn.
23731 (thumb1_output_casesi): Likewise.
23732 (thumb2_output_casesi): Likewise.
23733
23734 2013-04-03 Richard Biener <rguenther@suse.de>
23735
23736 PR tree-optimization/56817
23737 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
23738 Split out ...
23739 (tree_unroll_loops_completely_1): ... new function to manually
23740 walk the loop tree, properly defering outer loops of unrolled
23741 loops to later iterations.
23742
23743 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23744
23745 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
23746 (vectorizable_load): Likewise.
23747 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
23748 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
23749
23750 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23751
23752 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
23753 BIT_FIELD_REF.
23754
23755 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
23756
23757 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
23758
23759 2013-04-03 Bin Cheng <bin.cheng@arm.com>
23760
23761 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
23762
23763 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23764
23765 PR tree-optimization/56790
23766 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
23767 folding.
23768
23769 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23770
23771 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
23772 Handle VEC_MERGE.
23773 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
23774 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
23775 equal arguments.
23776
23777 2013-04-03 Jakub Jelinek <jakub@redhat.com>
23778
23779 PR c/19449
23780 * tree.h (force_folding_builtin_constant_p): New decl.
23781 * builtins.c (force_folding_builtin_constant_p): New variable.
23782 (fold_builtin_constant_p): Fold immediately also if
23783 force_folding_builtin_constant_p.
23784
23785 2013-04-03 Richard Biener <rguenther@suse.de>
23786
23787 PR tree-optimization/56812
23788 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
23789 DRs of the same interleaving chain are independent.
23790
23791 2013-04-02 Jason Merrill <jason@redhat.com>
23792
23793 * gdbinit.in (pbb): Use debug fn.
23794
23795 2013-04-02 Lawrence Crowl <crowl@google.com>
23796
23797 * sese.h (struct ivtype_map_elt_s): Remove unused.
23798 (extern debug_ivtype_map): Remove unused.
23799 (extern eq_ivtype_map_elts): Remove unused.
23800 * sese.c (debug_ivtype_map): Removed unused.
23801 (debug_ivtype_map_1): Removed unused.
23802 (debug_ivtype_elt): Remove unused.
23803 (eq_ivtype_map_elts): Remove unused.
23804
23805
23806 2013-04-02 Kai Tietz <ktietz@redhat.com>
23807
23808 PR target/52790
23809 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
23810 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
23811 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
23812 function.
23813 (legitimize_pe_coff_symbol): Likewise.
23814 (is_imported_p): New helper-function.
23815 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
23816 for Windows x64 targets.
23817 (ix86_expand_prologue): Optimize for pe-coff targets.
23818 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
23819 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
23820 medium/large code-model.
23821 (legitimize_pic_address): Likewise.
23822 (legitimize_tls_address): Likewise.
23823 (ix86_expand_call): Likewise.
23824 (x86_output_mi_thunk): Likewise.
23825 (get_dllimport_decl): Add new beimport argument.
23826 (construct_plt_address): Don't assert for x64 pe-coff targets.
23827 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
23828 targets.
23829 (SYMBOL_FLAG_STUBVAR): New macro.
23830 (SYMBOL_REF_STUBVAR_P): Likewise.
23831 * config/i386/winnt.c (stub_list): New structure.
23832 (stub_head): New local variable.
23833 (i386_pe_record_stub): New function.
23834 (i386_pe_file_end): Emit refptr-stubs.
23835
23836 2013-04-02 Jakub Jelinek <jakub@redhat.com>
23837
23838 PR rtl-optimization/56745
23839 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
23840 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
23841
23842 PR c++/34949
23843 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
23844 and both of them are MEM_REFs, just compare first argument for
23845 equality and attempt to deal even with differing offsets.
23846
23847 PR c++/34949
23848 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
23849 of gimple_clobber_p to be MEM_REF.
23850 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
23851 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
23852 after gimplification.
23853 * asan.c (get_mem_ref_of_assignment): Don't instrument
23854 gimple_clobber_p stmts.
23855 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
23856 gimple_clobber_p stmt if they have MEM_REF lhs and
23857 are dead because of another gimple_clobber_p stmt.
23858 * tree-ssa-live.c (clear_unused_block_pointer): Treat
23859 gimple_clobber_p stmts like debug stmts.
23860 (remove_unused_locals): Remove clobbers with MEM_REF lhs
23861 that refer to unused VAR_DECLs or uninitialized values.
23862 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
23863 gimple_clobber_p stmts if they refer to removed parameters.
23864 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
23865 formatting.
23866
23867 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
23868
23869 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
23870 using SWI48 mode attribute.
23871
23872 2013-04-02 Wei Mi <wmi@google.com>
23873
23874 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
23875 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
23876 *<rotate_insn><mode>3_mask in i386.md.
23877
23878 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
23879
23880 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
23881
23882 2013-04-02 Richard Biener <rguenther@suse.de>
23883
23884 PR tree-optimization/56778
23885 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
23886 Runtime alias tests are not supported for gather loads.
23887 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
23888 stmts referenced from SSA operands before updating SSA form.
23889
23890 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
23891 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23892
23893 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
23894 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
23895 * config/arm/cortex-a53.md: New file.
23896 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
23897 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
23898 * config/arm/arm.c (arm_issue_rate): Likewise.
23899 * config/arm/arm-tune.md: Regenerate
23900 * config/arm/arm-tables.opt: Regenerate.
23901 * config/arm/arm-cores.def: Add cortex-a53.
23902
23903 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
23904
23905 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
23906 non-static link.
23907
23908 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23909
23910 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
23911 scalar load/store operations using B/H registers.
23912 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
23913
23914 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23915
23916 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
23917 scalar move.
23918 * config/aarch64/aarch64.c
23919 (aarch64_simd_scalar_immediate_valid_for_move): New.
23920 * config/aarch64/aarch64-protos.h
23921 (aarch64_simd_scalar_immediate_valid_for_move): New.
23922 * config/aarch64/constraints.md (Dh, Dq): New.
23923 * config/aarch64/iterators.md (hq): New.
23924
23925 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
23926
23927 * reorg.c (get_branch_condition): Deal with conditional returns.
23928 (fill_simple_delay_slots): Remove dead code dealing with jumps.
23929
23930 2013-04-01 Wei Mi <wmi@google.com>
23931
23932 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
23933 Truncate operand 2 using %b asm operand modifier.
23934 (*<shift_insn><mode>3_mask): Ditto.
23935 (*<rotate_insn><mode>3_mask): Ditto.
23936
23937 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
23938
23939 PR middle-end/56798
23940 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
23941
23942 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
23943
23944 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
23945 of next_real_insn.
23946 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
23947
23948 2013-03-30 Lawrence Crowl <crowl@google.com>
23949
23950 * dse.c (clear_alias_sets): Remove never set.
23951 (disqualified_clear_alias_sets): Remove never set.
23952 (clear_alias_mode_pool): Remove never set.
23953 (dse_step0): Remove condition that is never true.
23954 (canon_address): Remove condition that is never true.
23955 (dse_step7): Remove condition that is never true.
23956 (rest_of_handle_dse): Remove condition that is never true.
23957 (rest_of_handle_dse::did_global): Remove never read from above.
23958 (dse_step2_spill): Remove never called from above.
23959 (dse_step5_spill): Remove never called from above.
23960
23961 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
23962
23963 * doc/md.texi (Standard Names) <casesi>: Update documentation for
23964 JUMP_TABLE_DATA changes.
23965 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
23966 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
23967 (Insns) <jump_table_data>: New entry.
23968 * doc/tm.texi: Regenerate.
23969
23970 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
23971
23972 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
23973 for table jump at the end of a basic block using tablejump_p.
23974 * targhooks.c (default_invalid_within_doloop): Likewise.
23975 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
23976 target hook implementation that is identical to the default hook.
23977 (rs6000_invalid_within_doloop): Remove.
23978
23979 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
23980 unused variable from tablejump_p call.
23981
23982 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
23983 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
23984 (INSN_DELETED_P): Likewise.
23985 (emit_jump_table_data): New prototype.
23986 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
23987 after 4th as unused.
23988 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
23989 * sched-vis.c (print_insn): Likewise.
23990 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
23991 insn for compatibility with back ends that use next_active_insn to
23992 identify jump table data.
23993 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
23994 (remove_insn): Likewise.
23995 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
23996 to be emitted.
23997 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
23998 (emit_jump_table_data): New function.
23999
24000 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
24001 basic block, a JUMP_TABLE_DATA never is.
24002 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
24003 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
24004 off from code handling real insns.
24005 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
24006 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
24007 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
24008 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
24009 is not a NONDEBUG_INSN_P.
24010 * ira-costs.c (scan_one_insn): Likewise.
24011 * jump.c (mark_all_labels): Likewise.
24012 (mark_jump_label_1): Likewise.
24013 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24014 * lra.c (get_insn_freq): Expect all insns reaching here to be in
24015 a basic block.
24016 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
24017 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
24018 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
24019 JUMP_TABLE_DATA_P insns.
24020 (calculate_elim_costs_all_insns): Likewise.
24021 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
24022 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
24023 (delete_output_reload): Code style fixups.
24024 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
24025 insn flags on this non-insn.
24026 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
24027 as scheduling barriers, for pre-change compatibility.
24028 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
24029 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
24030
24031 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
24032 redundant JUMP_TABLE_DATA_P test.
24033 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
24034 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24035 (frv_for_each_packet): Likewise.
24036 * config/i386/i386.c (min_insn_size): Likewise.
24037 (ix86_avoid_jump_mispredicts): Likewise.
24038 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24039 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
24040 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24041 (mips16_insn_length): Robustify.
24042 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
24043 (mips16_split_long_branches): Likewise.
24044 * config/pa/pa.c (pa_combine_instructions): Likewise.
24045 * config/rs6000/rs6000.c (get_next_active_insn): Treat
24046 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
24047 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
24048 as contributing to pool range lengths.
24049 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
24050 Remove redundant JUMP_TABLE_DATA_P test.
24051 (sh_loop_align): Likewise.
24052 (split_branches): Likewise.
24053 (sh_insn_length_adjustment): Likewise.
24054 * config/spu/spu.c (get_branch_target): Likewise.
24055
24056 2013-03-29 Jan Hubicka <jh@suse.cz>
24057
24058 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
24059 gcov streaming; stream hot bb threshold to ltrans.
24060 * predict.c (get_hot_bb_threshold): Break out from ....
24061 (maybe_hot_count_p): ... here.
24062 (set_hot_bb_threshold): New function.
24063 * lto-section-in.c (lto_section_name): Add profile.
24064 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
24065 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
24066 and data-streamer.h
24067 (histogram_entry): New structure.
24068 (histogram, histogram_pool): New global vars.
24069 (histogram_hash): New structure.
24070 (histogram_hash::hash): New method.
24071 (histogram_hash::equal): Likewise.
24072 (account_time_size): New function.
24073 (cmp_counts): New function.
24074 (dump_histogram): New function.
24075 (ipa_profile_generate_summary): New function.
24076 (ipa_profile_write_summary): New function.
24077 (ipa_profile_read_summary): New function.
24078 (ipa_profile): Decide on threshold.
24079 (pass_ipa_profile): Add ipa_profile_write_summary and
24080 ipa_profile_read_summary.
24081 * Makefile.in (ipa.o): Update dependencies.
24082 * lto-streamer.h (LTO_section_ipa_profile): New section.
24083
24084 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
24085
24086 * tree.h (VAR_P): New.
24087
24088 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
24089
24090 PR lto/56777
24091 * doc/invoke.texi ([-fwhole-program]): Fix typo.
24092
24093 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
24094
24095 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
24096 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24097 (control_flow_insn_p): Likewise.
24098 * cfgrtl.c (duplicate_insn_chain): Likewise.
24099 * final.c (get_attr_length_1): Likewise.
24100 (shorten_branches): Likewise.
24101 (final_scan_insn): Likewise.
24102 * function.c (instantiate_virtual_regs): Likewise.
24103 * gcse.c (insert_insn_end_basic_block): Likewise.
24104 * ira-costs.c (scan_one_insn): Likewise.
24105 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24106 * lra.c (check_rtl): Likewise.
24107 * reload1.c (elimination_costs_in_insn): Likewise.
24108 * reorg.c (follow_jumps): Likewise.
24109
24110 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
24111 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24112 (thumb_far_jump_used_p): Likewise.
24113 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
24114 (workaround_speculation): Likewise.
24115 (add_sched_insns_for_speculation): Likewise.
24116 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
24117 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24118 (frv_for_each_packet): Likewise.
24119 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
24120 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
24121 (final_emit_insn_group_barriers): Likewise.
24122 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24123 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24124 (mips16_insn_length): Likewise.
24125 * config/pa/pa.c (pa_reorg): Likewise.
24126 (pa_combine_instructions): Likewise.
24127 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
24128 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
24129 (sh_reorg): Likewise.
24130 (split_branches): Likewise.
24131 * config/spu/spu.c (get_branch_target): Likewise.
24132
24133 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
24134 JUMP_TABLE_DATA_P.
24135
24136 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
24137
24138 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
24139 Fix declaration name.
24140
24141 2013-03-28 Lawrence Crowl <crowl@google.com>
24142
24143 * graphds.h (struct graph.indicies): Remove unused.
24144 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
24145 (SCOP_ORIGINAL_PDDRS): Remove unused.
24146 * sese.h (extern insert_loop_close_phis): Removed unused.
24147 (extern insert_guard_phis): Removed unused.
24148 (extern ivtype_map_elt_info): Removed unused.
24149 (new_ivtype_map_elt): Removed unused.
24150 * sese.c (ivtype_map_elt_info): Removed unused.
24151
24152 2013-03-28 Lawrence Crowl <crowl@google.com>
24153
24154 * Makefile.in: Add several missing include dependences.
24155 (DUMPFILE_H): New.
24156 (test-dump.o): New. This object is not added to any executable,
24157 but is present for ad-hoc testing.
24158 * bitmap.c
24159 (debug (const bitmap_head_def &)): New.
24160 (debug (const bitmap_head_def *)): New.
24161 * bitmap.h
24162 (extern debug (const bitmap_head_def &)): New.
24163 (extern debug (const bitmap_head_def *)): New.
24164 * cfg.c
24165 (debug (edge_def &)): New.
24166 (debug (edge_def *)): New.
24167 * cfghooks.c
24168 (debug (basic_block_def &)): New.
24169 (debug (basic_block_def *)): New.
24170 * dumpfile.h
24171 (dump_node (const_tree, int, FILE *)): Correct source file.
24172 * dwarf2out.c
24173 (debug (die_struct &)): New.
24174 (debug (die_struct *)): New.
24175 * dwarf2out.h
24176 (extern debug (die_struct &)): New.
24177 (extern debug (die_struct *)): New.
24178 * gimple-pretty-print.c
24179 (debug (gimple_statement_d &)): New.
24180 (debug (gimple_statement_d *)): New.
24181 * gimple-pretty-print.h
24182 (extern debug (gimple_statement_d &)): New.
24183 (extern debug (gimple_statement_d *)): New.
24184 * ira-build.c
24185 (debug (ira_allocno_copy &)): New.
24186 (debug (ira_allocno_copy *)): New.
24187 (debug (ira_allocno &)): New.
24188 (debug (ira_allocno *)): New.
24189 * ira-int.h
24190 (extern debug (ira_allocno_copy &)): New.
24191 (extern debug (ira_allocno_copy *)): New.
24192 (extern debug (ira_allocno &)): New.
24193 (extern debug (ira_allocno *)): New.
24194 * ira-lives.c
24195 (debug (live_range &)): New.
24196 (debug (live_range *)): New.
24197 * lra-int.h
24198 (debug (lra_live_range &)): New.
24199 (debug (lra_live_range *)): New.
24200 * lra-lives.c
24201 (debug (lra_live_range &)): New.
24202 (debug (lra_live_range *)): New.
24203 * omega.c
24204 (debug (omega_pb_d &)): New.
24205 (debug (omega_pb_d *)): New.
24206 * omega.h
24207 (extern debug (omega_pb_d &)): New.
24208 (extern debug (omega_pb_d *)): New.
24209 * print-rtl.c
24210 (debug (const rtx_def &)): New.
24211 (debug (const rtx_def *)): New.
24212 * print-tree.c
24213 (debug_tree (tree): Move within file.
24214 (debug_raw (const tree_node &)): New.
24215 (debug_raw (const tree_node *)): New.
24216 (dump_tree_via_hooks (const tree_node *, int)): New.
24217 (debug (const tree_node &)): New.
24218 (debug (const tree_node *)): New.
24219 (debug_verbose (const tree_node &)): New.
24220 (debug_verbose (const tree_node *)): New.
24221 (debug_head (const tree_node &)): New.
24222 (debug_head (const tree_node *)): New.
24223 (debug_body (const tree_node &)): New.
24224 (debug_body (const tree_node *)): New.
24225 (debug_vec_tree (tree): Move and reimplement in terms of dump.
24226 (debug (vec<tree, va_gc> &)): New.
24227 (debug (vec<tree, va_gc> *)): New.
24228 * rtl.h
24229 (extern debug (const rtx_def &)): New.
24230 (extern debug (const rtx_def *)): New.
24231 * sbitmap.c
24232 (debug_raw (simple_bitmap_def &)): New.
24233 (debug_raw (simple_bitmap_def *)): New.
24234 (debug (simple_bitmap_def &)): New.
24235 (debug (simple_bitmap_def *)): New.
24236 * sbitmap.h
24237 (extern debug (simple_bitmap_def &)): New.
24238 (extern debug (simple_bitmap_def *)): New.
24239 (extern debug_raw (simple_bitmap_def &)): New.
24240 (extern debug_raw (simple_bitmap_def *)): New.
24241 * sel-sched-dump.c
24242 (debug (vinsn_def &)): New.
24243 (debug (vinsn_def *)): New.
24244 (debug_verbose (vinsn_def &)): New.
24245 (debug_verbose (vinsn_def *)): New.
24246 (debug (expr_def &)): New.
24247 (debug (expr_def *)): New.
24248 (debug_verbose (expr_def &)): New.
24249 (debug_verbose (expr_def *)): New.
24250 (debug (vec<rtx> &)): New.
24251 (debug (vec<rtx> *)): New.
24252 * sel-sched-dump.h
24253 (extern debug (vinsn_def &)): New.
24254 (extern debug (vinsn_def *)): New.
24255 (extern debug_verbose (vinsn_def &)): New.
24256 (extern debug_verbose (vinsn_def *)): New.
24257 (extern debug (expr_def &)): New.
24258 (extern debug (expr_def *)): New.
24259 (extern debug_verbose (expr_def &)): New.
24260 (extern debug_verbose (expr_def *)): New.
24261 (extern debug (vec<rtx> &)): New.
24262 (extern debug (vec<rtx> *)): New.
24263 * sel-sched-ir.h
24264 (_list_iter_cond_expr): Make inline instead of static.
24265 * sreal.c
24266 (debug (sreal &)): New.
24267 (debug (sreal *)): New.
24268 * sreal.h
24269 (extern debug (sreal &)): New.
24270 (extern debug (sreal *)): New.
24271 * tree.h
24272 (extern debug_raw (const tree_node &)): New.
24273 (extern debug_raw (const tree_node *)): New.
24274 (extern debug (const tree_node &)): New.
24275 (extern debug (const tree_node *)): New.
24276 (extern debug_verbose (const tree_node &)): New.
24277 (extern debug_verbose (const tree_node *)): New.
24278 (extern debug_head (const tree_node &)): New.
24279 (extern debug_head (const tree_node *)): New.
24280 (extern debug_body (const tree_node &)): New.
24281 (extern debug_body (const tree_node *)): New.
24282 (extern debug (vec<tree, va_gc> &)): New.
24283 (extern debug (vec<tree, va_gc> *)): New.
24284 * tree-cfg.c
24285 (debug (struct loop &)): New.
24286 (debug (struct loop *)): New.
24287 (debug_verbose (struct loop &)): New.
24288 (debug_verbose (struct loop *)): New.
24289 * tree-dump.c: Add header dependence.
24290 * tree-flow.h
24291 (extern debug (struct loop &)): New.
24292 (extern debug (struct loop *)): New.
24293 (extern debug_verbose (struct loop &)): New.
24294 (extern debug_verbose (struct loop *)): New.
24295 * tree-data-ref.c
24296 (debug (data_reference &)): New.
24297 (debug (data_reference *)): New.
24298 (debug (vec<data_reference_p> &)): New.
24299 (debug (vec<data_reference_p> *)): New.
24300 (debug (vec<ddr_p> &)): New.
24301 (debug (vec<ddr_p> *)): New.
24302 * tree-data-ref.h
24303 (extern debug (data_reference &)): New.
24304 (extern debug (data_reference *)): New.
24305 (extern debug (vec<data_reference_p> &)): New.
24306 (extern debug (vec<data_reference_p> *)): New.
24307 (extern debug (vec<ddr_p> &)): New.
24308 (extern debug (vec<ddr_p> *)): New.
24309 * tree-ssa-alias.c
24310 (debug (pt_solution &)): New.
24311 (debug (pt_solution *)): New.
24312 * tree-ssa-alias.h
24313 (extern debug (pt_solution &)): New.
24314 (extern debug (pt_solution *)): New.
24315 * tree-ssa-alias.c
24316 (debug (_var_map &)): New.
24317 (debug (_var_map *)): New.
24318 (debug (tree_live_info_d &)): New.
24319 (debug (tree_live_info_d *)): New.
24320 * tree-ssa-alias.h
24321 (extern debug (_var_map &)): New.
24322 (extern debug (_var_map *)): New.
24323 (extern debug (tree_live_info_d &)): New.
24324 (extern debug (tree_live_info_d *)): New.
24325
24326 2013-03-28 Jan Hubicka <jh@suse.cz>
24327
24328 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
24329
24330 2013-03-28 Ian Bolton <ian.bolton@arm.com>
24331
24332 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
24333 record only when desired or required.
24334
24335 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
24336
24337 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
24338 *vec_extractv2di_1_rex64. Use x64 isa attribute.
24339
24340 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24341
24342 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
24343 (*andsi3_compare0_uxtw): New pattern.
24344 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
24345 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
24346
24347 2013-03-28 Jan Hubicka <jh@suse.cz>
24348
24349 * data-streamer-in.c (streamer_read_gcov_count): New function.
24350 * gimple-streamer-out.c: Include value-prof.h.
24351 (output_gimple_stmt): Output histogram.
24352 (output_bb): Use streamer_write_gcov_count.
24353 * value-prof.c: Include data-streamer.h
24354 (dump_histogram_value): Add HIST_TYPE_MAX.
24355 (stream_out_histogram_value): New function.
24356 (stream_in_histogram_value): New function.
24357 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
24358 (stream_out_histogram_value, stream_in_histogram_value): Declare.
24359 * data-streamer-out.c (streamer_write_gcov_count): New function.
24360 (streamer_write_gcov_count_stream): New function.
24361 * lto-cgraph.c (lto_output_edge): Update counter streaming.
24362 (lto_output_node): Likewise.
24363 (input_node, input_edge): Likewise.
24364 * lto-streamer-out.c (output_cfg): Update streaming.
24365 * lto-streamer-in.c (input_cfg): Likewise.
24366 * data-streamer.h (streamer_write_gcov_count,
24367 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
24368 * gimple-streamer-in.c: Include value-prof.h
24369 (input_gimple_stmt): Input histograms.
24370 (input_bb): Update profile streaming.
24371
24372 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24373
24374 * genmodes.c (emit_max_int): New function.
24375 (emit_insn_modes_h): Added call to emit_max_function.
24376 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
24377 Added doc.
24378 * machmode.def: Fixed comment.
24379
24380 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24381
24382 * combine.c (try_combine): Removed useless assert.
24383 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
24384
24385 2013-03-28 Marek Polacek <polacek@redhat.com>
24386 Richard Biener <rguenther@suse.de>
24387
24388 PR tree-optimization/56695
24389 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
24390 build signed result of a vector comparison.
24391 * tree-cfg.c (verify_gimple_comparison): Check that a result
24392 of a vector comparison has signed type.
24393
24394 2013-03-28 Richard Biener <rguenther@suse.de>
24395
24396 PR tree-optimization/37021
24397 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
24398 do not restrict gaps between groups.
24399 * tree-vect-stmts.c (vectorizable_load): Properly account for
24400 a gap between groups.
24401
24402 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
24403
24404 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
24405 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
24406 is not enabled.
24407
24408 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
24409
24410 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
24411 * doc/extend.texi (Named Address Spaces): Ditto.
24412 (Variable Attributes): Ditto.
24413
24414 2013-03-27 Kai Tietz <ktietz@redhat.com>
24415
24416 * config.build: Add support for cygwin x64 target.
24417 * config.gcc: Likewise.
24418 * config.host: Likewise.
24419 * configure.ac: Likewise
24420 * configure: Regenerated.
24421
24422 2013-03-27 Kai Tietz <ktietz@redhat.com>
24423
24424 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
24425 * config/i386/t-cygwin-w64: New file.
24426 * config/i386/cygwin-w64.h: New file.
24427 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
24428 and add support for x64-cygwin target.
24429 (CPP_SPEC): Likewise.
24430 (CXX_WRAP_SPEC_LIST): Undefine before define.
24431 (LIBGCJ_SONAME): Use 15 as version.
24432
24433 2013-03-27 Richard Biener <rguenther@suse.de>
24434
24435 PR tree-optimization/56716
24436 * tree-ssa-structalias.c (perform_var_substitution): Adjust
24437 dumping for ref nodes.
24438
24439 2013-03-27 Martin Jambor <mjambor@suse.cz>
24440
24441 PR tree-optimization/55334
24442 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
24443 restricted pointers to arrays.
24444
24445 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
24446
24447 * Makefile.in (.SUFFIXES): Add .cc.
24448 (.c.o): Apply same recipe for implicit rule .cc.o.
24449
24450 2013-03-27 Richard Biener <rguenther@suse.de>
24451
24452 PR tree-optimization/37021
24453 * tree-vect-data-refs.c (vect_check_strided_load): Allow
24454 REALPART/IMAGPART_EXPRs around the supported refs.
24455 * tree-ssa-structalias.c (find_func_aliases): Assume that
24456 floating-point values are not used to transfer pointers.
24457
24458 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
24459
24460 * target.def (TARGET_HAS_IFUNC_P): New target hook.
24461 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
24462 * doc/tm.texi: Regenerate.
24463 * targhooks.h (default_has_ifunc_p): New.
24464 * targhooks.c (default_has_ifunc_p): Ditto.
24465 * config/linux-protos.h: New file.
24466 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
24467 hook for linux which disables support of indirect functions in android.
24468 * config/linux-android.c: New file.
24469 * config/t-linux-android.c: Ditto.
24470 * config.gcc: Added new object file linux-android.o.
24471 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
24472 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
24473 * varasm.c (do_assemble_alias): Likewise.
24474 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
24475 doesn't support indirect functions.
24476 * configure: Regenerate.
24477
24478 2013-03-27 Bin Cheng <bin.cheng@arm.com>
24479
24480 PR target/56102
24481 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
24482 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
24483 mult-word mode.
24484
24485 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24486
24487 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
24488
24489 2013-03-27 Terry Guo <terry.guo@arm.com>
24490
24491 * config/arm/arm-cores.def: Added core cortex-r7.
24492 * config/arm/arm-tune.md: Regenerated.
24493 * config/arm/arm-tables.opt: Regenerated.
24494 * doc/invoke.texi: Added entry for core cortex-r7.
24495
24496 2013-03-27 Walter Lee <walt@tilera.com>
24497
24498 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
24499 double-decrement of next_scratch_regno.
24500
24501 2013-03-27 Walter Lee <walt@tilera.com>
24502
24503 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
24504 input operands.
24505 (insn_v1mulus): Ditto.
24506 (insn_v2muls): Ditto.
24507
24508 2013-03-27 Walter Lee <walt@tilera.com>
24509
24510 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
24511 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
24512
24513 2013-03-27 Walter Lee <walt@tilera.com>
24514
24515 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
24516 (*sibcall_value): Ditto.
24517
24518 2013-03-27 Walter Lee <walt@tilera.com>
24519
24520 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
24521 (insn_mnz_v8qi): ... this ...
24522 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
24523 vector equivalent.
24524 (insn_v<n>mnz): Replaced by ...
24525 (insn_v1mnz): ... this ...
24526 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
24527 equivalent.
24528 (insn_mz_<mode>): Replaced by ...
24529 (insn_mz_v8qi): ... this ...
24530 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
24531 vector equivalent.
24532 (insn_v<n>mz): Replaced by ...
24533 (insn_v1mz): ... this ...
24534 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
24535 equivalent.
24536
24537 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
24538
24539 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
24540
24541 2013-03-26 Roland McGrath <mcgrathr@google.com>
24542
24543 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
24544 than fprintf with a non-constant, non-format string.
24545
24546 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24547
24548 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
24549 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
24550 operand 0 predicate.
24551 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
24552 attribute. Use general_x64nomem_operand as operand 1 predicate.
24553 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
24554 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24555 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
24556 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24557 (mov<mode>_insv_1): Remove expander. Merge insn with
24558 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
24559 Use general_x64nomem_operand as operand 1 predicate.
24560 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
24561 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
24562 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24563 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
24564 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
24565 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
24566 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24567 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
24568 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
24569 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
24570 (general_x64nomem_operand): Ditto.
24571
24572 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
24573
24574 * config/rtems.opt: Add -pthread option.
24575
24576 2013-03-26 Richard Biener <rguenther@suse.de>
24577
24578 * alias.c (find_base_term): Avoid redundant and not used recursion.
24579 (base_alias_check): Get the initial base term from the caller.
24580 (true_dependence_1): Compute and pass base terms to base_alias_check.
24581 (write_dependence_p): Likewise.
24582 (may_alias_p): Likewise.
24583
24584 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
24585
24586 * config/aarch64/aarch64.c (aarch64_classify_address): Support
24587 PC-relative load in SI modes and above only.
24588
24589 2013-03-26 Xinyu Qi <xyqi@marvell.com>
24590
24591 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
24592 * config/arm/iwmmxt.md (WCGR0): Update.
24593 (WCGR1, WCGR2, WCGR3): Likewise.
24594
24595 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24596
24597 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
24598 Use x64 and nox64 isa attributes.
24599
24600 2013-03-26 Richard Biener <rguenther@suse.de>
24601
24602 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
24603 alignment computations and rely on get_object_alignment_1
24604 for the !TYPE_P case.
24605 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
24606
24607 2013-03-26 Walter Lee <walt@tilera.com>
24608
24609 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
24610 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
24611
24612 2013-03-25 Jeff Law <law@redhat.com>
24613
24614 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
24615 check for INTEGRAL_TYPE_P that was missing due to checking in
24616 wrong version of prior patch.
24617
24618 2013-03-25 Walter Lee <walt@tilera.com>
24619
24620 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
24621 TILEGX_INSN_SHUFFLEBYTES1.
24622 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
24623 shufflebytes1.
24624 (tilegx_builtins): Ditto.
24625 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
24626
24627 2013-03-25 Walter Lee <walt@tilera.com>
24628
24629 * config/tilegx/tilegx.md (floatsisf2): New pattern.
24630 (floatunssisf2): New pattern.
24631 (floatsidf2): New pattern.
24632 (floatunssidf2): New pattern.
24633
24634 2013-03-25 Walter Lee <walt@tilera.com>
24635
24636 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
24637 tests for constraint J, K, N, P.
24638
24639 2013-03-25 Walter Lee <walt@tilera.com>
24640
24641 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
24642 Use indirect/pcrel encoding.
24643 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
24644 Ditto.
24645
24646 2013-03-25 Steve Ellcey <sellcey@mips.com>
24647
24648 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
24649 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
24650 * config/mips/mips.c (mips_option_override): Set IMADD default.
24651 * config/mips/mips.h (PTF_AVOID_IMADD): New.
24652 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
24653 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
24654 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
24655 * doc/invoke.texi (-mimadd/-mno-imadd): New.
24656
24657 2013-03-25 Jeff Law <law@redhat.com>
24658
24659 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
24660 slightly to avoid creating and folding useless trees. Simplify
24661 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
24662
24663 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
24664
24665 * config/i386/i386.md (*zero_extendsidi2): Merge with
24666 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
24667 * config/i386/predicates.md (x86_64_zext_operand): Rename from
24668 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
24669 targets. Clarify comment.
24670
24671 2013-03-25 Martin Jambor <mjambor@suse.cz>
24672
24673 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
24674 pass-through jump functions differently.
24675 (ipa_read_jump_function): Likewise. Also use setter functions to set
24676 up jump functions.
24677
24678 2013-03-25 Martin Jambor <mjambor@suse.cz>
24679
24680 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
24681 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
24682 process it.
24683 (ipa_get_indirect_edge_target): New function.
24684 (devirtualization_time_bonus): New parameter known_aggs, pass it to
24685 ipa_get_indirect_edge_target. Update all callers.
24686 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
24687 ipa_get_indirect_edge_target_1 instead of calling
24688 ipa_get_indirect_edge_target.
24689 (create_specialized_node): Pass aggvlas to
24690 ipcp_discover_new_direct_edges.
24691
24692 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24693
24694 * config/arm/arm.md (f_sels, f_seld): New types.
24695 (*cmov<mode>): New pattern.
24696 * config/arm/predicates.md (arm_vsel_comparison_operator): New
24697 predicate.
24698
24699 2013-03-25 Kai Tietz <ktietz@redhat.com>
24700
24701 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
24702 POSIX-printf for mingw-hosted builds.
24703
24704 2013-03-25 Richard Biener <rguenther@suse.de>
24705
24706 PR middle-end/56694
24707 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
24708 must-not-throw stmt location.
24709
24710 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24711
24712 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
24713 Emit load-acquire versions when acq is true.
24714 (arm_emit_store_exclusive): Add rel parameter.
24715 Emit store-release versions when rel is true.
24716 (arm_split_compare_and_swap): Use acquire-release instructions
24717 instead.
24718 of barriers when appropriate.
24719 (arm_split_atomic_op): Likewise.
24720 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
24721 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
24722 (VUNSPEC_SLX): Likewise.
24723 (VUNSPEC_LDA): Likewise.
24724 (VUNSPEC_STL): Likewise.
24725 * config/arm/sync.md (atomic_load<mode>): New pattern.
24726 (atomic_store<mode>): Likewise.
24727 (arm_load_acquire_exclusive<mode>): Likewise.
24728 (arm_load_acquire_exclusivesi): Likewise.
24729 (arm_load_acquire_exclusivedi): Likewise.
24730 (arm_store_release_exclusive<mode>): Likewise.
24731
24732 2013-03-25 Catherine Moore <clm@codesourcery.com>
24733
24734 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
24735 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
24736 * config/mip/predicates.md (lwsp_swsp_operand,
24737 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
24738 sb16_operand, db4_operand, db7_operand, ib3_operand,
24739 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
24740 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
24741 andi16_operand): New predicates.
24742 * config/mips/mips.md (compression): New attribute.
24743 (enabled): New attribute.
24744 (length): Consider compression in computing length.
24745 (shift_compression): New code attribute.
24746 (*add<mode>3): New operands. Record compression.
24747 (sub<mode>3): Likewise.
24748 (one_cmpl<mode>2): Likewise.
24749 (*and<mode>3): Likewise.
24750 (*ior<mode>3): Likewise.
24751 (unnamed pattern for xor): Likewise.
24752 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
24753 (*<optab><mode>3): Likewise.
24754 (*mov<mode>_internal: Likewise.
24755 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
24756 (mips_unsigned_immediate_p): New.
24757 (umips_lwsp_swsp_address_p): New.
24758 (m16_based_address_p): New.
24759 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
24760 (mips_unsigned_immediate_p): New prototype.
24761 (lwsp_swsp_address_p): New prototype.
24762 (m16_based_address_p): New prototype.
24763 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
24764 (mips_signed_immediate_p): New function.
24765 (m16_based_address_p): New function.
24766 (lwsp_swsp_address_p): New function.
24767 (mips_print_operand_punctuation): Recognize short delay slot insns
24768 for microMIPS.add<mode>3"
24769
24770 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24771
24772 PR target/56720
24773 * config/arm/iterators.md (v_cmp_result): New mode attribute.
24774 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
24775
24776 2013-03-25 Richard Biener <rguenther@suse.de>
24777
24778 PR tree-optimization/56689
24779 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
24780 any edge.
24781
24782 2013-03-25 Richard Biener <rguenther@suse.de>
24783
24784 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
24785 of bitmap.
24786 (memory_references): Likewise.
24787 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
24788 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
24789 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
24790 (gather_mem_refs_in_loops): Fold into ...
24791 (analyze_memory_references): ... this. Move initialization
24792 to tree_ssa_lim_initialize.
24793 (fill_always_executed_in): Rename to ...
24794 (fill_always_executed_in_1): ... this.
24795 (fill_always_executed_in): Move contains_call computation to
24796 this new function from ...
24797 (tree_ssa_lim_initialize): ... here.
24798 (tree_ssa_lim): Call fill_always_executed_in.
24799
24800 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24801
24802 * postreload.c (reload_combine): Fix code detecting returns.
24803
24804 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24805
24806 * function.c (emit_use_return_register_into_block): On cc0 targets,
24807 do not emit the sequence between cc0 setter and user.
24808
24809 2013-03-25 Kai Tietz <ktietz@redhat.com>
24810
24811 * config/i386/predicates.md (local_symbolic_operand): Interpret
24812 dll-imported symbols as none-local.
24813
24814 2013-03-25 Richard Biener <rguenther@suse.de>
24815
24816 * tree-ssa-loop-im.c (struct depend): Remove.
24817 (struct lim_aux_data): Make depends a vec of gimples.
24818 (free_lim_aux_data): Adjust.
24819 (add_dependency): Likewise.
24820 (set_level): Likewise.
24821
24822 2013-03-25 Richard Biener <rguenther@suse.de>
24823
24824 PR middle-end/56434
24825 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
24826 the pointer returned by calls with ECF_MALLOC set.
24827
24828 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24829
24830 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
24831
24832 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24833
24834 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
24835 using MMXMODE mode iterator.
24836 (*move<mode>_internal): Merge with *movv2sf_internal and
24837 *movv2sf_internal_rex64 using MMXMODE mode iterator.
24838
24839 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24840
24841 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
24842 (record_last_mem_set_info): Likewise.
24843
24844 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
24845 of XNEWVEC followed by memset.
24846 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
24847
24848 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24849
24850 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
24851 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
24852 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
24853 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
24854 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
24855 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
24856 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
24857 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
24858 BARRIER_P instead of GET_CODE.
24859
24860 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24861
24862 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
24863 inaccuracy in the probing code.
24864
24865 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
24866 (ctrapdi4): Likewise.
24867
24868 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24869
24870 * calls.c (expand_call): Add missing guard to code handling return
24871 of non-BLKmode structures in MSB.
24872 * function.c (expand_function_end): Likewise.
24873
24874 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24875
24876 * combine.c (try_combine): Adjust comment. Do not add the set of
24877 insn #0 if the destination indirectly is set or dies in insn #2.
24878 Tidy up code to distribute a new note.
24879
24880 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24881
24882 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
24883 also for alternatives 16 and 17.
24884
24885 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24886
24887 * config/i386/sse.md (*mov<mode>_internal): Merge with
24888 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
24889 Emit insn template depending on type attribute. Use
24890 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24891 movd instead of movq mnemonic for interunit moves. Rewrite mode
24892 attribute calculation. Remove unit attribute calculation.
24893 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
24894 Set prefix_data16 attribute for DImode ssemov types.
24895 Use Ym instead of y for SSE-MMX conversion alternatives.
24896 Reorder operand constraints.
24897
24898 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
24899
24900 * df.h (df_insn_delete): Adjust prototype.
24901 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
24902 and let it decide whether mark the basic block dirty.
24903 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
24904 * df-scan.c (df_insn_info_delete): New helper function, split
24905 off from df_insn_delete.
24906 (df_scan_free_bb_info): Use it.
24907 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
24908 Likewise.
24909 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
24910 that the insn is actually an insn and it has a non-NULL basic block.
24911 Do not mark basic block dirty if only deleting a DEBUG_INSN.
24912
24913 2013-03-22 Richard Biener <rguenther@suse.de>
24914
24915 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
24916 dep_ref members.
24917 (mem_ref_alloc): Do not allocate them.
24918 (refs_independent_p): Do not query or maintain a cache.
24919
24920 2013-03-22 Richard Biener <rguenther@suse.de>
24921
24922 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
24923 (gather_mem_refs_in_loops): Do not compute it.
24924 (analyze_memory_references): Do not allocate it.
24925 (tree_ssa_lim_finalize): Do not free it.
24926 (for_all_locs_in_loop): Do not query all_refs_in_loop.
24927
24928 2013-03-22 Richard Biener <rguenther@suse.de>
24929
24930 * is-a.h (as_a): Use gcc_checking_assert.
24931
24932 2013-03-22 Ian Bolton <ian.bolton@arm.com>
24933
24934 * config/aarch64/aarch64.c (aarch64_print_operand): New
24935 format specifier for printing a constant in hex.
24936 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
24937 format specifier for printing second operand.
24938
24939 2013-03-22 Richard Biener <rguenther@suse.de>
24940
24941 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
24942 bitmaps.
24943 (gather_mem_refs_in_loops): Perform store accumulation here.
24944 (create_vop_ref_mapping_loop): Remove.
24945 (create_vop_ref_mapping): Likewise.
24946 (analyze_memory_references): Initialize refs_stored_in_loop.
24947 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
24948 (record_indep_loop): Remove.
24949 (record_dep_loop): New function.
24950 (ref_indep_loop_p_1): Adjust to only walk over references
24951 in the loop, not its subloops.
24952 (ref_indep_loop_p): Rename to ...
24953 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
24954 maintaining a more fine-grained cache.
24955 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
24956 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
24957
24958 2013-03-22 Richard Biener <rguenther@suse.de>
24959
24960 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
24961 (struct mem_ref): Make accesses_in_loop a vec of a vec of
24962 aggregate mem_ref_loc.
24963 (free_mem_ref_locs): Inline into ...
24964 (memref_free): ... this and adjust.
24965 (mem_ref_alloc): Adjust.
24966 (mem_ref_locs_alloc): Remove.
24967 (record_mem_ref_loc): Adjust.
24968 (get_all_locs_in_loop): Rewrite into ...
24969 (for_all_locs_in_loop): ... this iterator.
24970 (rewrite_mem_ref_loc): New functor.
24971 (rewrite_mem_refs): Use for_all_locs_in_loop.
24972 (sm_set_flag_if_changed): New functor.
24973 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
24974 (ref_always_accessed): New functor.
24975 (ref_always_accessed_p): Use for_all_locs_in_loop.
24976
24977 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24978
24979 * tree-pass.h (PROP_gimple_lvec): New.
24980 * passes.c (dump_properties): Handle PROP_gimple_lvec.
24981 (init_optimization_passes): Move pass_lower_vector.
24982 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
24983 PROP_gimple_lvec.
24984 (pass_lower_vector): Provide PROP_gimple_lvec.
24985 (pass_lower_vector_ssa): Likewise.
24986 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
24987
24988 2013-03-21 Mark Wielaard <mjw@redhat.com>
24989
24990 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
24991
24992 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
24993
24994 * config/i386/i386.md (*movdi_internal): Disparage slightly
24995 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
24996 conversion alternatives.
24997
24998 2013-03-21 Jakub Jelinek <jakub@redhat.com>
24999
25000 PR middle-end/48087
25001 * diagnostic.def (DK_WERROR): New kind.
25002 * diagnostic.h (werrorcount): Define.
25003 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
25004 promoted to DK_ERROR, increment DK_WERROR counter instead of
25005 DK_ERROR counter.
25006 * toplev.c (toplev_main): Call print_ignored_options even if
25007 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
25008 even if just werrorcount is non-zero.
25009
25010 PR debug/55608
25011 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
25012 on failure.
25013 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
25014 (string_cst_pool_decl): New function.
25015 (optimize_one_addr_into_implicit_ptr): New function.
25016 (resolve_addr_in_expr): Optimize DWARF location expression
25017 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
25018 which doesn't live in memory, but has DW_AT_location or
25019 DW_AT_const_value, or refers to a string literal, into
25020 DW_OP_GNU_implicit_pointer.
25021 (optimize_location_into_implicit_ptr): New function.
25022 (resolve_addr): If removing DW_AT_location of a variable because
25023 it was DW_OP_addr of address of the variable, but the variable doesn't
25024 live in memory, try to emit const value attribute for the initializer.
25025
25026 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25027
25028 * tree.h (VECTOR_TYPE_P): New macro.
25029 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
25030 TYPE_MODE): Use it.
25031 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
25032 VEC_COND_EXPR cannot be lvalues.
25033 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
25034
25035 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25036
25037 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
25038 Restrict the transformation to equal modes.
25039
25040 2013-03-21 Richard Biener <rguenther@suse.de>
25041
25042 PR tree-optimization/39326
25043 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
25044 (MEM_ANALYZABLE): Adjust.
25045 (record_mem_ref_loc): Move bitmap ops ...
25046 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
25047 unanalyzable refs, do not record locations for it.
25048 (analyze_memory_references): Allocate ref zero as shared
25049 unanalyzable ref.
25050 (refs_independent_p): Do not test for unanalyzed mems here.
25051 (ref_indep_loop_p_1): Special-case disambiguation against
25052 the unanalyzed ref.
25053 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
25054
25055 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
25056
25057 * config/arm/arm-protos.h (tune_params): Add
25058 prefer_neon_for_64bits field.
25059 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
25060 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
25061 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
25062 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
25063 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
25064 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
25065 (arm_option_override): Handle -mneon-for-64bits new option.
25066 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
25067 (prefer_neon_for_64bits): Declare new variable.
25068 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
25069 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
25070 (arch_enabled): Handle new arch types. Remove support for onlya8
25071 and nota8.
25072 (one_cmpldi2): Use new arch names.
25073 (zero_extend<mode>di2, extend<mode>di2): Ditto.
25074 * config/arm/arm.opt (mneon-for-64bits): Add option.
25075 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
25076 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
25077 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
25078 of onlya8.
25079 * doc/invoke.texi (-mneon-for-64bits): Document.
25080
25081 2013-03-21 Richard Biener <rguenther@suse.de>
25082
25083 PR tree-optimization/39326
25084 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
25085 (sort_bbs_in_loop_postorder_cmp): New function.
25086 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
25087
25088 2013-03-21 Richard Biener <rguenther@suse.de>
25089
25090 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
25091 (vect_insert_into_interleaving_chain): Likewise.
25092 (vect_drs_dependent_in_basic_block): Inline ...
25093 (vect_slp_analyze_data_ref_dependence): ... here. New function,
25094 split out from ...
25095 (vect_analyze_data_ref_dependence): ... here. Simplify.
25096 (vect_check_interleaving): Simplify.
25097 (vect_analyze_data_ref_dependences): Likewise. Split out ...
25098 (vect_slp_analyze_data_ref_dependences): ... this new function.
25099 (dr_group_sort_cmp): New function.
25100 (vect_analyze_data_ref_accesses): Compute data-reference groups
25101 here instead of in vect_analyze_data_ref_dependence. Use
25102 a more efficient algorithm.
25103 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
25104 vect_slp_analyze_data_ref_dependences. Call
25105 vect_analyze_data_ref_accesses earlier.
25106 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
25107 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
25108 (vect_slp_analyze_data_ref_dependences): New prototype.
25109
25110 2013-03-21 Richard Biener <rguenther@suse.de>
25111
25112 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
25113 ref is stored in the loop.
25114 (find_refs_for_sm): Walk only over all stores.
25115 (store_motion_loop): Allocate from lim_bitmap_obstack.
25116 (store_motion): Likewise.
25117
25118 2013-03-21 Richard Biener <rguenther@suse.de>
25119
25120 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
25121 Update virtual SSA form.
25122
25123 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25124
25125 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
25126 * configure: Regenerate.
25127 * config.in: Regenerate.
25128 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
25129 if !HAVE_LD_EH_FRAME_CIEV3.
25130
25131 2013-03-21 Richard Biener <rguenther@suse.de>
25132
25133 * tree-cfg.c (verify_expr_no_block): New function.
25134 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
25135 nor DECL_VALUE_EXPR have locations with associated blocks.
25136 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
25137 (clear_unused_block_pointer): Remove code dealing with
25138 blocks in DECL_DEBUG_EXPR locations.
25139
25140 2013-03-21 Richard Biener <rguenther@suse.de>
25141
25142 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
25143 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
25144 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
25145 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
25146 instead of DECL_DEBUG_EXPR_IS_FROM.
25147 * gimplify.c (gimplify_modify_expr): Likewise.
25148 * tree-cfg.c (verify_expr_location_1): Likewise.
25149 * tree-complex.c (create_one_component_var): Likewise.
25150 * tree-sra.c (create_access_replacement): Likewise.
25151 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
25152 (clear_unused_block_pointer): Likewise.
25153 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
25154 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
25155 * var-tracking.c (var_debug_decl): Likewise.
25156 (track_expr_p): Likewise.
25157 * tree-inline.c (add_local_variables): Likewise. Set
25158 DECL_HAS_DEBUG_EXPR_P after copying it.
25159 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
25160 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
25161
25162 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25163
25164 PR bootstrap/56656
25165 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
25166 * configure: Regenerate.
25167 * config.in: Regenerate.
25168 * config/i386/i386.md (*movdf_internal): Use
25169 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25170 movd instead of movq mnemonic for interunit moves.
25171 (*movdi_internal): Ditto.
25172
25173 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25174
25175 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
25176 (abd<mode>_3): New pattern.
25177 (aba<mode>_3): New pattern.
25178 (fabd<mode>_3): New pattern.
25179
25180 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25181
25182 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
25183 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
25184 occurrence of REGISTER_PREFIX as its empty string.
25185
25186 2013-03-20 Jeff Law <law@redhat.com>
25187
25188 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
25189 addititional equivalences for equality comparisons between an SSA_NAME
25190 and a constant where the SSA_NAME was set from a widening conversion.
25191
25192 2013-03-20 Walter Lee <walt@tilera.com>
25193
25194 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
25195
25196 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25197
25198 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
25199 depending on type attribute.
25200 (*movti_internal): Ditto.
25201 (*movtf_internal): Ditto.
25202 (*movxf_internal): Ditto.
25203 (*movdf_internal): Ditto.
25204 (*movsf_internal): Ditto.
25205
25206 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25207
25208 * config/i386/i386.md (*movti_internal): Set prefix attribute to
25209 maybe_vex for sselog1 and ssemov types.
25210 (*movdi_internal): Reorder operand constraints.
25211 (*movsi_internal): Ditto. Set prefix attribute to
25212 maybe_vex for sselog1 and ssemov types.
25213 (*movtf_internal): Set prefix attribute to maybe_vex
25214 for sselog1 and ssemov types.
25215 (*movdf_internal): Ditto. Set prefix_data16 attribute for
25216 DImode ssemov types. Reorder operand constraints.
25217 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
25218 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
25219 attribute for SImode ssemov types. Reorder operand constraints.
25220
25221 2013-03-20 Martin Jambor <mjambor@suse.cz>
25222
25223 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
25224 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
25225
25226 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
25227
25228 * config/rs6000/predicates.md (indexed_address, update_address_mem
25229 update_indexed_address_mem): New predicates.
25230 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
25231 attribute for load/store instructions.
25232 * config/rs6000/dfp.md (movsd_store): Likewise.
25233 (movsd_load): Likewise.
25234 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
25235 (unnamed HI->DI extend define_insn): Likewise.
25236 (unnamed SI->DI extend define_insn): Likewise.
25237 (unnamed QI->SI extend define_insn): Likewise.
25238 (unnamed QI->HI extend define_insn): Likewise.
25239 (unnamed HI->SI extend define_insn): Likewise.
25240 (unnamed HI->SI extend define_insn): Likewise.
25241 (extendsfdf2_fpr): Likewise.
25242 (movsi_internal1): Likewise.
25243 (movsi_internal1_single): Likewise.
25244 (movhi_internal): Likewise.
25245 (movqi_internal): Likewise.
25246 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
25247 attribute for load/store instructions.
25248 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
25249 instructions.
25250 (mov<mode>_softfloat): Likewise.
25251 (mov<mode>_hardfloat32): Likewise.
25252 (mov<mode>_hardfloat64): Likewise.
25253 (mov<mode>_softfloat64): Likewise.
25254 (movdi_internal32): Likewise.
25255 (movdi_internal64): Likewise.
25256 (probe_stack_<mode>): Likewise.
25257
25258 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
25259
25260 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
25261 floating point, and decimal floating point to reload iterator.
25262
25263 * config/rs6000/constraints.md (wl constraint): New constraints to
25264 return FLOAT_REGS if certain options are used to reduce the number
25265 of separate patterns that exist in the file.
25266 (wx constraint): Likewise.
25267 (wz constraint): Likewise.
25268
25269 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25270 -mdebug=reg, print wg, wl, wx, and wz constraints.
25271 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25272 Initialize the reload functions for 64-bit binary/decimal floating
25273 point types.
25274 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25275 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25276 create the buffer on the stack to overcome not having a 32-bit
25277 load and store.
25278 (rs6000_emit_move): Likewise.
25279 (rs6000_secondary_memory_needed_rtx): Likewise.
25280 (rs6000_alloc_sdmode_stack_slot): Likewise.
25281 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25282 via xxlxor, just like DFmode 0.0.
25283
25284 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
25285 define as 1 if we are running on a power7 or newer.
25286 (enum r6000_reg_class_enum): Add new constraints.
25287
25288 * config/rs6000/dfp.md (movsd): Delete, combine with binary
25289 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25290 with other moves by using conditional constraits (wg). Use LFIWZX
25291 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
25292 (movsd splitter): Likewise.
25293 (movsd_hardfloat): Likewise.
25294 (movsd_softfloat): Likewise.
25295
25296 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
25297 binary and decimal floating point moves.
25298 (fmove_ok): New attributes to combine binary and decimal floating
25299 point moves, and to combine power6x (mfpgpr) moves along normal
25300 floating moves.
25301 (real_value_to_target): Likewise.
25302 (f32_lr): Likewise.
25303 (f32_lm): Likewise.
25304 (f32_li): Likewise.
25305 (f32_sr): Likewise.
25306 (f32_sm): Likewise.
25307 (f32_si): Likewise.
25308 (movsf): Combine binary and decimal floating point moves. Combine
25309 power6x (mfpgpr) moves with other moves by using conditional
25310 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
25311 (mov<mode> for SFmode/SDmode); Likewise.
25312 (SFmode/SDmode splitters): Likewise.
25313 (movsf_hardfloat): Likewise.
25314 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
25315 (movsf_softfloat): Likewise.
25316 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
25317
25318 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
25319 wx and wz constraints.
25320
25321 * config/rs6000/constraints.md (wg constraint): New constraint to
25322 return FLOAT_REGS if -mmfpgpr (power6x) was used.
25323
25324 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
25325 constraint.
25326
25327 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25328 -mdebug=reg, print wg, wl, wx, and wz constraints.
25329 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25330 Initialize the reload functions for 64-bit binary/decimal floating
25331 point types.
25332 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25333 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25334 create the buffer on the stack to overcome not having a 32-bit
25335 load and store.
25336 (rs6000_emit_move): Likewise.
25337 (rs6000_secondary_memory_needed_rtx): Likewise.
25338 (rs6000_alloc_sdmode_stack_slot): Likewise.
25339 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25340 via xxlxor, just like DFmode 0.0.
25341
25342 * config/rs6000/dfp.md (movdd): Delete, combine with binary
25343 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25344 with other moves by using conditional constraits (wg). Use LFIWZX
25345 and STFIWX for loading SDmode on power7.
25346 (movdd splitters): Likewise.
25347 (movdd_hardfloat32): Likewise.
25348 (movdd_softfloat32): Likewise.
25349 (movdd_hardfloat64_mfpgpr): Likewise.
25350 (movdd_hardfloat64): Likewise.
25351 (movdd_softfloat64): Likewise.
25352
25353 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
25354 64-bit binary and decimal floating point moves.
25355 (FMOVE64X): Likewise.
25356 (movdf): Combine 64-bit binary and decimal floating point moves.
25357 Combine power6x (mfpgpr) moves with other moves by using
25358 conditional constraits (wg).
25359 (mov<mode> for DFmode/DDmode): Likewise.
25360 (DFmode/DDmode splitters): Likewise.
25361 (movdf_hardfloat32): Likewise.
25362 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
25363 (movdf_softfloat32): Likewise.
25364 (movdf_hardfloat64_mfpgpr): Likewise.
25365 (movdf_hardfloat64): Likewise.
25366 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
25367 (movdf_softfloat64): Likewise.
25368 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
25369 (reload_<mode>_load): Move to later in the file so they aren't in
25370 the middle of the floating point move insns.
25371 (reload_<mode>_store): Likewise.
25372
25373 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
25374 constraint.
25375
25376 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
25377 constraint if -mdebug=reg.
25378 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
25379 Enable using dd reload support if needed.
25380
25381 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
25382 binary and decimal floating point moves in rs6000.md.
25383 (movtd_internal): Likewise.
25384
25385 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
25386 decimal floating point moves.
25387 (movtf): Likewise.
25388 (movtf_internal): Likewise.
25389 (mov<mode>_internal, TDmode/TFmode): Likewise.
25390 (movtf_softfloat): Likewise.
25391 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
25392
25393 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
25394 movdi_internal64, using wg constraint for move direct operations.
25395 (movdi_internal64): Likewise.
25396
25397 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
25398 MODES_TIEABLE_P for selected modes. Print the numerical value of
25399 the various virtual registers. Use GPR/FPR first/last values,
25400 instead of hard coding the register numbers. Print which modes
25401 have reload functions registered.
25402 (rs6000_option_override_internal): If -mdebug=reg, trace the options
25403 settings before/after setting cpu, target and subtarget settings.
25404 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
25405 and for secondary reload failures in rs6000_secondary_reload_inner.
25406 (rs6000_secondary_reload_fail): Likewise.
25407 (rs6000_secondary_reload_inner): Likewise.
25408
25409 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
25410 macros for first/last GPR and FPR registers.
25411 (LAST_GPR_REGNO): Likewise.
25412 (FIRST_FPR_REGNO): Likewise.
25413 (LAST_FPR_REGNO): Likewise.
25414
25415 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
25416 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
25417 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
25418 (vcond<mode><mode>): Likewise.
25419 (vcondu<mode><mode>): Likewise.
25420 (vector_gtu<mode>): Likewise.
25421 (vector_gte<mode>): Likewise.
25422 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
25423 to prevent the compiler from converting DImode operations to TImode.
25424 (ior<mode>3): Likewise.
25425 (and<mode>3): Likewise.
25426 (one_cmpl<mode>2): Likewise.
25427 (nor<mode>3): Likewise.
25428 (andc<mode>3): Likewise.
25429
25430 * config/rs6000/constraints.md (wt constraint): New constraint
25431 that returns VSX_REGS if TImode is allowed in VSX registers.
25432
25433 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
25434 constant under VSX.
25435
25436 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
25437 similar to TImode, but it is restricted to being in the GPRs.
25438
25439 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
25440 TImode to occupy a single VSX register.
25441
25442 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
25443 -mvsx-timode for power7/power8.
25444 (power7 cpu): Likewise.
25445 (power8 cpu): Likewise.
25446
25447 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
25448 sure that TFmode/TDmode take up two registers if they are ever
25449 allowed in the upper VSX registers.
25450 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
25451 registers.
25452 (rs6000_init_hard_regno_mode_ok): Likewise.
25453 (rs6000_debug_reg_global): Add debugging for PTImode and wt
25454 constraint. Print if LRA is turned on.
25455 (rs6000_option_override_internal): Give an error if -mvsx-timode
25456 and VSX is not enabled.
25457 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
25458 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
25459 to reg+offset addressing. Use PTImode when checking offset
25460 addresses for validity.
25461 (reg_offset_addressing_ok_p): Likewise.
25462 (rs6000_legitimate_offset_address_p): Likewise.
25463 (rs6000_legitimize_address): Likewise.
25464 (rs6000_legitimize_reload_address): Likewise.
25465 (rs6000_legitimate_address_p): Likewise.
25466 (rs6000_eliminate_indexed_memrefs): Likewise.
25467 (rs6000_emit_move): Likewise.
25468 (rs6000_secondary_reload): Likewise.
25469 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
25470 reloads to fpr registers to continue to use reg+offset addressing,
25471 but 64-bit reloads to altivec registers need reg+reg addressing.
25472 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
25473 it. Treat LO_SUM like a PLUS operation.
25474 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
25475 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
25476 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
25477 registers to share a register with a smaller sized type, since VSX
25478 puts scalars in the upper 64-bits.
25479 (print_operand): Add support for PTImode.
25480 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
25481 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
25482 registers, but don't have arithmetic support.
25483 (rs6000_memory_move_cost): Add test for VSX.
25484 (rs6000_opt_masks): Add -mvsx-timode.
25485
25486 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
25487 for TImode.
25488 (VSs): Likewise.
25489 (VSr): Use wt constraint for TImode.
25490 (VSv): Drop TImode support.
25491 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
25492 (vsx_movti_64bit): Likewise.
25493 (vsx_movti_32bit): Likewise.
25494 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
25495 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
25496 one '?' on the appropriate output constraint. Do not allow TImode
25497 logical operations on 32-bit systems.
25498 (vsx_ior<mode>3): Likewise.
25499 (vsx_xor<mode>3): Likewise.
25500 (vsx_one_cmpl<mode>2): Likewise.
25501 (vsx_nor<mode>3): Likewise.
25502 (vsx_andc<mode>3): Likewise.
25503 (vsx_concat_<mode>): Likewise.
25504 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
25505
25506 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
25507 OPTION_MASK_VSX_TIMODE.
25508 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
25509 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
25510
25511 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
25512 (TI2 iterator): New iterator for TImode, PTImode.
25513 (wd mode attribute): Add values for vector types.
25514 (movti_string): Replace TI move operations with operations for TImode
25515 and PTImode. Add support for TImode being allowed in VSX registers.
25516 (mov<mode>_string, TImode/PTImode): Likewise.
25517 (movti_ppc64): Likewise.
25518 (mov<mode>_ppc64, TImode/PTImode): Likewise.
25519 (TI mode splitters): Likewise.
25520
25521 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
25522 constraint.
25523
25524 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
25525
25526 PR tree-optimization/56355
25527 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
25528 Also handle integers with undefined overflow.
25529
25530 2013-03-20 Catherine Moore <clm@codesourcery.com>
25531 Maciej W. Rozycki <macro@codesourcery.com>
25532 Tom de Vries <tom@codesourcery.com>
25533 Nathan Sidwell <nathan@codesourcery.com>
25534 Iain Sandoe <iain@codesourcery.com>
25535 Nathan Froyd <froydnj@codesourcery.com>
25536 Chao-ying Fu <fu@mips.com>
25537
25538 * doc/extend.texi: (micromips, nomicromips, nocompression):
25539 Document new function attributes.
25540 * doc/invoke.texi (minterlink-compressed, mmicromips,
25541 m14k, m14ke, m14kec): Document new options.
25542 (minterlink-mips16): Update documentation.
25543 * doc/md.texi (ZC, ZD): Document new constraints.
25544 * configure.ac (gcc_cv_as_micromips): Check if linker
25545 supports the .set micromips directive.
25546 * configure: Regenerate.
25547 * config.in: Regenerate.
25548 * config/mips/mips-tables.opt: Regenerate.
25549 * config/mips/micromips.md: New file.
25550 * constraints.md (ZC, ZD): New constraints.
25551 * config/mips/predicates.md (movep_src_register): New predicate.
25552 (movep_src_operand): New predicate.
25553 (non_volatile_mem_operand): New predicate.
25554 * config/mips/mips.md (multimem): New type.
25555 (length): Differentiate between 17-bit and 18-bit branch offsets.
25556 (MOVEP1, MOVEP2): New mode iterator.
25557 (mov_<load>l): Use ZC constraint.
25558 (mov_<load>r): Likewise.
25559 (mov_<store>l): Likewise.
25560 (mov_<store>r): Likewise.
25561 (*branch_equality<mode>_inverted): Add microMIPS support.
25562 (*branch_equality<mode>): Likewise.
25563 (*jump_absolute): Likewise.
25564 (indirect_jump_<mode>): Likewise.
25565 (tablejump_<mode>): Likewise.
25566 (<optab>_internal): Likewise.
25567 (sibcall_internal): Likewise.
25568 (sibcall_value_internal): Likewise.
25569 (prefetch): Use constraint ZD.
25570 * config/mips/mips.opt (minterlink-compressed): New option.
25571 (minterlink-mips16): Now an alias for minterlink-compressed.
25572 (mmicromips): New option.
25573 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
25574 (compare_and_swap_12): Likewise.
25575 (sync_add<mode>): Likewise.
25576 (sync_<optab>_12): Likewise.
25577 (sync_old_<optab>_12): Likewise.
25578 (sync_new_<optab>_12): Likewise.
25579 (sync_nand_12): Likewise.
25580 (sync_old_nand_12): Likewise.
25581 (sync_new_nand_12): Likewise.
25582 (sync_sub<mode>): Likewise.
25583 (sync_old_add<mode>): Likewise.
25584 (sync_old_sub<mode>): Likewise.
25585 (sync_new_add<mode>): Likewise.
25586 (sync_new_sub<mode>): Likewise.
25587 (sync_<optab><mode>): Likewise.
25588 (sync_old_<optab><mode>): Likewise.
25589 (sync_new_<optab><mode>): Likewise.
25590 (sync_nand<mode>): Likewise.
25591 (sync_old_nand<mode>): Likewise.
25592 (sync_new_nand<mode>): Likewise.
25593 (sync_lock_test_and_set<mode>): Likewise.
25594 (test_and_set_12): Likewise.
25595 (atomic_compare_and_swap<mode>): Likewise.
25596 (atomic_exchange<mode>_llsc): Likewise.
25597 (atomic_fetch_add<mode>_llsc): Likewise.
25598 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
25599 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
25600 (umips_save_restore_pattern_p): Likewise.
25601 (umips_load_store_pair_p): Likewise.
25602 (umips_output_load_store_pair): Likewise.
25603 (umips_movep_target_p): Likewise.
25604 (umips_12bit_offset_address_p): Likewise.
25605 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
25606 (mips_base_mips16): Rename this...
25607 (mips_base_compression_flags): ...to this. Update all uses.
25608 (mips_attribute_table): Add micromips, nomicromips and nocompression.
25609 (mips_mips16_decl_p): Delete.
25610 (mips_nomips16_decl_p): Delete.
25611 (mips_get_compress_on_flags): New function.
25612 (mips_get_compress_off_flags): New function.
25613 (mips_get_compress_mode): New function.
25614 (mips_get_compress_on_name): New function.
25615 (mips_get_compress_off_name): New function.
25616 (mips_insert_attributes): Support multiple compression types.
25617 (mips_merge_decl_attributes): Likewise.
25618 (umips_12bit_offset_address_p): New function.
25619 (mips_start_function_definition): Emit .set micromips directive.
25620 (mips_call_may_need_jalx_p): New function.
25621 (mips_function_ok_for_sibcall): Add microMIPS support.
25622 (mips_print_operand_punctuation): Support short delay slots and
25623 compact jumps.
25624 (umips_swm_mask, umips_swm_encoding): New.
25625 (umips_build_save_restore): New function.
25626 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
25627 (was_mips16_p): Remove.
25628 (old_compression_mode): New.
25629 (mips_set_compression_mode): New function.
25630 (mips_set_current_function): Add microMIPS support.
25631 (mips_option_override): Likewise.
25632 (umips_save_restore_pattern_p): New function.
25633 (umips_output_save_restore): New function.
25634 (umips_load_store_pair_p_1): New function.
25635 (umips_load_store_pair_p): New function.
25636 (umips_output_load_store_pair_1): New function.
25637 (umips_output_load_store_pair): New function.
25638 (umips_movep_target_p) New function.
25639 (mips_prepare_pch_save): Add microMIPS support.
25640 * config/mips/mips.h (TARGET_COMPRESSION): New.
25641 (TARGET_CPU_CPP_BUILTINS): Update macro
25642 to use new compression flags and to support microMIPS.
25643 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
25644 (MIPS_ARCH_FLOAT_SPEC): Likewise.
25645 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
25646 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
25647 (ASM_SPEC): Support mmicromips and mno-micromips.
25648 (M16STORE_REG_P): New macro.
25649 (MIPS_CALL): Support TARGET_MICROMIPS.
25650 (MICROMIPS_J): New macro.
25651 (mips_base_mips16): Rename this...
25652 (mips_base_compression_flags): ...to this.
25653 (UMIPS_12BIT_OFFSET_P): New macro.
25654 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
25655 (MULTILIB_DIRNAMES): Likewise.
25656 2013-03-20 Richard Biener <rguenther@suse.de>
25657
25658 PR tree-optimization/56661
25659 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
25660 the result does not have to be distinct.
25661
25662 2013-03-20 Richard Biener <rguenther@suse.de>
25663
25664 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
25665 remap_gimple_op_r.
25666
25667 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
25668 Steven Bosscher <steven@gcc.gnu.org>
25669
25670 PR rtl-optimization/56605
25671 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
25672
25673 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25674
25675 PR bootstrap/56656
25676 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
25677 that require movd instead of movq.
25678
25679 2013-03-20 Richard Biener <rguenther@suse.de>
25680
25681 * tree-ssa-structalias.c (struct variable_info): Add pointer
25682 to the first field of an aggregate with sub-vars. Make
25683 this and the pointer to the next subfield its ID.
25684 (vi_next): New function.
25685 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
25686 storedanything_id, integer_id): Increment by one.
25687 (new_var_info, get_call_vi, lookup_call_clobber_vi,
25688 get_call_clobber_vi): Adjust.
25689 (solution_set_expand): Simplify and speedup.
25690 (solution_set_add): Inline into ...
25691 (set_union_with_increment): ... this. Adjust accordingly.
25692 (do_sd_constraint): Likewise.
25693 (do_ds_constraint): Likewise.
25694 (do_complex_constraint): Simplify.
25695 (build_pred_graph): Adjust.
25696 (solve_graph): Likewise. Simplify and speedup.
25697 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
25698 get_constraint_for_component_ref, get_constraint_for_1,
25699 first_vi_for_offset, first_or_preceding_vi_for_offset,
25700 create_function_info_for, create_variable_info_for_1,
25701 create_variable_info_for, intra_create_variable_infos): Adjust.
25702 (init_base_vars): Push NULL for ID zero.
25703 (compute_points_to_sets): Adjust.
25704
25705 2013-03-20 Richard Biener <rguenther@suse.de>
25706
25707 * cfgloop.c (verify_loop_structure): Streamline and avoid
25708 ICEing on corrupt loop tree.
25709 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
25710 loop tree.
25711
25712 2013-03-20 Richard Biener <rguenther@suse.de>
25713
25714 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
25715 check whether an SSA update is needed.
25716
25717 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
25718
25719 * config/mips/constraints.md (T): Rename to...
25720 (Yf): ...this.
25721 (U): Rename to...
25722 (Yd): ...this.
25723 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
25724 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
25725
25726 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25727
25728 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
25729 (*subsi3_carryin_uxtw): Likewise.
25730
25731 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25732
25733 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
25734 (*rorsi3_insn_uxtw): Likewise.
25735
25736 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25737
25738 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
25739 (*extrsi5_insn_uxtw): Likewise.
25740
25741 2013-03-19 Richard Biener <rguenther@suse.de>
25742
25743 PR tree-optimization/56273
25744 * passes.c (init_optimization_passes): Move second VRP after DOM.
25745
25746 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
25747
25748 * config/i386/i386.md (*movti_internal): Merge from
25749 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
25750 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
25751 nox64 isa attributes.
25752
25753 2013-03-18 Richard Biener <rguenther@suse.de>
25754
25755 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
25756 (unite): Likewise.
25757 (merge_node_constraints): Likewise.
25758 (build_succ_graph): Likewise.
25759 (valid_graph_edge): Inline into single caller.
25760 (unify_nodes): Likewise. Use bitmap_set_bit return value
25761 and cache varinfo.
25762 (scc_visit): Fix formatting and variable use.
25763 (do_sd_constraint): Use gcc_checking_assert.
25764 (do_ds_constraint): Likewise.
25765 (do_complex_constraint): Likewise.
25766 (condense_visit): Likewise. Cleanup.
25767 (dump_pred_graph): New function.
25768 (perform_var_substitution): Dump the pred-graph before
25769 variable substitution.
25770 (find_equivalent_node): Use gcc_checking_assert.
25771 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
25772
25773 2013-03-18 Richard Biener <rguenther@suse.de>
25774
25775 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
25776 Remove cond_expr_stmt_list argument and do not gimplify the
25777 built expression.
25778 (vect_loop_versioning): Adjust.
25779 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
25780 Cleanup to use less temporaries.
25781 (vect_create_data_ref_ptr): Cleanup.
25782
25783 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25784
25785 PR tree-optimization/56635
25786 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
25787 require types_compatible_p types.
25788
25789 2013-03-18 Nick Clifton <nickc@redhat.com>
25790
25791 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
25792 spurious backslash.
25793
25794 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
25795 Add missing line to comment describing function.
25796
25797 2013-03-18 Richard Biener <rguenther@suse.de>
25798
25799 PR tree-optimization/56210
25800 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
25801 Handle string / character search functions.
25802 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
25803
25804 2013-03-18 Richard Biener <rguenther@suse.de>
25805
25806 PR middle-end/56483
25807 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
25808 and implement properly.
25809 * gimple.h (gimple_cond_single_var_p): Remove.
25810
25811 2013-03-18 Richard Biener <rguenther@suse.de>
25812
25813 * tree-data-ref.h (find_data_references_in_loop): Declare.
25814 * tree-data-ref.c (get_references_in_stmt): Use a stack
25815 vector pre-allocated in the callers.
25816 (find_data_references_in_stmt): Adjust.
25817 (graphite_find_data_references_in_stmt): Likewise.
25818 (create_rdg_vertices): Likewise.
25819 (find_data_references_in_loop): Export.
25820 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
25821 Compute dependences here...
25822 (vect_analyze_data_refs): ...not here. When we encounter
25823 a non-vectorizable data reference in basic-block vectorization
25824 truncate the data reference vector. Do not bother to
25825 fixup data-dependence information for gather loads.
25826 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
25827 of data references, as reported.
25828
25829 2013-03-18 Richard Biener <rguenther@suse.de>
25830
25831 PR tree-optimization/3713
25832 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
25833 has_constants and expr.
25834 (stmt_has_constants): Properly valueize SSA names when deciding
25835 whether the stmt has constants.
25836
25837 2013-03-18 Richard Biener <rguenther@suse.de>
25838
25839 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
25840 whole function when there is nothing to do.
25841 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
25842 * tree-vectorizer.c (vectorize_loops): Update virtual and
25843 loop-closed SSA once.
25844 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
25845
25846 2013-03-18 Richard Biener <rguenther@suse.de>
25847
25848 PR middle-end/56113
25849 * domwalk.c (bb_postorder): New global static.
25850 (cmp_bb_postorder): New function.
25851 (walk_dominator_tree): Replace scheme imposing an order for
25852 visiting dominator sons by one sorting them at the time they
25853 are pushed on the stack.
25854
25855 2013-03-18 Richard Biener <rguenther@suse.de>
25856
25857 PR tree-optimization/39326
25858 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
25859 (struct mem_ref): Replace mem member with ao_ref typed member.
25860 (MEM_ANALYZABLE): Adjust.
25861 (memref_eq): Likewise.
25862 (mem_ref_alloc): Likewise.
25863 (gather_mem_refs_stmt): Likewise.
25864 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
25865 (execute_sm_if_changed_flag_set): Adjust.
25866 (execute_sm): Likewise.
25867 (ref_always_accessed_p): Likewise.
25868 (refs_independent_p): Likewise.
25869 (can_sm_ref_p): Likewise.
25870
25871 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25872
25873 PR c/56566
25874 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
25875 return 1 even for !unsignedp.
25876
25877 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25878
25879 * config/i386/i386.md (isa): Add x64 and nox64.
25880 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
25881 (*pushtf): Enable *roF alternative for x64 isa only.
25882 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
25883 mode attribute of integer alternatives to DImode for TARGET_64BIT.
25884 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
25885 (*movtf_internal): Merge from *movtf_internal_rex64 and
25886 *movtf_internal_sse. Use x64 and nox64 isa attributes.
25887 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
25888 nox64 isa attributes.
25889 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
25890 nox64 isa attributes.
25891 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
25892
25893 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25894
25895 * config/alpha/alpha.c (TARGET_LRA_P): New define.
25896
25897 2013-03-17 Jakub Jelinek <jakub@redhat.com>
25898
25899 PR target/56640
25900 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
25901 class names. Remove trailing comma after "ALL_REGS".
25902
25903 2013-03-16 Jan Hubicka <jh@suse.cz>
25904
25905 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
25906 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
25907 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
25908 of cgraph_get_create_node.
25909 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
25910
25911 2013-03-16 Jason Merrill <jason@redhat.com>
25912
25913 PR debug/49090
25914 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
25915 with DW_AT_default_value.
25916
25917 2013-03-16 Jakub Jelinek <jakub@redhat.com>
25918
25919 * BASE-VER: Set to 4.9.0.
25920
25921 2013-03-14 Andi Kleen <ak@linux.intel.com>
25922
25923 PR target/56619
25924 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
25925 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
25926 Document _x* TSX intrinsics.
25927
25928 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
25929 David Holsgrove <david.holsgrove@xilinx.com>
25930
25931 * configure.ac: Add MicroBlaze TLS support detection.
25932 * configure: Regenerate.
25933 * config/microblaze/microblaze-protos.h
25934 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
25935 symbol_mentioned_p, label_mentioned_p): Add prototypes.
25936 * config/microblaze/microblaze.c (microblaze_address_type): Add
25937 ADDRESS_TLS and tls_reloc address types.
25938 (microblaze_address_info): Add tls_reloc.
25939 (TARGET_HAVE_TLS): Define.
25940 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
25941 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
25942 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
25943 load_tls_operand, microblaze_call_tls_get_addr,
25944 microblaze_legitimize_tls_address): New functions.
25945 (microblaze_classify_unspec): Handle UNSPEC_TLS.
25946 (get_base_reg): Use microblaze_tls_symbol_p.
25947 (microblaze_classify_address): Handle TLS.
25948 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
25949 label_mentioned_p and microblaze_tls_referenced_p.
25950 (microblaze_legitimize_address): Handle TLS.
25951 (microblaze_address_insns): Handle ADDRESS_TLS.
25952 (pic_address_needs_scratch): Handle TLS.
25953 (print_operand_address): Handle TLS.
25954 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
25955 (microblaze_expand_move): Handle TLS.
25956 (microblaze_legitimate_constant_p): Check
25957 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
25958 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
25959 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
25960 (PIC_OFFSET_TABLE_REGNUM): Set.
25961 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
25962 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
25963 (addsi3, movsi_internal2, movdf_internal): Update constraints
25964 * config/microblaze/predicates.md (arith_plus_operand): Define
25965 (move_operand): Redefine as move_src_operand,
25966 check microblaze_tls_referenced_p.
25967
25968 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25969
25970 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
25971 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
25972
25973 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25974
25975 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
25976 CC mode for AND.
25977
25978 2013-03-14 Jakub Jelinek <jakub@redhat.com>
25979
25980 PR tree-optimization/53265
25981 * common.opt (Waggressive-loop-optimizations): New option.
25982 * tree-ssa-loop-niter.c: Include tree-pass.h.
25983 (do_warn_aggressive_loop_optimizations): New function.
25984 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
25985 if number_of_latch_executions returned constant.
25986 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
25987 early. If number_of_latch_executions returned constant, set
25988 nb_iterations_upper_bound back to it.
25989 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
25990 field.
25991 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
25992 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
25993
25994 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
25995 (MULTILIB_OSDIRNAMES): Set.
25996 * genmultilib: If defaultosdirname doesn't start with :: , set
25997 defaultosdirname2 instead, clear it and emit two . multilib_raw
25998 entries instead of just one.
25999
26000 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
26001
26002 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26003 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
26004 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26005 (SUBTARGET_OVERRIDE_OPTIONS): New.
26006
26007 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
26008
26009 PR target/49880
26010 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
26011 (musermode): Convert to Var(TARGET_USERMODE).
26012 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
26013 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
26014 * config/sh/sh.c (sh_option_override): Use
26015 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
26016 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
26017 condition.
26018 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
26019 TARGET_SH4.
26020 (udivsi3_i4_single, divsi3_i4_single): Use
26021 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
26022
26023 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
26024
26025 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
26026 default setting.
26027
26028 2013-03-13 Richard Biener <rguenther@suse.de>
26029
26030 PR tree-optimization/56608
26031 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
26032 calls when vectorizing basic-blocks.
26033
26034 2013-03-13 Jakub Jelinek <jakub@redhat.com>
26035
26036 PR plugins/45078
26037 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
26038 tm_file.
26039
26040 2013-03-12 Jakub Jelinek <jakub@redhat.com>
26041
26042 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
26043
26044 2013-03-11 Jan Hubicka <jh@suse.cz>
26045
26046 PR lto/56557
26047 * lto-streamer-out.c (output_symbol_p): Skip references from
26048 constructors of external variables.
26049
26050 2013-03-11 Jan Hubicka <jh@suse.cz>
26051
26052 PR middle-end/56571
26053 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
26054 from pseudos.
26055 * emit-rtl.c (verify_rtx_sharing): Likewise.
26056 (copy_insn_1): Likewise.
26057 * rtl.c (copy_rtx): Likewise.
26058
26059 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
26060
26061 PR target/56591
26062 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
26063 output_operand_lossage message.
26064
26065 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
26066
26067 PR target/56470
26068 * arm.c (shift_op): Validate RTL pattern on the fly.
26069 (arm_print_operand, case 'S'): Don't use shift_operator to validate
26070 the RTL.
26071
26072 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26073
26074 PR target/56347
26075 * config/pa/pa.md (call_value): Check for calls to powf and direct to
26076 new call patterns that clobber %fr12.
26077 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
26078 split and postreload patterns.
26079 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
26080 registers %fr12 and %fr12R as call used.
26081
26082 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
26083
26084 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
26085 (canon_address, record_store, replace_read, check_mem_read_rtx,
26086 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
26087 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
26088 rest_of_handle_dse): Likewise.
26089
26090 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
26091
26092 PR middle-end/56524
26093 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
26094 Add base_optabs.
26095 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
26096 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
26097 (save_optabs_if_changed): Replace with...
26098 (init_tree_optimization_optabs): ...this.
26099 * optabs.c (save_optabs_if_changed): Rename to...
26100 (init_tree_optimization_optabs): ...this. Take the optimization node
26101 as argument. Do nothing if the base optabs are already correct.
26102 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
26103 to recompute optabs.
26104 * function.h (function): Remove optabs field.
26105 * function.c (invoke_set_current_function_hook): Call
26106 init_tree_optimization_optabs. Use the result to initialize
26107 this_fn_optabs.
26108
26109 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
26110
26111 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
26112 if GTMA_HAS_NO_INSTRUMENTATION.
26113 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
26114 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
26115 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
26116 * gimple-pretty-print.c (dump_gimple_transaction): Handle
26117 GTMA_HAS_NO_INSTRUMENTATION.
26118
26119 2013-03-08 Jakub Jelinek <jakub@redhat.com>
26120
26121 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
26122 libasan_preinit.o.
26123
26124 2013-03-08 Marek Polacek <polacek@redhat.com>
26125 Jakub Jelinek <jakub@redhat.com>
26126
26127 PR tree-optimization/56478
26128 * predict.c (is_comparison_with_loop_invariant_p): Change the
26129 type of loop_step to tree.
26130 (predict_loops): Adjust.
26131 (predict_iv_comparison): Perform the computations on double_ints.
26132
26133 2013-03-08 Richard Biener <rguenther@suse.de>
26134
26135 PR tree-optimization/56570
26136 * tree-cfg.c (verify_expr_location_1): Verify locations for
26137 DECL_DEBUG_EXPR.
26138 * tree-sra.c (create_access_replacement): Strip locations
26139 from DECL_DEBUG_EXPRs.
26140
26141 2013-03-08 Richard Biener <rguenther@suse.de>
26142
26143 * tree-inline.c (expand_call_inline): Do not associate
26144 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
26145 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
26146
26147 2013-03-08 Richard Biener <rguenther@suse.de>
26148
26149 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
26150 or block changes with -Og. Fix for location / block encoding
26151 changes and PHI arguments with locations.
26152
26153 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
26154
26155 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
26156 for all counters.
26157 (struct output_info): Likewise.
26158 (register_overhead): Remove bad gcc_assert.
26159 (bitmap_find_bit): If there is only a single bitmap element, do not
26160 count a miss as a search.
26161 (print_statistics): Update for counter type changes.
26162 (dump_bitmap_statistics): Likewise. Print headers such that they
26163 are properly lined up with the printed counters.
26164
26165 2013-03-07 Jakub Jelinek <jakub@redhat.com>
26166
26167 PR tree-optimization/56559
26168 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
26169 check that it has only a single use.
26170
26171 2013-03-07 Richard Biener <rguenther@suse.de>
26172
26173 * doc/invoke.texi (fwhole-program): Discourage use in combination
26174 with -flto.
26175
26176 2013-03-06 Jakub Jelinek <jakub@redhat.com>
26177
26178 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
26179
26180 PR tree-optimization/56539
26181 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
26182 instead of GSI_CONTINUE_LINKING as last argument to
26183 force_gimple_operand_gsi. Adjust function comment.
26184
26185 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
26186 aarch64-cores.def.
26187
26188 PR middle-end/56548
26189 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
26190 promoted mode, convert the result back to the original mode.
26191
26192 2013-03-06 Richard Biener <rguenther@suse.de>
26193
26194 PR middle-end/56294
26195 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
26196 (insert_updated_phi_nodes_compare_uids): New function.
26197 (update_ssa): Sort symbols_to_rename after UID before
26198 traversing it to insert PHI nodes.
26199
26200 2013-03-06 Richard Biener <rguenther@suse.de>
26201
26202 PR middle-end/50494
26203 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
26204 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
26205
26206 Revert
26207 2013-02-13 Richard Biener <rguenther@suse.de>
26208
26209 PR lto/50494
26210 * varasm.c (output_constant_def_1): Get the decl representing
26211 the constant as argument.
26212 (output_constant_def): Wrap output_constant_def_1.
26213 (make_decl_rtl): Use output_constant_def_1 with the decl
26214 representing the constant.
26215 (build_constant_desc): Optionally re-use a decl already
26216 representing the constant.
26217 (tree_output_constant_def): Adjust.
26218
26219 2013-03-06 Joey Ye <joey.ye@arm.com>
26220
26221 PR lto/50293
26222 * gcc.c (convert_white_space): New function.
26223 (main): Handles white space in function name.
26224
26225 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
26226
26227 PR target/56529
26228 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
26229 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
26230 to SH_DIV_CALL_TABLE for TARGET_SH2.
26231 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
26232 list.
26233 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
26234 call-table options.
26235
26236 2013-03-05 Sterling Augustine <saugustine@google.com>
26237 Cary Coutant <ccoutant@google.com>
26238
26239 PR debug/55364
26240 * dwarf2out.c (resolve_addr): Don't call
26241 remove_loc_list_addr_table_entries a second time for the same
26242 expression.
26243
26244 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26245
26246 PR debug/56510
26247 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
26248 (avoid_complex_debug_insns): New function.
26249 (expand_debug_locations): Call it.
26250
26251 PR rtl-optimization/56484
26252 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
26253 lifetimes of hard registers on small register class machines.
26254
26255 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26256
26257 * config/microblaze/microblaze-protos.h: Rename
26258 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
26259 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
26260 fast_interrupt.
26261 (microblaze_fast_interrupt_function_p): New function.
26262 (microblaze_is_interrupt_handler): Rename to
26263 microblaze_is_interrupt_variant and add fast_interrupt check.
26264 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
26265 (save_restore_insns): Likewise.
26266 (compute_frame_size): Likewise.
26267 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
26268 (microblaze_globalize_label): Likewise.
26269 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
26270 * config/microblaze/microblaze.md: Use wrapper
26271 microblaze_is_interrupt_variant.
26272
26273 2013-03-05 Kai Tietz <ktietz@redhat.com>
26274
26275 * sdbout.c (sdbout_one_type): Switch to current function's section
26276 supporting cold/hot.
26277
26278 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26279
26280 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
26281 -mxl-reorder.
26282
26283 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26284
26285 PR middle-end/56461
26286 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
26287 if VALGRIND_GET_VBITS is defined, temporarily make object
26288 memory all defined, and restore previous valgrind addressability
26289 and definability afterwards. Free this_object at the end.
26290
26291 PR middle-end/56461
26292 * lra.c (lra): Call lra_clear_live_ranges if live_p,
26293 right before calling lra_create_live_ranges, also call it
26294 when clearing live_p. Only call lra_clear_live_ranges
26295 at the end if live_p.
26296
26297 PR middle-end/56461
26298 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
26299
26300 2013-03-05 Richard Biener <rguenther@suse.de>
26301
26302 PR tree-optimization/56521
26303 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
26304 value-id.
26305
26306 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
26307
26308 PR c++/55135
26309 * except.h (remove_unreachable_eh_regions): New prototype.
26310 * except.c (remove_eh_handler_splicer): New function, split out
26311 of remove_eh_handler.
26312 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
26313 warning about running it on many EH regions one at a time.
26314 (remove_unreachable_eh_regions_worker): New function, walk the
26315 EH tree in depth-first order and remove non-marked regions.
26316 (remove_unreachable_eh_regions): New function.
26317 * tree-eh.c (mark_reachable_handlers): New function, split out
26318 from remove_unreachable_handlers.
26319 (remove_unreachable_handlers): Use mark_reachable_handlers and
26320 remove_unreachable_eh_regions.
26321 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
26322 and remove_unreachable_eh_regions.
26323
26324 2013-03-05 Richard Biener <rguenther@suse.de>
26325
26326 PR middle-end/56525
26327 * loop-init.c (fix_loop_structure): Remove loops in two stages,
26328 not freeing them until the end.
26329
26330 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26331
26332 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
26333
26334 2013-03-05 Richard Biener <rguenther@suse.de>
26335
26336 PR tree-optimization/56270
26337 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
26338 of loads after scheduling an SLP instance.
26339
26340 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26341
26342 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
26343 tic6x.exp.
26344 (check_gcc_parallelize): Run guality.exp as a separate job from
26345 vect.exp with unsorted.exp and $(dg_target_exps) separately from
26346 struct-layout-1.exp with stackalign.exp.
26347
26348 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
26349
26350 PR middle-end/56461
26351 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
26352 load_index sbitmap even if some bit in it isn't set.
26353
26354 PR middle-end/56461
26355 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
26356 (discover_iteration_bound_by_body_walk): Change queues to
26357 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
26358 spelling in comment. Call safe_push on queues[bound_index] directly.
26359 Release queues[queue_index] in every iteration unconditionally.
26360 Release bounds vector.
26361
26362 PR middle-end/56461
26363 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
26364 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
26365 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
26366 inner_phis vector.
26367
26368 2013-03-05 Richard Biener <rguenther@suse.de>
26369
26370 PR lto/56515
26371 * tree-inline.c (remap_blocks_to_null): New function.
26372 (expand_call_inline): When expanding a call stmt without
26373 an associated block inline remap all callee blocks to NULL.
26374
26375 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26376
26377 PR rtl-optimization/56494
26378 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
26379 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
26380 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
26381
26382 PR middle-end/56461
26383 * sel-sched-ir.c (free_sched_pools): Release
26384 succs_info_pool.stack[succs_info_pool.max_top] vectors too
26385 if succs_info_pool.max_top isn't -1.
26386
26387 PR bootstrap/56509
26388 * opts.c (opts_obstack, opts_concat): Moved to...
26389 * opts-common.c (opts_obstack, opts_concat): ... here.
26390
26391 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26392
26393 PR middle-end/56461
26394 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
26395
26396 2013-03-04 Martin Jambor <mjambor@suse.cz>
26397
26398 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
26399 all appropriate places.
26400
26401 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
26402
26403 PR tree-optimization/56424
26404 * ipa-split.c (split_function): Do not set the RSO flag if result is
26405 not by reference and its type is a register type.
26406
26407 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
26408
26409 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
26410 (microblaze_legitimate_pic_operand): Likewise
26411 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
26412 new function microblaze_legitimate_pic_operand
26413 * config/microblaze/microblaze-protos.h
26414 (microblaze_legitimate_pic_operand): Declare.
26415
26416 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26417
26418 * config/microblaze/predicates.md (call_insn_simple_operand):
26419 New predicate for supported rtx code types.
26420 * config/microblaze/microblaze.md (call_internal1): Use
26421 call_insn_simple_operand predicate.
26422
26423 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26424
26425 PR middle-end/56461
26426 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
26427 partitions.ordered_remove.
26428
26429 PR middle-end/56461
26430 * tree-vect-stmts.c (vectorizable_conversion): Don't call
26431 vec_oprnds0.create (1) for modifier == NONE.
26432
26433 PR middle-end/56461
26434 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
26435 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
26436 vec_oprnds1 right before pushing anything to it for
26437 scalar_shift_arg.
26438
26439 PR middle-end/56461
26440 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
26441 set nbbs to 0 instead of having separate code path.
26442 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
26443 instead of false as last argument if returning NULL.
26444
26445 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
26446
26447 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
26448 the attribute is now called "target" instead of "option".
26449 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
26450 * doc/tm.texi.in (Target Attributes): Likewise document the correct
26451 attribute/pragma name for TARGET_OPTION_VALID_P and
26452 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
26453 * doc/tm.texi: Regenerated.
26454
26455 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
26456
26457 * config/microblaze/microblaze.c:
26458 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
26459 * config/microblaze/microblaze.h: Add -mxl-reorder to
26460 DRIVER_SELF_SPECS.
26461 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
26462 instructions emitted if TARGET_REORDER.
26463 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
26464 or 0 for -m/-mno case, but initialises as 2 to detect default use case
26465 separately.
26466
26467 2013-03-01 Xinliang David Li <davidxl@google.com>
26468
26469 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
26470 walk length.
26471
26472 2013-03-01 Jakub Jelinek <jakub@redhat.com>
26473
26474 PR middle-end/56461
26475 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
26476 vector even when returning true. Fix up function comment formatting.
26477
26478 PR middle-end/56461
26479 * ira-build.c (ira_loop_nodes_count): New variable.
26480 (create_loop_tree_nodes): Initialize it.
26481 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
26482
26483 PR middle-end/56461
26484 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
26485 method on dr_chain and result_chain.
26486 * tree-vect-stmts.c (vectorizable_store): Only call
26487 result_chain.create if j == 0.
26488
26489 PR middle-end/56461
26490 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
26491 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
26492 before overwriting it.
26493
26494 2013-03-01 Tobias Burnus <burnus@net-b.de>
26495
26496 * doc/extended.texi (C Extensions): Change order in @menu
26497 to match @node.
26498 (Other MIPS Built-in Functions): Move last MIPS entry before
26499 "picoChip Built-in Functions".
26500 (SH Built-in Functions): Move after RX Built-in Functions.
26501 * doc/gcc.texi (Introduction): Change order in @menu
26502 to match @node.
26503 * doc/md.texi (Constraints): Ditto.
26504 * gty.texi (Type Information): Ditto.
26505 (User-provided marking routines for template types): Make
26506 subsection.
26507 * doc/invoke.texi (AArch64 Options): Move before
26508 "Adapteva Epiphany Options".
26509
26510 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
26511 Jakub Jelinek <jakub@redhat.com>
26512
26513 PR sanitizer/56454
26514 * asan.c (gate_asan): Lookup no_sanitize_address instead of
26515 no_address_safety_analysis attribute.
26516 * doc/extend.texi (no_address_safety_attribute): Rename to
26517 no_sanitize_address attribute, mention no_address_safety_analysis
26518 attribute as deprecated alias.
26519
26520 2013-02-28 Jakub Jelinek <jakub@redhat.com>
26521
26522 PR middle-end/56461
26523 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
26524 type to vec<vec<tree> > *.
26525 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
26526 to be vec<tree> instead of vec<tree> *, set vec_defs
26527 to vNULL and call vec_defs.create (number_of_vects), adjust other
26528 uses of vec_defs.
26529 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
26530 vectorizable_condition): Adjust vect_get_slp_defs callers.
26531
26532 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26533
26534 * config/aarch64/aarch64.c
26535 (aarch64_float_const_representable): Remove unused variable.
26536
26537 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26538
26539 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
26540
26541 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26542
26543 * config/aarch64/aarch64-builtins.c
26544 (aarch64_init_simd_builtins): Make static.
26545
26546 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26547
26548 * config/aarch64/aarch64.c
26549 (aarch64_simd_make_constant): Make static.
26550
26551 2013-02-28 Martin Jambor <mjambor@suse.cz>
26552
26553 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
26554 with no initialization to the RHS of debug statements.
26555
26556 2013-02-28 Martin Jambor <mjambor@suse.cz>
26557
26558 PR tree-optimization/56294
26559 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
26560 Adjust dumping.
26561 (get_access_replacement): Do not call create_access_replacement.
26562 Assert a replacement exists.
26563 (get_repl_default_def_ssa_name): Create the replacement declaration
26564 itself.
26565
26566 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26567
26568 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
26569 final_end_function.
26570
26571 2013-02-28 Marek Polacek <polacek@redhat.com>
26572
26573 PR rtl-optimization/56466
26574 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
26575 if we're changing a loop.
26576 (peel_loops_completely): Likewise.
26577
26578 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
26579
26580 PR c++/55813
26581 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
26582
26583 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26584
26585 PR target/56445
26586 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
26587 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
26588 INTX_FTYPE_FX, FX_FTYPE_INTX.
26589 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
26590
26591 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26592
26593 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
26594 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
26595 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
26596 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
26597 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
26598 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
26599 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
26600 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
26601 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
26602 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
26603 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
26604 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
26605 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
26606 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
26607 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
26608 (avrxmega6): Increase max flash segments from 5 to 6.
26609 * config/avr/t-multilib: Regenerate.
26610 * config/avr/avr-tables.opt: Regenerate.
26611 * doc/avr-mmcu.texi: Regenerate.
26612
26613 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26614
26615 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
26616 (avr_device_to_arch): Rename to avr_device_to_ld.
26617 (avr_device_to_as): New prototype.
26618 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
26619 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
26620 * config/avr/driver-avr.c (avr_device_to_as): New.
26621 (avr_device_to_arch): Rename to avr_device_to_ld.
26622
26623 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26624
26625 PR middle-end/56461
26626 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
26627 method on dr_chain and result_chain.
26628
26629 PR middle-end/56461
26630 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
26631 pointer_set_destroy on not_executed_last_iteration.
26632
26633 PR middle-end/56461
26634 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
26635
26636 PR middle-end/56461
26637 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
26638 FOR_EACH_DEFINED_FUNCTION when freeing state.
26639
26640 PR middle-end/56461
26641 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
26642 pool_free.
26643 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
26644 overwriting it.
26645
26646 PR middle-end/56461
26647 * ipa-cp.c (decide_whether_version_node): Call vec_free on
26648 known_aggs[i].items and release known_aggs vector.
26649
26650 PR middle-end/56461
26651 * ipa-reference.c (propagate): Free node_info even for alias nodes.
26652
26653 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26654
26655 * config/microblaze/microblaze.c (microblaze_emit_compare):
26656 Use xor for EQ/NE comparisions.
26657 * config/microblaze/microblaze.md (cstoresf4): Add constraints
26658 (cbranchsf4): Adjust operator to comparison_operator.
26659
26660 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26661
26662 PR middle-end/56461
26663 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
26664 vector.
26665 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
26666 vec_safe_push, always update *slot.
26667 (redirect_edge_var_map_clear): Use vec_free.
26668 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
26669 (free_var_map_entry): Use vec_free.
26670 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
26671 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
26672
26673 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
26674
26675 PR middle-end/45472
26676 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
26677 when the may_trap_p bit of the exprs being merged differs.
26678 Reorder tests for speculativeness in the logical and operator.
26679
26680 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26681
26682 * incpath.c (add_standard_paths): Use reconcat instead of concat
26683 where appropriate and avoid leaking memory.
26684
26685 * opts.h: Include obstack.h.
26686 (opts_concat): New prototype.
26687 (opts_obstack): New declaration.
26688 * opts.c (opts_concat): New function.
26689 (opts_obstack): New variable.
26690 (init_options_struct): Call gcc_init_obstack on opts_obstack.
26691 (finish_options): Use opts_concat instead of concat
26692 and XOBNEWVEC instead of XNEWVEC.
26693 * opts-common.c (generate_canonical_option, decode_cmdline_option,
26694 generate_option): Likewise.
26695 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
26696 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
26697
26698 PR target/56455
26699 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
26700 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
26701
26702 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26703
26704 PR middle-end/56461
26705 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
26706
26707 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26708
26709 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
26710 (arm_block_move_unaligned_straight): Likewise.
26711 (arm_adjust_block_mem): Likewise.
26712
26713 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26714
26715 PR target/48901
26716 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
26717 temp, cond and label.
26718 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
26719
26720 PR target/52500
26721 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
26722 * config/c6x/c6x.h (dbx_register_map): Update declaration.
26723
26724 PR target/52501
26725 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
26726 of prologue/epilogue functions.
26727
26728 PR target/52550
26729 * config/tilegx/tilegx.c (tilegx_expand_prologue):
26730 Remove unused variable cfa_offset.
26731 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
26732
26733 PR target/54639
26734 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
26735 type promotion to unsigned.
26736
26737 PR target/54640
26738 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
26739 for HOST_WIDE_INT of 32 bit / same size as int.
26740 (arm_block_move_unaligned_straight): Likewise.
26741 (arm_adjust_block_mem): Likewise.
26742
26743 PR target/54662
26744 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
26745 ALL_CFLAGS.
26746
26747 2013-02-26 Marek Polacek <polacek@redhat.com>
26748
26749 PR tree-optimization/56426
26750 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
26751
26752 2013-02-26 Richard Biener <rguenther@suse.de>
26753
26754 PR target/56444
26755 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
26756 unused variable loops.
26757
26758 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26759
26760 PR tree-optimization/56448
26761 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
26762 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
26763 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
26764 later operands of the references, or even first operand for
26765 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
26766
26767 PR tree-optimization/56443
26768 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
26769 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
26770 to type_for_mode langhook.
26771
26772 2013-02-25 Matt Turner <mattst88@gmail.com>
26773
26774 * doc/invoke.texi: Document r4700.
26775
26776 2013-02-25 Richard Biener <rguenther@suse.de>
26777
26778 PR tree-optimization/56175
26779 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
26780 split out from ...
26781 (simplify_bitwise_binary): ... here. Also guard the conversion
26782 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
26783
26784 2013-02-25 Catherine Moore <clm@codesourcery.com>
26785
26786 Revert:
26787 2013-02-24 Catherine Moore <clm@codesourcery.com>
26788 Maciej W. Rozycki <macro@codesourcery.com>
26789 Tom de Vries <tom@codesourcery.com>
26790 Nathan Sidwell <nathan@codesourcery.com>
26791 Iain Sandoe <iain@codesourcery.com>
26792 Nathan Froyd <froydnj@codesourcery.com>
26793 Chao-ying Fu <fu@mips.com>
26794
26795 * doc/extend.texi: (micromips, nomicromips, nocompression):
26796 Document new function attributes.
26797 * doc/invoke.texi (minterlink-compressed, mmicromips,
26798 m14k, m14ke, m14kec): Document new options.
26799 (minterlink-mips16): Update documentation.
26800 * doc/md.texi (ZC, ZD): Document new constraints.
26801 * configure.ac (gcc_cv_as_micromips): Check if linker
26802 supports the .set micromips directive.
26803 * configure: Regenerate.
26804 * config.in: Regenerate.
26805 * config/mips/mips-tables.opt: Regenerate.
26806 * config/mips/micromips.md: New file.
26807 * constraints.md (ZC, AD): New constraints.
26808 * config/mips/predicates.md (movep_src_register): New predicate.
26809 (movep_src_operand): New predicate.
26810 (non_volatile_mem_operand): New predicate.
26811 * config/mips/mips.md (multimem): New type.
26812 (length): Differentiate between 17-bit and 18-bit branch offsets.
26813 (MOVEP1, MOVEP2): New mode iterator.
26814 (mov_<load>l): Use ZC constraint.
26815 (mov_<load>r): Likewise.
26816 (mov_<store>l): Likewise.
26817 (mov_<store>r): Likewise.
26818 (*branch_equality<mode>_inverted): Add microMIPS support.
26819 (*branch_equality<mode>): Likewise.
26820 (*jump_absolute): Likewise.
26821 (indirect_jump_<mode>): Likewise.
26822 (tablejump_<mode>): Likewise.
26823 (<optab>_internal): Likewise.
26824 (sibcall_internal): Likewise.
26825 (sibcall_value_internal): Likewise.
26826 (prefetch): Use constraint ZD.
26827 * config/mips/mips.opt (minterlink-compressed): New option.
26828 (minterlink-mips16): Now an alias for minterlink-compressed.
26829 (mmicromips): New option.
26830 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26831 (compare_and_swap_12): Likewise.
26832 (sync_add<mode>): Likewise.
26833 (sync_<optab>_12): Likewise.
26834 (sync_old_<optab>_12): Likewise.
26835 (sync_new_<optab>_12): Likewise.
26836 (sync_nand_12): Likewise.
26837 (sync_old_nand_12): Likewise.
26838 (sync_new_nand_12): Likewise.
26839 (sync_sub<mode>): Likewise.
26840 (sync_old_add<mode>): Likewise.
26841 (sync_old_sub<mode>): Likewise.
26842 (sync_new_add<mode>): Likewise.
26843 (sync_new_sub<mode>): Likewise.
26844 (sync_<optab><mode>): Likewise.
26845 (sync_old_<optab><mode>): Likewise.
26846 (sync_new_<optab><mode>): Likewise.
26847 (sync_nand<mode>): Likewise.
26848 (sync_old_nand<mode>): Likewise.
26849 (sync_new_nand<mode>): Likewise.
26850 (sync_lock_test_and_set<mode>): Likewise.
26851 (test_and_set_12): Likewise.
26852 (atomic_compare_and_swap<mode>): Likewise.
26853 (atomic_exchange<mode>_llsc): Likewise.
26854 (atomic_fetch_add<mode>_llsc): Likewise.
26855 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26856 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26857 (umips_save_restore_pattern_p): Likewise.
26858 (umips_load_store_pair_p): Likewise.
26859 (umips_output_load_store_pair): Likewise.
26860 (umips_movep_target_p): Likewise.
26861 (umips_12bit_offset_address_p): Likewise.
26862 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26863 (mips_base_mips16): Rename this...
26864 (mips_base_compression_flags): ...to this. Update all uses.
26865 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26866 (mips_mips16_decl_p): Delete.
26867 (mips_nomips16_decl_p): Delete.
26868 (mips_get_compress_on_flags): New function.
26869 (mips_get_compress_off_flags): New function.
26870 (mips_get_compress_mode): New function.
26871 (mips_get_compress_on_name): New function.
26872 (mips_get_compress_off_name): New function.
26873 (mips_insert_attributes): Support multiple compression types.
26874 (mips_merge_decl_attributes): Likewise.
26875 (umips_12bit_offset_address_p): New function.
26876 (mips_start_function_definition): Emit .set micromips directive.
26877 (mips_call_may_need_jalx_p): New function.
26878 (mips_function_ok_for_sibcall): Add microMIPS support.
26879 (mips_print_operand_punctuation): Support short delay slots and
26880 compact jumps.
26881 (umips_swm_mask, umips_swm_encoding): New.
26882 (umips_build_save_restore): New function.
26883 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26884 (was_mips16_p): Remove.
26885 (old_compression_mode): New.
26886 (mips_set_compression_mode): New function.
26887 (mips_set_current_function): Add microMIPS support.
26888 (mips_option_override): Likewise.
26889 (umips_save_restore_pattern_p): New function.
26890 (umips_output_save_restore): New function.
26891 (umips_load_store_pair_p_1): New function.
26892 (umips_load_store_pair_p): New function.
26893 (umips_output_load_store_pair_1): New function.
26894 (umips_output_load_store_pair): New function.
26895 (umips_movep_target_p) New function.
26896 (mips_prepare_pch_save): Add microMIPS support.
26897 * config/mips/mips.h (TARGET_COMPRESSION): New.
26898 (TARGET_CPU_CPP_BUILTINS): Update macro
26899 to use new compression flags and to support microMIPS.
26900 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26901 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26902 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26903 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26904 (ASM_SPEC): Support mmicromips and mno-micromips.
26905 (M16STORE_REG_P): New macro.
26906 (MIPS_CALL): Support TARGET_MICROMIPS.
26907 (MICROMIPS_J): New macro.
26908 (mips_base_mips16): Rename this...
26909 (mips_base_compression_flags): ...to this.
26910 (UMIPS_12BIT_OFFSET_P): New macro.
26911 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26912 (MULTILIB_DIRNAMES): Likewise.
26913
26914 2013-02-25 Tom de Vries <tom@codesourcery.com>
26915
26916 PR rtl-optimization/56131
26917 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
26918 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
26919 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
26920
26921 2013-02-25 Tobias Burnus <burnus@net-b.de>
26922
26923 * doc/invoke.texi (-fsanitize=): Move from optimization
26924 to debugging options.
26925
26926 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26927
26928 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
26929
26930 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26931 Alexander Monakov <amonakov@ispras.ru>
26932
26933 PR middle-end/56077
26934 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
26935 flush pending lists also on non-jumps. Adjust comment.
26936
26937 2013-02-24 Catherine Moore <clm@codesourcery.com>
26938 Maciej W. Rozycki <macro@codesourcery.com>
26939 Tom de Vries <tom@codesourcery.com>
26940 Nathan Sidwell <nathan@codesourcery.com>
26941 Iain Sandoe <iain@codesourcery.com>
26942 Nathan Froyd <froydnj@codesourcery.com>
26943 Chao-ying Fu <fu@mips.com>
26944
26945 * doc/extend.texi: (micromips, nomicromips, nocompression):
26946 Document new function attributes.
26947 * doc/invoke.texi (minterlink-compressed, mmicromips,
26948 m14k, m14ke, m14kec): Document new options.
26949 (minterlink-mips16): Update documentation.
26950 * doc/md.texi (ZC, ZD): Document new constraints.
26951 * configure.ac (gcc_cv_as_micromips): Check if linker
26952 supports the .set micromips directive.
26953 * configure: Regenerate.
26954 * config.in: Regenerate.
26955 * config/mips/mips-tables.opt: Regenerate.
26956 * config/mips/micromips.md: New file.
26957 * constraints.md (ZC, AD): New constraints.
26958 * config/mips/predicates.md (movep_src_register): New predicate.
26959 (movep_src_operand): New predicate.
26960 (non_volatile_mem_operand): New predicate.
26961 * config/mips/mips.md (multimem): New type.
26962 (length): Differentiate between 17-bit and 18-bit branch offsets.
26963 (MOVEP1, MOVEP2): New mode iterator.
26964 (mov_<load>l): Use ZC constraint.
26965 (mov_<load>r): Likewise.
26966 (mov_<store>l): Likewise.
26967 (mov_<store>r): Likewise.
26968 (*branch_equality<mode>_inverted): Add microMIPS support.
26969 (*branch_equality<mode>): Likewise.
26970 (*jump_absolute): Likewise.
26971 (indirect_jump_<mode>): Likewise.
26972 (tablejump_<mode>): Likewise.
26973 (<optab>_internal): Likewise.
26974 (sibcall_internal): Likewise.
26975 (sibcall_value_internal): Likewise.
26976 (prefetch): Use constraint ZD.
26977 * config/mips/mips.opt (minterlink-compressed): New option.
26978 (minterlink-mips16): Now an alias for minterlink-compressed.
26979 (mmicromips): New option.
26980 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26981 (compare_and_swap_12): Likewise.
26982 (sync_add<mode>): Likewise.
26983 (sync_<optab>_12): Likewise.
26984 (sync_old_<optab>_12): Likewise.
26985 (sync_new_<optab>_12): Likewise.
26986 (sync_nand_12): Likewise.
26987 (sync_old_nand_12): Likewise.
26988 (sync_new_nand_12): Likewise.
26989 (sync_sub<mode>): Likewise.
26990 (sync_old_add<mode>): Likewise.
26991 (sync_old_sub<mode>): Likewise.
26992 (sync_new_add<mode>): Likewise.
26993 (sync_new_sub<mode>): Likewise.
26994 (sync_<optab><mode>): Likewise.
26995 (sync_old_<optab><mode>): Likewise.
26996 (sync_new_<optab><mode>): Likewise.
26997 (sync_nand<mode>): Likewise.
26998 (sync_old_nand<mode>): Likewise.
26999 (sync_new_nand<mode>): Likewise.
27000 (sync_lock_test_and_set<mode>): Likewise.
27001 (test_and_set_12): Likewise.
27002 (atomic_compare_and_swap<mode>): Likewise.
27003 (atomic_exchange<mode>_llsc): Likewise.
27004 (atomic_fetch_add<mode>_llsc): Likewise.
27005 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27006 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27007 (umips_save_restore_pattern_p): Likewise.
27008 (umips_load_store_pair_p): Likewise.
27009 (umips_output_load_store_pair): Likewise.
27010 (umips_movep_target_p): Likewise.
27011 (umips_12bit_offset_address_p): Likewise.
27012 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27013 (mips_base_mips16): Rename this...
27014 (mips_base_compression_flags): ...to this. Update all uses.
27015 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27016 (mips_mips16_decl_p): Delete.
27017 (mips_nomips16_decl_p): Delete.
27018 (mips_get_compress_on_flags): New function.
27019 (mips_get_compress_off_flags): New function.
27020 (mips_get_compress_mode): New function.
27021 (mips_get_compress_on_name): New function.
27022 (mips_get_compress_off_name): New function.
27023 (mips_insert_attributes): Support multiple compression types.
27024 (mips_merge_decl_attributes): Likewise.
27025 (umips_12bit_offset_address_p): New function.
27026 (mips_start_function_definition): Emit .set micromips directive.
27027 (mips_call_may_need_jalx_p): New function.
27028 (mips_function_ok_for_sibcall): Add microMIPS support.
27029 (mips_print_operand_punctuation): Support short delay slots and
27030 compact jumps.
27031 (umips_swm_mask, umips_swm_encoding): New.
27032 (umips_build_save_restore): New function.
27033 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27034 (was_mips16_p): Remove.
27035 (old_compression_mode): New.
27036 (mips_set_compression_mode): New function.
27037 (mips_set_current_function): Add microMIPS support.
27038 (mips_option_override): Likewise.
27039 (umips_save_restore_pattern_p): New function.
27040 (umips_output_save_restore): New function.
27041 (umips_load_store_pair_p_1): New function.
27042 (umips_load_store_pair_p): New function.
27043 (umips_output_load_store_pair_1): New function.
27044 (umips_output_load_store_pair): New function.
27045 (umips_movep_target_p) New function.
27046 (mips_prepare_pch_save): Add microMIPS support.
27047 * config/mips/mips.h (TARGET_COMPRESSION): New.
27048 (TARGET_CPU_CPP_BUILTINS): Update macro
27049 to use new compression flags and to support microMIPS.
27050 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27051 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27052 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27053 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27054 (ASM_SPEC): Support mmicromips and mno-micromips.
27055 (M16STORE_REG_P): New macro.
27056 (MIPS_CALL): Support TARGET_MICROMIPS.
27057 (MICROMIPS_J): New macro.
27058 (mips_base_mips16): Rename this...
27059 (mips_base_compression_flags): ...to this.
27060 (UMIPS_12BIT_OFFSET_P): New macro.
27061 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27062 (MULTILIB_DIRNAMES): Likewise.
27063
27064 2013-02-24 Jakub Jelinek <jakub@redhat.com>
27065
27066 PR target/52555
27067 * target-globals.c (save_target_globals): For init_reg_sets and
27068 target_reinit remporarily set this_fn_optabs to this_target_optabs.
27069
27070 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
27071
27072 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
27073 * config/aarch64/t-aarch64
27074 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
27075
27076 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
27077
27078 PR inline-asm/56148
27079 * lra-constraints.c (process_alt_operands): Reload operand
27080 conflicting with earlier clobber only if no more other conflicting
27081 operands.
27082
27083 2013-02-22 Jakub Jelinek <jakub@redhat.com>
27084
27085 PR sanitizer/56393
27086 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
27087 if not linking a shared library.
27088
27089 2013-02-22 Seth LaForge <sethml@google.com>
27090
27091 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
27092
27093 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
27094
27095 * config/arm/arm.md (split for extendsidi): Update condition.
27096 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
27097 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
27098 (qhs_zextenddi_cstr): Likewise.
27099
27100 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27101
27102 PR middle-end/56420
27103 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
27104 avoid signed wrapping.
27105 (expand_mult): Handle properly multiplication by
27106 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
27107 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
27108 in the compiler if coeff is HOST_WIDE_INT_MIN.
27109 (expand_divmod): Don't make ext_op1 static, change it's type to
27110 uhwi. Avoid undefined behavior in -INTVAL (op1).
27111
27112 PR rtl-optimization/50339
27113 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
27114 field.
27115 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
27116 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
27117 into splitting_ashiftrt field.
27118 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
27119 ASHIFTRT.
27120 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
27121 choices.
27122
27123 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
27124
27125 PR middle-end/56108
27126 * trans-mem.c (execute_tm_mark): Do not expand transactions that
27127 are sure to go irrevocable.
27128
27129 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
27130
27131 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
27132 scalars are valid operands.
27133
27134 2013-02-21 Martin Jambor <mjambor@suse.cz>
27135
27136 PR tree-optimization/56310
27137 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
27138 only matching indices and non-negative final offsets.
27139 (intersect_aggregates_with_edge): Pass src_idx to
27140 agg_replacements_to_vector. Pass src_idx insstead of index to
27141 intersect_with_agg_replacements.
27142
27143 2013-02-21 Martin Jambor <mjambor@suse.cz>
27144
27145 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
27146 instead of hard-wired defaults.
27147
27148 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
27149
27150 * doc/invoke.texi (MIPS Options): Update documentation of the
27151 floating-point multiply-accumulate instruction restrictions.
27152
27153 2013-02-21 Kostya Serebryany <kcc@google.com>
27154
27155 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
27156 asan_shadow_offset on x86_64 linux.
27157
27158 2013-02-21 Richard Biener <rguenther@suse.de>
27159
27160 PR tree-optimization/56415
27161 Revert
27162 2013-02-11 Richard Biener <rguenther@suse.de>
27163
27164 PR tree-optimization/56273
27165 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27166 first VRP run.
27167
27168 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27169
27170 PR bootstrap/56258
27171 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
27172 instead of @itemx.
27173
27174 PR inline-asm/56405
27175 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
27176 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
27177
27178 2013-02-20 Jan Hubicka <jh@suse.cz>
27179
27180 PR tree-optimization/56265
27181 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
27182 when target is referenced for first time.
27183
27184 2013-02-20 Richard Biener <rguenther@suse.de>
27185
27186 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
27187 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
27188 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
27189 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
27190 not return anything.
27191 (rename_ssa_copies): Do not remove unused locals.
27192 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
27193 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
27194 * passes.c (execute_function_todo): Do not schedule unused locals
27195 removal if cleanup_tree_cfg did something.
27196 * tree-ssa-live.c (remove_unused_locals): Dump statistics
27197 about the number of removed locals.
27198
27199 2013-02-20 Richard Biener <rguenther@suse.de>
27200
27201 PR tree-optimization/56398
27202 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
27203
27204 2013-02-20 Martin Jambor <mjambor@suse.cz>
27205
27206 PR tree-optimization/55334
27207 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
27208 restricted pointers to arrays.
27209
27210 2013-02-20 Richard Biener <rguenther@suse.de>
27211 Jakub Jelinek <jakub@redhat.com>
27212
27213 PR tree-optimization/56396
27214 * tree-ssa-ccp.c (n_const_val): New static variable.
27215 (get_value): Return NULL for SSA names we don't have a lattice
27216 entry for.
27217 (ccp_initialize): Initialize n_const_val.
27218 * tree-ssa-copy.c (n_copy_of): New static variable.
27219 (init_copy_prop): Initialize n_copy_of.
27220 (get_value): Return NULL_TREE for SSA names we don't have a
27221 lattice entry for.
27222
27223 2013-02-20 Martin Jambor <mjambor@suse.cz>
27224
27225 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
27226
27227 2013-02-20 Richard Biener <rguenther@suse.de>
27228
27229 * genpreds.c (write_lookup_constraint): Do not compare first
27230 letter of the constraint again.
27231
27232 2013-02-20 Richard Biener <rguenther@suse.de>
27233
27234 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
27235 and ceil_log2.
27236 (get_use_iv_cost): Terminate hashtable walk when coming across
27237 an empty entry.
27238
27239 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
27240
27241 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
27242 reassociation for avx2 targets.
27243
27244 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27245
27246 * config/microblaze/microblaze.c: microblaze_has_clz = 0
27247 Add version check for v8.10.a to enable microblaze_has_clz
27248 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
27249 version and TARGET_PATTERN_COMPARE check
27250 * config/microblaze/microblaze.md: New clzsi2 instruction
27251
27252 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27253
27254 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
27255 function before branching.
27256
27257 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27258
27259 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
27260 DUMP_INSN_RTX_UID.
27261 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
27262
27263 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27264
27265 PR middle-end/55889
27266 * sel-sched.c: Include ira.h.
27267 (implicit_clobber_conflict_p): New function.
27268 (moveup_expr): Use it.
27269 * Makefile.in (sel-sched.o): Depend on ira.h.
27270
27271 2013-02-19 Richard Biener <rguenther@suse.de>
27272
27273 PR tree-optimization/56384
27274 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
27275 (vn_hash_type): Split out from ...
27276 (vn_hash_constant_with_type): ... here.
27277 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
27278 (vn_phi_eq): Compare types from vn_phi_s structure.
27279 (vn_phi_lookup): Populate vn_phi_s type.
27280 (vn_phi_insert): Likewise.
27281
27282 2013-02-19 Jakub Jelinek <jakub@redhat.com>
27283
27284 PR tree-optimization/56350
27285 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
27286 if haven't found reduction or nested cycle operand, rather than
27287 asserting we must find it.
27288
27289 PR tree-optimization/56381
27290 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
27291 to fold_build3.
27292
27293 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
27294 Jakub Jelinek <jakub@redhat.com>
27295
27296 PR target/52555
27297 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
27298 (swap_optab_enable): Same.
27299 (init_all_optabs): Use argument instead of global.
27300 * tree.h (struct tree_optimization_option): New field target_optabs.
27301 * expr.h (init_all_optabs): Add argument to prototype.
27302 (TREE_OPTIMIZATION_OPTABS): New.
27303 (save_optabs_if_changed): Protoize.
27304 * optabs.h: Declare this_fn_optabs.
27305 * optabs.c (save_optabs_if_changed): New.
27306 Declare this_fn_optabs.
27307 (init_optabs): Add argument to init_all_optabs() call.
27308 * function.c (invoke_set_current_function_hook): Handle per
27309 function optabs.
27310 * function.h (struct function): New field optabs.
27311 * config/mips/mips.c (mips_set_mips16_mode): Handle when
27312 optimization_current_node has changed.
27313 * target-globals.h (save_target_globals_default_opts): Protoize.
27314 * target-globals.c (save_target_globals_default_opts): New.
27315
27316 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27317
27318 PR target/56347
27319 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
27320 registers %fr12 and %fr12R as call used.
27321
27322 PR target/56214
27323 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
27324 and HImode, require all displacements to be an integer multiple of
27325 their mode size.
27326 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
27327 only allow QImode and HImode when reload is in progress and strict is
27328 true. Likewise for symbolic addresses. Use base14_operand to check
27329 displacements in REG+BASE addresses.
27330
27331 2013-02-18 Richard Biener <rguenther@suse.de>
27332
27333 PR tree-optimization/56366
27334 * tree-vect-loop.c (get_initial_def_for_induction): Properly
27335 handle sign-conversion of outer-loop initial induction value.
27336
27337 2013-02-18 Richard Biener <rguenther@suse.de>
27338
27339 PR middle-end/56349
27340 * cfghooks.c (merge_blocks): If we merge a latch into another
27341 block adjust references to it.
27342 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
27343 (verify_loop_structure): Verify that a recorded latch is in fact
27344 a latch.
27345
27346 2013-02-18 Richard Biener <rguenther@suse.de>
27347
27348 PR tree-optimization/56321
27349 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
27350 order SSA name release and virtual operand unlinking.
27351
27352 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27353
27354 * config/microblaze/microblaze.md (save_stack_block): Define.
27355 (restore_stack_block): Likewise.
27356
27357 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27358
27359 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
27360 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
27361 * config/microblaze/microblaze.c (microblaze_option_override):
27362 Bail out early for PIC modes when target does not support PIC.
27363
27364 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27365
27366 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
27367 Replace with a microblaze version.
27368 (microblaze_trampoline_init): Adapt for microblaze.
27369 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
27370 microblaze.
27371
27372 2013-02-16 Jakub Jelinek <jakub@redhat.com>
27373 Dodji Seketeli <dodji@redhat.com>
27374
27375 PR asan/56330
27376 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
27377 (instrument_mem_region_access): Do not forget to always put
27378 instrumentation of the of 'base' and 'base + len' in a "if (len !=
27379 0) statement, even for cases where either 'base' or 'base + len'
27380 are not instrumented -- because they have been previously
27381 instrumented. Simplify the logic by putting all the statements
27382 instrument 'base + len' inside a sequence, and then insert that
27383 sequence right before the current insertion point. Then, to
27384 instrument 'base + len', just get an iterator on that statement.
27385 And do not forget to update the pointer to iterator the function
27386 received as argument.
27387
27388 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
27389
27390 PR rtl-optimization/56348
27391 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
27392
27393 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
27394
27395 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
27396 (clean_graph_dump_file): Pass base to start_graph_dump.
27397
27398 2013-02-14 Richard Henderson <rth@redhat.com>
27399
27400 PR target/55941
27401 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
27402
27403 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
27404
27405 * collect2-aix.h: Define F_LOADONLY.
27406
27407 2013-02-14 Richard Biener <rguenther@suse.de>
27408
27409 PR lto/50494
27410 * varasm.c (output_constant_def_1): Get the decl representing
27411 the constant as argument.
27412 (output_constant_def): Wrap output_constant_def_1.
27413 (make_decl_rtl): Use output_constant_def_1 with the decl
27414 representing the constant.
27415 (build_constant_desc): Optionally re-use a decl already
27416 representing the constant.
27417 (tree_output_constant_def): Adjust.
27418
27419 2013-02-14 Dodji Seketeli <dodji@redhat.com>
27420
27421 Fix an asan crash
27422 * asan.c (instrument_builtin_call): Really put the length of the
27423 second source argument into src1_len.
27424
27425 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27426
27427 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
27428 argument. If it is false, don't create edge from then_bb to
27429 fallthru_bb.
27430 (insert_if_then_before_iter): Pass true to it.
27431 (build_check_stmt): Pass false to it.
27432 (transform_statements): Flush hash table only on extended basic
27433 block boundaries, rather than at the beginning of every bb.
27434 Don't flush hash table on nonfreeing_call_p calls.
27435 * tree-flow.h (nonfreeing_call_p): New prototype.
27436 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
27437
27438 2013-02-13 David S. Miller <davem@davemloft.net>
27439
27440 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
27441
27442 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
27443
27444 PR target/56184
27445 * ira.c (max_regno_before_ira): Move from ...
27446 (ira): ... here.
27447 (fix_reg_equiv_init): Use max_regno_before_ira instead of
27448 vec_safe_length.
27449
27450 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27451
27452 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
27453
27454 2013-02-13 Richard Biener <rguenther@suse.de>
27455
27456 PR lto/56295
27457 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
27458 globals in MEM_REFs.
27459
27460 2013-02-13 Richard Biener <rguenther@suse.de>
27461
27462 * loop-init.c (loop_optimizer_init): Clear loop state when
27463 re-initializing preserved loops.
27464 * loop-unswitch.c (unswitch_single_loop): Return whether
27465 we unswitched the loop. Do not verify loop state here.
27466 (unswitch_loops): When we unswitched a loop discover new loops.
27467
27468 2013-02-13 Kostya Serebryany <kcc@google.com>
27469
27470 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
27471 on x86_64 linux.
27472 * sanitizer.def: Rename __asan_init to __asan_init_v1.
27473
27474 2013-02-12 Dodji Seketeli <dodji@redhat.com>
27475
27476 Avoid instrumenting duplicated memory access in the same basic block
27477 * Makefile.in (asan.o): Add new dependency on hash-table.h
27478 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
27479 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
27480 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
27481 (free_mem_ref_resources, has_mem_ref_been_instrumented)
27482 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
27483 (get_mem_ref_of_assignment): New functions.
27484 (get_mem_refs_of_builtin_call): Extract from
27485 instrument_builtin_call and tweak a little bit to make it fit with
27486 the new signature.
27487 (instrument_builtin_call): Use the new
27488 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
27489 of is_gimple_builtin_call.
27490 (instrument_derefs, instrument_mem_region_access): Insert the
27491 instrumented memory reference into the hash table.
27492 (maybe_instrument_assignment): Renamed instrument_assignment into
27493 this, and change it to advance the iterator when instrumentation
27494 actually happened and return true in that case. This makes it
27495 homogeneous with maybe_instrument_assignment, and thus give a
27496 chance to callers to be more 'regular'.
27497 (transform_statements): Clear the memory reference hash table
27498 whenever we enter a new BB, when we cross a function call, or when
27499 we are done transforming statements. Use
27500 maybe_instrument_assignment instead of instrumentation. No more
27501 need to special case maybe_instrument_assignment and advance the
27502 iterator after calling it; it's now handled just like
27503 maybe_instrument_call. Update comment.
27504
27505 2013-02-13 Richard Biener <rguenther@suse.de>
27506
27507 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
27508 Fix loop discovery code.
27509
27510 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
27511
27512 PR inline-asm/56148
27513 * lra-constraints.c (process_alt_operands): Match early clobber
27514 operand with itself. Check conflicts with earlyclobber only if
27515 the operand is not reloaded. Prefer to reload conflicting operand
27516 if earlyclobber and matching operands are the same.
27517
27518 2013-02-12 Richard Biener <rguenther@suse.de>
27519
27520 PR lto/56297
27521 * lto-streamer-out.c (write_symbol): Do not output symbols
27522 for hard register variables.
27523
27524 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
27525
27526 PR target/54222
27527 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
27528 (umulsidi3_insn, mulsidi3_insn): New insns.
27529
27530 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
27531
27532 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
27533 (struct tune_params): Add vec_costs field.
27534 * config/arm/arm.c (arm_builtin_vectorization_cost)
27535 (arm_add_stmt_cost): New functions.
27536 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
27537 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
27538 (arm_default_vec_cost): New struct of type cpu_vec_costs.
27539 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
27540 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
27541 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
27542 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
27543
27544 2013-02-12 Richard Biener <rguenther@suse.de>
27545
27546 PR lto/56295
27547 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
27548 decls again if possible.
27549
27550 2013-02-12 Richard Biener <rguenther@suse.de>
27551
27552 PR middle-end/56288
27553 * tree-ssa.c (verify_ssa_name): Fix check, move
27554 SSA_NAME_IN_FREE_LIST check up.
27555
27556 2013-02-12 Jakub Jelinek <jakub@redhat.com>
27557 Steven Bosscher <steven@gcc.gnu.org>
27558
27559 PR rtl-optimization/56151
27560 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
27561 equal to op0 or op1, and last_insn pattern is CODE operation
27562 with MEM dest and one of the operands matches that MEM.
27563
27564 2013-02-11 Sriraman Tallam <tmsriram@google.com>
27565
27566 * doc/extend.texi: Document Function Multiversioning and "default"
27567 parameter string to target attribute.
27568 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
27569 target attribute parameter is "default".
27570 (ix86_compare_version_priority): Remove checks for target attribute.
27571 (ix86_mangle_function_version_assembler_name): Change error to sorry.
27572 Remove check for target attribute equal to NULL. Add assert.
27573 (ix86_generate_version_dispatcher_body): Change error to sorry.
27574
27575 2013-02-11 Iain Sandoe <iain@codesourcery.com>
27576 Jack Howarth <howarth@bromo.med.uc.edu>
27577 Patrick Marlier <patrick.marlier@gmail.com>
27578
27579 PR libitm/55693
27580 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
27581 define ENDFILE_SPEC as TM_DESTRUCTOR.
27582 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
27583
27584 2013-02-11 Alexander Potapenko <glider@google.com>
27585 Jack Howarth <howarth@bromo.med.uc.edu>
27586 Jakub Jelinek <jakub@redhat.com>
27587
27588 PR sanitizer/55617
27589 * config/darwin.c (cdtor_record): Rename ctor_record.
27590 (sort_cdtor_records): Rename sort_ctor_records.
27591 (finalize_dtors): New routine to sort destructors by
27592 priority before use in assemble_integer.
27593 (machopic_asm_out_destructor): Use finalize_dtors if needed.
27594
27595 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
27596
27597 PR rtl-optimization/56275
27598 * simplify-rtx.c (avoid_constant_pool_reference): Check that
27599 offset is non-negative and less than cmode size before
27600 calling simplify_subreg.
27601
27602 2013-02-11 Richard Biener <rguenther@suse.de>
27603
27604 PR tree-optimization/56264
27605 * cfgloop.h (fix_loop_structure): Adjust prototype.
27606 * loop-init.c (fix_loop_structure): Return the number of
27607 newly discovered loops.
27608 * tree-cfgcleanup.c (repair_loop_structures): When new loops
27609 are discovered, do a full loop-closed SSA rewrite.
27610
27611 2013-02-11 Richard Biener <rguenther@suse.de>
27612
27613 PR tree-optimization/56273
27614 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27615 first VRP run.
27616 (check_array_ref): Fix missing newline in dumps.
27617 (search_for_addr_array): Likewise.
27618
27619 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
27620
27621 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
27622
27623 2013-02-09 Jakub Jelinek <jakub@redhat.com>
27624
27625 PR target/56256
27626 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
27627
27628 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
27629
27630 PR rtl-optimization/56246
27631 * lra-constraints.c (simplify_operand_subreg): Try to reuse
27632 reload pseudo.
27633 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
27634 constraints are satisfied.
27635
27636 2013-02-08 Jeff Law <law@redhat.com>
27637
27638 PR debug/53948
27639 * emit-rtl.c (reg_is_parm_p): New function.
27640 * regs.h (reg_is_parm_p): New prototype.
27641 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
27642 callee-clobbered registers.
27643
27644 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
27645
27646 PR target/56043
27647 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
27648 If there is no implicit builtin declaration, just return NULL.
27649
27650 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27651
27652 * config/i386/sse.md (FMAMODEM): New mode iterator.
27653 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
27654 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
27655
27656 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27657
27658 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
27659 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
27660 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
27661
27662 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27663
27664 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
27665 (microblaze*-*-elf): Likewise.
27666 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
27667 LINK_SPEC.
27668 * config/microblaze/microblaze-c.c: Add builtin defines for
27669 _LITTLE_ENDIAN and _BIG_ENDIAN.
27670 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
27671 add to TARGET_DEFAULT flags.
27672 Expand ASM_SPEC and LINK_SPEC.
27673 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
27674 * config/microblaze/microblaze.md: Update extendsidi2 and
27675 movdi_internal instructions to use low-order / high-order reg
27676 print_operands.
27677 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
27678 options and inversemask / mask of LITTLE_ENDIAN.
27679 * config/microblaze/t-microblaze: Expand multilib options to
27680 include mlittle-endian (le) and update exceptions patterns.
27681
27682 2013-02-08 Jakub Jelinek <jakub@redhat.com>
27683
27684 PR rtl-optimization/56195
27685 * lra-constraints.c (get_reload_reg): Don't reuse regs
27686 if they have smaller mode than requested, if they have
27687 wider mode than requested, try to return a SUBREG.
27688
27689 PR tree-optimization/56250
27690 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
27691 if type is unsigned and code isn't MULT_EXPR.
27692
27693 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27694
27695 PR tree-optimization/56064
27696 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
27697 bits according to mode.
27698 * fixed-value.h (fixed_from_double_int)
27699 (const_fixed_from_double_int): Adjust comments.
27700
27701 2013-02-08 Richard Biener <rguenther@suse.de>
27702
27703 PR lto/56231
27704 * lto-streamer.h (struct data_in): Remove current_file, current_line
27705 and current_col members.
27706 * lto-streamer-out.c (lto_output_location): Stream changed bits
27707 en-block for efficiency.
27708 * lto-streamer-in.c (clear_line_info): Remove.
27709 (lto_input_location): Cache current file, line and column
27710 globally via local statics. Read changed bits en-block.
27711 (input_function): Do not call clear_line_info.
27712 (lto_read_body): Likewise.
27713 (lto_input_toplevel_asms): Likewise.
27714
27715 2013-02-08 Michael Matz <matz@suse.de>
27716
27717 PR tree-optimization/52448
27718 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
27719 (nt_call_phase): New static.
27720 (add_or_mark_expr): Only mark accesses with newer phase than any
27721 call seen.
27722 (nonfreeing_call_p): New.
27723 (nt_init_block): Update nt_call_phase, mark blocks as visited.
27724 (nt_fini_block): Keep blocks marked as visited.
27725 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
27726
27727 2013-02-08 Richard Biener <rguenther@suse.de>
27728
27729 * ira.c (ira): Free broken dominator information.
27730
27731 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27732
27733 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
27734
27735 2013-02-08 Marek Polacek <polacek@redhat.com>
27736
27737 * cfgloop.c (verify_loop_structure): Add more checking of headers.
27738
27739 2013-02-08 Richard Biener <rguenther@suse.de>
27740
27741 PR middle-end/56181
27742 * cfgloop.h (flow_loops_find): Adjust.
27743 (bb_loop_header_p): Declare.
27744 * cfgloop.c (bb_loop_header_p): New function split out from ...
27745 (flow_loops_find): ... here. Adjust function signature,
27746 support incremental loop structure update.
27747 (verify_loop_structure): Cleanup. Verify a loop is a loop.
27748 * cfgloopmanip.c (fix_loop_structure): Move ...
27749 * loop-init.c (fix_loop_structure): ... here.
27750 (apply_loop_flags): Split out from ...
27751 (loop_optimizer_init): ... here.
27752 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
27753 in incremental mode, only remove dead loops here.
27754
27755 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27756
27757 PR target/54222
27758 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
27759 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
27760 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
27761 (*round<mode>3.libgcc): New insns for fixed-modes.
27762 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
27763 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
27764 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
27765 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
27766 implementations. Define to __builtin_avr_absFX,
27767 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
27768 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
27769 __builtin_avr_countlsFX, respectively.
27770 * config/avr/avr-c.c (target.h): Include it.
27771 (enum avr_builtin_id): New enum.
27772 (avr_resolve_overloaded_builtin): New static function.
27773 (avr_register_target_pragmas): Use it to set
27774 targetm.resolve_overloaded_builtin.
27775 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
27776 tree nodes used by DEF_BUILTIN.
27777 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
27778 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
27779 <AVR_BUILTIN_xxBITS>: Same.
27780
27781 2013-02-08 Richard Biener <rguenther@suse.de>
27782
27783 * cfgloop.c (verify_loop_structure): Properly handle
27784 a loop exiting to another loop header.
27785 * ira-int.h (ira_loops): Remove.
27786 * ira.c (ira_loops): Remove.
27787 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
27788 (do_reload): Use loop_optimizer_finalize.
27789 * ira-build.c (create_loop_tree_nodes): Use get_loops and
27790 number_of_loops to access the loop tree.
27791 (more_one_region_p): Likewise.
27792 (finish_loop_tree_nodes): Likewise.
27793 (rebuild_regno_allocno_maps): Likewise.
27794 (mark_loops_for_removal): Likewise.
27795 (mark_all_loops_for_removal): Likewise.
27796 (remove_unnecessary_regions): Likewise.
27797 (ira_build): Likewise.
27798 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
27799
27800 2013-02-08 Richard Biener <rguenther@suse.de>
27801
27802 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
27803 * ipa-pure-const.c (analyze_function): Avoid calling
27804 mark_irreducible_loops twice.
27805 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
27806
27807 2013-02-07 David S. Miller <davem@davemloft.net>
27808
27809 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
27810 on 'reg'.
27811 * var-tracking.c (vt_add_function_parameter): Test the presence of
27812 HAVE_window_save properly and do not remap argument registers when
27813 we have a leaf function.
27814
27815 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
27816
27817 PR bootstrap/56227
27818 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
27819 instead of "ll".
27820 * config/i386/i386.c (ix86_print_operand): Ditto.
27821
27822 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27823
27824 * lra-constraints.c (process_alt_operands): Fix recently added comment.
27825
27826 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27827
27828 PR rtl-optimization/56225
27829 * lra-constraints.c (process_alt_operands): Check that reload hard
27830 reg can hold value for strict_low_part.
27831
27832 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27833
27834 PR debug/56154
27835 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
27836 dwarf2out_end_function.
27837 (in_first_function_p, maybe_at_text_label_p,
27838 first_loclabel_num_not_at_text_label): New variables.
27839 (dwarf2out_var_location): In the first function find out
27840 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
27841 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
27842 functions.
27843
27844 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
27845
27846 PR rtl-optimization/56178
27847 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
27848 SUBREG of a register. Tidy up related block of code.
27849 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
27850 note if the source is a register or a SUBREG of a register.
27851
27852 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27853
27854 PR target/56228
27855 * config/rs6000/rs6000.md (ptrm): New mode attr.
27856 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
27857 call_value_indirect_aix<pttrsize>,
27858 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
27859 m in constraints.
27860
27861 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
27862
27863 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
27864 if -bnortl. Convert to strcmp and strncmp.
27865
27866 2013-02-07 Alan Modra <amodra@gmail.com>
27867
27868 PR target/54009
27869 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
27870 addresses won't wrap when offsetting.
27871 (rs6000_secondary_reload): Provide secondary reloads needed for
27872 wrapping LO_SUM addresses.
27873
27874 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
27875
27876 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
27877 MACH, just __MACH__.
27878
27879 2013-02-06 Richard Biener <rguenther@suse.de>
27880
27881 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
27882 instead of calling fix_loop_structure.
27883
27884 2013-02-06 Jakub Jelinek <jakub@redhat.com>
27885
27886 PR middle-end/56217
27887 * omp-low.c (use_pointer_for_field): Return false if
27888 lower_send_shared_vars doesn't generate any copy-out code.
27889
27890 2013-02-06 Tom de Vries <tom@codesourcery.com>
27891
27892 PR rtl-optimization/56131
27893 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
27894 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
27895 of the label is NULL. Add comment.
27896
27897 2013-02-05 Jakub Jelinek <jakub@redhat.com>
27898
27899 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
27900
27901 PR sanitizer/55374
27902 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
27903 (STATIC_LIBTSAN_LIBS): Likewise.
27904 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
27905 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
27906 is defined, don't add anything else beyond that.
27907 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
27908 (LINK_COMMAND_SPEC): Use them.
27909
27910 PR tree-optimization/56205
27911 * tree-stdarg.c (check_all_va_list_escapes): Return true if
27912 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
27913 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
27914
27915 2013-02-05 Richard Biener <rguenther@suse.de>
27916
27917 PR tree-optimization/53342
27918 PR tree-optimization/53185
27919 * tree-vectorizer.h (vect_check_strided_load): Remove.
27920 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
27921 not disallow peeling for vectorized strided loads.
27922 (vect_check_strided_load): Make static and simplify.
27923 (vect_analyze_data_refs): Adjust.
27924 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
27925 correctly when vectorizing strided loads.
27926
27927 2013-02-05 Richard Biener <rguenther@suse.de>
27928
27929 * doc/install.texi: Refer to ISL, not PPL.
27930
27931 2013-02-05 Jan Hubicka <jh@suse.cz>
27932
27933 PR tree-optimization/55789
27934 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
27935
27936 2013-02-05 Jan Hubicka <jh@suse.cz>
27937
27938 PR tree-optimization/55789
27939 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
27940 the dead call anyway.
27941
27942 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
27943
27944 PR sanitizer/55374
27945 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
27946
27947 2013-02-04 Alexander Potapenko <glider@google.com>
27948 Jack Howarth <howarth@bromo.med.uc.edu>
27949 Jakub Jelinek <jakub@redhat.com>
27950
27951 PR sanitizer/55617
27952 * config/darwin.c (sort_ctor_records): Stabilized qsort
27953 on constructor priority by using original position.
27954 (finalize_ctors): New routine to sort constructors by
27955 priority before use in assemble_integer.
27956 (machopic_asm_out_constructor): Use finalize_ctors if needed.
27957
27958 2013-02-04 Jakub Jelinek <jakub@redhat.com>
27959
27960 PR libstdc++/54314
27961 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
27962 about visibility on artificial decls.
27963 * config/sol2.c (solaris_assemble_visibility): Likewise.
27964
27965 2013-02-04 Kai Tietz <ktietz@redhat.com>
27966
27967 PR target/56186
27968 * config/i386/i386.c (function_value_ms_64): Add additional valtype
27969 argument and improve checking of return-argument types for 16-byte
27970 modes.
27971 (ix86_function_value_1): Add additional valtype argument on call
27972 of function_value_64.
27973 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
27974 handling infunction_value_64 function.
27975
27976 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
27977
27978 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
27979
27980 2013-02-04 Richard Biener <rguenther@suse.de>
27981
27982 PR tree-optimization/56188
27983 * tree-ssa-structalias.c (label_visit): Consider case with
27984 initially non-empty points-to set.
27985 (perform_var_substitution): Dump node mapping and clean up.
27986
27987 2013-02-04 Richard Guenther <rguenther@suse.de>
27988
27989 PR lto/56168
27990 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
27991 node prevail as last resort.
27992 (lto_symtab_merge_decls): Remove guard on LTRANS here.
27993 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
27994
27995 2013-02-04 Richard Biener <rguenther@suse.de>
27996
27997 PR tree-optimization/56113
27998 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
27999 Merge into ...
28000 (equiv_class_lookup_or_add): ... this.
28001 (label_visit): Adjust and fix error in previous patch.
28002 (perform_var_substitution): Adjust.
28003
28004 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
28005
28006 * config/sh/divtab.c: Fix formatting and comments throughout the file.
28007 * config/sh/sh4-300.md: Likewise.
28008 * config/sh/sh4a.md: Likewise.
28009 * config/sh/constraints.md: Likewise.
28010 * config/sh/sh.md: Likewise.
28011 * config/sh/netbsd-elf.h: Likewise.
28012 * config/sh/predicates.md: Likewise.
28013 * config/sh/sh-protos.h: Likewise.
28014 * config/sh/ushmedia.h: Likewise.
28015 * config/sh/linux.h: Likewise.
28016 * config/sh/sh.c: Likewise.
28017 * config/sh/superh.h: Likewise.
28018 * config/sh/elf.h: Likewise.
28019 * config/sh/sh4.md: Likewise.
28020 * config/sh/sh.h: Likewise.
28021
28022 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28023
28024 * config/pa/constraints.md: Adjust unused letters. Change "T"
28025 constraint to match_test floating_point_store_memory_operand().
28026 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
28027 (base14_operand): New.
28028 (floating_point_store_memory_operand): New.
28029 (integer_store_memory_operand): Revise to use base14_operand and
28030 reg_plus_base_memory_operand.
28031 (move_dest_operand): Allow symbolic_memory_operands.
28032 (symbolic_memory_operand): Check for LO_SOM.
28033 (symbolic_operand): Change default case to break.
28034 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
28035 CONST_DOUBLE values to be reloaded by putting them into memory when
28036 the destination is a floating point register.
28037 (movdf): Remove code to handle CONST_DOUBLE.
28038 (movsf): Likewise.
28039 (reload_indf_r1): New.
28040 (reload_insf_r1): New.
28041 Consistently use "Q" and "T" constraints with integer and floating
28042 point move instructions, respectively.
28043 (movdi): Remove FAIL.
28044 Change predicate for source operand unamed DImode move from
28045 general_operand to move_src_operand.
28046 (umulsidi3): Change predicate for destination operand to
28047 register_operand.
28048 Likewise for similar unamed patterns.
28049 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
28050 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
28051 (hppa_legitimize_address): Simplify mask calculation.
28052 (pa_emit_move_sequence): Revised handling of secondary reloads from
28053 REG+D addresses for floating point loads and stores. Directly handle
28054 loading CONST0_RTX (mode) to a floating point register.
28055 (pa_secondary_reload): Handle reloading DF and SFmode constant values
28056 to floating point registers. Don't restrict secondary reloads to
28057 floating point registers to integer modes. Revise some comments and
28058 cleanup some code.
28059 (TARGET_LEGITIMATE_ADDRESS_P): Define.
28060 (pa_legitimate_address_p): New.
28061 (pa_legitimize_reload_address): New.
28062 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
28063 (STRICT_REG_OK_FOR_BASE_P): New.
28064 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
28065 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
28066
28067 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
28068 Andrew Dixie <andrewd@gentrack.com>
28069
28070 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
28071 flag set.
28072
28073 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
28074
28075 * expmed.c (extract_bit_field_1): Pass the full width of the
28076 structure to get_best_reg_extraction_insn.
28077
28078 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28079
28080 PR target/54601
28081 * configure.ac (use_cxa_atexit): Add AIX.
28082 * configure: Regenerate.
28083
28084 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
28085
28086 2013-02-01 Jakub Jelinek <jakub@redhat.com>
28087
28088 PR debug/54793
28089 * final.c (need_profile_function): New variable.
28090 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
28091 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
28092 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
28093 notes, targetm.asm_out.function_prologue doesn't emit anything,
28094 HAVE_prologue and profiler should be emitted before prologue,
28095 set need_profile_function instead of emitting it.
28096 (final_scan_insn): If need_profile_function, emit
28097 profile_function on the first NOTE_INSN_BASIC_BLOCK or
28098 NOTE_INSN_FUNCTION_BEG note.
28099
28100 2013-02-01 Richard Henderson <rth@redhat.com>
28101
28102 * config/rs6000/rs6000.md (smulditi3): New.
28103 (umulditi3): New.
28104
28105 * config/alpha/alpha.md (umulditi3): New.
28106
28107 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28108
28109 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
28110 (ASM_OUTPUT_ALIGNED_LOCAL): New.
28111
28112 2013-02-01 Richard Biener <rguenther@suse.de>
28113
28114 PR tree-optimization/56113
28115 * tree-ssa-structalias.c (label_visit): Reduce work for
28116 single-predecessor nodes.
28117
28118 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
28119
28120 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
28121 range isn't testing for zero.
28122
28123 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
28124
28125 PR middle-end/56113
28126 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
28127
28128 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
28129 Nick Clifton <nickc@redhat.com>
28130
28131 * config/v850/constraints.md (Q): Define as a memory constraint.
28132 * config/v850/predicates.md (label_ref_operand): New predicate.
28133 (e3v5_shift_operand): New predicate.
28134 (ior_operator): New predicate.
28135 * config/v850/t-v850: Add e3v5 multilib.
28136 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
28137 (v850_gen_movdi): Prototype.
28138 * config/v850/v850.c: Add support for e3v5 architecture.
28139 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
28140 TARGET_V850E_UP.
28141 (construct_save_jarl): Add e3v5 long JARL support.
28142 (v850_adjust_insn_length): New function. Adjust length of call
28143 insns when using e3v5 instructions.
28144 (v850_gen_movdi): New function: Generate instructions to move a
28145 DImode value.
28146 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
28147 (CPP_SPEC): Define __v850e3v5__ as appropriate.
28148 (TARGET_USE_FPU): Enable for e3v5.
28149 (CONST_OK_FOR_W): New macro.
28150 (ADJUST_INSN_LENGTH): Define.
28151 * config/v850/v850.md (UNSPEC_LOOP): Define.
28152 (attr cpu): Add v850e3v5.
28153 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
28154 (movdi): New pattern.
28155 (movdi_internal): New pattern.
28156 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
28157 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
28158 (cstoresf4): Likewise.
28159 (cstoredf4): Likewise.
28160 (insv): New pattern.
28161 (rotlso3_a): New pattern.
28162 (rotlsi3_b): New pattern
28163 (rotlsi3_v850e3v5): New pattern.
28164 (doloop_begin): New pattern.
28165 (fix_loop_counter): New pattern.
28166 (doloop_end): New pattern.
28167 (branch_normal): Add e3v5 long branch support.
28168 (branch_invert): Likewise.
28169 (branch_z_normal): Likewise.
28170 (branch_z_invert): Likewise.
28171 (branch_nz_normal): Likewise.
28172 (branch_nz_invert): Likewise.
28173 (call_internal_short): Add e3v5 register-indirect JARL support.
28174 (call_internal_long): Likewise.
28175 (call_value_internal_short): Likewise.
28176 (call_value_internal_long): Likewise.
28177 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
28178 (mloop): New option.
28179 * config.gcc: Add support for configuring v840e3v5 target.
28180 * doc/invoke.texi: Document new v850 specific command line options.
28181
28182 2013-01-31 Paul Koning <ni1d@arrl.net>
28183
28184 PR debug/55059
28185 PR debug/54508
28186 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
28187 children if parent is a class.
28188 (prune_unused_types_prune): Don't add DW_AT_declaration.
28189
28190 2013-01-31 Richard Biener <rguenther@suse.de>
28191
28192 PR tree-optimization/56157
28193 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
28194 match up operand with SLP child.
28195
28196 2013-01-31 Jason Merrill <jason@redhat.com>
28197
28198 PR debug/54410
28199 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
28200 parameters the first time.
28201 (gen_scheduled_generic_parms_dies): Check completeness here.
28202
28203 2013-01-31 Richard Biener <rguenther@suse.de>
28204
28205 PR middle-end/53073
28206 * common.opt (faggressive-loop-optimizations): New flag,
28207 enabled by default.
28208 * doc/invoke.texi (faggressive-loop-optimizations): Document.
28209 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
28210 infer_loop_bounds_from_undefined by it.
28211
28212 2013-01-31 Richard Biener <rguenther@suse.de>
28213
28214 PR tree-optimization/56150
28215 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
28216 visit virtual operands.
28217 (find_uses_to_rename_bb): Likewise.
28218
28219 2013-01-31 Richard Biener <rguenther@suse.de>
28220
28221 PR tree-optimization/56150
28222 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
28223 mixed store non-store stmts.
28224
28225 2013-01-30 Jakub Jelinek <jakub@redhat.com>
28226
28227 PR sanitizer/55374
28228 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
28229 LIBASAN_EARLY_SPEC is defined.
28230 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
28231 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
28232 before %o.
28233 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
28234
28235 PR c++/55742
28236 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
28237 invalid args instead of ICEing on it.
28238 (ix86_valid_target_attribute_tree): Return error_mark_node if
28239 ix86_valid_target_attribute_inner_p failed.
28240 (ix86_valid_target_attribute_p): Return false only if
28241 ix86_valid_target_attribute_tree returned error_mark_node. Allow
28242 target("default") attribute.
28243 (sorted_attr_string): Change argument from const char * to tree,
28244 merge in all target attribute arguments rather than just one.
28245 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
28246 instead of free. Avoid using strcat.
28247 (ix86_mangle_function_version_assembler_name): Mangle
28248 target("default") as if no target attribute is present. Adjust
28249 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
28250 instead of xmalloc and XDELETEVEC instead of free.
28251 (ix86_function_versions): Don't return true if one of the decls
28252 doesn't have target attribute. If they don't and one of the decls
28253 is DECL_FUNCTION_VERSIONED, report an error. Adjust
28254 sorted_attr_string caller. Use XDELETEVEC instead of free.
28255 (ix86_supports_function_versions): Remove.
28256 (make_name): Fix up formatting.
28257 (make_dispatcher_decl): Remove resolver_name and its initialization.
28258 Avoid leaking memory.
28259 (is_function_default_version): Return true if there is
28260 target("default") attribute rather than no target attribute at all.
28261 (make_resolver_func): Avoid leaking memory.
28262 (ix86_generate_version_dispatcher_body): Likewise.
28263 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
28264 * target.def (supports_function_versions): Remove.
28265 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
28266 * doc/tm.texi: Regenerated.
28267
28268 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
28269
28270 PR rtl-optimization/56144
28271 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
28272 for values with side effects.
28273
28274 2013-01-30 Richard Biener <rguenther@suse.de>
28275
28276 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
28277 (sparseset_pop): Likewise.
28278 * cfganal.c (compute_idf): Likewise. Increase work-stack size
28279 to be able to use quick_push in the worker loop.
28280
28281 2013-01-30 Marek Polacek <polacek@redhat.com>
28282
28283 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
28284
28285 2013-01-30 Richard Biener <rguenther@suse.de>
28286
28287 PR lto/56147
28288 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
28289
28290 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
28291
28292 PR tree-optimization/56064
28293 * fixed-value.c (fixed_from_double_int): New function.
28294 * fixed-value.h (fixed_from_double_int): New prototype.
28295 (const_fixed_from_double_int): New static inline function.
28296 * fold-const.c (native_interpret_fixed): New static function.
28297 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
28298 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
28299 (native_encode_fixed): New static function.
28300 (native_encode_expr) <FIXED_CST>: Use it.
28301 (native_interpret_int): Move double_int worker code to...
28302 * double-int.c (double_int::from_buffer): ...this new static method.
28303 * double-int.h (double_int::from_buffer): Prototype it.
28304
28305 2013-01-30 Richard Biener <rguenther@suse.de>
28306
28307 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
28308 New pointer-map and obstack.
28309 (init_alias_vars): Allocate pointer-map and obstack.
28310 (delete_points_to_sets): Free them.
28311 (find_what_var_points_to): Cache result.
28312 (find_what_p_points_to): Adjust for changed interface of
28313 find_what_var_points_to.
28314 (compute_points_to_sets): Likewise.
28315 (ipa_pta_execute): Likewise.
28316
28317 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
28318
28319 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
28320 * configure: Regenerate.
28321 * config.in: Regenerate.
28322 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
28323 #nobits/#progbits if supported.
28324
28325 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
28326
28327 PR target/56121
28328 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
28329 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
28330 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
28331
28332 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28333
28334 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
28335 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
28336
28337 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28338
28339 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
28340 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
28341
28342 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28343
28344 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
28345 declaration.
28346 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
28347 * config/arm/cortex-a7.md: New bypasses using
28348 arm_mac_accumulator_is_result.
28349
28350 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28351
28352 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
28353 (cortex_a7_neon_mla): Likewise.
28354 (cortex_a7_fpfmad): New reservation.
28355 (cortex_a7_fpmacs): Use ffmas and update required units.
28356 (cortex_a7_fpmuld): Update required units and latency.
28357 (cortex_a7_fpmacd): Likewise.
28358 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
28359 (cortex_a7_neon). Likewise.
28360 (bypass) Update participating units.
28361
28362 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28363
28364 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
28365 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
28366 from fmac to ffma.
28367 * config/arm/vfp11.md (vfp_farith): Use ffmas.
28368 (vfp_fmul): Use ffmad.
28369 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
28370 (cortex_r4_fmacd): Use ffmad.
28371 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
28372 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
28373 (cortex_a9_fmacd): Use ffmad.
28374 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
28375 (cortex_a8_vfp_macd): Use ffmad.
28376 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
28377 (cortex_a5_fpmacd): Use ffmad.
28378 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
28379 (cortex_a15_vfp_macd): Use ffmad.
28380 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
28381
28382 2013-01-29 Jason Merrill <jason@redhat.com>
28383
28384 PR libstdc++/54314
28385 * varasm.c (default_assemble_visibility): Don't warn about
28386 visibility on artificial decls.
28387
28388 2013-01-29 Richard Biener <rguenther@suse.de>
28389
28390 PR tree-optimization/56113
28391 * tree-ssa-structalias.c (equiv_class_lookup): Also return
28392 the bitmap leader.
28393 (label_visit): Free duplicate bitmaps and record the leader instead.
28394 (perform_var_substitution): Adjust.
28395
28396 2013-01-29 Richard Biener <rguenther@suse.de>
28397
28398 PR tree-optimization/55270
28399 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
28400 the CFG, schedule loops for fixup.
28401
28402 2013-01-29 Nick Clifton <nickc@redhat.com>
28403
28404 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
28405 SP_REG.
28406
28407 2013-01-28 Leif Ekblad <leif@rdos.net>
28408
28409 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
28410 * config/i386/i386.h (TARGET_RDOS): New macro.
28411 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
28412 * config/i386/i386.c (ix86_option_override_internal): For 64bit
28413 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
28414 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
28415 DEFAULT_LARGE_SECTION_THRESHOLD.
28416 * config/i386/i386.md (R14_REG, R15_REG): New constants.
28417 * config/i386/rdos.h: New file.
28418 * config/i386/rdos64.h: New file.
28419
28420 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
28421
28422 PR other/54814
28423 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
28424 TEST_HARD_REG_BIT.
28425
28426 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28427
28428 PR rtl-optimization/56117
28429 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
28430 call cselib_lookup_from_insn on the MEM before calling
28431 add_insn_mem_dependence.
28432
28433 2013-01-28 Richard Biener <rguenther@suse.de>
28434
28435 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
28436 to a stmt that didn't have one.
28437 (copy_phis_for_bb): Likewise for PHI arguments.
28438 (copy_debug_stmt): Likewise for debug stmts.
28439
28440 2013-01-28 Richard Biener <rguenther@suse.de>
28441
28442 PR tree-optimization/56034
28443 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
28444 (partition_builtin_p): Adjust.
28445 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
28446 it is the last partition.
28447 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
28448 up the vertex for the definition.
28449 (classify_partition): Classify whether a partition is a
28450 PKIND_REDUCTION, thus has uses outside of the loop.
28451 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
28452 Merge all PKIND_REDUCTION partitions into the last partition.
28453 (tree_loop_distribution): Seed partitions from reductions as well.
28454
28455 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28456
28457 PR tree-optimization/56125
28458 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
28459 pow(x,c) into sqrt(x) * powi(x, n/2) or
28460 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
28461 optimizing for size.
28462 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
28463 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
28464 integer.
28465
28466 PR tree-optimization/56094
28467 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
28468 to UNKNOWN_LOCATION while gimplifying expr.
28469
28470 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
28471
28472 PR target/56114
28473 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
28474 operand 0 in movabs insn template for -masm=intel asm alternative.
28475 (*movabs<mode>_2): Ditto for operand 1.
28476
28477 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
28478
28479 PR target/54663
28480 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
28481 of microblaze-c.o
28482
28483 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28484
28485 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
28486 tm_file.
28487
28488 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
28489
28490 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
28491 Undef to avoid warning.
28492
28493 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28494
28495 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
28496 * configure: Regenerate.
28497
28498 2013-01-25 Jakub Jelinek <jakub@redhat.com>
28499
28500 PR tree-optimization/56098
28501 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
28502 for stmts with volatile ops.
28503 (cond_store_replacement): Don't optimize if assign has volatile ops.
28504 (cond_if_else_store_replacement_1): Don't optimize if either
28505 then_assign or else_assign have volatile ops.
28506 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
28507 volatile ops.
28508
28509 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28510
28511 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
28512
28513 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28514
28515 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
28516 missing ':' in asm example.
28517
28518 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
28519
28520 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
28521 entries into lane and laneq entries.
28522 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
28523 Remove AdvSIMD scalar modes.
28524 (aarch64_sq<r>dmulh_laneq<mode>): New.
28525 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
28526 modes.
28527 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
28528 builtin implementations to relfect changes in RTL in aarch64-simd.md.
28529 * config/aarch64/iterators.md (VCOND): New.
28530 (VCONQ): New.
28531
28532 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28533
28534 PR target/54222
28535 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
28536 Add NULL LIBNAME argument to existing definitions.
28537 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
28538 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
28539 * config/avr/avr.c (DEF_BUILTIN): Same.
28540 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
28541 (avr_expand_builtin): Expand to a vanilla call if a libgcc
28542 implementation is available (DECL_ASSEMBLER_NAME is set).
28543 (avr_fold_absfx): New static function.
28544 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
28545 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
28546 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
28547 AVR_BUILTIN_ABSLLK.
28548 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
28549 (abshk, absk, abslk, absllk): Provide as static inline functions.
28550
28551 2013-01-25 Marek Polacek <polacek@redhat.com>
28552
28553 PR tree-optimization/56035
28554 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
28555
28556 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
28557
28558 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
28559 (*movtf_internal_rex64): Add (!o,C) alternative
28560 (*movxf_internal_rex64): Ditto.
28561 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
28562
28563 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
28564
28565 * doc/invoke.texi: fix typo.
28566 * doc/objc.texi: fix typo.
28567
28568 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
28569
28570 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
28571 for the first two alternatives.
28572
28573 2013-01-24 Diego Novillo <dnovillo@google.com>
28574
28575 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
28576 (ggc-zone.o): Remove.
28577 * configure.ac: Remove option --with-gc.
28578 * configure: Re-generate.
28579 * doc/install.texi: Remove documentation for --with-gc.
28580 * gengtype.c (write_enum_defn): Remove. Update all users.
28581 (write_Types_process_field): Remove generation of gt_e_* argument.
28582 (output_type_enum): Remove. Update all users.
28583 (write_enum_defn): Remove. Update all users.
28584 (enum alloc_zone): Remove. Update all users.
28585 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
28586 * ggc-common.c (ggc_splay_alloc): Remove first argument.
28587 Update all callers.
28588 (struct ptr_data): Remove field TYPE. Update all users.
28589 (gt_pch_note_object): Remove argument TYPE. Update all users.
28590 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
28591 Update all users.
28592 * ggc-none.c (ggc_alloc_typed_stat): Remove.
28593 (struct alloc_zone): Remove.
28594 (ggc_internal_alloc_zone_stat): Remove.
28595 (ggc_internal_cleared_alloc_zone_stat): Remove.
28596 * ggc-page.c (ggc_alloc_typed_stat): Remove.
28597 (ggc_pch_count_object): Remove last argument. Update all users.
28598 (ggc_pch_alloc_object): Remove last argument. Update all users.
28599 (struct alloc_zone): Remove.
28600 * ggc-zone.c: Remove.
28601 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
28602 (struct alloc_zone): Remove.
28603 (ggc_alloc_typed_stat): Remove.
28604 (ggc_alloc_typed): Remove.
28605 (ggc_splay_alloc): Remove first argument.
28606 (rtl_zone): Remove. Update all users.
28607 (tree_zone): Remove. Update all users.
28608 (tree_id_zone): Remove. Update all users.
28609 (ggc_internal_zone_alloc_stat): Remove. Update all users.
28610 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
28611 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
28612 * tree-ssanames.c: Remove references to zone allocator in comments.
28613
28614 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
28615
28616 * config/avr/avr.c (avr_out_fract): Make register numbers that
28617 might be outside of source operand signed.
28618
28619 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
28620
28621 * config/i386/constraints.md (Yf): New constraint.
28622 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
28623 of f constraint to conditionaly disable x87 register preferences.
28624 (*movdf_internal): Ditto.
28625 (*movsf_internal): Ditto.
28626
28627 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
28628
28629 PR inline-asm/55934
28630 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
28631 that have operands with impossible constraints.
28632 Add a FIXME for a speed-up opportunity.
28633 * lra-constraints.c (process_alt_operands): Verify that a class
28634 selected from constraints on asms is valid for the operand mode.
28635 (curr_insn_transform): Remove incorrect comment.
28636
28637 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
28638
28639 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
28640 TOC operand is a valid symbol ref in the constant pool.
28641
28642 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28643
28644 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
28645
28646 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
28647
28648 PR target/54222
28649 * config/avr/stdfix.h: New file.
28650 * t-avr (stdfix-gcc.h): New rule to build it.
28651 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
28652
28653 2013-01-23 Kostya Serebryany <kcc@google.com>
28654
28655 * config/darwin.h: remove dependency on
28656 CoreFoundation (asan on Mac OS).
28657
28658 2013-01-23 Jakub Jelinek <jakub@redhat.com>
28659
28660 PR target/49069
28661 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
28662 instead of cmpdi_operand for first comparison operand.
28663 Don't assert that comparison operands aren't both constants.
28664
28665 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28666
28667 * doc/install.texi (Downloading the Source): Update references to
28668 downloading separate components.
28669
28670 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28671
28672 * doc/extend.texi (__int128): Improve grammar.
28673
28674 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
28675
28676 PR target/56028
28677 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
28678 alternative to (o,r).
28679 (*movdi_internal_rex64): Remove (!o,n) alternative.
28680 (DImode immediate->memory splitter): Remove.
28681 (DImode immediate->memory peephole2): Remove.
28682 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
28683 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
28684 alternative to (!o,*r).
28685 (*movtf_internal_sse): New pattern.
28686 (*movxf_internal_rex64): New pattern.
28687 (*movxf_internal): Disable for TARGET_64BIT.
28688 (*movdf_internal_rex64): Remove (!o,F) alternative.
28689
28690 2013-01-22 Jakub Jelinek <jakub@redhat.com>
28691
28692 PR middle-end/56074
28693 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
28694 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
28695 * tree-vect-loop-manip.c (find_loop_location): Also ignore
28696 stmt locations where LOCATION_LOCUS of the stmt location is
28697 UNKNOWN_LOCATION or BUILTINS_LOCATION.
28698
28699 PR target/55686
28700 * config/i386/i386.md (UNSPEC_STOS): New.
28701 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
28702 *strsetqi_1): Add UNSPEC_STOS.
28703
28704 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
28705
28706 PR c++/56067
28707 * doc/invoke.texi: Remove left over -Wsynth example.
28708
28709 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28710
28711 PR tree-optimization/56051
28712 * fold-const.c (fold_binary_loc): Don't fold
28713 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
28714 a narrowing conversion, or widening conversion from signed
28715 to unsigned.
28716
28717 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28718
28719 PR rtl-optimization/56023
28720 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
28721 dependent on debug instruction.
28722
28723 2013-01-21 Martin Jambor <mjambor@suse.cz>
28724
28725 PR middle-end/56022
28726 * function.c (allocate_struct_function): Call
28727 invoke_set_current_function_hook earlier.
28728
28729 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28730
28731 * reload1.c (init_reload): Only initialize reload_obstack
28732 during the first call.
28733
28734 2013-01-21 Marek Polacek <polacek@redhat.com>
28735
28736 * cfgloop.c (verify_loop_structure): Fix up grammar.
28737
28738 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
28739
28740 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
28741 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
28742
28743 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28744
28745 PR target/56058
28746 * config/arm/marvell-pj4.md: Update copyright year.
28747 Fix up use of alu to alu_reg and simple_alu_imm.
28748
28749 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28750
28751 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
28752
28753 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
28754
28755 PR target/55433
28756 * lra-constraints.c (curr_insn_transform): Don't reuse original
28757 insn for secondary memory move when memory mode should be different.
28758
28759 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28760
28761 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
28762 atomic_storedi_1): New patterns.
28763
28764 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
28765
28766 btver2 pipeline descriptions.
28767 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
28768 descriptions.
28769 * config/i386/i386.md (btver2_decode): New type attributes.
28770 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
28771 type attributes.
28772 * config/i386/btver2.md: New file describing btver2 pipelines.
28773
28774 2013-01-19 Andrew Pinski <apinski@cavium.com>
28775
28776 PR tree-optimization/52631
28777 * tree-ssa-sccvn (visit_use): Before looking up the original
28778 statement, try looking up the simplified expression.
28779
28780 2013-01-19 Anthony Green <green@moxielogic.com>
28781
28782 * config/moxie/moxie.c (moxie_expand_prologue): Set
28783 current_function_static_stack_size.
28784
28785 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28786
28787 PR tree-optimization/56029
28788 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
28789 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
28790
28791 2013-01-18 Sharad Singhai <singhai@google.com>
28792
28793 PR tree-optimization/55995
28794 * dumpfile.c (dump_loc): Print location only if available.
28795 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
28796
28797 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
28798
28799 PR target/55433
28800 * lra-constraints.c (curr_insn_transform): Reuse original insn for
28801 secondary memory move.
28802 (inherit_reload_reg): Use rclass instead of cl for
28803 check_secondary_memory_needed_p.
28804
28805 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28806
28807 PR middle-end/56015
28808 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
28809 the case where writing real complex part of target modifies op1.
28810
28811 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
28812
28813 * config/aarch64/aarch64-simd.md
28814 (aarch64_vcond_internal<mode>): Handle unordered cases.
28815 * config/aarch64/iterators.md (v_cmp_result): New.
28816
28817 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
28818 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28819
28820 * config/arm/marvell-pj4.md: New file.
28821 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
28822 * config/arm/arm.md (generic_sched): Add marvell_pj4.
28823 (generic_vfp): Likewise.
28824 * config/arm/arm-cores.def: Add marvell-pj4.
28825 * config/arm/arm-tune.md: Regenerate.
28826 * config/arm/arm-tables.opt: Regenerate.
28827 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
28828 * doc/invoke.texi: Document marvell-pj4.
28829
28830 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
28831
28832 * config/aarch64/arm_neon.h: Map scalar types to standard types.
28833
28834 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28835
28836 PR debug/54114
28837 PR debug/54402
28838 PR debug/49888
28839 * var-tracking.c (negative_power_of_two_p): New.
28840 (global_get_addr_cache, local_get_addr_cache): New.
28841 (get_addr_from_global_cache, get_addr_from_local_cache): New.
28842 (vt_canonicalize_addr): Rewrite using the above. Adjust the
28843 heading comment.
28844 (vt_stack_offset_p): Remove.
28845 (vt_canon_true_dep): Always canonicalize loc's address.
28846 (clobber_overlapping_mems): Make sure we have a MEM.
28847 (local_get_addr_clear_given_value): New.
28848 (val_reset): Clear local cached entries.
28849 (compute_bb_dataflow): Create and release the local cache.
28850 Disable duplicate MEMs clobbering.
28851 (emit_notes_in_bb): Clobber MEMs likewise.
28852 (vt_emit_notes): Create and release the local cache.
28853 (vt_initialize, vt_finalize): Create and release the global
28854 cache, respectively.
28855 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
28856
28857 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28858
28859 PR libmudflap/53359
28860 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
28861 not found in the symtab.
28862
28863 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28864
28865 PR debug/56006
28866 PR rtl-optimization/55547
28867 PR rtl-optimization/53827
28868 PR debug/53671
28869 PR debug/49888
28870 * alias.c (offset_overlap_p): New, factored out of...
28871 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
28872 the conservative special case for symbolic constants. Don't
28873 adjust zero sizes on alignment.
28874
28875 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
28876
28877 PR rtl-optimization/52573
28878 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
28879 REG_UNUSED for the same register.
28880
28881 2013-01-17 Richard Biener <rguenther@suse.de>
28882 Marek Polacek <polacek@redhat.com>
28883
28884 PR rtl-optimization/55833
28885 * loop-unswitch.c (unswitch_loops): Move loop verification...
28886 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
28887 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
28888 Set it to true when we're removing a loop from hierarchy tree in
28889 an irreducible region.
28890 (fix_bb_placements): Adjust caller.
28891 (fix_loop_placements): Likewise.
28892
28893 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
28894
28895 * config/avr/builtins.def (DEF_BUILTIN): Factor out
28896 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
28897 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
28898 Remove ID. Adjust comments.
28899 * config/avr/avr-c.c (avr_builtin_name): Remove.
28900 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
28901 * config/avr/avr.c (avr_tolower): New static function.
28902 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
28903 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
28904 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
28905 default expansion.
28906
28907 2013-01-17 Jan Hubicka <jh@suse.cz>
28908
28909 PR tree-optimization/55273
28910 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
28911
28912 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
28913
28914 PR target/55981
28915 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
28916 store through atomic_store<mode>_1.
28917 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
28918
28919 2013-01-17 Martin Jambor <mjambor@suse.cz>
28920
28921 PR tree-optimizations/55264
28922 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
28923 for virtual methods.
28924 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
28925 virtual methods before inlining is over.
28926 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
28927 virtual functions.
28928 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
28929 non-virtual.
28930
28931 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
28932
28933 PR rtl-optimization/56005
28934 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
28935 pending reads for prefetch.
28936
28937 2013-01-16 Ian Bolton <ian.bolton@arm.com>
28938
28939 * config/aarch64/aarch64.md
28940 (*cstoresi_neg_uxtw): New pattern.
28941 (*cmovsi_insn_uxtw): New pattern.
28942 (*<optab>si3_uxtw): New pattern.
28943 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
28944 (*<optab>si3_insn_uxtw): New pattern.
28945 (*bswapsi2_uxtw): New pattern.
28946
28947 2013-01-16 Richard Biener <rguenther@suse.de>
28948
28949 * tree-inline.c (tree_function_versioning): Remove set but
28950 never used variable.
28951
28952 2013-01-16 Richard Biener <rguenther@suse.de>
28953
28954 PR tree-optimization/55964
28955 * tree-flow.h (rename_variables_in_loop): Remove.
28956 (rename_variables_in_bb): Likewise.
28957 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
28958 (copy_loop_before): Adjust and delete update-ssa status.
28959 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
28960 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
28961 (rename_variables_in_loop): Remove.
28962 (slpeel_update_phis_for_duplicate_loop): Likewise.
28963 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
28964 use available cfg machinery instead of duplicating it.
28965 Update PHI nodes and perform poor-mans SSA update here.
28966 (slpeel_tree_peel_loop_to_edge): Adjust.
28967
28968 2013-01-16 Richard Biener <rguenther@suse.de>
28969
28970 PR tree-optimization/54767
28971 PR tree-optimization/53465
28972 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
28973 (vrp_visit_phi_node): For PHI arguments coming via backedges
28974 drop all symbolical range information.
28975 (execute_vrp): Compute backedges.
28976
28977 2013-01-16 Richard Biener <rguenther@suse.de>
28978
28979 * doc/install.texi: Update CLooG and ISL requirements to
28980 0.18.0 and 0.11.1.
28981
28982 2013-01-16 Christian Bruel <christian.bruel@st.com>
28983
28984 PR target/55301
28985 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
28986 (broken_move): Handle UNSPECV_SP_SWITCH_B.
28987 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
28988
28989 2013-01-16 DJ Delorie <dj@redhat.com>
28990
28991 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
28992 (UNSPECV_SP_SWITCH_E): New.
28993 (sp_switch_1): Change to an unspec.
28994 (sp_switch_2): Change to an unspec. Don't use post-inc when we
28995 replace $r15.
28996
28997 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
28998
28999 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
29000 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
29001 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
29002 (expand_mem_thread_fence): Ditto.
29003 (expand_mem_signal_fence): Ditto.
29004 (expand_atomic_load): Ditto.
29005 (expand_atomic_store): Ditto.
29006
29007 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
29008
29009 PR rtl-optimization/55547
29010 PR rtl-optimization/53827
29011 PR debug/53671
29012 PR debug/49888
29013 * alias.c (memrefs_conflict_p): Set sizes to negative after
29014 AND adjustments.
29015
29016 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29017
29018 PR target/55940
29019 * function.c (thread_prologue_and_epilogue_insns): Always
29020 add crtl->drap_reg to set_up_by_prologue.set, even if
29021 stack_realign_drap is false.
29022
29023 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
29024
29025 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
29026 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
29027 *call): Fix indention.
29028
29029 2013-01-15 Tom de Vries <tom@codesourcery.com>
29030
29031 PR target/55876
29032 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
29033 Update comment.
29034
29035 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
29036
29037 PR rtl-optimization/55153
29038 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
29039
29040 2013-01-15 Martin Jambor <mjambor@suse.cz>
29041
29042 PR tree-optimization/55920
29043 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
29044 accesses as grp_to_be_debug_replaced.
29045
29046 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29047
29048 PR tree-optimization/55920
29049 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
29050 there is non-useless type conversion needed from debug rhs to lhs,
29051 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
29052
29053 2013-01-15 Joseph Myers <joseph@codesourcery.com>
29054 Mikael Pettersson <mikpe@it.uu.se>
29055
29056 PR target/43961
29057 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
29058 Thumb.
29059 (ASM_OUTPUT_CASE_LABEL): Remove.
29060 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
29061 * final.c (shorten_branches): Update alignment of labels before
29062 jump tables if CASE_VECTOR_SHORTEN_MODE.
29063
29064 2013-01-15 Richard Biener <rguenther@suse.de>
29065
29066 PR bootstrap/55961
29067 * system.h: Do not include gmp.h for building host tools.
29068
29069 2013-01-15 Richard Biener <rguenther@suse.de>
29070
29071 PR middle-end/55882
29072 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
29073 account for bitpos when computing alignment.
29074
29075 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
29076
29077 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
29078 (ix86_target_macros_internal): Likewise.
29079
29080 * config/i386/i386.c (m_CORE2I7): Removed.
29081 (m_CORE_HASWELL): New macro.
29082 (m_CORE_ALL): Likewise.
29083 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
29084 (initial_ix86_arch_features): Likewise.
29085 (processor_target_table): Initializations for Core avx2.
29086 (cpu_names): New names "core-avx2".
29087 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
29088 PROCESSOR_CORE_HASWELL.
29089 (ix86_issue_rate): New case.
29090 (ia32_multipass_dfa_lookahead): Likewise.
29091 (ix86_sched_init_global): Likewise.
29092
29093 * config/i386/i386.h (TARGET_HASWELL): New macro.
29094 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
29095 (processor_type): New PROCESSOR_HASWELL.
29096
29097 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29098
29099 PR tree-optimization/55955
29100 * tree-vect-loop.c (vectorizable_reduction): Give up early on
29101 *SHIFT_EXPR and *ROTATE_EXPR codes.
29102
29103 PR tree-optimization/48766
29104 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
29105 -ftrapv disable -fwrapv.
29106
29107 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29108
29109 PR target/55974
29110 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
29111 etc. to 1 and not to __flash.
29112 Use LL suffix for __INT24_MAX__ with -mint8.
29113 Use ULL suffix for __UINT24_MAX__ with -mint8.
29114
29115 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29116
29117 * config/avr/avr-arch.h
29118 (struct base_arch_s): Use typedef avr_arch_t instead.
29119 (struct arch_info_s): Use typedef avr_arch_info_t instead.
29120 (struct mcu_type_s): Use typedef avr_mcu_t instead.
29121 * config/avr/avr.c: Same.
29122 * config/avr/avr-devices.c: Same.
29123 * config/avr/driver-avr.c: Same.
29124 * config/avr/gen-avr-mmcu-texi.c: Same.
29125 * config/avr/avr-mcus.def: Adjust comment.
29126
29127 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
29128
29129 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
29130 * config/aarch64/iterators.md (VALLDI): New.
29131
29132 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
29133 Andi Kleen <ak@linux.intel.com>
29134
29135 PR target/55948
29136 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
29137 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
29138 memmodel flag.
29139
29140 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29141
29142 * config/avr/avr-stdint.h: Remove trailing blanks.
29143 * config/avr/avr-log.h: Same.
29144 * config/avr/avr-arch.h: Same.
29145 * config/avr/avr-devices.c: Same.
29146 * config/avr/avr-dimode.md: Same.
29147 * config/avr/predicates.md: Same.
29148 * config/avr/avr-c.c: Same. And fix typo.
29149
29150 * config/avr/avr-protos.h: Same. And:
29151 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
29152 (init_cumulative_args): Rename to avr_init_cumulative_args.
29153 (expand_prologue): Rename to avr_expand_prologue.
29154 (expand_epilogue): Rename to avr_expand_epilogue.
29155 (adjust_insn_length): Rename to avr_adjust_insn_length.
29156 (notice_update_cc): Rename to avr_notice_update_cc.
29157 (final_prescan_insn): Rename to avr_final_prescan_insn.
29158 * config/avr/avr.c: Same.
29159 * config/avr/avr.h: Same.
29160 * config/avr/avr.md: Remove trailing blanks.
29161 (prologue): Use avr_expand_prologue.
29162 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
29163
29164 2013-01-14 Richard Biener <rguenther@suse.de>
29165
29166 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
29167 verify_location, collect_subblocks): New functions.
29168 (verify_gimple_in_cfg): Verify that locations only reference
29169 BLOCKs in the functions BLOCK tree.
29170
29171 2013-01-14 Richard Biener <rguenther@suse.de>
29172
29173 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
29174 PHI argument.
29175 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
29176 unshare reference.
29177 (insert_out_of_ssa_copy_on_edge): Likewise.
29178 (rewrite_close_phi_out_of_ssa): Likewise.
29179 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
29180 debug expressions.
29181 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
29182 propagated constants.
29183 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
29184 can not be shared.
29185
29186 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29187
29188 * config/avr/avr-modes.def: Add GPL copyright notice.
29189
29190 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
29191
29192 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
29193 MEMMODEL_MASK to determine memory model.
29194 (atomic_store<mode>): Ditto from operands[2].
29195 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
29196
29197 2013-01-13 Jakub Jelinek <jakub@redhat.com>
29198
29199 PR fortran/55935
29200 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
29201 (fold_gimple_assign): Don't call unshare_expr here.
29202 (fold_ctor_reference): Call unshare_expr.
29203
29204 2013-01-13 Terry Guo <terry.guo@arm.com>
29205
29206 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
29207 * doc/fragments.texi: Document MULTILIB_REUSE.
29208 * gcc.c (multilib_reuse): New internal spec.
29209 (set_multilib_dir): Also search multilib from multilib_reuse.
29210 * genmultilib (tmpmultilib3): Refactor code.
29211 (tmpmultilib4): Ditto.
29212 (multilib_reuse): New multilib argument.
29213
29214 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
29215
29216 * Makefile.in: Update copyright.
29217
29218 2013-01-12 Tom de Vries <tom@codesourcery.com>
29219
29220 PR middle-end/55890
29221 * calls.c (expand_call): Check if arg_nr is valid.
29222
29223 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
29224
29225 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
29226 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
29227 documentation. Add missing '__' in front of
29228 __builtin_ia32_packssdw256.
29229
29230 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29231
29232 PR target/55719
29233 * config/s390/s390.c (s390_preferred_reload_class): Do not return
29234 NO_REGS for larl operands.
29235 (s390_reload_larl_operand): Use s390_load_address instead of
29236 emit_move_insn.
29237
29238 2013-01-11 Richard Biener <rguenther@suse.de>
29239
29240 * tree-cfg.c (verify_node_sharing_1): Split out from ...
29241 (verify_node_sharing): ... here.
29242 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
29243
29244 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
29245
29246 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
29247 Substitute TREECHECKING.
29248 * configure: Regenerate.
29249 * Makefile.in (TREECHECKING): New.
29250
29251 2013-01-11 Richard Guenther <rguenther@suse.de>
29252
29253 PR tree-optimization/44061
29254 * tree-vrp.c (extract_range_basic): Compute zero as
29255 value-range for __builtin_constant_p of function parameters.
29256
29257 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
29258
29259 Update copyright years.
29260
29261 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
29262
29263 PR rtl-optimization/55672
29264 * lra-eliminations.c (mark_not_eliminable): Permit addition with
29265 const to be eliminable.
29266
29267 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
29268
29269 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
29270 * configure: Regenerate.
29271
29272 2013-01-10 Richard Biener <rguenther@suse.de>
29273
29274 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
29275
29276 2013-01-10 Richard Biener <rguenther@suse.de>
29277
29278 PR bootstrap/55792
29279 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
29280 locations for virtual PHI arguments.
29281 (rewrite_update_phi_arguments): Likewise.
29282
29283 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
29284
29285 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
29286 on to assembler.
29287
29288 2013-01-10 Jakub Jelinek <jakub@redhat.com>
29289
29290 PR tree-optimization/55921
29291 * tree-complex.c (expand_complex_asm): New function.
29292 (expand_complex_operations_1): Call it for GIMPLE_ASM.
29293
29294 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29295
29296 PR target/55718
29297 * config/s390/s390.c (s390_symref_operand_p)
29298 (s390_loadrelative_operand_p): Merge the two functions.
29299 (s390_check_qrst_address, print_operand_address): Add parameters
29300 to s390_loadrelative_operand_p invokation.
29301 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
29302 (s390_reload_larl_operand, s390_secondary_reload): Use
29303 s390_loadrelative_operand_p instead of s390_symref_operand_p.
29304 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
29305
29306 2013-01-09 Mike Stump <mikestump@comcast.net>
29307
29308 * dse.c (record_store): Remove unnecessary assert.
29309
29310 2013-01-09 Jan Hubicka <jh@suse.cz>
29311
29312 PR tree-optimization/55569
29313 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
29314 * cfgloop.h (scale_loop_profile): Likewise.
29315
29316 2013-01-09 Jan Hubicka <jh@suse.cz>
29317
29318 PR lto/45375
29319 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
29320 functions.
29321 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
29322
29323 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
29324
29325 PR middle-end/55114
29326 * expr.h (maybe_emit_group_store): Declare.
29327 * expr.c (maybe_emit_group_store): New function.
29328 * builtins.c (expand_builtin_int_roundingfn): Call it.
29329 (expand_builtin_int_roundingfn_2): Likewise.
29330
29331 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
29332
29333 PR rtl-optimization/55829
29334 * lra-constraints.c (match_reload): Add code for absent output.
29335 (curr_insn_transform): Add code for reloads of matched inputs
29336 without output.
29337
29338 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
29339
29340 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
29341 attribute of movddup insn to DF.
29342 (*vec_interleave_lowv2df): Ditto.
29343 (vec_dupv2df): Ditto.
29344
29345 2013-01-09 Jan Hubicka <jh@suse.cz>
29346
29347 PR tree-optimiation/55875
29348 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
29349 EVERY_ITERATION parameter.
29350 (number_of_iterations_exit): Check if exit is executed every iteration.
29351 (idx_infer_loop_bounds): Similarly here.
29352 (n_of_executions_at_most): Simplify
29353 to only test for cases where statement is dominated by the
29354 particular bound; handle correctly the "postdominance" test.
29355 (scev_probably_wraps_p): Use max loop iterations info
29356 as a global bound first.
29357
29358 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
29359 Nick Clifton <nickc@redhat.com>
29360
29361 * config/v850/v850.md (cbranchsf4): New pattern.
29362 (cstoresf4): New pattern.
29363 (cbranchdf4): New pattern.
29364 (cstoredf4): New pattern.
29365 (movsicc): Disallow floating point comparisons.
29366 (cmpsf_le_insn): Fix order of operators.
29367 (cmpsf_lt_insn): Likewise.
29368 (cmpsf_eq_insn): Likewise.
29369 (cmpdf_le_insn): Likewise.
29370 (cmpdf_lt_insn): Likewise.
29371 (cmpdf_eq_insn): Likewise.
29372 (cmpsf_ge_insn): Use LE comparison.
29373 (cmpdf_ge_insn): Likewise.
29374 (cmpsf_gt_insn): Use LT comparison.
29375 (cmpdf_gt_insn): Likewise.
29376 (cmpsf_ne_insn): Delete pattern.
29377 (cmpdf_ne_insn): Delete pattern.
29378 * config/v850/v850.c (v850_gen_float_compare): Use
29379 gen_cmpdf_eq_insn for NE comparison.
29380 (v850_float_z_comparison_operator)
29381 (v850_float_nz_comparison_operator): Move from here ...
29382 * config/v850/predicates.md: ... to here. Move GT and GE
29383 comparisons into v850_float_z_comparison_operator.
29384 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
29385 Delete prototype.
29386 (v850_float_nz_comparison_operator): Likewise.
29387
29388 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29389
29390 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
29391 with calls to gen_insvsi/gen_insvdi.
29392
29393 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29394
29395 * config/i386/i386.c (initial_ix86_tune_features): Set up
29396 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
29397
29398 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
29399 Jakub Jelinek <jakub@redhat.com>
29400
29401 PR tree-optimization/48189
29402 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
29403 If nitercst is 0, don't predict the exit edge.
29404
29405 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29406
29407 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
29408 in asm_fprintf with reg_names.
29409 (aarch64_print_operand_address): Likewise.
29410 (aarch64_return_addr): Likewise.
29411 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
29412
29413 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29414
29415 * config/pa/pa.h (VAL_U6_BITS_P): Define.
29416 (INT_U6_BITS): Likewise.
29417 * config/pa/predicates.md (uint6_operand): New predicate.
29418 (shift5_operand, shift6_operand): Likewise.
29419 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
29420 arith32_operand.
29421 (lshrdi3): Use shift6_operand.
29422 (shrpsi4, shrpdi4): New insn patterns.
29423 (extzv): Delete expander.
29424 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
29425 predicates in unamed zero extract patterns. Tighten common constraint.
29426 (extv): Delete expander.
29427 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
29428 predicates in unamed sign extract patterns. Tighten common constraint.
29429 (insv): Delete expander.
29430 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
29431 predicates in unamed insert patterns. Tighten common constraint.
29432 Change uint32_operand predicate to uint6_operand predicate in unamed
29433 DImode pattern to insert constant values of type 1...1xxxx.
29434
29435 2013-01-04 Jan Hubicka <jh@suse.cz>
29436
29437 PR tree-optimization/55823
29438 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
29439 issue.
29440
29441 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29442 Uros Bizjak <ubizjak@gmail.com>
29443
29444 PR rtl-optimization/55845
29445 * df-problems.c (can_move_insns_across): Stop scanning at
29446 volatile_insn_p source instruction or give up if
29447 across_from .. across_to range contains any volatile_insn_p
29448 instructions.
29449
29450 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29451
29452 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
29453 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
29454 Declare.
29455 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
29456 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
29457
29458 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29459
29460 PR fortran/55341
29461 * asan.c (asan_clear_shadow): New function.
29462 (asan_emit_stack_protection): Use it.
29463
29464 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29465
29466 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
29467 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
29468 with tab instead of space.
29469
29470 2013-01-08 Nick Clifton <nickc@redhat.com>
29471
29472 * config/rl78/rl78.c (rl78_expand_prologue): Always select
29473 register bank 0 at the start of an interrupt handler.
29474 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
29475 MDBH registers.
29476
29477 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29478
29479 * config/aarch64/aarch64-simd.md
29480 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
29481 (aarch64_simd_bsl): Likewise.
29482 (aarch64_vcond_internal<mode>): Likewise.
29483 (vcond<mode><mode>): Likewise.
29484 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
29485 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
29486
29487 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29488
29489 * config/aarch64/aarch64-builtins.c
29490 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
29491
29492 2013-01-08 Martin Jambor <mjambor@suse.cz>
29493
29494 PR debug/55579
29495 * tree-sra.c (analyze_access_subtree): Return true also after
29496 potentially creating a debug-only replacement.
29497
29498 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29499
29500 PR middle-end/55890
29501 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
29502
29503 PR tree-optimization/54120
29504 * tree-vrp.c (range_fits_type_p): Don't allow
29505 src_precision < precision from signed vr to unsigned_p
29506 if vr->min or vr->max is negative.
29507 (simplify_float_conversion_using_ranges): Test can_float_p
29508 against CODE_FOR_nothing.
29509
29510 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29511 Richard Biener <rguenther@suse.de>
29512
29513 PR middle-end/55851
29514 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
29515 types instead of just INTEGER_TYPE types.
29516
29517 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
29518
29519 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
29520 TF_SIZE): Define.
29521
29522 2013-01-07 Steve Ellcey <sellcey@mips.com>
29523
29524 PR target/42661
29525 * config/mips/mips.opt: Change mad to mmad to match documentation.
29526
29527 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29528
29529 PR target/55897
29530 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
29531 .progmemx.data now.
29532
29533 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29534
29535 PR target/55897
29536 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
29537 (avr_addrspace_t): Add .section_name field.
29538 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
29539 array size.
29540 (avr_addrspace): Same. Initialize .section_name. Remove last
29541 NULL entry. Put __memx into .progmemx.data.
29542 (progmem_section_prefix): Remove.
29543 (avr_asm_init_sections): No need to initialize progmem_section.
29544 (avr_asm_named_section): Use avr_addrspace[].section_name to get
29545 section name prefix.
29546 (avr_asm_select_section): Ditto. And use get_unnamed_section to
29547 retrieve the progmem section.
29548 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
29549 boundary to run over avr_addrspace[].
29550 (avr_register_target_pragmas): Ditto.
29551
29552 2013-01-06 Jakub Jelinek <jakub@redhat.com>
29553
29554 * varasm.c (output_constant_def_contents): For asan_protect_global
29555 protected strings, adjust DECL_ALIGN if needed, before testing for
29556 anchored symbols.
29557 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
29558 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
29559 normal decls.
29560 (output_object_block): For asan protected decls, emit asan padding
29561 after their contents.
29562 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
29563 (asan_finish_file): Test it here instead.
29564
29565 2013-01-07 Nick Clifton <nickc@redhat.com>
29566 Matthias Klose <doko@debian.org>
29567 Doug Kwan <dougkwan@google.com>
29568 H.J. Lu <hongjiu.lu@intel.com>
29569
29570 PR driver/55470
29571 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
29572
29573 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
29574
29575 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
29576
29577 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
29578
29579 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
29580
29581 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29582
29583 PR target/54461
29584 * doc/install.texi (Cross-Compiler-Specific Options): Document
29585 --with-avrlibc.
29586
29587 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
29588
29589 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
29590 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
29591 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
29592 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
29593 vqmovun_high_s64): Fix source operand number and update copyright.
29594
29595 2013-01-07 Richard Biener <rguenther@suse.de>
29596
29597 PR middle-end/55890
29598 * gimple.h (gimple_call_builtin_p): New overload.
29599 * gimple.c (validate_call): New function.
29600 (gimple_call_builtin_p): Likewise.
29601 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
29602 Use gimple_call_builtin_p.
29603 (find_func_clobbers): Likewise.
29604 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
29605 (strlen_optimize_stmt): Likewise.
29606
29607 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29608
29609 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
29610 (vld1q_dup_*): Likewise.
29611 (vld1_*): Likewise.
29612 (vld1q_*): Likewise.
29613 (vld1_lane_*): Likewise.
29614 (vld1q_lane_*): Likewise.
29615
29616 2013-01-07 Richard Biener <rguenther@suse.de>
29617
29618 * lto-streamer.h (LTO_minor_version): Bump to 2.
29619
29620 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29621
29622 * config/aarch64/aarch64-protos.h
29623 (aarch64_const_double_zero_rtx_p): Rename to...
29624 (aarch64_float_const_zero_rtx_p): ...this.
29625 (aarch64_float_const_representable_p): New.
29626 (aarch64_output_simd_mov_immediate): Likewise.
29627 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
29628 move immediate case.
29629 * config/aarch64/aarch64.c
29630 (aarch64_const_double_zero_rtx_p): Rename to...
29631 (aarch64_float_const_zero_rtx_p): ...this.
29632 (aarch64_print_operand): Allow printing of new constants.
29633 (aarch64_valid_floating_const): New.
29634 (aarch64_legitimate_constant_p): Check for valid floating-point
29635 constants.
29636 (aarch64_simd_valid_immediate): Likewise.
29637 (aarch64_vect_float_const_representable_p): New.
29638 (aarch64_float_const_representable_p): Likewise.
29639 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
29640 (aarch64_output_simd_mov_immediate): New.
29641 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
29642 (*movdf_aarch64): Likewise.
29643 * config/aarch64/constraints.md (Ufc): New.
29644 (Y): call aarch64_float_const_zero_rtx.
29645 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
29646
29647 2013-01-07 Richard Biener <rguenther@suse.de>
29648
29649 PR tree-optimization/55888
29650 PR tree-optimization/55862
29651 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
29652 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
29653 not if it is contained therein.
29654
29655 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29656
29657 * config/avr/t-avr: Typo.
29658
29659 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29660
29661 PR55243
29662 * config/avr/t-avr: Don't automatically rebuild
29663 $(srcdir)/config/avr/t-multilib
29664 $(srcdir)/config/avr/avr-tables.opt
29665 $(srcdir)/doc/avr-mmcu.texi
29666 (avr-mcus): New phony target to build them on request.
29667 (s-avr-mlib, s-avr-mmcu-texi): Remove.
29668 * avr/avr-mcus.def: Adjust comments.
29669
29670 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
29671
29672 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
29673
29674 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29675
29676 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
29677
29678 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29679
29680 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
29681
29682 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
29683
29684 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
29685 to generate profiling.
29686 * config/rs6000/aix64.h (LIB_SPEC): Same.
29687
29688 2013-01-04 Andrew Pinski <apinski@cavium.com>
29689
29690 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
29691 New function.
29692 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
29693
29694 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
29695
29696 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
29697 unconditionally.
29698 (ix86_expand_move): Ditto.
29699 (ix86_zero_extend_to_Pmode): Ditto.
29700 (ix86_expand_call): Ditto.
29701 (ix86_expand_special_args_builtin): Ditto.
29702 (ix86_expand_builtin): Ditto.
29703
29704 2013-01-04 Richard Biener <rguenther@suse.de>
29705
29706 PR tree-optimization/55862
29707 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
29708 translating them through PHI nodes.
29709
29710 2013-01-04 Martin Jambor <mjambor@suse.cz>
29711
29712 PR tree-optimization/55755
29713 * tree-sra.c (sra_modify_assign): Do not check that an access has no
29714 children when trying to avoid producing a VIEW_CONVERT_EXPR.
29715
29716 2013-01-04 Marek Polacek <polacek@redhat.com>
29717
29718 PR middle-end/55859
29719 * opts.c (default_options_optimization): Clarify error message.
29720
29721 2013-01-04 Richard Biener <rguenther@suse.de>
29722
29723 PR middle-end/55863
29724 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
29725 reassociation.
29726
29727 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29728
29729 PR target/53789
29730 * config/pa/pa.md (movsi): Revert previous change.
29731 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
29732 references.
29733
29734 2013-01-03 Richard Henderson <rth@redhat.com>
29735
29736 * config/i386/i386.c (ix86_expand_move): Always assign to op1
29737 after eliminating TLS symbols.
29738
29739 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
29740
29741 PR bootstrap/50167
29742 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
29743 * graphite-poly.c (debug_gmp_value): Likewise.
29744
29745 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
29746
29747 PR target/55712
29748 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
29749 selected code model, define __code_mode_small__, __code_model_medium__,
29750 __code_model_large__, __code_model_32__ or __code_model_kernel__.
29751 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
29752 xchg temporary register with %k. Declare temporary register as
29753 early clobbered.
29754 [__x86_64__]: For medium and large code models, preserve %rbx register.
29755
29756 2013-01-03 Richard Biener <rguenther@suse.de>
29757
29758 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
29759 (dump_subscript): Adjust.
29760 (finalize_ddr_dependent): Do not dump redundant info.
29761 (analyze_siv_subscript): Adjust.
29762 (subscript_dependence_tester): Likewise.
29763 (compute_affine_dependence): Likewise.
29764
29765 2013-01-03 Richard Biener <rguenther@suse.de>
29766
29767 Revert
29768 2013-01-03 Richard Biener <rguenther@suse.de>
29769
29770 PR tree-optimization/55857
29771 * tree-vect-stmts.c (vectorizable_load): Do not setup
29772 re-alignment for invariant loads.
29773
29774 2013-01-02 Richard Biener <rguenther@suse.de>
29775
29776 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29777 invariant load do not generate a vector load from the scalar location.
29778
29779 2013-01-03 Richard Biener <rguenther@suse.de>
29780
29781 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
29782 for not vectorizing.
29783 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
29784 not build INDIRECT_REFs, call get_name once only.
29785 (vect_create_data_ref_ptr): Likewise. Dump base object kind
29786 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
29787
29788 2013-01-03 Richard Biener <rguenther@suse.de>
29789
29790 PR tree-optimization/55857
29791 * tree-vect-stmts.c (vectorizable_load): Do not setup
29792 re-alignment for invariant loads.
29793
29794 2013-01-03 Richard Biener <rguenther@suse.de>
29795
29796 PR lto/55848
29797 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
29798 prefer a built-in decl.
29799
29800 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29801
29802 * gcc.c (process_command): Update copyright notice dates.
29803 * gcov.c (print_version): Likewise.
29804 * gcov-dump.c (print_version): Likewise.
29805
29806 PR rtl-optimization/55838
29807 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
29808 iv0.step, iv1.step and step.
29809
29810 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29811 Marc Glisse <marc.glisse@inria.fr>
29812
29813 PR tree-optimization/55832
29814 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
29815 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
29816 integer_{one,zero}_node.
29817
29818 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29819
29820 PR debug/54402
29821 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
29822 * var-tracking.c (reverse_op): Don't add reverse ops to
29823 VALUEs that have already
29824 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
29825
29826 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
29827
29828 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
29829
29830 2013-01-02 Teresa Johnson <tejohnson@google.com>
29831
29832 * dumpfile.c (dump_loc): Print filename with location.
29833 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
29834 new location_t parameter to emit complete unroll message with
29835 new dump framework.
29836 (canonicalize_loop_induction_variables): Compute loops location
29837 and pass to try_unroll_loop_completely.
29838 * loop-unroll.c (report_unroll_peel): New function.
29839 (peel_loops_completely): Use new dump format with location
29840 for main dumpfile message, and invoke report_unroll_peel on success.
29841 (decide_unrolling_and_peeling): Ditto.
29842 (decide_peel_once_rolling): Remove old dumpfile message subsumed
29843 by report_unroll_peel.
29844 (decide_peel_completely): Ditto.
29845 (decide_unroll_constant_iterations): Ditto.
29846 (decide_unroll_runtime_iterations): Ditto.
29847 (decide_peel_simple): Ditto.
29848 (decide_unroll_stupid): Ditto.
29849 * cfgloop.c (get_loop_location): New function.
29850 * cfgloop.h (get_loop_location): Declare.
29851
29852 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29853
29854 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
29855 NULL.
29856
29857 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29858
29859 PR middle-end/55198
29860 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
29861 BLKmode objects when EXPAND_MEMORY is specified.
29862
29863 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29864
29865 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
29866 in loop predicate.
29867 (fold_builtin_cpu): Do not share cpu model decls across statements.
29868
29869 2013-01-02 Jason Merrill <jason@redhat.com>
29870
29871 PR c++/55804
29872 * tree.c (build_array_type_1): Revert earlier change.
29873
29874 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
29875
29876 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
29877 "cortex-a57".
29878 * config/aarch64/aarch64-tune.md: Re-generate.
29879
29880 2013-01-02 Richard Biener <rguenther@suse.de>
29881
29882 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29883 invariant load do not generate a vector load from the scalar location.
29884
29885 2013-01-02 Richard Biener <rguenther@suse.de>
29886
29887 PR bootstrap/55784
29888 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
29889 * configure: Regenerate.
29890
29891 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29892
29893 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
29894 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
29895 (expand_builtin_int_roundingfn_2): Keep the original target around
29896 for the fallback case.
29897
29898 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29899
29900 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
29901 to be clear for sign changes.
29902
29903 2013-01-01 Jan Hubicka <jh@suse.cz>
29904
29905 * ipa-inline-analysis.c: Fix formatting.
29906
29907 2013-01-01 Jakub Jelinek <jakub@redhat.com>
29908
29909 PR tree-optimization/55831
29910 * tree-vect-loop.c (get_initial_def_for_induction): Use
29911 gsi_after_labels instead of gsi_start_bb.
29912 \f
29913 Copyright (C) 2013 Free Software Foundation, Inc.
29914
29915 Copying and distribution of this file, with or without modification,
29916 are permitted in any medium without royalty provided the copyright
29917 notice and this notice are preserved.