1 2013-10-02 Richard Biener <rguenther@suse.de>
3 * tree-loop-distribution.c: Include tree-vectorizer.h for
5 (enum partition_kind): Remove PKIND_REDUCTION.
6 (struct partition_s): Remove has_writes member, add reduction_p
8 (partition_alloc): Adjust.
9 (partition_builtin_p): Likewise.
10 (partition_has_writes): Remove.
11 (partition_reduction_p): New function.
12 (partition_merge_into): Likewise.
13 (generate_code_for_partition): Commonize builtin partition
15 (rdg_cannot_recompute_vertex_p): Remove.
16 (already_processed_vertex_p): Likewise.
17 (rdg_flag_vertex): Do not set has_writes.
18 (classify_partition): Adjust.
19 (rdg_build_partitions): Do not set has_writes, treat all
21 (distribute_loop): Record number of library calls generated.
23 (tree_loop_distribution): Report number of loops and library
24 calls generated as opt-info.
26 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
28 * tree-flow.h: Include new .h files. Move prototypes.
29 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
30 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
31 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
32 * tree-pretty-print.h: Add prototypes from tree-flow.h.
33 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
34 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
35 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
36 * tree.h (get_ref_base_and_extent): Move prototype out.
37 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
39 * gimple-low.h: New File. Add prototypes from tree-flow.h.
40 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
41 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
42 * tree-scalar-evolution.c: Include tree.h.
43 * sese.c: Include tree.h.
44 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
45 * dwarf2out.c: Include tree-dfa.h.
46 * tree-chrec.c: Include tree.h.
47 * tree-data-ref.c: Include tree.h.
49 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
51 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
54 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
56 * config/t-sol2 (sol2-c.o): Remove header dependencies.
57 Use $(COMPILE) and $(POSTCOMPILE).
58 (sol2-cxx.o): Likewise.
59 (sol2-stubs.o): Likewise.
61 * config/x-solaris (host-solaris.o): Likewise.
63 * config/sparc/t-sparc (sparc.o): Remove.
64 (sparc-c.o): Remove header dependencies.
65 Use $(COMPILE) and $(POSTCOMPILE).
66 * config/sparc/x-sparc: Likewise.
68 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
70 * config/arc/arc-opts.h: Add 2013 to Copyright years.
71 * config/arc/arc700.md: Likewise.
72 * config/arc/arc-modes.def: Likewise.
73 * config/arc/arc-simd.h: Likewise.
74 * config/arc/t-arc-uClibc: Likewise.
75 * config/arc/t-arc-newlib: Likewise.
77 2013-10-02 Renlin Li <renlin.li@arm.com>
79 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
80 (aarch64_expand_epilogue): Likewise.
82 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
83 Yufeng Zhang <yufeng.zhang@arm.com>
85 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
87 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
88 'base_in' represent a conversion and legal_cast_p_1 holds; set
89 'base_in' with the returned value from get_unwidened.
91 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
93 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
94 plus_constant is not used.
96 2013-10-01 Wei Mi <wmi@google.com>
98 * config/i386/x86-tune.def (DEF_TUNE): Remove
100 * config/i386/i386.md: Add define_peephole2 to
101 break partial reg stall for cvtss2sd/cvtsd2ss.
103 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
105 * config/arc/arc.c (pass_arc_ifcvt::clone):
106 Update for ctxt_ ->m_ctxt change.
108 2013-10-01 Jeff Law <law@redhat.com>
110 * tree-ssa-threadupdate.c (struct redirection_data): Delete
111 outgoing_edge and intermediate_edge fields. Instead store the path.
112 (redirection_data::hash): Hash on the last edge's destination index.
113 (redirection_data::equal): Check the entire thread path.
114 (lookup_redirectio_data): Corresponding changes.
115 (create_edge_and_update_destination_phis): Likewise.
116 (thread_single_edge): Likewise.
118 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
119 Diego Novillo <dnovillo@google.com>
121 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
122 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
123 (vld32wh_insn, vld32wl_insn): Delete commented-out old
124 versions of these patterns.
126 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
127 (__builtin_arc_aligned): Likewise.
129 * config/arc/arc.md: Expand adc_0 comment stating the intended
130 purpose and why it isn't ready.
131 Replace commented out call_value_via_label_mixed with a
132 plain comment about bl_s.
134 * config/arc/arc.c (stdio.h): Don't include directly.
135 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
136 Remove if (1) condition.
137 (arc_encode_section_info): Fix comment.
139 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
141 * config/arc/arc.c (arc_conditional_register_usage):
142 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
143 Also set reg_alloc_order for DMA config regs.
145 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
146 Jeremy Bennett <jeremy.bennett@embecosm.com>
148 * doc/install.texi (--with-cpu): Mention ARC.
149 (arc-*-elf32): New paragraph.
150 (arc-linux-uclibc): Likewise.
151 * doc/md.texi (Machine Constraints): Add ARC part.
152 * doc/invoke.texi: (menu): Add ARC Options.
153 (Machine Dependent Options) <ARC Options>: Add synopsis.
154 (node ARC Options): Add.
155 * doc/extend.texi (long_call / short_call attribute): Add ARC.
156 (ARC Built-in Functions): New section defining
157 generic ARC built-in functions.
158 (ARC SIMD Built-in Functions): New section defining SIMD specific
160 (Declaring Attributes of Functions): Extended
161 description of short_call and long_call attributes for ARC and
164 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
165 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
166 Joern Rennecke <joern.rennecke@embecosm.com>
167 Muhammad Khurram Riaz <khurram.riaz@arc.com>
168 Brendan Kehoe <brendan@zen.org>
169 Michael Eager <eager@eagercon.com>
170 Simon Cook <simon.cook@embecosm.com>
171 Jeremy Bennett <jeremy.bennett@embecosm.com>
173 * config/arc, common/config/arc: New directories.
175 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
176 Brendan Kehoe <brendan@zen.org>
177 Simon Cook <simon.cook@embecosm.com>
179 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
181 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
183 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
184 * tree-ssa-coalesce.h: New. Move prototype to here.
185 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
186 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
187 (gimple_can_coalesce_p): Move to...
188 * gimple.c (gimple_can_coalesce_p): Here.
190 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
192 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
193 (dump_decl_set): Move to gimple.c.
194 * gimple.h: Don't include tree-ssa-operands.h.
195 (dump_decl_set): Add prototype.
196 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
197 Move to gimple-ssa.h.
198 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
199 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
200 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
201 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
202 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
204 (dump_decl_set): Relocate here.
205 * gimple-ssa.h: New file.
206 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
207 Relocate from gimple.h.
208 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
209 * tree-ssa-operands.c (swap_ssa_operands): Rename from
210 swap_tree_operands and remove non-ssa path.
211 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
212 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
214 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
215 vect_is_simple_reduction_1): Use swap_ssa_operands.
216 * tree-flow.h: Move various prototypes to tree-phinodes.h.
217 (enum need_phi_state): Move to tree-into-ssa.c.
218 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
219 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
220 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
221 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
222 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
223 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
224 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
225 num_imm_uses): Move to ssa-iterators.h.
226 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
227 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
229 (op_iter_done, op_iter_next_def, op_iter_next_tree,
230 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
231 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
232 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
233 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
234 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
235 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
236 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
237 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
238 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
239 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
240 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
241 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
242 (set_phi_nodes): Move to tree-phinodes.h.
243 * tree-ssa-operands.h (enum ssa_op_iter_type,
244 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
245 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
246 (dump_decl_set): Remove prototype.
247 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
248 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
249 (set_phi_nodes): Relocate from tree-flow-inline.h.
250 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
252 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
253 include list. Temporarily add gimple.h to include list.
254 * ssa-iterators.h: New file.
255 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
256 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
257 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
258 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
259 Relocate from tree-ssa-operands.h.
260 (delink_imm_use, link_imm_use_to_list, link_imm_use,
261 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
262 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
263 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
264 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
265 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
266 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
267 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
268 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
269 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
270 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
271 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
272 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
273 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
274 Relocate from tree-flow-inline.h.
275 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
277 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
280 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
281 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
282 Insert '\t' to output template.
283 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
284 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
285 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
286 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
288 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
290 * doc/install.texi (Host/target specific installation notes for GCC):
291 Put @anchor before @heading.
292 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
293 Use @email for email addresses.
295 2013-10-01 Jeff Law <law@redhat.com>
297 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
298 a vec. Only delete the path if we create one without successfully
299 registering a jump thread.
300 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
302 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
303 (paths): New vector of jump threading paths.
304 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
305 (THREAD_PATH): New accessor macro for the entire thread path.
306 (lookup_redirection_data): Get intermediate and final outgoing edge
307 from the thread path.
308 (create_edge_and_update_destination_phis): Copy the threading path.
309 (ssa_fix_duplicate_block_edges): Get edges and block types from the
311 (ssa_redirect_edges): Get edges and block types from the jump threading
312 path. Free the path vector.
313 (thread_block): Get edges from the jump threading path. Look at the
314 entire path to see if we thread to a loop exit. If we cancel a jump
315 thread request, then free the path vector.
316 (thread_single_edge): Get edges and block types from the jump threading
317 path. Free the path vector.
318 (thread_through_loop_header): Get edges and block types from the jump
319 threading path. Free the path vector.
320 (mark_threaded_blocks): Iterate over the vector of paths and store
321 the path on the appropriate edge. Get edges and block types from the
323 (mark_threaded_blocks): Get edges and block types from the jump
324 threading path. Free the path vector.
325 (thread_through_all_blocks): Use the vector of paths rather than
326 a vector of 3-edge sets.
327 (register_jump_thread): Accept pointer to a path vector rather
328 than the path vector itself. Store the path vector for later use.
331 2013-10-01 Jakub Jelinek <jakub@redhat.com>
332 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
335 * config/s390/s390.c (s390_split_branches): Modify check for table
337 (s390_chunkify_start): Rearrange table jump insn check in order to
338 deal with compare and branch insns correctly.
340 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
344 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
345 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
346 define_insn_and_split.
347 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
348 (shiftsi3_compare): New pattern.
350 * config/arm/unspecs.md (UNSPEC_RRX): New.
352 2013-10-01 Alan Modra <amodra@gmail.com>
354 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
355 casing inout operands.
357 2013-10-01 Richard Biener <rguenther@suse.de>
359 PR tree-optimization/58553
360 * tree-loop-distribution.c (struct partition_s): Add niter member.
361 (classify_partition): Populate niter member for the partition
362 and properly identify whether the relevant store happens before
363 or after the loop exit.
364 (generate_memset_builtin): Use niter member from the partition.
365 (generate_memcpy_builtin): Likewise.
367 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
369 * vec.h (vec_prefix, vec): Prefix member names with "m_".
370 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
372 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
374 * basic-block.h (edge_list): Prefix member names with "m_".
375 * context.h (context): Likewise.
376 * domwalk.h (dom_walker): Likewise.
377 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
378 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
379 * hash-table.h (hash_table): Likewise.
380 * machmode.h (bit_field_mode_iterator): Likewise.
381 * pass_manager.h (pass_list): Likewise.
382 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
383 * tree-pass.h (pass_data): Likewise.
384 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
385 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
386 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
387 * asan.c (pass_data_asan): Update accordingly.
388 * cfganal.c (control_dependences::find_control_dependence): Likewise.
389 (control_dependences::control_dependences): Likewise.
390 (control_dependences::~control_dependences): Likewise.
391 (control_dependences::~control_dependences): Likewise.
392 (control_dependences::get_edges_dependent_on): Likewise.
393 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
394 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
395 * context.c (gcc::context::context): Likewise.
396 * cprop.c (pass_rtl_cprop::clone): Likewise.
397 * domwalk.c (dom_walker::walk): Likewise.
398 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
399 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
400 * mode-switching.c (pass_mode_switching::clone): Likewise.
401 * passes.c (opt_pass::opt_pass): Likewise.
402 (pass_manager::pass_manager): Likewise.
403 * predict.c (pass_strip_predict_hints::clone): Likewise.
404 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
405 (pass_split_all_insns::clone): Likewise.
406 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
408 (bit_field_mode_iterator::next_mode): Likewise.
409 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
410 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
411 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
412 * tree-complex.c (pass_lower_complex::clone): Likewise.
413 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
414 * tree-object-size.c (pass_object_sizes::clone): Likewise.
415 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
416 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
417 (pass_fold_builtins::clone): Likewise.
418 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
419 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
420 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
421 (pass_cd_dce::clone): Likewise.
422 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
423 (pass_phi_only_cprop::clone): Likewise.
424 * tree-ssa-dse.c (pass_dse::clone): Likewise.
425 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
426 * tree-ssa-loop.c (pass_lim::clone): Likewise.
427 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
428 * tree-ssa-pre.c (pass_fre::clone): Likewise.
429 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
430 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
431 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
432 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
433 * tree-vrp.c (pass_vrp::clone): Likewise.
434 * tsan.c (pass_tsan::clone): Likewise.
436 2013-09-30 Jakub Jelinek <jakub@redhat.com>
439 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
440 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
443 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
444 optimization, punt if sign_bit_p looked through any zero extension.
446 2013-09-30 Teresa Johnson <tejohnson@google.com>
448 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
449 Update redirected out edge count in joiner case.
450 (ssa_redirect_edges): Common the joiner and non-joiner cases
451 so that joiner case gets profile updates.
453 2013-09-30 Richard Biener <rguenther@suse.de>
455 PR tree-optimization/58554
456 * tree-loop-distribution.c (classify_partition): Require
457 unconditionally executed stores for memcpy and memset recognition.
458 (tree_loop_distribution): Calculate dominance info.
460 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
462 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
463 (NO_PROFILE_COUNTERS): Likewise.
464 (PROFILE_HOOK): Likewise.
465 (FUNCTION_PROFILER): Likewise.
466 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
468 2013-09-30 Iain Sandoe <iain@codesourcery.com>
470 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
471 calls and defines in TARGET_MACHO conditional.
472 (load_macho_picbase_di): Likewise.
473 (reload_macho_picbase): Likewise.
474 (reload_macho_picbase_si): Likewise.
475 (reload_macho_picbase_di): Likewise.
476 (nonlocal_goto_receiver): Likewise.
478 2013-09-30 Nick Clifton <nickc@redhat.com>
480 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
481 that use the MSP430X ISA.
482 (msp430_option_override): Scan -mmcu command line option for any
483 MCU name that supports the MSP430X ISA.
484 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
485 -mmcu options which enable the MSP430X ISA.
487 2013-09-30 Richard Biener <rguenther@suse.de>
490 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
491 before looking for setjmp-like calls.
493 2013-09-29 Iain Sandoe <iain@codesourcery.com>
496 * config/darwin-protos.h (machopic_get_function_picbase): New.
497 * config/darwin.c (machopic_get_function_picbase): New.
498 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
499 label for a new func. (load_macho_picbase_di): Likewise.
500 (reload_macho_picbase): New expand.
501 (reload_macho_picbase_si): New insn.
502 (reload_macho_picbase_di): New insn.
503 (nonlocal_goto_receiver): New define and split.
504 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
505 (unspecv enum): Add UNSPECV_NLGR.
507 2013-09-29 Iain Sandoe <iain@codesourcery.com>
509 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
510 that altivec registers are correctly sized on Darwin.
512 2013-09-29 Iain Sandoe <iain@codesourcery.com>
514 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
515 darwin-driver.o): Use COMPILE and POSTCOMPILE.
516 * config/x-darwin (host-darwin.o): Likewise.
517 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
518 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
519 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
521 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
523 * doc/invoke.texi: Fix usage of @tie{} command.
525 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
527 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
529 2013-09-28 Jan Hubicka <jh@suse.cz>
531 * config/i386/x86-tune.def: Add documentation for each of the options;
534 2013-09-28 Jan Hubicka <jh@suse.cz>
536 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
538 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
539 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
540 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
541 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
542 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
544 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
546 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
547 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
548 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
549 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
550 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
551 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
552 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
553 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
554 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
555 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
556 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
557 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
558 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
559 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
560 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
561 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
562 gimple.h, godump.c, graphite-clast-to-gimple.c,
563 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
564 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
565 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
566 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
567 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
568 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
569 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
570 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
571 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
572 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
573 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
574 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
575 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
576 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
577 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
578 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
579 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
580 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
581 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
582 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
583 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
584 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
585 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
586 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
587 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
588 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
589 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
590 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
591 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
592 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
593 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
594 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
595 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
596 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
597 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
598 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
599 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
600 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
601 whitespace before "(".
603 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
605 * expr.h (extract_bit_field): Remove packedp parameter.
606 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
607 from forward declaration.
608 (store_split_bit_field): Remove packedp arg from calls to
609 extract_fixed_bit_field.
610 (extract_bit_field_1): Remove packedp parameter and packedp
611 argument from recursive calls and calls to extract_fixed_bit_field.
612 (extract_bit_field): Remove packedp parameter and corresponding
613 arg to extract_bit_field_1.
614 (extract_fixed_bit_field): Remove packedp parameter. Remove code
616 (extract_split_bit_field): Remove packedp arg from call to
617 extract_fixed_bit_field.
618 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
619 (copy_blkmode_from_reg): Likewise.
620 (copy_blkmode_to_reg): Likewise.
621 (read_complex_part): Likewise.
622 (store_field): Likewise.
623 (expand_expr_real_1): Likewise.
624 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
625 to extract_bit_field.
626 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
627 call to extract_bit_field.
628 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
629 call to extract_bit_field.
630 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
631 and special packedp behavior from -fstrict-volatile-bitfields
634 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
636 * lra-eliminations.c (init_elim_table): Guard value_p.
638 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
640 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
641 DFmode, DImode, and SFmode in the upper VSX registers based on the
642 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
643 if -mpower8-vector. Combine -mvsx-timode handling with the rest
644 of the VSX register handling.
646 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
648 (zero_extendsidi2_lfiwzx): Add support for loading into the
649 Altivec registers with -mpower8-vector. Use wu/wv constraints to
650 only do VSX memory options on Altivec registers.
651 (extendsidi2_lfiwax): Likewise.
652 (extendsfdf2_fpr): Likewise.
653 (mov<mode>_hardfloat, SF/SD modes): Likewise.
654 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
655 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
656 (movdi_internal64): Likewise.
658 2013-09-27 Xinliang David Li <davidxl@google.com>
660 * opts.c (finish_options): Adjust parameters
661 according to vect cost model.
662 (common_handle_option): Set dynamic vect cost
664 targhooks.c (default_add_stmt_cost): Compute stmt cost
666 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
668 * tree-vectorizer.h (unlimited_cost_model): New function.
669 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
670 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
672 (vect_enhance_data_refs_alignment): Ditto.
673 * flag-types.h: New enum.
674 * common/config/i386/i386-common.c (ix86_option_init_struct):
675 No need to initialize vect_cost_model flag.
676 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
679 2013-09-27 Diego Novillo <dnovillo@google.com>
681 * gimple.h (enum ssa_mode): Remove.
683 2013-09-27 Paulo Matos <pmatos@broadcom.com>
686 * gcc.dg/pr58463.c: New test.
688 2013-09-27 Paulo Matos <pmatos@broadcom.com>
690 * cfgloop.h (number_of_loops): Fix typo in check for null.
692 2013-09-27 Jakub Jelinek <jakub@redhat.com>
695 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
696 are children of outermost saved_cfun's loop, and set it up to
697 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
698 if loop != loop0 and SESE region contains bbs that belong to loop0.
700 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
702 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
703 (binary_scale_code_p, get_base_term, get_index_term): New functions.
704 (set_address_segment, set_address_base, set_address_index)
705 (set_address_disp): Accept the argument unconditionally.
706 (baseness): Remove must_be_base_p and must_be_index_p checks.
707 (decompose_normal_address): Classify as much as possible in the
710 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
712 * cse.c (count_reg_usage): Handle INT_LIST.
713 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
714 * reginfo.c (reg_scan_mark_refs): Likewise.
715 * reload1.c (eliminate_regs_1): Likewise.
717 2013-09-27 Iain Sandoe <iain@codesourcery.com>
720 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
723 2013-09-27 Richard Biener <rguenther@suse.de>
725 PR tree-optimization/58459
726 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
727 restriction not propagating into loops.
729 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
731 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
732 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
733 * doc/tree-ssa.texi (Walking use-def chains): Delete.
735 2013-09-26 Richard Biener <rguenther@suse.de>
737 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
739 2013-09-26 Richard Biener <rguenther@suse.de>
741 * alias.h (component_uses_parent_alias_set): Rename to ...
742 (component_uses_parent_alias_set_from): ... this.
743 * alias.c (component_uses_parent_alias_set): Rename to ...
744 (component_uses_parent_alias_set_from): ... this and return
746 (reference_alias_ptr_type_1): Use the result from
747 component_uses_parent_alias_set_from instead of stripping
748 components one at a time.
749 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
751 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
753 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
754 Move prototypes to...
755 * tree-ssa-ter.h: New File. Move prototypes here.
756 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
757 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
758 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
759 from is_replaceable_p.
760 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
761 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
762 (process_replaceable): Use ter_is_replaceable_p.
763 (find_replaceable_in_bb): Use ter_is_replaceable_p.
764 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
765 newly refactored ssa_is_replaceable_p.
766 * cfgexpand.c: Include tree-outof-ssa.h.
767 * ssaexpand.h: Delete.
769 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
771 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
772 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
774 (create_gimple_tmp): Delete.
775 (get_expr_type, build_assign, build_type_cast): Move to...
776 * gimple-builder.c: New File.
777 (get_expr_type): Relocate from gimple.c.
778 (build_assign, build_type_cast): Change to only create ssanames.
779 * gimple.h: Move prototypes to...
780 * gimple-builder.h: New File. Here.
781 * tree-ssa.h: And here.
782 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
783 count_uses_and_derefs): Relocate from gimple.c.
784 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
785 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
786 * tree-ssa-math-opts (execute_cse_reciprocals): Use
787 gimple_replace_ssa_lhs.
788 * asan.c: Include gimple-builder.h.
789 * Makefile.in: Add gimple-builder.o.
791 2013-09-26 Richard Biener <rguenther@suse.de>
793 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
794 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
795 (loe_visit_block): Use gcc_checking_assert.
796 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
797 gimple_assign_ssa_name_copy_p.
798 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
801 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
803 * config/rs6000/t-rs6000 (rs6000.o): Remove.
804 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
806 2013-09-26 Richard Biener <rguenther@suse.de>
808 PR tree-optimization/58539
809 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
810 the fact that debug statements are not taking part in loop-closed
813 2013-09-26 Nick Clifton <nickc@redhat.com>
815 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
816 time warning message.
817 (msp430_print_operand_raw): Delete unused letter parameter.
818 (TARGET_PRINT_OPERAND_ADDRESS): Define.
819 (msp430_print_operand_address): New function.
820 (msp430_print_operand): Move address printing code from here to
822 * config/msp430/msp430.md (movsipsi2): Add comment in generated
824 (zero_extendpsisi2): Likewise.
825 (extendpsisi2): New pattern.
826 (andneghi3): New pattern.
828 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
830 * config/aarch64/aarch64.opt (mlra): New option.
831 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
832 (TARGET_LRA_P): Define.
834 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
836 * expr.c (expand_assignment): Remove obsolete comment.
838 2013-09-25 Jeff Law <law@redhat.com>
840 * tree-flow.h (thread_through_all_blocks): Prototype moved into
841 tree-ssa-threadupdate.h.
842 (register_jump_thread): Similarly.
843 * tree-ssa-threadupdate.h: New header file.
844 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
845 * tree-vrp.c: Likewise.
846 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
847 (thread_around_empty_blocks): Change type of path vector argument to
848 an edge,type pair from just an edge. Initialize both elements when
849 appending to a jump threading path. Tweak references to elements
851 (thread_across_edge): Similarly. Release memory for the elements
853 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
854 (dump_jump_thread_path): New function broken out from
855 register_jump_thread.
856 (register_jump_thread): Use dump_jump_thread_path. Change type of
857 path vector entries. Search the path for NULL edges and dump
858 the path if one is found. Tweak the conversion of path to 3-edge
859 form to use the block copy type information embedded in the path.
861 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
863 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
865 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
866 Vladimir Makarov <vmakarov@redhat.com>
868 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
869 from the least significant bit.
870 (strip_address_mutations): Add bitfield operations handling.
871 (must_be_index_p): Add shifting and rotate operations handling.
872 (set_address_base): Use must_be_base_p predicate.
873 (set_address_index): Use must_be_index_p predicate.
875 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
876 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
877 Sergey Lega <sergey.s.lega@intel.com>
878 Anna Tikhonova <anna.tikhonova@intel.com>
879 Ilya Tocar <ilya.tocar@intel.com>
880 Andrey Turetskiy <andrey.turetskiy@intel.com>
881 Ilya Verbin <ilya.verbin@intel.com>
882 Kirill Yukhin <kirill.yukhin@intel.com>
883 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
885 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
887 (ix86_expand_vector_move_misalign): Support new unaligned load and
888 stores and use new names.
889 (CODE_FOR_sse2_storedqu): Rename to ...
890 (CODE_FOR_sse2_storedquv16qi): ... this.
891 (CODE_FOR_sse2_loaddqu): Rename to ...
892 (CODE_FOR_sse2_loaddquv16qi): ... this.
893 (CODE_FOR_avx_loaddqu256): Rename to ...
894 (CODE_FOR_avx_loaddquv32qi): ... this.
895 (CODE_FOR_avx_storedqu256): Rename to ...
896 (CODE_FOR_avx_storedquv32qi): ... this.
897 * config/i386/i386.md (fpint_logic): New.
898 * config/i386/sse.md (VMOVE): Extend for AVX512.
902 (VI_UNALIGNED_LOADSTORE): Ditto.
903 (sse2_avx_avx512f): Ditto.
904 (sse2_avx2): Extend for AVX512.
905 (sse4_1_avx2): Ditto.
907 (sse): Extend for AVX512.
910 (avxsizesuffix): Ditto.
911 (sseintvecmode): Ditto.
913 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
914 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
915 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
916 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
917 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
918 (<sse2_avx_avx512f>_storedqu<mode): ... this.
919 (<sse>_movnt<mode>): Replace constraint "x" with "v".
920 (STORENT_MODE): Extend for AVX512.
921 (*absneg<mode>2): Replace constraint "x" with "v".
922 (*mul<mode>3): Ditto.
923 (*ieee_smin<mode>3): Ditto.
924 (*ieee_smax<mode>3): Ditto.
925 (avx_cmp<mode>3): Replace VF with VF_128_256.
926 (*<sse>_maskcmp<mode>3_comm): Ditto.
927 (<sse>_maskcmp<mode>3): Ditto.
928 (<sse>_andnot<mode>3): Extend for AVX512.
929 (<code><mode>3, anylogic): Replace VF with VF_128_256.
930 (<code><mode>3, fpint_logic): New.
931 (*<code><mode>3): Extend for AVX512.
932 (avx512flogicsuff): New.
933 (avx512f_<logic><mode>): Ditto.
934 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
936 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
937 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
938 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
939 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
940 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
941 (xop_vpermil2<mode>3): Ditto.
942 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
943 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
944 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
945 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
947 2013-09-25 Tom Tromey <tromey@redhat.com>
949 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
950 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
951 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
952 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
953 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
954 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
955 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
956 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
957 (GRAPHITE_HTAB_H): Remove.
959 2013-09-25 Tom Tromey <tromey@redhat.com>
961 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
963 2013-09-25 Tom Tromey <tromey@redhat.com>
965 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
967 2013-09-25 Tom Tromey <tromey@redhat.com>
969 * config/i386/t-i386 (i386.o): Remove.
970 (i386-c.o): Use COMPILE and POSTCOMPILE.
972 2013-09-25 Tom Tromey <tromey@redhat.com>
974 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
976 2013-09-25 Tom Tromey <tromey@redhat.com>
978 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
979 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
980 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
981 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
982 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
983 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
984 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
985 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
986 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
987 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
988 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
989 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
990 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
991 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
992 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
993 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
994 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
995 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
996 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
997 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
998 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
999 (tree-ssa-pre.o, tree-ssa-sccvn.o)
1000 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
1001 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
1002 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
1003 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
1004 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
1005 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
1006 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
1007 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
1008 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
1009 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
1010 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
1011 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
1012 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
1013 (graphite-clast-to-gimple.o, graphite-dependences.o)
1014 (graphite-interchange.o, graphite-poly.o)
1015 (graphite-scop-detection.o, graphite-sese-to-poly.o)
1016 (graphite-optimize-isl.o, tree-vect-loop.o)
1017 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
1018 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
1019 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
1020 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
1021 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
1022 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
1023 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
1024 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
1025 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
1026 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
1027 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
1028 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
1029 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
1030 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
1031 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
1032 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
1033 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
1034 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
1035 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
1036 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
1037 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
1038 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
1039 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
1040 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
1041 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
1042 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
1043 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
1044 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
1045 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
1046 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
1047 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
1048 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
1049 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
1050 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
1051 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
1052 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
1053 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
1054 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
1055 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
1056 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
1057 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
1058 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
1059 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
1060 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
1061 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
1062 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
1063 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
1064 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
1065 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
1066 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
1067 (gcov-dump.o): Remove.
1068 (default-c.o): Use COMPILE and POSTCOMPILE.
1069 (CFLAGS-gcc.o): New variable.
1070 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
1072 2013-09-25 Tom Tromey <tromey@redhat.com>
1074 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
1075 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
1076 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
1077 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
1078 (c-family/c-pragma.o, c-family/c-pretty-print.o)
1079 (c-family/c-semantics.o, c-family/c-ada-spec.o)
1080 (c-family/array-notation-common.o, c-family/stub-objc.o)
1081 (c-family/c-ubsan.o): Remove.
1083 2013-09-25 Tom Tromey <tromey@redhat.com>
1085 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
1087 2013-09-25 Tom Tromey <tromey@redhat.com>
1089 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
1090 to add -DENABLE_SHARED_LIBGCC.
1091 (gcc.o): Don't use subshell.
1093 2013-09-25 Tom Tromey <tromey@redhat.com>
1095 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
1096 * configure.ac: Don't invoke AM_PROG_CC_C_O.
1097 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
1098 * configure, config.in: Rebuild.
1100 2013-09-25 Tom Tromey <tromey@redhat.com>
1102 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
1103 (COMPILE, POSTCOMPILE): New variables.
1104 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
1105 (DEPFILES): New variable.
1106 Include ".Po" files.
1107 * configure.ac: Add checks for dependency checking.
1108 * configure, aclocal.m4: Regenerate.
1110 2013-09-25 Tom Tromey <tromey@redhat.com>
1112 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
1113 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
1115 2013-09-25 Tom Tromey <tromey@redhat.com>
1117 * Makefile.in (generated_files): Add options.h,
1118 target-hooks-def.h, insn-opinit.h,
1119 common/common-target-hooks-def.h, pass-instances.def,
1120 c-family/c-target-hooks-def.h.
1122 2013-09-25 Jeff Law <law@redhat.com>
1124 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
1125 than foo[foo.length () - 1] to access last member in a vec.
1126 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
1128 2013-09-25 Richard Biener <rguenther@suse.de>
1131 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
1133 2013-09-25 Jan Hubicka <jh@suse.cz>
1135 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
1138 2013-09-25 Marek Polacek <polacek@redhat.com>
1141 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
1142 when determining the type name.
1144 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
1146 * config/sh/sh.md: Fix formatting.
1148 2013-09-24 Xinliang David Li <davidxl@google.com>
1150 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
1151 max peel iterations parameter.
1152 * param.def: New parameter.
1153 * doc/invoke.texi: Document New parameter.
1155 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
1157 * gimple-pretty-print.c: Various whitespace tweaks.
1158 * tree-core.h: Likewise.
1159 * tree-pretty-print.c: Likewise.
1160 * tree-ssa-alias.c: Likewise.
1161 * tree-ssa-copy.c: Likewise.
1162 * tree-ssanames.c: Likewise.
1163 * tree-ssanames.h: Likewise.
1164 * tree-vrp.c: Likewise.
1166 2013-09-24 Alan Modra <amodra@gmail.com>
1170 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
1171 for output operands that disallow regs. Don't use EXPAND_WRITE on
1174 2013-09-24 Richard Biener <rguenther@suse.de>
1177 * tree.c (reference_alias_ptr_type): Move ...
1178 * alias.c (reference_alias_ptr_type): ... here and implement
1179 in terms of the new reference_alias_ptr_type_1.
1180 (ref_all_alias_ptr_type_p): New helper.
1181 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
1182 use ref_all_alias_ptr_type_p.
1183 (get_deref_alias_set): Add flag_strict_aliasing check here.
1184 (reference_alias_ptr_type_1): New function, split out from ...
1185 (get_alias_set): ... here.
1186 (alias_ptr_types_compatible_p): New function.
1187 * alias.h (reference_alias_ptr_type): Declare.
1188 (alias_ptr_types_compatible_p): Likewise.
1189 * tree.h (reference_alias_ptr_type): Remove.
1190 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
1191 to compare MEM_REF alias types.
1193 2013-09-24 Richard Biener <rguenther@suse.de>
1195 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
1197 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
1199 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
1200 reload helper function arrays into a single array reg_addr.
1201 (reload_fpr_gpr): Likewise.
1202 (reload_gpr_vsx): Likewise.
1203 (reload_vsx_gpr): Likewise.
1204 (struct rs6000_reg_addr): Likewise.
1205 (reg_addr): Likewise.
1206 (rs6000_debug_reg_global): Change rs6000_vector_reload,
1207 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
1208 (rs6000_init_hard_regno_mode_ok): Likewise.
1209 (rs6000_secondary_reload_direct_move): Likewise.
1210 (rs6000_secondary_reload): Likewise.
1212 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
1213 constraints: wu, ww, and wy. Repurpose wv constraint added during
1214 power8 changes. Put wg constraint in alphabetical order.
1216 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
1217 for future work to add ISA 2.07 VSX single precision support.
1218 (-mvsx-scalar-double): Change default from -1 to 1, update
1219 documentation comment.
1220 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
1221 (-mupper-regs-df): New debug switch to control whether DF values
1222 can go in the traditional Altivec registers.
1223 (-mupper-regs-sf): New debug switch to control whether SF values
1224 can go in the traditional Altivec registers.
1226 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
1228 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
1229 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
1230 Add new constraints, wu/ww/wy. Repurpose wv constraint.
1231 (rs6000_debug_legitimate_address_p): Print if we are running
1232 before, during, or after reload.
1233 (rs6000_secondary_reload): Add a comment.
1234 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
1236 * config/rs6000/constraints.md (wa constraint): Sort w<x>
1237 constraints. Update documentation string.
1238 (wd constraint): Likewise.
1239 (wf constraint): Likewise.
1240 (wg constraint): Likewise.
1241 (wn constraint): Likewise.
1242 (ws constraint): Likewise.
1243 (wt constraint): Likewise.
1244 (wx constraint): Likewise.
1245 (wz constraint): Likewise.
1246 (wu constraint): New constraint for ISA 2.07 SFmode scalar
1248 (ww constraint): Likewise.
1249 (wy constraint): Likewise.
1250 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
1251 the previous submissions.
1252 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
1254 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
1256 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
1257 (REG_BR_PROB): Say that the probability is stored in an int_list.
1258 * reg-notes.def: Update commentary to mention INT_LIST.
1259 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
1260 (INT_LIST): New rtx.
1261 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
1262 * rtlanal.c (int_reg_note_p): New function.
1263 (alloc_reg_note): Assert that the note does not have an int argument.
1264 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
1265 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
1266 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
1267 rather than an INSN_LIST. Handle INT_LIST.
1268 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
1269 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
1270 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
1271 Manipulate them as ints rather than rtxes.
1272 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
1273 * regmove.c (copy_src_to_dest): Likewise.
1274 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
1276 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
1277 into the cases that need it.
1278 * config/arm/arm.c (arm_unwind_emit): Likewise.
1280 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
1281 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
1282 * loop-doloop.c (add_test, doloop_modify): Likewise.
1283 * loop-unswitch.c (compare_and_jump_seq): Likewise.
1284 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
1285 * predict.c (combine_predictions_for_insn): Likewise.
1286 * print-rtl.c (print_rtx): Handle INT_LIST.
1287 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
1288 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
1289 * config/arm/arm.c (emit_unlikely_jump): Likewise.
1290 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
1291 (ix86_split_fp_branch, predict_jump): Likewise.
1292 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
1293 * config/sh/sh.c (expand_cbranchsi4): Likewise.
1294 * config/spu/spu.c (ea_load_store_inline): Likewise.
1296 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
1297 value of a REG_BR_PROB note.
1298 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
1299 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
1300 * emit-rtl.c (try_split): Likewise.
1301 * predict.c (br_prob_note_reliable_p): Likewise.
1302 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
1303 * reorg.c (mostly_true_jump): Likewise.
1304 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
1305 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
1306 * config/i386/i386.c (ix86_print_operand): Likewise.
1307 * config/ia64/ia64.c (ia64_print_operand): Likewise.
1308 * config/mmix/mmix.c (mmix_print_operand): Likewise.
1309 * config/rs6000/rs6000.c (output_cbranch): Likewise.
1310 * config/s390/s390.c (s390_expand_tbegin): Likewise.
1311 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
1312 * config/sparc/sparc.c (output_cbranch): Likewise.
1313 * config/spu/spu.c (get_branch_target): Likewise.
1314 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
1315 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
1317 2013-09-23 Jan Hubicka <jh@suse.cz>
1319 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
1321 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
1322 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
1324 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
1325 (gimple_fold_call): Dump inconsistent devirtualizations; add
1326 sanity check for type based devirtualizations.
1327 * ipa-prop.c: Include ipa-utils.h
1328 (ipa_intraprocedural_devirtualization): Add sanity check.
1329 (try_make_edge_direct_virtual_call): Likewise.
1331 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
1333 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
1334 assignment statements.
1336 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
1338 * gimple-pretty-print.c (dump_ssaname_info): New function.
1339 (dump_gimple_phi): Call it.
1340 (pp_gimple_stmt_1): Likewise.
1341 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
1342 (range_info_def): Declare.
1343 * tree-pretty-print.c (pp_double_int): New function.
1344 (dump_generic_node): Call it.
1345 * tree-pretty-print.h (pp_double_int): Declare.
1346 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
1347 * tree-ssanames.h (range_info_def): New structure.
1348 (value_range_type): Move definition here.
1349 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
1351 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
1353 (set_range_info): New function.
1354 (get_range_info): Likewise.
1355 (duplicate_ssa_name_range_info): Likewise.
1356 (duplicate_ssa_name_fn): Check pointer type and call
1357 duplicate_ssa_name_range_info.
1358 * tree-ssa-copy.c (fini_copy_prop): Likewise.
1359 * tree-vrp.c (value_range_type): Remove definition, now in
1361 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
1362 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
1363 (SSA_NAME_RANGE_INFO): New macro.
1365 2013-09-23 Richard Biener <rguenther@suse.de>
1367 PR tree-optimization/58464
1368 * tree-ssa-pre.c (phi_trans_lookup): Remove.
1369 (phi_trans_add): Change to add conditionally on being not present.
1370 (phi_translate_1): Remove recursion detection here.
1371 (phi_translate): Pre-seed the cache with NULL to catch
1372 recursion here in a more generic way.
1373 (bitmap_find_leader): Adjust comment.
1374 (get_representative_for): Dump value-numbers.
1375 (create_expression_by_pieces): Likewise.
1376 (insert_into_preds_of_block): Likewise.
1378 2013-09-23 Christian Bruel <christian.bruel@st.com>
1381 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
1382 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
1384 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
1387 2013-09-20 Renlin Li <renlin.li@arm.com>
1389 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
1391 (aarch64_expand_epilogue): Likewise.
1392 (aarch64_legitimize_reload_address): Likewise.
1394 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
1396 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
1397 NULL_TREE before pushing them onto the vector. Likewise for labels.
1399 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
1401 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
1402 comparison operators when -fno-trapping-math is in effect.
1403 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
1404 comparison operators in TFmode and assert that unsupported operators
1406 (ia64_print_operand): Likewise.
1408 2013-09-21 Jan Hubicka <jh@suse.cz>
1410 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
1411 (sse_typeless_stores): Enable for core
1412 (sse_load0_by_pxor): Likewise.
1413 (four_jump_limit): Disable for core.
1414 (pad_returns): Likewise.
1415 (avoid_vector_decode): Likewise.
1416 (fuse_cmp_and_branch): Enable for cores.
1417 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
1419 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
1422 * config/pa/pa.c (pa_option_override): Disable auto increment and
1423 decrement instructions until reload is completed.
1425 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
1426 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
1427 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
1429 2013-09-20 DJ Delorie <dj@redhat.com>
1430 Nick Clifton <nickc@redhat.com>
1432 * config/rl78/rl78.c: Various whitespace and comment tweaks.
1433 (need_to_save): Save bank 0 on interrupts.
1434 (characterize_address): Strip far address wrappers.
1435 (rl78_as_legitimate_address): Likewise.
1436 (transcode_memory_rtx): Likewise.
1437 (rl78_peep_movhi_p): Disable this peephole after devirt.
1438 (rl78_propogate_register_origins): Forget all origins when a
1440 * config/rl78/rl78-virt.md: Various whitespace tweaks.
1441 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
1443 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
1444 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
1445 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
1447 * config/rl78/constraints.md: Various whitespace and paren tweaks.
1449 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
1451 * config/pa/pa.md: In "scc" insn patterns, change output template to
1452 handle const0_rtx in reg_or_0_operand operands.
1454 2013-09-20 Martin Husemann <martin@NetBSD.org>
1457 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
1458 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
1460 2013-09-20 Richard Biener <rguenther@suse.de>
1463 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
1464 remembering SSA name version and block index.
1465 (new_scev_info_str): Adjust.
1466 (hash_scev_info): Likewise. Also hash the block index.
1467 (eq_scev_info): Adjust.
1468 (find_var_scev_info): Likewise.
1469 (struct instantiate_cache_entry): Remove.
1470 (struct instantiate_cache_type): Use a htab to map name, block
1472 (instantiate_cache_type::~instantiate_cache_type): Adjust.
1473 (get_instantiated_value_entry): Likewise.
1474 (hash_idx_scev_info, eq_idx_scev_info): New functions.
1475 (instantiate_scev_name): Adjust.
1477 2013-09-20 Jeff Law <law@redhat.com>
1479 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
1481 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
1483 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
1484 Call aarch64_simd_expand_args to update op[argc].
1486 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
1488 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
1491 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
1493 * gengtype.c (file_rules): Added rule for *.cc files.
1494 (get_output_file_with_visibility): Give fatal message when no
1497 2013-09-20 Renlin Li <renlin.li@arm.com>
1499 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
1500 (aarch64_expand_epilogue): Likewise.
1501 (aarch64_legitimize_reload_address): Likewise.
1503 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
1506 * expr.c (expand_assignment): Remove misalignp code path.
1508 2013-09-20 Marek Polacek <polacek@redhat.com>
1511 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
1512 TYPE_PRECISION. Add asserts.
1514 2013-09-20 Richard Biener <rguenther@suse.de>
1516 PR tree-optimization/58453
1517 * tree-loop-distribution.c (distribute_loop): Apply the cost
1518 model for -ftree-loop-distribute-patterns, too.
1520 2013-09-20 Richard Biener <rguenther@suse.de>
1523 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
1524 make type comparison less strict.
1526 2013-09-20 Alan Modra <amodra@gmail.com>
1528 * configure: Regenerate.
1529 * aclocal.m4: Regenerate.
1531 2013-09-20 Marek Polacek <polacek@redhat.com>
1534 * doc/extend.texi: Document that attribute used is meant to be used
1535 on variables with static storage duration.
1537 2013-09-19 Jakub Jelinek <jakub@redhat.com>
1539 PR tree-optimization/58472
1540 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
1541 simd_lane_access set inv_p = false.
1542 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
1543 the simduid magic VAR_DECL.
1545 2013-09-19 Jan Hubicka <jh@suse.cz>
1547 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
1549 2013-09-17 Jeff Law <law@redhat.com>
1551 * tree-ssa-dom.c (record_temporary_equivalences): New function
1552 split out of dom_opt_dom_walker::after_dom_children.
1553 (dom_opt_dom_walker::thread_across_edge): Move common code
1554 in here from dom_opt_dom_walker::after_dom_children.
1555 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
1557 2013-09-19 Jan Hubicka <jh@suse.cz>
1559 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
1560 (TARGET_GENERIC): Use PROCESOR_GENERIC
1561 (enum processor_type): Unify generic32 and 64.
1562 * i386.md (cpu): Likewise.
1563 * x86-tune.def (use_leave): Enable for generic32.
1564 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
1565 * athlon.md: Change generic64 to generic in all occurences.
1566 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
1567 (ix86_target_macros_internal): Likewise.
1568 * driver-i386.c (host_detect_local_cpu): Likewise.
1569 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
1571 (generic_memcpy, generic_memset, generic_cost): This one.
1572 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
1573 (m_GENERIC32, m_GENERIC64): Remove.
1574 (m_GENERIC): Turn into one flag.
1575 (processor_target): Unify generic tunnings.
1576 (ix86_option_override_internal): Replace generic32/64 by generic.
1577 (ix86_issue_rate): Likewise.
1578 (ix86_adjust_cost): Likewise.
1580 2013-09-19 Jan Hubicka <jh@suse.cz>
1582 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
1583 of speculative flag.
1585 2013-09-19 Jakub Jelinek <jakub@redhat.com>
1587 * omp-low.c (expand_omp_sections): Always pass len - 1 to
1588 GOMP_sections_start, even if !exit_reachable.
1590 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
1592 * lra-constraints.c (need_for_all_save_p): Use macro
1593 HARD_REGNO_CALL_PART_CLOBBERED.
1594 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
1595 set up pseudo conflict hard regs.
1597 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
1600 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
1603 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
1605 PR rtl-optimization/58438
1606 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
1607 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
1608 from previous subpasses.
1610 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
1612 * arm.c (arm_get_frame_offsets): Validate architecture supports
1613 LDRD/STRD before accepting the tuning preference.
1614 (arm_expand_prologue): Likewise.
1615 (arm_expand_epilogue): Likewise.
1617 2013-09-18 Richard Biener <rguenther@suse.de>
1619 PR tree-optimization/58417
1620 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
1621 have chrecs with symbols defined in the loop as operands.
1622 (chrec_fold_multiply): Likewise.
1623 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
1624 parameters before folding binary operations.
1625 (struct instantiate_cache_entry_hasher): Remove.
1626 (struct instantiate_cache_type): Use a pointer-map.
1627 (instantiate_cache_type::instantiate_cache_type): New function.
1628 (instantiate_cache_type::get): Likewise.
1629 (instantiate_cache_type::set): Likewise.
1630 (instantiate_cache_type::~instantiate_cache_type): Adjust.
1631 (get_instantiated_value_entry): Likewise.
1632 (global_cache): New global.
1633 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
1634 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
1635 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
1636 (instantiate_scev_name): Adjust.
1637 (instantiate_scev): Construct global instead of local cache.
1638 (resolve_mixers): Likewise.
1640 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
1641 Paolo Carlini <paolo.carlini@oracle.com>
1644 * doc/implement-cxx.texi: Fix references to the C++ standards.
1646 2013-09-18 Jakub Jelinek <jakub@redhat.com>
1648 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
1649 * tree-vect-data-refs.c (vect_analyze_data_refs): For
1650 simd_lane_access drs, update also DR_ALIGNED_TO.
1652 2013-09-18 Marek Polacek <polacek@redhat.com>
1655 * doc/extend.texi: Document no_sanitize_undefined attribute.
1656 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
1657 no_sanitize_undefined attribute.
1659 2013-09-18 Nick Clifton <nickc@redhat.com>
1661 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
1662 (ASM_DECLARE_FUNCTION_NAME): Define.
1664 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
1666 * compare-elim.c (find_comparison_dom_walker): New class
1667 (find_comparisons_in_bb): Rename to
1668 find_comparison_dom_walker::before_dom_children
1669 (find_comparisons): Adjust
1670 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
1672 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
1673 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
1674 (init_walk_dominator_tree): Remove declaration.
1675 (fini_walk_dominator_tree): Remove declaration.
1676 * fwprop.c (single_def_use_dom_walker): New class
1677 (single_def_use_enter_block): Convert to
1678 single_def_use_dom_walker::before_dom_children.
1679 (single_def_use_leave_block): Convert to
1680 single_def_use_dom_walker::after_dom_children.
1681 (build_single_def_use_links): Adjust.
1682 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
1684 (find_candidates_in_block): Convert to
1685 find_candidates_dom_walker::before_dom_children.
1686 (execute_strength_reduction): Adjust.
1687 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
1688 (sese_dom_walker): New class.
1689 (sese_dom_walker::sese_dom_walker): New constructor.
1690 (sese_dom_walker::~sese_dom_walker): New destructor.
1691 (build_sese_conditions_before): Convert to
1692 sese_dom_walker::before_dom_children.
1693 (build_sese_conditions_after): Convert to
1694 sese_dom_walker::after_dom_children.
1695 (build_poly_scop): Adjust
1696 * tree-into-ssa.c (rewrite_dom_walker): New class
1697 (rewrite_enter_block): Convert to
1698 rewrite_dom_walker::before_dom_children.
1699 (rewrite_leave_block): Convert to
1700 rewrite_dom_walker::after_dom_children.
1701 (rewrite_update_dom_walker): New class.
1702 (rewrite_update_enter_block): Convert to
1703 rewrite_update_dom_walker::before_dom_children.
1704 (rewrite_update_leave_block): Convert to
1705 rewrite_update_dom_walker::after_dom_children.
1706 (rewrite_blocks, rewrite_into_ssa): Adjust.
1707 (mark_def_dom_walker): New class.
1708 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
1709 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
1710 (mark_def_sites_blocks): Convert to
1711 mark_def_dom_walker::before_dom_children.
1712 (mark_def_site_blocks): Remove.
1713 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
1714 (tree_ssa_dominator_optimize): Adjust.
1715 (dom_thread_across_edge): Convert to method
1716 dom_opt_dom_walker::thread_across_edge.
1717 (dom_opt_enter_block): Convert to member function
1718 dom_opt_dom_walker::before_dom_children.
1719 (dom_opt_leave_block): Convert to member function
1720 dom_opt_dom_walker::after_dom_children.
1721 * tree-ssa-dse.c (dse_dom_walker): New class.
1722 (dse_enter_block): Convert to member function
1723 dse_dom_walker::before_dom_children.
1724 (tree_ssa_dse): Adjust.
1725 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
1726 (determine_invariantness_stmt): Convert to method
1727 invariantness_dom_walker::before_dom_children.
1728 (determine_invariantness): Remove
1729 (move_computations_dom_walker): New class.
1730 (move_computations_stmt): Convert to method
1731 move_computations_dom_walker::before_dom_children.
1732 (move_computations, tree_ssa_lim): Adjust.
1733 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
1734 (nt_init_block): Make method
1735 notrappping_dom_walker::before_dom_children.
1736 (nt_fini_block): Make
1737 method nontrapping_dom_walker::after_dom_children.
1738 (get_non_trapping): Adjust.
1739 * tree-ssa-pre.c (eliminate_dom_walker): New class.
1740 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
1741 (eliminate_leave_block): Make method.
1742 eliminate_dom_walker::after_dom_children.
1744 * tree-ssa-strlen.c (strlen_dom_walker): New class.
1745 (strlen_enter_block): Make method
1746 strlen_dom_walker::before_dom_children.
1747 (strlen_leave_block): Make
1748 method strlen_dom_walker::after_dom_children.
1749 (tree_ssa_strlen): Adjust.
1750 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
1751 (tree_ssa_uncprop): Adjust.
1752 (uncprop_leave_block): Make method
1753 uncprop_dom_walker::after_dom_children.
1754 (uncprop_leave_block): Make method
1755 uncprop_dom_walker::before_dom_children.
1757 2013-09-18 Bin Cheng <bin.cheng@arm.com>
1759 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
1760 branch in same basic block. Check both src and dest of the move insn.
1762 2013-09-17 Nick Clifton <nickc@redhat.com>
1764 * config/rl78/rl78-real.md (bf): New pattern.
1766 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
1767 (rl78_print_operand): Do not put a # before a %B.
1768 * config/rl78/rl78.opt: Tweak doc strings.
1770 2013-09-17 DJ Delorie <dj@redhat.com>
1772 * config/rl78/constraints.md (Wcv): Allow up to $r31.
1773 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
1774 (rl78_option_override): Likewise, if -mallregs.
1775 (is_virtual_register): Likewise.
1776 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
1777 (REGNO_OK_FOR_BASE_P): Likewise.
1778 * config/rl78/rl78.opt (-mallregs): New.
1780 2013-09-17 Nick Clifton <nickc@redhat.com>
1782 * config/rl78/rl78.c (need_to_save): Change return type to bool.
1783 For interrupt functions: save all call clobbered registers if the
1784 interrupt handler is not a leaf function.
1785 (rl78_expand_prologue): Always recompute the frame information.
1786 For interrupt functions: only select bank 0 if one of the bank 0
1787 registers is going to be psuhed.
1789 2013-09-17 DJ Delorie <dj@redhat.com>
1791 * config/rl78/constraints.md: For each W* constraint, rename to C*
1792 and create a W* constraint that checks for an optional ES: prefix
1794 * config/rl78/rl78.md (UNS_ES_ADDR): New.
1795 (es_addr): New. Used to wrap far addresses.
1796 * config/rl78/rl78-protos.h (rl78_es_addr): New.
1797 (rl78_es_base): New.
1798 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
1799 wrapped far addresses.
1800 (rl78_print_operand_1): Unwrap far addresses before processing.
1801 (rl78_lo16): Wrap far addresses in unspecs.
1802 (rl78_es_addr): New.
1803 (rl78_es_base): New.
1804 (insn_ok_now): Check for not-yet-wrapped far addresses.
1805 (transcode_memory_rtx): Properly re-wrap far addresses.
1807 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
1809 * config/sparc/t-rtems: Add leon3 multilibs.
1811 2013-09-17 Cong Hou <congh@google.com>
1813 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
1814 when checking the dot production pattern. The type of rhs operand
1815 of multiply is now checked correctly.
1817 2013-09-17 Jeff Law <law@redhat.com>
1819 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1820 edge implied equivalences into successor phis.
1821 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
1822 here from tree-ssa-threadedge.c.
1823 (mark_threaded_blocks): When threading through a joiner, if both
1824 successors of the joiner's clone reach the same block, verify the
1825 PHI arguments are equal. If not, cancel the jump threading request.
1826 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
1827 tree-ssa-threadupdate.c
1828 (thread_across_edge): Don't check PHI argument equality when
1829 threading through joiner block here.
1831 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
1833 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
1834 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
1835 (warn_uninit, warn_uninitialized_vars,
1836 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
1837 Move to tree-ssa-uninit.c.
1838 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c
1839 (has_undefined_value_p): New. Pass dependant parts of
1840 ssa_undefined_value_p.
1841 (uninit_undefined_value_p): Use has_undefined_value_p.
1842 (warn_uninit, warn_uninitialized_vars,
1843 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
1844 Move from tree-ssa.c.
1845 * tree-ssa.h: Adjust prototypes.
1847 2013-09-17 Jan Hubicka <jh@suse.cz>
1850 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
1851 * ipa-inline.c (can_inline_edge_p): Do not downgrade
1852 FUNCTION_NOT_OPTIMIZED.
1853 * ipa-inline-analysis.c (compute_inline_parameters): Function
1854 not optimized is not inlinable unless it is alwaysinline.
1855 (inline_analyze_function): Force calls in not optimized
1856 function not inlinable.
1858 2013-09-17 Jan Hubicka <jh@suse.cz>
1861 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
1863 * ipa.c (function_and_variable_visibility): Likewise.
1864 * ipa-profile.c (ipa_profile): Likewise.
1866 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
1869 * cgraph.c (cgraph_function_body_availability): Check for ifunc
1870 attribute, and don't inline the resolver in this case.
1872 2013-09-17 Teresa Johnson <tejohnson@google.com>
1874 * coverage.c (get_coverage_counts): Add missing newline.
1876 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1878 PR tree-optimization/58088
1879 * fold-const.c (mask_with_trailing_zeros): New function.
1880 (fold_binary_loc): Make sure we don't recurse infinitely
1881 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
1882 Use mask_with_trailing_zeros where appropriate.
1884 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
1886 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
1887 of 'prev' var to get better distance estimation.
1889 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
1891 * tree-inline.h (struct copy_body_data): Add transform_parameter.
1892 * tree-inline.c (is_parameter_of): New predicate.
1893 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
1894 a parameter has been remapped.
1895 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
1896 (optimize_inline_calls): Initialize transform_parameter.
1897 (copy_gimple_seq_and_replace_locals): Likewise.
1898 (tree_function_versioning): Likewise.
1899 (maybe_inline_call_in_expr): Likewise.
1901 2013-09-17 Nick Clifton <nickc@redhat.com>
1903 * config/msp430/msp430-protos.h: Add prototypes for new functions.
1904 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
1906 (is_attr_func): New function.
1907 (msp430_is_interrupt_func): New function.
1908 (is_naked_func): New function.
1909 (is_reentrant_func): New function.
1910 (is_critical_func): New function.
1911 (msp430_start_function): Add annotations for function attributes.
1912 (msp430_attr): New function.
1913 (msp430_attribute_table): New.
1914 (msp430_function_section): New function.
1915 (TARGET_ASM_FUNCTION_SECTION): Define.
1916 (msp430_builtin): New enum.
1917 (msp430_init_builtins): New function.
1918 (msp430_builtin_devl): New function.
1919 (msp430_expand_builtin): New function.
1920 (TARGET_INIT_BUILTINS): Define.
1921 (TARGET_EXPAND_BUILTINS): Define.
1922 (TARGET_BUILTIN_DECL): Define.
1923 (msp430_expand_prologue): Add support for naked, interrupt,
1924 critical and reentrant functions.
1925 (msp430_expand_epilogue): Likewise.
1926 (msp430_print_operand): Handle 'O' character.
1927 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
1929 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
1930 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
1931 (pushm): Use a 'n' rather than an 'i' constraint.
1932 (msp_return): Add generation of the interrupt return instruction.
1933 (disable_interrupts): New pattern.
1934 (enable_interrupts): New pattern.
1935 (push_intr_state): New pattern.
1936 (pop_intr_state): New pattern.
1937 (bic_SR): New pattern.
1938 (bis_SR): New pattern.
1939 * doc/extend.texi: Document MSP430 function attributes and builtin
1942 2013-09-17 Richard Biener <rguenther@suse.de>
1944 PR tree-optimization/58432
1945 * tree-loop-distribution.c (tree_loop_distribution): Also
1946 scan PHIs for outside loop uses and seed a partition from them.
1948 2013-09-17 Bin Cheng <bin.cheng@arm.com>
1950 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
1951 (restructure_reference): Call backtrace_base_for_ref.
1953 2013-09-17 Alan Modra <amodra@gmail.com>
1956 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
1959 2013-09-16 DJ Delorie <dj@redhat.com>
1961 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
1962 vregs location for RL78/G10.
1963 (rl78_expand_prologue): Avoid SEL on G10.
1964 (rl78_expand_epilogue): Likewise.
1965 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
1966 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
1967 __RL78_G10__ when appropriate.
1968 (ASM_SPEC): Pass -mg10 along to the assembler.
1969 * config/rl78/rl78.md (sel_rb): Disable for G10.
1970 * config/rl78/rl78.opt: Add -mg10 option.
1971 * config/rl78/t-rl78: Add -mg10 multilib.
1973 2013-09-16 Xinliang David Li <davidxl@google.com>
1975 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
1976 * omp-low.c (omp_max_vf): Ditto.
1977 (expand_omp_simd): Ditto.
1978 * tree-vectorizer.c (vectorize_loops): Ditto.
1979 (gate_vect_slp): Ditto.
1980 (gate_increase_alignment): Ditto.
1981 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
1982 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
1983 (gate_tree_vectorize): Name change.
1984 (tree_vectorize): Ditto.
1985 (pass_vectorize::gate): Call new function.
1986 (pass_vectorize::execute): Ditto.
1987 * opts.c: O3 default setting change.
1988 (finish_options): Check new flag.
1989 * doc/invoke.texi: Document new flags.
1990 * common.opt: New flags.
1992 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
1994 * doc/tm.texi.in (Cond Exec Macros): Remove node.
1995 (Condition Code): Don't reference it.
1996 * doc/tm.texi: Regenerate.
1998 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
2001 * lra-constraints.c (undo_optional_reloads): Consider all optional
2002 reload even if it did not get a hard reg.
2004 2013-09-16 Teresa Johnson <tejohnson@google.com>
2006 * dumpfile.c (dump_loc): Remove newline emission.
2007 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
2008 emission to dump_printf_loc calls where missing.
2009 (vect_mark_for_runtime_alias_test): Ditto.
2010 (vect_analyze_data_ref_dependence): Ditto.
2011 (vect_analyze_data_ref_dependences): Ditto.
2012 (vect_slp_analyze_data_ref_dependence): Ditto.
2013 (vect_slp_analyze_data_ref_dependences): Ditto.
2014 (vect_compute_data_ref_alignment): Ditto.
2015 (vect_update_misalignment_for_peel): Ditto.
2016 (vect_verify_datarefs_alignment): Ditto.
2017 (vector_alignment_reachable_p): Ditto.
2018 (vect_get_data_access_cost): Ditto.
2019 (vect_enhance_data_refs_alignment): Ditto.
2020 (vect_find_same_alignment_drs): Ditto.
2021 (vect_analyze_data_refs_alignment): Ditto.
2022 (vect_analyze_group_access): Ditto.
2023 (vect_analyze_data_ref_access): Ditto.
2024 (vect_analyze_data_ref_accesses): Ditto.
2025 (vect_prune_runtime_alias_test_list): Ditto.
2026 (vect_analyze_data_refs): Ditto.
2027 (vect_create_addr_base_for_vector_ref): Ditto.
2028 (vect_create_data_ref_ptr): Ditto.
2029 (vect_grouped_store_supported): Ditto.
2030 (vect_grouped_load_supported): Ditto.
2031 * value-prof.c (check_counter): Ditto.
2032 (check_ic_target): Ditto.
2033 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
2034 (vect_recog_widen_mult_pattern): Ditto.
2035 (vect_recog_widen_sum_pattern): Ditto.
2036 (vect_recog_over_widening_pattern): Ditto.
2037 (vect_recog_widen_shift_pattern): Ditto.
2038 (vect_recog_rotate_pattern): Ditto.
2039 (vect_recog_vector_vector_shift_pattern): Ditto.
2040 (vect_recog_divmod_pattern): Ditto.
2041 (vect_recog_mixed_size_cond_pattern): Ditto.
2042 (vect_recog_bool_pattern): Ditto.
2043 (vect_pattern_recog_1): Ditto.
2044 (vect_pattern_recog): Ditto.
2045 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
2046 (vect_is_simple_iv_evolution): Ditto.
2047 (vect_analyze_scalar_cycles_1): Ditto.
2048 (vect_get_loop_niters): Ditto.
2049 (vect_analyze_loop_1): Ditto.
2050 (vect_analyze_loop_form): Ditto.
2051 (vect_analyze_loop_operations): Ditto.
2052 (vect_analyze_loop_2): Ditto.
2053 (vect_analyze_loop): Ditto.
2054 (report_vect_op): Ditto.
2055 (vect_is_slp_reduction): Ditto.
2056 (vect_is_simple_reduction_1): Ditto.
2057 (vect_get_known_peeling_cost): Ditto.
2058 (vect_estimate_min_profitable_iters): Ditto.
2059 (vect_model_reduction_cost): Ditto.
2060 (vect_model_induction_cost): Ditto.
2061 (get_initial_def_for_induction): Ditto.
2062 (vect_create_epilog_for_reduction): Ditto.
2063 (vectorizable_reduction): Ditto.
2064 (vectorizable_induction): Ditto.
2065 (vectorizable_live_operation): Ditto.
2066 (vect_loop_kill_debug_uses): Ditto.
2067 (vect_transform_loop): Ditto.
2068 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
2069 (vect_stmt_relevant_p): Ditto.
2070 (process_use): Ditto.
2071 (vect_mark_stmts_to_be_vectorized): Ditto.
2072 (vect_model_simple_cost): Ditto.
2073 (vect_model_promotion_demotion_cost): Ditto.
2074 (vect_model_store_cost): Ditto.
2075 (vect_get_store_cost): Ditto.
2076 (vect_model_load_cost): Ditto.
2077 (vect_get_load_cost): Ditto.
2078 (vect_init_vector_1): Ditto.
2079 (vect_get_vec_def_for_operand): Ditto.
2080 (vect_finish_stmt_generation): Ditto.
2081 (vectorizable_call): Ditto.
2082 (vectorizable_conversion): Ditto.
2083 (vectorizable_assignment): Ditto.
2084 (vectorizable_shift): Ditto.
2085 (vectorizable_operation): Ditto.
2086 (vectorizable_store): Ditto.
2087 (vectorizable_load): Ditto.
2088 (vectorizable_condition): Ditto.
2089 (vect_analyze_stmt): Ditto.
2090 (vect_transform_stmt): Ditto.
2091 (vect_is_simple_use): Ditto.
2092 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
2093 (vect_can_advance_ivs_p): Ditto.
2094 (vect_update_ivs_after_vectorizer): Ditto.
2095 (vect_do_peeling_for_loop_bound): Ditto.
2096 (vect_gen_niters_for_prolog_loop): Ditto.
2097 (vect_update_inits_of_drs): Ditto.
2098 (vect_create_cond_for_alias_checks): Ditto.
2099 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
2100 (vect_build_slp_tree_1): Ditto.
2101 (vect_supported_load_permutation_p): Ditto.
2102 (vect_analyze_slp_instance): Ditto.
2103 (vect_analyze_slp): Ditto.
2104 (vect_make_slp_decision): Ditto.
2105 (vect_detect_hybrid_slp): Ditto.
2106 (vect_bb_vectorization_profitable_p): Ditto.
2107 (vect_slp_analyze_bb_1): Ditto.
2108 (vect_update_slp_costs_according_to_vf): Ditto.
2109 (vect_get_mask_element): Ditto.
2110 (vect_transform_slp_perm_load): Ditto.
2111 (vect_schedule_slp_instance): Ditto.
2112 (vect_schedule_slp): Ditto.
2113 (vect_slp_transform_bb): Ditto.
2114 * profile.c (read_profile_edge_counts): Ditto.
2115 (compute_branch_probabilities): Ditto.
2116 * coverage.c (get_coverage_counts): Ditto.
2118 2013-09-16 Diego Novillo <dnovillo@google.com>
2120 * tree-core.h: Add missing comment lines from refactoring of tree.h.
2122 2013-09-16 Jan Hubicka <jh@suse.cz>
2124 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
2125 abstract functions; for static functions check the presence of body.
2127 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
2129 * config/aarch64/aarch64-simd-builtins.def (fma): New.
2130 * config/aarch64/aarch64-simd.md
2131 (aarch64_mla_elt<mode>): New.
2132 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
2133 (aarch64_mls_elt<mode>): Likewise.
2134 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
2135 (aarch64_fma4_elt<mode>): Likewise.
2136 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
2137 (aarch64_fma4_elt_to_128v2df): Likewise.
2138 (aarch64_fma4_elt_to_64df): Likewise.
2139 (fnma<mode>4): Likewise.
2140 (aarch64_fnma4_elt<mode>): Likewise.
2141 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
2142 (aarch64_fnma4_elt_to_128v2df): Likewise.
2143 (aarch64_fnma4_elt_to_64df): Likewise.
2144 * config/aarch64/iterators.md (VDQSF): New.
2145 * config/aarch64/arm_neon.h
2146 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
2147 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
2149 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
2151 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
2152 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
2153 (aarch64_mul3_elt_to_128df): Likewise.
2154 (aarch64_mul3_elt_to_64v2df): Likewise.
2155 * config/aarch64/iterators.md (VEL): Also handle DFmode.
2157 (VMUL_CHANGE_NLANES) Likewise.
2160 * config/aarch64/arm_neon.h
2161 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
2163 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
2165 * config/aarch64/arm_neon.h
2166 (vcvtx_high_f32_f64): Fix parameters.
2168 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2169 Uros Bizjak <ubizjak@gmail.com>
2171 * config/alpha.c: Include tree-ssanames.h.
2173 2013-09-16 Richard Biener <rguenther@suse.de>
2175 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
2176 (dot_rdg_1): Handle control_dd.
2177 (create_edge_for_control_dependence): New function.
2178 (create_rdg_edges): Add control dependences if asked for.
2179 (build_rdg): Likewise.
2180 (generate_loops_for_partition): If there are not necessary
2181 control stmts remove all their dependencies.
2182 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
2183 (distribute_loop): Pass on control dependences.
2184 (tree_loop_distribution): Compute control dependences and remove
2185 restriction on number of loop nodes.
2187 2013-09-16 Jakub Jelinek <jakub@redhat.com>
2189 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
2192 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
2194 * cse.c (try_const_anchors): Punt on CC modes.
2196 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2198 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
2200 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
2203 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
2206 2013-09-14 Iain Sandoe <iain@codesourcery.com>
2209 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
2211 (darwin_objc1_section): Likewise.
2212 (darwin_file_end): Emit Image Info section when required.
2214 2013-09-14 Jan Hubicka <jh@suse.cz>
2216 * tree-into-ssa.c (gate_into_ssa): New.
2217 (pass_data_build_ssa): Use it.
2218 * cgraph.h (expand_thunk): Update prototype.
2219 * cgraphunit.c (analyze_function): Expand thunks early.
2220 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
2221 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
2222 set lowered flag; do not add new function.
2223 (assemble_thunks_and_aliases): Update.
2224 * tree-ssa.c (gate_init_datastructures): New gate.
2225 (pass_data_init_datastructures): Use it.
2227 2013-09-14 Iain Sandoe <iain@codesourcery.com>
2230 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
2231 xmm register set described in the psABI.
2233 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
2235 * dwarf2out.c (should_emit_struct_debug): Add check
2236 for type_decl variable is not NULL.
2238 2013-09-13 Jacek Caban <jacek@codeweavers.com>
2240 * config.gcc: Use new winnt-c.c target hooks
2241 * config/t-winnt: New file
2242 * config/winnt-c.c: New file
2243 * doc/tm.texi.in: Document new hook
2244 * doc/tm.texi: Regenerated
2246 2013-09-13 Jan Hubicka <jh@suse.cz>
2249 * ipa-inline.c (check_callers): New function.
2250 (check_caller_edge): Remove.
2251 (want_inline_function_to_all_callers_p): Also permit alises that are
2253 (inline_to_all_callers): Terminate the walk when devirtualization
2254 introduce new calls.
2256 2013-09-13 Jan Hubicka <jh@suse.cz>
2258 * ipa-inline-analysis.c (struct growth_data): Add node.
2259 (do_estimate_growth_1): Fix detection of recursion.
2261 2013-09-13 Jakub Jelinek <jakub@redhat.com>
2263 PR tree-optimization/58392
2264 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
2265 to avoid shadowing of outer loop variable. If
2266 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
2267 replace_by_duplicate_decl simduid of loops that have it set and
2268 set dest_cfun->has_simduid_loops and/or
2269 dest_cfun->has_force_vect_loops.
2270 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
2271 instead of maybe_lookup_decl.
2272 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
2273 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
2274 call. Copy over force_vect and copy and remap simduid. Set
2275 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
2276 (copy_cfg_body): Remove blocks_to_copy argument. Use
2277 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
2278 caller. Don't set cfun->has_simduid_loops and/or
2279 cfun->has_force_vect_loops here.
2280 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
2282 (expand_call_inline, tree_function_versioning): Adjust copy_body
2285 2013-09-13 Martin Jambor <mjambor@suse.cz>
2288 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
2289 the assert if the edge was a speculative one.
2291 2013-09-13 Richard Biener <rguenther@suse.de>
2293 * tree-data-ref.h (known_dependences_p): Move here ...
2294 * tree-loop-distribution.c (known_dependences_p): ... from here.
2295 (dump_rdg_component, debug_rdg_component): Remove.
2297 (generate_loops_for_partition): Use gimple_uid instead of
2298 relying on matching stmt visit order.
2299 (rdg_build_partitions): Take starting stmt vector.
2300 (ldist_gen): Merge into ...
2301 (distribute_loop): ... this function. Do not compute starting
2303 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
2305 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2307 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
2308 Set type attribute correctly. Set predicable_short_it attribute.
2309 (cmpsi_shiftsi): Remove %? from output template.
2311 2013-09-13 Richard Biener <rguenther@suse.de>
2313 * tree-loop-distribution.c (struct rdg_component,
2314 rdg_defs_used_in_other_loops_p, free_rdg_components,
2315 rdg_build_components): Remove.
2316 (stmts_from_loop): Do not record virtual PHIs.
2317 (generate_loops_for_partition): Skip virtual PHIs.
2318 (build_rdg_partition_for_component): Rename to ...
2319 (build_rdg_partition_for_vertex): ... this and adjust.
2320 (rdg_build_partitions): Take a vector of starting vertices
2321 instead of components. Remove unnecessary leftover handling.
2322 (ldist_gen): Do not build components or record other stores.
2323 (distribute_loop): Do not distribute loops containing stmts
2326 2013-09-13 Christian Bruel <christian.bruel@st.com>
2329 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
2331 2013-09-13 Kai Tietz <ktietz@redhat.com>
2333 * config.gcc: Separate cases for mingw and cygwin targets,
2334 and add 64-bit cygwin target case.
2336 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
2337 dll-export inline-functions.
2338 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
2341 2013-09-13 Jeff Law <law@redhat.com>
2345 2013-09-06 Jeff Law <law@redhat.com>
2347 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2348 edge implied equivalences into successor phis.
2350 2013-09-12 DJ Delorie <dj@redhat.com>
2352 * config/rl78/rl78-virt.md: Change from | to \; for asm line
2355 2013-09-12 Brooks Moses <bmoses@google.com>
2358 * Makefile.in: Do not install driver binaries in $(target)/bin.
2360 2013-09-12 DJ Delorie <dj@redhat.com>
2362 * config/rl78/rl78.opt (mrelax): New.
2363 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
2364 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
2366 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
2368 (rl78_expand_epilogue): Likewise.
2370 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
2373 * lra-eliminations.c (remove_reg_equal_offset_note): New.
2374 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
2375 pointer elimination with using remove_reg_equal_offset_note.
2377 2013-09-12 DJ Delorie <dj@redhat.com>
2379 * config/msp430/: New port.
2380 * config.gcc (msp430): Added.
2381 * doc/invoke.texi: Document MSP430 options.
2382 * doc/install.texi: Document msp430-elf
2383 * doc/md.texi: Document msp430-elf
2384 * doc/contrib.texi: Document msp430-elf
2386 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
2389 2013-09-12 Martin Jambor <mjambor@suse.cz>
2392 * ipa-prop.c (remove_described_reference): Give up if the edge in the
2393 reference descriptor is NULL.
2394 (ipa_edge_removal_hook): If owning a reference descriptor, set its
2397 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
2399 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
2400 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
2401 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
2402 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
2403 make_temp_ssa_name): move to tree-ssanames.h
2404 * tree-ssa-alias.h: Move prototype.
2405 * tree-ssa.h: Include tree-ssanames.h.
2406 * tree-ssanames.c (FREE_SSANAMES): Move to here.
2407 * tree-ssanames.h: New. Move items from tree-flow*.h
2408 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
2410 2013-09-12 Richard Biener <rguenther@suse.de>
2412 PR tree-optimization/58404
2413 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
2414 propagate non-invariant addresses into dereferences wrapped
2415 in component references.
2417 2013-09-12 Richard Biener <rguenther@suse.de>
2419 PR tree-optimization/58402
2420 * passes.def: Move pass_late_warn_uninitialized later.
2422 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
2424 * tree-ssa.h: New. Move content from tree-flow.h and
2426 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
2427 Move prototypes belonging to tree-ssa.c.
2428 * tree-flow-inline.h (redirect_edge_var_map_def,
2429 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
2431 * gimple.h: Adjust prototypes.
2432 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
2434 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
2435 * tree.h: Move prototype to tree-ssa.h.
2436 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
2437 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
2438 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
2439 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
2440 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
2441 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
2442 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
2443 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
2444 graphite-blocking.c, graphite-clast-to-gimple.c,
2445 graphite-dependences.c, graphite-interchange.c,
2446 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
2447 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
2448 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
2449 ipa-reference.c, ipa-split.c, ipa-utils.c,
2450 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
2451 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
2452 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
2453 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2454 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
2455 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
2456 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
2457 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
2458 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
2459 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
2460 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
2461 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
2462 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
2463 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
2464 value-prof.c, var-tracking.c,
2465 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
2467 2013-09-12 Richard Biener <rguenther@suse.de>
2469 PR tree-optimization/58396
2470 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
2471 (build_rdg): Take a loop-nest parameter, fix memleaks.
2472 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
2474 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
2476 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
2479 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
2481 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
2482 The FMA instruction names should have a 'v' prefix.
2484 2013-09-12 Richard Biener <rguenther@suse.de>
2486 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
2487 (dot_rdg): Use popen instead of system in optional code.
2488 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
2489 (already_processed_vertex_p): Adjust.
2490 (has_anti_or_output_dependence, predecessor_has_mem_write,
2491 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
2492 rdg_flag_uses): Remove.
2493 (rdg_flag_vertex): Simplify.
2494 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
2496 (build_rdg_partition_for_component): Process the first vertex
2497 of a component only.
2498 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
2500 2013-09-12 Alan Modra <amodra@gmail.com>
2502 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
2504 2013-09-11 DJ Delorie <dj@redhat.com>
2505 Nick Clifton <nickc@redhat.com>
2507 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
2508 (rl78_stack_based_mem): New.
2509 * config/rl78/constraints.md (Iv08): New.
2518 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
2520 (movsi): Change from expand to insn-and-split.
2521 (ashrsi3): Clobber AX.
2525 * config/rl78/rl78.md (CC_REG): Fix.
2526 (addsi3): Allow memory and immediate operands.
2527 (addsi3_internal): Split into...
2528 (addsi3_internal_virt): ...new, and ...
2529 (addsi3_internal_real): ...new.
2531 (subsi3_internal_virt): New.
2532 (subsi3_internal_real): New.
2533 (mulsi3): Add memory operand.
2534 (mulsi3_rl78): Likewise.
2535 (mulsi3_g13): Likewise.
2536 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
2537 (cbranchqi4_real): Add more constraint options.
2538 (cbranchhi4_real): Expand pattern.
2539 (cbranchhi4_real_signed): New.
2540 (cbranchhi4_real_inverted): New.
2541 (cbranchsi4_real_lt): New.
2542 (cbranchsi4_real_ge): New.
2543 (cbranchsi4_real_signed): New.
2544 (cbranchsi4_real): New.
2546 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
2548 (lshrsi3_virt): Likewise.
2549 (ashlsi3_virt): Likewise.
2550 (cbranchqi4_virt_signed): New.
2551 (cbranchhi4_virt_signed): New.
2552 (cbranchsi4_virt): New.
2553 * config/rl78/rl78.c: Whitespace fixes throughout.
2554 (move_elim_pass): New.
2555 (pass_data_rl78_move_elim): New.
2556 (pass_rl78_move_elim): New.
2557 (make_pass_rl78_move_elim): New.
2558 (rl78_devirt_info): Run devirt earlier.
2559 (rl78_move_elim_info): New.
2560 (rl78_asm_file_start): Register it.
2561 (rl78_split_movsi): New.
2562 (rl78_as_legitimate_address): Allow virtual base registers when
2564 (rl78_addr_space_convert): Remove spurious debug stuff.
2565 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
2566 (rl78_print_operand): More cases for not printing '#'.
2567 (rl78_expand_compare): Remove most of the logic.
2568 (content_memory): New.
2569 (clear_content_memory): New.
2570 (get_content_index): New.
2571 (get_content_name): New.
2572 (display_content_memory): New.
2573 (update_content): New.
2574 (record_content): New.
2575 (already_contains): New.
2576 (insn_ok_now): Re-recog insns with virtual registers.
2577 (add_postponed_content_update): New.
2578 (process_postponed_content_update): New.
2579 (gen_and_emit_move): New.
2580 (transcode_memory_rtx): Record new location content.
2581 Use gen_and_emit_move.
2582 (force_into_acc): New.
2583 (move_to_acc): Use gen_and_emit_move.
2584 (move_from_acc): Likewise.
2585 (move_acc_to_reg): Likewise.
2586 (move_to_x): Likewise.
2587 (move_to_hl): Likewise.
2588 (move_to_de): Likewise.
2589 (rl78_alloc_physical_registers_op1): Record location content.
2590 (has_constraint): New.
2591 (rl78_alloc_physical_registers_op2): Record location content.
2593 (rl78_alloc_physical_registers_ro1): Likewise.
2594 (rl78_alloc_physical_registers_cmp): Likewise.
2595 (rl78_alloc_physical_registers_umul): Likewise.
2596 (rl78_alloc_address_registers_macax): New.
2597 (rl78_alloc_physical_registers): Initialize and set location
2598 content memory as needed.
2599 (rl78_reorg): Make sure split2 is called.
2600 (rl78_rtx_costs): New.
2602 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
2604 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
2605 for (not (neg ...)) and (neg (not ...)) cases.
2607 2013-09-11 Richard Biener <rguenther@suse.de>
2610 * passes.def: Split critical edges before late uninit warning passes.
2611 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
2613 2013-09-11 Jakub Jelinek <jakub@redhat.com>
2615 PR tree-optimization/58385
2616 * fold-const.c (build_range_check): If both low and high are NULL,
2617 use omit_one_operand_loc to preserve exp side-effects.
2619 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2621 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
2623 2013-09-11 Richard Biener <rguenther@suse.de>
2625 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
2626 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
2627 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
2628 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
2629 stmts_from_loop, known_dependences_p, build_empty_rdg,
2630 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
2631 * tree-loop-distribution.c: ... here.
2632 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
2633 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
2634 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
2635 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
2636 * tree-loop-distribution.c: ... here.
2637 * tree-loop-distribution.c: Include gimple-pretty-print.h.
2638 (struct partition_s): Add loops member.
2639 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
2640 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
2641 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
2643 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2644 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2645 Sergey Lega <sergey.s.lega@intel.com>
2646 Anna Tikhonova <anna.tikhonova@intel.com>
2647 Ilya Tocar <ilya.tocar@intel.com>
2648 Andrey Turetskiy <andrey.turetskiy@intel.com>
2649 Ilya Verbin <ilya.verbin@intel.com>
2650 Kirill Yukhin <kirill.yukhin@intel.com>
2651 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2653 * config/i386/constraints.md (k): New.
2655 * config/i386/i386.c (const regclass_map): Add new mask registers.
2656 (dbx_register_map): Ditto.
2657 (dbx64_register_map): Ditto.
2658 (svr4_dbx_register_map): Ditto.
2659 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
2661 (ix86_preferred_reload_class): Disable constants for mask registers.
2662 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
2663 (ix86_hard_regno_mode_ok): Support new mask registers.
2664 (x86_order_regs_for_local_alloc): Ditto.
2665 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
2666 (FIXED_REGISTERS): Add new mask registers.
2667 (CALL_USED_REGISTERS): Ditto.
2668 (REG_ALLOC_ORDER): Ditto.
2669 (VALID_MASK_REG_MODE): New.
2670 (FIRST_MASK_REG): Ditto.
2671 (LAST_MASK_REG): Ditto.
2672 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
2673 (MAYBE_MASK_CLASS_P): New.
2674 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
2675 (REG_CLASS_CONTENTS): Ditto.
2676 (MASK_REGNO_P): New.
2677 (ANY_MASK_REG_P): Ditto.
2678 (HI_REGISTER_NAMES): Add new mask registers.
2679 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
2680 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
2682 (attribute "type"): Add mskmov, msklog.
2683 (attribute "length_immediate"): Support them.
2684 (attribute "memory"): Ditto.
2685 (attribute "prefix_0f"): Ditto.
2686 (*movhi_internal): Support new mask registers.
2687 (*movqi_internal): Ditto.
2688 (define_split): Split out clobber pattern is a logic
2689 insn on mask registers.
2690 (*k<logic><mode>): New.
2691 (*andhi_1): Extend to support mask regs.
2692 (*andqi_1): Extend to support mask regs.
2694 (define_split): Split and-not to and and not if operands
2696 (*<code><mode>_1): Separate HI mode to new pattern...
2697 (*<code>hi_1): This.
2698 (*<code>qi_1): Extend to support mask regs.
2700 (kortestzhi): Ditto.
2701 (kortestchi): Ditto.
2703 (*one_cmpl<mode>2_1): Remove HImode and handle it...
2704 (*one_cmplhi2_1): ...Here, now with mask registers support.
2705 (*one_cmplqi2_1): Support new mask registers.
2706 (HI/QImode arithmetics splitter): Don't split if mask registers
2708 (HI/QImode not splitter): Ditto.
2709 * config/i386/predicated.md (mask_reg_operand): New.
2710 (general_reg_operand): Ditto.
2712 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2714 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
2715 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
2717 2013-09-10 Jeff Law <law@redhat.com>
2719 PR tree-optimization/58380
2720 * tree-ssa-threadupdate.c (thread_block): Recognize another case
2721 of threading through a buried loop header.
2723 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
2724 return value for single successor case.
2726 2013-09-10 Jan Hubicka <jh@suse.cz>
2728 * ipa-devirt.c (ipa_devirt): Enable with LTO.
2730 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
2733 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
2734 support conditional execution.
2735 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
2737 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
2739 * lra.c (lra): Clear lra_optional_reload_pseudos before every
2741 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
2742 Check destination too to check move insn.
2743 (undo_optional_reloads): Add check that the original peudo did not
2744 changed its allocation and the optional reload was inherited on last
2745 inheritance pass. Break loop after deciding to keep optional reload.
2746 (lra_undo_inheritance): Add check that inherited pseudo still in
2749 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
2751 * config/aarch64/aarch64.md (generic_sched): New.
2752 * config/aarch64/aarch64-generic.md (load): Make conditional
2753 on generic_sched attribute.
2754 (nonload): Likewise.
2756 2013-09-10 Jan Hubicka <jh@suse.cz>
2758 * lto-cgraph.c: Include ipa-utils.h.
2759 (compute_ltrans_boundary): Also add possible targets into the boundary.
2761 2013-09-10 Jan Hubicka <jh@suse.cz>
2763 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
2764 VAR_DECL of vtable rather than full expression.
2766 2013-09-10 Jan Hubicka <jh@suse.cz>
2767 Paolo Carlini <paolo.carlini@oracle.com>
2769 * cgraphunit.c (analyze_functions): Save input_location, set it
2770 to UNKNOWN_LOCATION and restore it at the end.
2772 2013-09-10 Martin Jambor <mjambor@suse.cz>
2774 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
2775 represented by a thunk.
2777 2013-09-10 Jeff Law <law@redhat.com>
2779 PR tree-optimization/58343
2780 * tree-ssa-threadupdate.c (thread_block): Identify and disable
2781 jump threading requests through loop headers buried in the middle
2782 of a jump threading path.
2784 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
2785 in return value/type.
2787 2013-09-10 Jakub Jelinek <jakub@redhat.com>
2789 PR rtl-optimization/58365
2790 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
2791 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
2794 2013-09-10 Richard Biener <rguenther@suse.de>
2796 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
2797 * tree-data-ref.c (create_rdg_vertices): Collect all data
2798 references, signal failure to the caller, use data-ref API.
2799 (build_rdg): Compute data references only once. Maintain lifetime
2800 of data references and data dependences from within RDG.
2801 (free_rdg): Free dependence relations.
2802 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
2803 inventing extra dependences.
2804 (distribute_loop): Update for RDG API changes.
2806 2013-09-10 Kai Tietz <ktietz@redhat.com>
2808 * doc/invoke.texi (fms-extensions): Document changed
2809 behavior for ms-abi targets.
2810 * config/i386/i386.c (ix86_option_override_internal):
2811 Set default value of option -fms-extension for ms-abi targets.
2813 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2815 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
2817 2013-09-10 Alan Modra <amodra@gmail.com>
2820 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
2822 2013-09-10 Alan Modra <amodra@gmail.com>
2824 * config/rs6000/predicates.md (add_cint_operand): New.
2825 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
2826 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
2827 using add_cint_operand.
2828 (largetoc_high_plus_aix): Likewise.
2830 2013-09-09 Jakub Jelinek <jakub@redhat.com>
2832 PR tree-optimization/58364
2833 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
2834 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
2835 the current range can't be an unconditional true or false.
2837 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
2839 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
2841 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
2843 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
2845 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
2848 * doc/invoke.texi (-Wdelete-incomplete): Document it.
2850 2013-09-09 Ian Bolton <ian.bolton@arm.com>
2852 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
2853 NO_REGS for immediate that can't be moved directly into FP_REGS.
2855 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2857 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
2858 comparison with negated operand.
2859 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
2862 2013-09-09 Richard Biener <rguenther@suse.de>
2865 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
2866 of a subloop record all its block as affecting loop-closed SSA form.
2868 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2870 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
2871 of an rtx/bitpos pair.
2872 (store_fixed_bit_field): Update accordingly.
2874 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2876 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
2878 * builtins.c (expand_errno_check): Likewise.
2879 * dwarf2cfi.c (init_return_column_size): Likewise.
2880 * except.c (sjlj_mark_call_sites): Likewise.
2881 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
2882 * lra-constraints.c (emit_inc): Likewise.
2883 * ree.c (combine_set_extension): Likewise.
2884 * regmove.c (fixup_match_2): Likewise.
2885 * reload1.c (inc_for_reload): Likewise.
2887 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2889 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
2890 (force_to_mode, simplify_shift_const_1, simplify_comparison):
2891 Use gen_int_mode with the mode of the associated simplify_* call.
2892 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
2893 * expmed.c (expand_shift_1): Likewise.
2894 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2895 * loop-iv.c (iv_number_of_iterations): Likewise.
2896 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2897 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
2899 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2901 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
2902 of the associated expand_* call.
2903 (asan_emit_stack_protection): Likewise.
2904 * builtins.c (round_trampoline_addr): Likewise.
2905 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
2906 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
2907 (emit_store_flag): Likewise.
2908 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
2909 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
2911 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2912 * ifcvt.c (noce_try_store_flag_constants): Likewise.
2913 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2914 * modulo-sched.c (generate_prolog_epilog): Likewise.
2915 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
2916 (expand_ctz, expand_ffs, expand_unop): Likewise.
2918 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2920 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
2921 of the associated gen_rtx_* call.
2922 * caller-save.c (init_caller_save): Likewise.
2923 * combine.c (find_split_point, make_extraction): Likewise.
2924 (make_compound_operation): Likewise.
2925 * dwarf2out.c (mem_loc_descriptor): Likewise.
2926 * explow.c (plus_constant, probe_stack_range): Likewise.
2927 * expmed.c (expand_mult_const): Likewise.
2928 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
2929 * reload1.c (init_reload): Likewise.
2930 * valtrack.c (cleanup_auto_inc_dec): Likewise.
2931 * var-tracking.c (adjust_mems): Likewise.
2932 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
2933 rather than gen_rtx_fmt_ee.
2935 2013-09-09 Jan Hubicka <jh@suse.cz>
2938 * value-prof.c (gimple_ic): Copy also abnormal edges.
2940 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2942 * asan.c (asan_shadow_cst): Use gen_int_mode.
2944 2013-09-08 Jan Hubicka <jh@suse.cz>
2946 * ipa-profile.c: Add toplevel comment.
2947 (ipa_propagate_frequency_1): Be more conservative when profile is read.
2948 (contains_hot_call_p): New function.
2949 (ipa_propagate_frequency): Set frequencies based on counts when
2951 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
2952 profile; do not tamper with profile after inlining if it is read.
2954 2013-09-08 Jan Hubicka <jh@suse.cz>
2956 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
2959 2013-09-08 Jan Hubicka <jh@suse.cz>
2961 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
2964 2013-09-08 Jeff Law <law@redhat.com>
2967 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
2970 2013-09-08 Andi Kleen <ak@linux.intel.com>
2972 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
2974 2013-09-08 Jan Hubicka <jh@suse.cz>
2976 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
2978 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
2980 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
2981 for non-debug insns.
2982 * lra.c (new_insn_reg): Take the containing insn as a parameter.
2983 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
2984 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
2987 2013-09-08 Jan Hubicka <jh@suse.cz>
2989 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
2990 targets and devirtualize to BUILT_IN_UNREACHABLE.
2991 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
2992 * ipa.c (walk_polymorphic_call_targets): New function.
2993 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
2994 functions; use the new timevar.
2995 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
2996 was removed from the program.
2997 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
2998 not consider it in the walk when its vtable is dead.
2999 (possible_polymorphic_call_targets_1): Pass anonymous flag to
3001 (devirt_variable_node_removal_hook): New function.
3002 (possible_polymorphic_call_targets): Also register
3003 devirt_variable_node_removal_hook.
3004 (ipa_devirt): Do not do non-speculative devirtualization.
3005 (gate_ipa_devirt): One execute if devirtualizing speculatively.
3007 2013-09-08 Jan Hubicka <jh@suse.cz>
3009 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
3010 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
3011 varpool_remove_variable_insertion_hook): Declare.
3012 * varpool.c (varpool_node_hook_list): New structure.
3013 (first_varpool_node_removal_hook,
3014 first_varpool_variable_insertion_hook): New variables.
3015 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
3016 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
3017 varpool_remove_variable_insertion_hook,
3018 varpool_call_variable_insertion_hooks): New functions.
3019 (varpool_remove_node): Use it.
3021 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
3024 * diagnostic.c (diagnostic_build_prefix): When s.file is
3025 "<built-in>" don't output line and column numbers.
3027 2013-09-06 Jan Hubicka <jh@suse.cz>
3029 * cgraphunit.c (expand_thunk): Get body before touching arguments.
3030 * lto-streamer-out.c: Stream thunks, too.
3031 * lto-streamer-in.c (input_function): Pop cfun here
3032 (lto_read_body): Instead of here.
3034 2013-09-06 Caroline Tice <cmtice@google.com>
3036 * doc/install.texi: Add documentation for the --enable-vtable-verify
3037 and the --disable-libvtv configure options.
3039 2013-09-06 Jeff Law <law@redhat.com>
3041 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
3042 edge implied equivalences into successor phis.
3044 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
3046 * resource.c (mark_referenced_resources): Handle COND_EXEC.
3048 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
3050 * resource.c (mark_target_live_regs): Compute resources taking
3051 into account if a call is predicated or not.
3053 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
3055 * toplev.c (output_stack_usage): Be prepared for suffixes created by
3056 the compiler in the function names.
3058 2013-09-06 Jan Hubicka <jh@suse.cz>
3061 * ipa-inline.c (has_caller_p): New function.
3062 (want_inline_function_to_all_callers_p): Use it.
3063 (sum_callers, inline_to_all_callers): Break out from ...
3064 (ipa_inline): ... here.
3066 2013-09-06 Jan Hubicka <jh@suse.cz>
3068 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
3069 only when AVX is enabled.
3071 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3073 * config/aarch64/aarch64.md
3074 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
3075 is fpsimd_<load/store>2.
3076 (load_pair<mode>): Likewise.
3077 (store_pair<mode>): Likewise.
3079 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3081 * config/arm/types.md (type): Add "mrs" type.
3082 * config/aarch64/aarch64.md
3083 (aarch64_load_tp_hard): Make type "mrs".
3085 (load_tp_hard): Make type "mrs".
3086 * config/arm/cortex-a15.md: Update with new attributes.
3087 * config/arm/cortex-a5.md: Update with new attributes.
3088 * config/arm/cortex-a53.md: Update with new attributes.
3089 * config/arm/cortex-a7.md: Update with new attributes.
3090 * config/arm/cortex-a8.md: Update with new attributes.
3091 * config/arm/cortex-a9.md: Update with new attributes.
3092 * config/arm/cortex-m4.md: Update with new attributes.
3093 * config/arm/cortex-r4.md: Update with new attributes.
3094 * config/arm/fa526.md: Update with new attributes.
3095 * config/arm/fa606te.md: Update with new attributes.
3096 * config/arm/fa626te.md: Update with new attributes.
3097 * config/arm/fa726te.md: Update with new attributes.
3099 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3101 * config/aarch64/aarch64.md
3102 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
3103 (*movtf_aarch64): Likewise.
3105 (thumb1_movdi_insn): Use "multiple" for type where more than one
3106 instruction is used for a move.
3107 (*arm32_movhf): Likewise.
3108 (*thumb_movdf_insn): Likewise.
3110 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3112 * config/arm/types.md (type): Rename fcpys to fmov.
3114 (*arm_movsi_vfp): Rename type fcpys as fmov.
3115 (*thumb2_movsi_vfp): Likewise
3116 (*movhf_vfp_neon): Likewise
3117 (*movhf_vfp): Likewise
3118 (*movsf_vfp): Likewise
3119 (*thumb2_movsf_vfp): Likewise
3120 (*movsfcc_vfp): Likewise
3121 (*thumb2_movsfcc_vfp): Likewise
3122 * config/aarch64/aarch64-simd.md
3123 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
3124 * config/aarch64/aarch64.md
3125 (*movsi_aarch64): Replace type mov_reg with fmovs.
3126 (*movdi_aarch64): Likewise
3127 (*movsf_aarch64): Likewise
3128 (*movdf_aarch64): Likewise
3130 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
3131 * config/arm/iwmmxt.md
3132 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
3133 * config/arm/arm1020e.md: Update with new attributes.
3134 * config/arm/cortex-a15-neon.md: Update with new attributes.
3135 * config/arm/cortex-a5.md: Update with new attributes.
3136 * config/arm/cortex-a53.md: Update with new attributes.
3137 * config/arm/cortex-a7.md: Update with new attributes.
3138 * config/arm/cortex-a8-neon.md: Update with new attributes.
3139 * config/arm/cortex-a9.md: Update with new attributes.
3140 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3141 * config/arm/cortex-r4f.md: Update with new attributes.
3142 * config/arm/marvell-pj4.md: Update with new attributes.
3143 * config/arm/vfp11.md: Update with new attributes.
3145 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3147 * config/aarch64/aarch64.md
3148 (*madd<mode>): Fix type attribute.
3149 (*maddsi_uxtw): Likewise.
3150 (*msub<mode>): Likewise.
3151 (*msubsi_uxtw): Likewise.
3152 (<su_optab>maddsidi4): Likewise.
3153 (<su_optab>msubsidi4): Likewise.
3155 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3157 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
3158 * config/arm/arm.md (core_cycles): Remove fdiv.
3159 * config/arm/vfp.md:
3160 (*sqrtsf2_vfp): Update for attribute changes.
3161 (*sqrtdf2_vfp): Likewise.
3162 * config/aarch64/aarch64.md:
3163 (sqrt<mode>2): Update for attribute changes.
3164 * config/arm/arm1020e.md: Update with new attributes.
3165 * config/arm/cortex-a15-neon.md: Update with new attributes.
3166 * config/arm/cortex-a5.md: Update with new attributes.
3167 * config/arm/cortex-a53.md: Update with new attributes.
3168 * config/arm/cortex-a7.md: Update with new attributes.
3169 * config/arm/cortex-a8-neon.md: Update with new attributes.
3170 * config/arm/cortex-a9.md: Update with new attributes.
3171 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3172 * config/arm/cortex-r4f.md: Update with new attributes.
3173 * config/arm/marvell-pj4.md: Update with new attributes.
3174 * config/arm/vfp11.md: Update with new attributes.
3176 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3178 * config/arm/types.md
3179 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
3180 * config/aarch64/aarch64.md
3181 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
3183 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
3184 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
3185 (float<GPI:mode><GPF:mode>2): Likewise.
3187 (*truncsisf2_vfp): Update with new attributes.
3188 (*truncsidf2_vfp): Likewise.
3189 (fixuns_truncsfsi2): Likewise.
3190 (fixuns_truncdfsi2): Likewise.
3191 (*floatsisf2_vfp): Likewise.
3192 (*floatsidf2_vfp): Likewise.
3193 (floatunssisf2): Likewise.
3194 (floatunssidf2): Likewise.
3195 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
3196 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
3197 * config/arm/arm1020e.md: Update with new attributes.
3198 * config/arm/cortex-a15-neon.md: Update with new attributes.
3199 * config/arm/cortex-a5.md: Update with new attributes.
3200 * config/arm/cortex-a53.md: Update with new attributes.
3201 * config/arm/cortex-a7.md: Update with new attributes.
3202 * config/arm/cortex-a8-neon.md: Update with new attributes.
3203 * config/arm/cortex-a9.md: Update with new attributes.
3204 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3205 * config/arm/cortex-r4f.md: Update with new attributes.
3206 * config/arm/marvell-pj4.md: Update with new attributes.
3207 * config/arm/vfp11.md: Update with new attributes.
3209 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3211 * config/aarch64/arm_neon.h
3212 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
3213 (vqtbx<1,2,3,4><q>_s8): Likewise.
3215 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3217 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
3218 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
3219 (add<mode>3): Add type attribute.
3220 (add<mode>3): Likewise.
3221 (usadd<mode>3): Likewise.
3222 (ssadd<mode>3): Likewise.
3223 (sub<mode>3): Likewise.
3224 (sub<mode>3): Likewise.
3225 (ussub<mode>3): Likewise.
3226 (sssub<mode>3): Likewise.
3227 (ssmulsa3): Likewise.
3228 (usmulusa3): Likewise.
3229 (arm_usatsihi): Likewise.
3231 (*movdi_vfp): Add types for all instructions.
3232 (*movdi_vfp_cortexa8): Likewise.
3233 (*movhf_vfp_neon): Likewise.
3234 (*movhf_vfp): Likewise.
3235 (*movdf_vfp): Likewise.
3236 (*thumb2_movdf_vfp): Likewise.
3237 (*thumb2_movdfcc_vfp): Likewise.
3238 * config/arm/arm.md: Add type attribute to all insn patterns.
3239 (*thumb1_adddi3): Add type attribute.
3240 (*arm_adddi3): Likewise.
3241 (*adddi_sesidi_di): Likewise.
3242 (*adddi_zesidi_di): Likewise.
3243 (*thumb1_addsi3): Likewise.
3244 (addsi3_compare0): Likewise.
3245 (*addsi3_compare0_scratch): Likewise.
3246 (*compare_negsi_si): Likewise.
3247 (cmpsi2_addneg): Likewise.
3248 (*addsi3_carryin_<optab>): Likewise.
3249 (*addsi3_carryin_alt2_<optab>): Likewise.
3250 (*addsi3_carryin_clobercc_<optab>): Likewise.
3251 (*subsi3_carryin): Likewise.
3252 (*subsi3_carryin_const): Likewise.
3253 (*subsi3_carryin_compare): Likewise.
3254 (*subsi3_carryin_compare_const): Likewise.
3255 (*arm_subdi3): Likewise.
3256 (*thumb_subdi3): Likewise.
3257 (*subdi_di_zesidi): Likewise.
3258 (*subdi_di_sesidi): Likewise.
3259 (*subdi_zesidi_di): Likewise.
3260 (*subdi_sesidi_di): Likewise.
3261 (*subdi_zesidi_ze): Likewise.
3262 (thumb1_subsi3_insn): Likewise.
3263 (*arm_subsi3_insn): Likewise.
3264 (*anddi3_insn): Likewise.
3265 (*anddi_zesidi_di): Likewise.
3266 (*anddi_sesdi_di): Likewise.
3267 (*ne_zeroextracts): Likewise.
3268 (*ne_zeroextracts): Likewise.
3269 (*ite_ne_zeroextr): Likewise.
3270 (*ite_ne_zeroextr): Likewise.
3271 (*anddi_notdi_di): Likewise.
3272 (*anddi_notzesidi): Likewise.
3273 (*anddi_notsesidi): Likewise.
3274 (andsi_notsi_si): Likewise.
3275 (thumb1_bicsi3): Likewise.
3276 (*iordi3_insn): Likewise.
3277 (*iordi_zesidi_di): Likewise.
3278 (*iordi_sesidi_di): Likewise.
3279 (*thumb1_iorsi3_insn): Likewise.
3280 (*xordi3_insn): Likewise.
3281 (*xordi_zesidi_di): Likewise.
3282 (*xordi_sesidi_di): Likewise.
3283 (*arm_xorsi3): Likewise.
3284 (*andsi_iorsi3_no): Likewise.
3285 (*smax_0): Likewise.
3286 (*smax_m1): Likewise.
3287 (*arm_smax_insn): Likewise.
3288 (*smin_0): Likewise.
3289 (*arm_smin_insn): Likewise.
3290 (*arm_umaxsi3): Likewise.
3291 (*arm_uminsi3): Likewise.
3292 (*minmax_arithsi): Likewise.
3293 (*minmax_arithsi_): Likewise.
3294 (*satsi_<SAT:code>): Likewise.
3295 (arm_ashldi3_1bit): Likewise.
3296 (arm_ashrdi3_1bit): Likewise.
3297 (arm_lshrdi3_1bit): Likewise.
3298 (*arm_negdi2): Likewise.
3299 (*thumb1_negdi2): Likewise.
3300 (*arm_negsi2): Likewise.
3301 (*thumb1_negsi2): Likewise.
3302 (*negdi_extendsid): Likewise.
3303 (*negdi_zero_extend): Likewise.
3304 (*arm_abssi2): Likewise.
3305 (*thumb1_abssi2): Likewise.
3306 (*arm_neg_abssi2): Likewise.
3307 (*thumb1_neg_abss): Likewise.
3308 (one_cmpldi2): Likewise.
3309 (extend<mode>di2): Likewise.
3310 (*compareqi_eq0): Likewise.
3311 (*arm_extendhisi2addsi): Likewise.
3312 (*arm_movdi): Likewise.
3313 (*thumb1_movdi_insn): Likewise.
3314 (*arm_movt): Likewise.
3315 (*thumb1_movsi_insn): Likewise.
3316 (pic_add_dot_plus_four): Likewise.
3317 (pic_add_dot_plus_eight): Likewise.
3318 (tls_load_dot_plus_eight): Likewise.
3319 (*thumb1_movhi_insn): Likewise.
3320 (*thumb1_movsf_insn): Likewise.
3321 (*movdf_soft_insn): Likewise.
3322 (*thumb_movdf_insn): Likewise.
3323 (cbranchsi4_insn): Likewise.
3324 (cbranchsi4_scratch): Likewise.
3325 (*negated_cbranchsi4): Likewise.
3326 (*tbit_cbranch): Likewise.
3327 (*tlobits_cbranch): Likewise.
3328 (*tstsi3_cbranch): Likewise.
3329 (*cbranchne_decr1): Likewise.
3330 (*addsi3_cbranch): Likewise.
3331 (*addsi3_cbranch_scratch): Likewise.
3332 (*arm_cmpdi_insn): Likewise.
3333 (*arm_cmpdi_unsig): Likewise.
3334 (*arm_cmpdi_zero): Likewise.
3335 (*thumb_cmpdi_zero): Likewise.
3336 (*deleted_compare): Likewise.
3337 (*mov_scc): Likewise.
3338 (*mov_negscc): Likewise.
3339 (*mov_notscc): Likewise.
3340 (*cstoresi_eq0_thumb1_insn): Likewise.
3341 (cstoresi_nltu_thumb1): Likewise.
3342 (cstoresi_ltu_thu): Likewise.
3343 (thumb1_addsi3_addgeu): Likewise.
3344 (*arm_jump): Likewise.
3345 (*thumb_jump): Likewise.
3346 (*check_arch2): Likewise.
3347 (arm_casesi_internal): Likewise.
3348 (thumb1_casesi_dispatch): Likewise.
3349 (*arm_indirect_jump): Likewise.
3350 (*thumb1_indirect_jump): Likewise.
3352 (*and_scc): Likewise.
3353 (*ior_scc): Likewise.
3354 (*compare_scc): Likewise.
3355 (*cond_move): Likewise.
3356 (*cond_arith): Likewise.
3357 (*cond_sub): Likewise.
3358 (*cmp_ite0): Likewise.
3359 (*cmp_ite1): Likewise.
3360 (*cmp_and): Likewise.
3361 (*cmp_ior): Likewise.
3362 (*ior_scc_scc): Likewise.
3363 (*ior_scc_scc_cmp): Likewise.
3364 (*and_scc_scc): Likewise.
3365 (*and_scc_scc_cmp): Likewise.
3366 (*and_scc_scc_nod): Likewise.
3367 (*negscc): Likewise.
3368 (movcond_addsi): Likewise.
3369 (movcond): Likewise.
3370 (*ifcompare_plus_move): Likewise.
3371 (*if_plus_move): Likewise.
3372 (*ifcompare_move_plus): Likewise.
3373 (*if_move_plus): Likewise.
3374 (*ifcompare_arith_arith): Likewise.
3375 (*if_arith_arith): Likewise.
3376 (*ifcompare_arith_move): Likewise.
3377 (*if_arith_move): Likewise.
3378 (*ifcompare_move_arith): Likewise.
3379 (*if_move_arith): Likewise.
3380 (*ifcompare_move_not): Likewise.
3381 (*if_move_not): Likewise.
3382 (*ifcompare_not_move): Likewise.
3383 (*if_not_move): Likewise.
3384 (*ifcompare_shift_move): Likewise.
3385 (*if_shift_move): Likewise.
3386 (*ifcompare_move_shift): Likewise.
3387 (*if_move_shift): Likewise.
3388 (*ifcompare_shift_shift): Likewise.
3389 (*ifcompare_not_arith): Likewise.
3390 (*ifcompare_arith_not): Likewise.
3391 (*if_arith_not): Likewise.
3392 (*ifcompare_neg_move): Likewise.
3393 (*if_neg_move): Likewise.
3394 (*ifcompare_move_neg): Likewise.
3395 (*if_move_neg): Likewise.
3396 (prologue_thumb1_interwork): Likewise.
3397 (*cond_move_not): Likewise.
3398 (*sign_extract_onebit): Likewise.
3399 (*not_signextract_onebit): Likewise.
3400 (stack_tie): Likewise.
3401 (align_4): Likewise.
3402 (align_8): Likewise.
3403 (consttable_end): Likewise.
3404 (consttable_1): Likewise.
3405 (consttable_2): Likewise.
3406 (consttable_4): Likewise.
3407 (consttable_8): Likewise.
3408 (consttable_16): Likewise.
3409 (*thumb1_tablejump): Likewise.
3410 (prefetch): Likewise.
3411 (force_register_use): Likewise.
3412 (thumb_eh_return): Likewise.
3413 (load_tp_hard): Likewise.
3414 (load_tp_soft): Likewise.
3415 (tlscall): Likewise.
3416 (*arm_movtas_ze): Likewise.
3417 (*arm_rev): Likewise.
3418 (*arm_revsh): Likewise.
3419 (*arm_rev16): Likewise.
3420 * config/arm/thumb2.md
3421 (*thumb2_smaxsi3): Likewise.
3422 (*thumb2_sminsi3): Likewise.
3423 (*thumb32_umaxsi3): Likewise.
3424 (*thumb2_uminsi3): Likewise.
3425 (*thumb2_negdi2): Likewise.
3426 (*thumb2_abssi2): Likewise.
3427 (*thumb2_neg_abss): Likewise.
3428 (*thumb2_movsi_insn): Likewise.
3429 (tls_load_dot_plus_four): Likewise.
3430 (*thumb2_movhi_insn): Likewise.
3431 (*thumb2_mov_scc): Likewise.
3432 (*thumb2_mov_negs): Likewise.
3433 (*thumb2_mov_negs): Likewise.
3434 (*thumb2_mov_nots): Likewise.
3435 (*thumb2_mov_nots): Likewise.
3436 (*thumb2_movsicc_): Likewise.
3437 (*thumb2_movsfcc_soft_insn): Likewise.
3438 (*thumb2_indirect_jump): Likewise.
3439 (*thumb2_and_scc): Likewise.
3440 (*thumb2_ior_scc): Likewise.
3441 (*thumb2_ior_scc_strict_it): Likewise.
3442 (*thumb2_cond_move): Likewise.
3443 (*thumb2_cond_arith): Likewise.
3444 (*thumb2_cond_ari): Likewise.
3445 (*thumb2_cond_sub): Likewise.
3446 (*thumb2_negscc): Likewise.
3447 (*thumb2_movcond): Likewise.
3448 (thumb2_casesi_internal): Likewise.
3449 (thumb2_casesi_internal_pic): Likewise.
3450 (*thumb2_alusi3_short): Likewise.
3451 (*thumb2_mov<mode>_shortim): Likewise.
3452 (*thumb2_addsi_short): Likewise.
3453 (*thumb2_subsi_short): Likewise.
3454 (thumb2_addsi3_compare0): Likewise.
3455 (*thumb2_cbz): Likewise.
3456 (*thumb2_cbnz): Likewise.
3457 (*thumb2_one_cmplsi2_short): Likewise.
3458 (*thumb2_negsi2_short): Likewise.
3459 (*orsi_notsi_si): Likewise.
3460 * config/arm/arm1020e.md: Update with new attributes.
3461 * config/arm/arm1026ejs.md: Update with new attributes.
3462 * config/arm/arm1136jfs.md: Update with new attributes.
3463 * config/arm/arm926ejs.md: Update with new attributes.
3464 * config/arm/cortex-a15.md: Update with new attributes.
3465 * config/arm/cortex-a5.md: Update with new attributes.
3466 * config/arm/cortex-a53.md: Update with new attributes.
3467 * config/arm/cortex-a7.md: Update with new attributes.
3468 * config/arm/cortex-a8.md: Update with new attributes.
3469 * config/arm/cortex-a9.md: Update with new attributes.
3470 * config/arm/cortex-m4.md: Update with new attributes.
3471 * config/arm/cortex-r4.md: Update with new attributes.
3472 * config/arm/fa526.md: Update with new attributes.
3473 * config/arm/fa606te.md: Update with new attributes.
3474 * config/arm/fa626te.md: Update with new attributes.
3475 * config/arm/fa726te.md: Update with new attributes.
3477 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3479 * config/aarch64/aarch64-simd.md
3480 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
3481 <vwx> iterator to ensure correct register choice.
3482 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
3483 (aarch64_sqdmull_n<mode>): Likewise.
3484 (aarch64_sqdmull2_n<mode>_internal): Likewise.
3485 * config/aarch64/arm_neon.h
3486 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
3487 (vml<as><q>_n_<su>16): Likewise.
3488 (vml<as>l_high_lane<q>_<su>16): Likewise.
3489 (vml<as>l_high_n_<su>16): Likewise.
3490 (vml<as>l_lane<q>_<su>16): Likewise.
3491 (vml<as>l_n_<su>16): Likewise.
3492 (vmul<q>_lane<q>_<su>16): Likewise.
3493 (vmul<q>_n_<su>16): Likewise.
3494 (vmull_lane<q>_<su>16): Likewise.
3495 (vmull_n_<su>16): Likewise.
3496 (vmull_high_lane<q>_<su>16): Likewise.
3497 (vmull_high_n_<su>16): Likewise.
3498 (vqrdmulh<q>_n_s16): Likewise.
3500 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
3502 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
3503 have the correct lane parameter.
3505 2013-09-06 Richard Biener <rguenther@suse.de>
3507 * cfganal.c (control_dependences::~control_dependences):
3508 Properly free all of the vector.
3510 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
3513 * config/i386/i386.c (ix86_conditional_register_usage):
3514 Proper initialize extended SSE registers.
3516 2013-09-06 Jan Hubicka <jh@suse.cz>
3518 PR tree-optimization/58311
3519 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
3521 2013-09-06 Jan Hubicka <jh@suse.cz>
3523 * Makefile.in (tree-sra.o): Update dependencies.
3524 * tree-sra.c: Include ipa-utils.h
3525 (scan_function): Use recursive_call_p.
3526 (has_caller_p): New function.
3527 (cgraph_for_node_and_aliases): Count also callers of aliases.
3529 2013-09-06 Jan Hubicka <jh@suse.cz>
3532 * cgraph.h (symtab_semantically_equivalent_p): Declare.
3533 * tree-tailcall.c: Include ipa-utils.h.
3534 (find_tail_calls): Use it.
3535 * ipa-pure-const.c (check_call): Likewise.
3536 * ipa-utils.c (recursive_call_p): New function.
3537 * ipa-utils.h (recursive_call_p): Dclare.
3538 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
3539 (symtab_semantically_equivalent_p): New function.
3540 * Makefile.in (tree-tailcall.o): Update dependencies.
3542 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
3544 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
3547 2013-09-06 Richard Biener <rguenther@suse.de>
3549 * lto-streamer.h (lto_global_var_decls): Remove.
3550 * Makefile.in (OBJS): Remove lto-symtab.o.
3551 (lto-symtab.o): Remove.
3552 (GTFILES): Remove lto-symtab.c
3553 * lto-symtab.c: Move to lto/
3555 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3557 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
3558 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
3559 (UNSPEC_FPINT_RINT): New constant definitions.
3560 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
3561 definition with 2 attributes.
3562 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
3563 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
3566 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3568 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
3569 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
3570 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
3571 attribute to "z196_alone".
3572 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
3575 2013-09-06 Richard Biener <rguenther@suse.de>
3577 * basic-block.h (class control_dependences): New.
3578 * tree-ssa-dce.c (control_dependence_map): Remove.
3580 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
3581 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
3582 find_pdom, find_control_dependence, find_all_control_dependences):
3584 (mark_control_dependent_edges_necessary,
3585 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
3586 tree_dce_done, perform_tree_ssa_dce): Adjust.
3587 * cfganal.c (set_control_dependence_map_bit,
3588 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
3589 find_all_control_dependences): Move from tree-ssa-dce.c and
3590 implement as methods of control_dependences class.
3591 (control_dependences::control_dependences): New.
3592 (control_dependences::~control_dependences): Likewise.
3593 (control_dependences::get_edges_dependent_on): Likewise.
3594 (control_dependences::get_edge): Likewise.
3596 2013-09-04 Jan Hubicka <jh@suse.cz>
3598 * tree.c (types_same_for_odr): Drop overactive check.
3599 * ipa-devirt.c (hash_type_name): Likewise.
3601 2013-09-04 Jan Hubicka <jh@suse.cz>
3603 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
3604 (analyze_functions): ... here.
3606 2013-09-04 Jan Hubicka <jh@suse.cz>
3609 * cgraphunit.c (analyze_functions): Clear AUX fields
3610 after processing; initialize assembler name has.
3612 2013-09-05 Jeff Law <law@redhat.com>
3614 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
3615 from thread_around_empty_block. Record threading path into PATH.
3616 Recurse if threading through the initial block is successful.
3617 (thread_across_edge): Corresponding changes to slightly simplify.
3619 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
3621 * config/aarch64/aarch64.md
3622 (type): Remove frecpe, frecps, frecpx.
3623 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
3624 fix to be a TARGET_SIMD instruction.
3625 (aarch64_frecps): Remove.
3626 * config/aarch64/aarch64-simd.md
3627 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
3628 (aarch64_frecps<mode>): Handle all float/vector of float modes.
3630 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
3631 Sofiane Naci <sofiane.naci@arm.com>
3633 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
3634 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
3635 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
3636 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
3637 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
3638 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
3639 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
3640 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
3641 into "clz, "rbit". Rename "shift" to "shift_imm".
3642 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
3643 changes. Update for attribute changes all occurrences of arlo_* and
3645 * config/arm/arm-fixed.md: Update for attribute changes
3646 all occurrences of arlo_* types.
3647 * config/arm/thumb2.md: Update for attribute changes all occurrences
3649 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
3650 (cortexa7_older_only): Likewise.
3651 (cortexa7_younger): Likewise.
3652 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
3653 (1020alu_shift_op): Likewise.
3654 (1020alu_shift_reg_op): Likewise.
3655 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
3656 (alu_shift_op): Likewise.
3657 (alu_shift_reg_op): Likewise.
3658 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
3659 (11_alu_shift_op): Likewise.
3660 (11_alu_shift_reg_op): Likewise.
3661 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
3662 (9_alu_shift_reg_op): Likewise.
3663 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
3665 (cortex_a15_alu_shift): Likewise.
3666 (cortex_a15_alu_shift_reg): Likewise.
3667 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
3669 (cortex_a5_alu_shift): Likewise.
3670 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
3672 (cortex_a53_alu_shift): Likewise.
3673 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
3675 (cortex_a7_alu_reg): Likewise.
3676 (cortex_a7_alu_shift): Likewise.
3677 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
3679 (cortex_a8_alu_shift): Likewise.
3680 (cortex_a8_alu_shift_reg): Likewise.
3681 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
3682 (cortex_a9_dp_shift): Likewise.
3683 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
3685 * config/arm/cortex-r4.md
3686 (cortex_r4_alu): Update for attribute changes.
3687 (cortex_r4_mov): Likewise.
3688 (cortex_r4_alu_shift_reg): Likewise.
3689 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
3690 (526_alu_shift_op): Likewise.
3691 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
3692 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
3693 (626te_alu_shift_op): Likewise.
3694 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
3695 (726te_alu_shift_op): Likewise.
3696 (726te_alu_shift_reg_op): Likewise.
3697 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
3698 (mp626_alu_shift_op): Likewise.
3699 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
3700 (pj4_alu_conds): Likewise.
3701 (pj4_shift): Likewise.
3702 (pj4_shift_conds): Likewise.
3703 (pj4_alu_shift): Likewise.
3704 (pj4_alu_shift_conds): Likewise.
3705 * config/aarch64/aarch64.md: Update for attribute change
3706 all occurrences of arlo_* and shift* types.
3708 2013-09-05 Mike Stump <mikestump@comcast.net>
3710 * tree.h: Move documentation for tree_function_decl to tree-core.h
3711 with the declaration.
3713 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
3716 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
3717 looking for widest mode.
3719 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
3721 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
3723 2013-09-05 Richard Biener <rguenther@suse.de>
3725 PR tree-optimization/58137
3726 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
3727 Do not create vectors of pointers.
3728 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
3729 types for the components of the vector initializer.
3730 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
3731 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
3733 2013-09-05 Martin Jambor <mjambor@suse.cz>
3735 * ipa-prop.c (remove_described_reference): Accept missing references,
3736 return false if that hppens, otherwise return true.
3737 (cgraph_node_for_jfunc): New function.
3738 (try_decrement_rdesc_refcount): Likewise.
3739 (try_make_edge_direct_simple_call): Use them.
3740 (ipa_edge_removal_hook): Remove references from rdescs.
3741 (ipa_edge_duplication_hook): Clone rdescs and their references
3742 when the new edge has the same caller as the old one.
3743 * cgraph.c (cgraph_resolve_speculation): Remove speculative
3744 reference before removing any edges.
3746 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
3748 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
3750 * thumb2.md (thumb2_storewb_parisi): New pattern.
3752 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
3754 * config/aarch64/aarch64-option-extensions.def: Add
3755 AARCH64_OPT_EXTENSION of 'crc'.
3756 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
3757 (AARCH64_ISA_CRC): Ditto.
3758 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
3759 description of the CRC extension.
3761 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
3763 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
3764 * config/rs6000/linux.h: Ditto.
3765 * alpha/linux.h: Ditto.
3766 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
3767 no_c99_libc_has_function.
3768 * config/c6x/uclinux-elf.h: Ditto.
3769 * config/lm32/uclinux-elf.h: Ditto.
3770 * config/m68k/uclinux.h: Ditto.
3771 * config/moxie/uclinux.h: Ditto.
3772 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
3773 (crisv32-*-linux*, cris-*-linux*): Ditto.
3774 * config/bfin/bfin.c: Include "tm_p.h".
3776 2013-09-05 Richard Biener <rguenther@suse.de>
3778 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
3779 check for a definition without a basic-block.
3781 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
3782 Sofiane Naci <sofiane.naci@arm.com>
3784 * config/aarch64/aarch64.md
3785 (*movti_aarch64): Rename r_2_f and f_2_r.
3786 (*movsf_aarch64): Likewise.
3787 (*movdf_aarch64): Likewise.
3788 (*movtf_aarch64): Likewise.
3789 (aarch64_movdi_<mode>low): Likewise.
3790 (aarch64_movdi_<mode>high): Likewise.
3791 (aarch64_mov<mode>high_di): Likewise.
3792 (aarch64_mov<mode>low_di): Likewise.
3793 (aarch64_movtilow_tilow): Likewise.
3794 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
3795 values to config/arm/types.md
3796 (attribute "conds"): Update for attribute change.
3797 (anddi3_insn): Likewise.
3798 (iordi3_insn): Likewise.
3799 (xordi3_insn): Likewise.
3800 (one_cmpldi2): Likewise.
3801 * config/arm/types.md (type): Add Neon types.
3802 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
3803 use "type" attribute.
3804 (movmisalign<mode>_neon_store): Likewise.
3805 (movmisalign<mode>_neon_load): Likewise.
3806 (vec_set<mode>_internal): Likewise.
3807 (vec_setv2di_internal): Likewise.
3808 (vec_extract<mode>): Likewise.
3809 (vec_extractv2di): Likewise.
3810 (add<mode>3_neon): Likewise.
3811 (adddi3_neon): Likewise.
3812 (sub<mode>3_neon): Likewise.
3813 (subdi3_neon): Likewise.
3814 (mul<mode>3_neon): Likewise.
3815 (mul<mode>3add<mode>_neon): Likewise.
3816 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
3817 (fma<VCVTF:mode>4)): Likewise.
3818 (fma<VCVTF:mode>4_intrinsic): Likewise.
3819 (fmsub<VCVTF:mode>4)): Likewise.
3820 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3821 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3822 (ior<mode>3): Likewise.
3823 (and<mode>3): Likewise.
3824 (anddi3_neon): Likewise.
3825 (orn<mode>3_neon): Likewise.
3826 (orndi3_neon): Likewise.
3827 (bic<mode>3_neon): Likewise.
3828 (bicdi3_neon): Likewise.
3829 (xor<mode>3): Likewise.
3830 (one_cmpl<mode>2): Likewise.
3831 (abs<mode>2): Likewise.
3832 (neg<mode>2): Likewise.
3833 (umin<mode>3_neon): Likewise.
3834 (umax<mode>3_neon): Likewise.
3835 (smin<mode>3_neon): Likewise.
3836 (smax<mode>3_neon): Likewise.
3837 (vashl<mode>3): Likewise.
3838 (vashr<mode>3_imm): Likewise.
3839 (vlshr<mode>3_imm): Likewise.
3840 (ashl<mode>3_signed): Likewise.
3841 (ashl<mode>3_unsigned): Likewise.
3842 (neon_load_count): Likewise.
3843 (ashldi3_neon_noclobber): Likewise.
3844 (signed_shift_di3_neon): Likewise.
3845 (unsigned_shift_di3_neon): Likewise.
3846 (ashrdi3_neon_imm_noclobber): Likewise.
3847 (lshrdi3_neon_imm_noclobber): Likewise.
3848 (widen_ssum<mode>3): Likewise.
3849 (widen_usum<mode>3): Likewise.
3850 (quad_halves_<code>v4si): Likewise.
3851 (quad_halves_<code>v4sf): Likewise.
3852 (quad_halves_<code>v8hi): Likewise.
3853 (quad_halves_<code>v16qi): Likewise.
3854 (reduc_splus_v2di): Likewise.
3855 (neon_vpadd_internal<mode>): Likewise.
3856 (neon_vpsmin<mode>): Likewise.
3857 (neon_vpsmax<mode>): Likewise.
3858 (neon_vpumin<mode>): Likewise.
3859 (neon_vpumax<mode>): Likewise.
3860 (ss_add<mode>_neon): Likewise.
3861 (us_add<mode>_neon): Likewise.
3862 (ss_sub<mode>_neon): Likewise.
3863 (us_sub<mode>_neon): Likewise.
3864 (neon_vadd<mode>_unspec): Likewise.
3865 (neon_vaddl<mode>): Likewise.
3866 (neon_vaddw<mode>): Likewise.
3867 (neon_vhadd<mode>): Likewise.
3868 (neon_vqadd<mode>): Likewise.
3869 (neon_vaddhn<mode>): Likewise.
3870 (neon_vmul<mode>): Likewise.
3871 (neon_vmla<mode>): Likewise.
3872 (neon_vmlal<mode>): Likewise.
3873 (neon_vmls<mode>): Likewise.
3874 (neon_vmlsl<mode>): Likewise.
3875 (neon_vqdmulh<mode>): Likewise.
3876 (neon_vqdmlal<mode>): Likewise.
3877 (neon_vqdmlsl<mode>): Likewise.
3878 (neon_vmull<mode>): Likewise.
3879 (neon_vqdmull<mode>): Likewise.
3880 (neon_vsub<mode>_unspec): Likewise.
3881 (neon_vsubl<mode>): Likewise.
3882 (neon_vsubw<mode>): Likewise.
3883 (neon_vqsub<mode>): Likewise.
3884 (neon_vhsub<mode>): Likewise.
3885 (neon_vsubhn<mode>): Likewise.
3886 (neon_vceq<mode>): Likewise.
3887 (neon_vcge<mode>): Likewise.
3888 (neon_vcgeu<mode>): Likewise.
3889 (neon_vcgt<mode>): Likewise.
3890 (neon_vcgtu<mode>): Likewise.
3891 (neon_vcle<mode>): Likewise.
3892 (neon_vclt<mode>): Likewise.
3893 (neon_vcage<mode>): Likewise.
3894 (neon_vcagt<mode>): Likewise.
3895 (neon_vtst<mode>): Likewise.
3896 (neon_vabd<mode>): Likewise.
3897 (neon_vabdl<mode>): Likewise.
3898 (neon_vaba<mode>): Likewise.
3899 (neon_vabal<mode>): Likewise.
3900 (neon_vmax<mode>): Likewise.
3901 (neon_vmin<mode>): Likewise.
3902 (neon_vpaddl<mode>): Likewise.
3903 (neon_vpadal<mode>): Likewise.
3904 (neon_vpmax<mode>): Likewise.
3905 (neon_vpmin<mode>): Likewise.
3906 (neon_vrecps<mode>): Likewise.
3907 (neon_vrsqrts<mode>): Likewise.
3908 (neon_vqabs<mode>): Likewise.
3909 (neon_vqneg<mode>): Likewise.
3910 (neon_vcls<mode>): Likewise.
3911 (clz<mode>2): Likewise.
3912 (popcount<mode>2): Likewise.
3913 (neon_vrecpe): Likewise.
3914 (neon_vrsqrte): Likewise.
3915 (neon_vget_lane<mode>_sext_internal): Likewise.
3916 (neon_vget_lane<mode>_zext_internal): Likewise.
3917 (neon_vdup_n<mode>): Likewise.
3918 (neon_vdup_nv2di): Likewise.
3919 (neon_vdpu_lane<mode>_internal): Likewise.
3920 (neon_vswp<mode>): Likewise.
3921 (float<mode><V_cvtto>2): Likewise.
3922 (floatuns<mode><V_cvtto>2): Likewise.
3923 (fix_trunc<mode><V_cvtto>)2): Likewise
3924 (fixuns_trunc<mode><V_cvtto)2): Likewise.
3925 (neon_vcvt<mode>): Likewise.
3926 (neon_vcvtv4sfv4hf): Likewise.
3927 (neon_vcvtv4hfv4sf): Likewise.
3928 (neon_vcvt_n<mode>): Likewise.
3929 (neon_vmovn<mode>): Likewise.
3930 (neon_vqmovn<mode>): Likewise.
3931 (neon_vqmovun<mode>): Likewise.
3932 (neon_vmovl<mode>): Likewise.
3933 (neon_vmul_lane<mode>): Likewise.
3934 (neon_vmull_lane<mode>): Likewise.
3935 (neon_vqdmull_lane<mode>): Likewise.
3936 (neon_vqdmulh_lane<mode>): Likewise.
3937 (neon_vmla_lane<mode>): Likewise.
3938 (neon_vmlal_lane<mode>): Likewise.
3939 (neon_vqdmlal_lane<mode>): Likewise.
3940 (neon_vmls_lane<mode>): Likewise.
3941 (neon_vmlsl_lane<mode>): Likewise.
3942 (neon_vqdmlsl_lane<mode>): Likewise.
3943 (neon_vext<mode>): Likewise.
3944 (neon_vrev64<mode>): Likewise.
3945 (neon_vrev32<mode>): Likewise.
3946 (neon_vrev16<mode>): Likewise.
3947 (neon_vbsl<mode>_internal): Likewise.
3948 (neon_vshl<mode>): Likewise.
3949 (neon_vqshl<mode>): Likewise.
3950 (neon_vshr_n<mode>): Likewise.
3951 (neon_vshrn_n<mode>): Likewise.
3952 (neon_vqshrn_n<mode>): Likewise.
3953 (neon_vqshrun_n<mode>): Likewise.
3954 (neon_vshl_n<mode>): Likewise.
3955 (neon_vqshl_n<mode>): Likewise.
3956 (neon_vqshlu_n<mode>): Likewise.
3957 (neon_vshll_n<mode>): Likewise.
3958 (neon_vsra_n<mode>): Likewise.
3959 (neon_vsri_n<mode>): Likewise.
3960 (neon_vsli_n<mode>): Likewise.
3961 (neon_vtbl1v8qi): Likewise.
3962 (neon_vtbl2v8qi): Likewise.
3963 (neon_vtbl3v8qi): Likewise.
3964 (neon_vtbl4v8qi): Likewise.
3965 (neon_vtbx1v8qi): Likewise.
3966 (neon_vtbx2v8qi): Likewise.
3967 (neon_vtbx3v8qi): Likewise.
3968 (neon_vtbx4v8qi): Likewise.
3969 (neon_vtrn<mode>_internal): Likewise.
3970 (neon_vzip<mode>_internal): Likewise.
3971 (neon_vuzp<mode>_internal): Likewise.
3972 (neon_vld1<mode>): Likewise.
3973 (neon_vld1_lane<mode>): Likewise.
3974 (neon_vld1_dup<mode>): Likewise.
3975 (neon_vld1_dupv2di): Likewise.
3976 (neon_vst1<mode>): Likewise.
3977 (neon_vst1_lane<mode>): Likewise.
3978 (neon_vld2<mode>): Likewise.
3979 (neon_vld2_lane<mode>): Likewise.
3980 (neon_vld2_dup<mode>): Likewise.
3981 (neon_vst2<mode>): Likewise.
3982 (neon_vst2_lane<mode>): Likewise.
3983 (neon_vld3<mode>): Likewise.
3984 (neon_vld3qa<mode>): Likewise.
3985 (neon_vld3qb<mode>): Likewise.
3986 (neon_vld3_lane<mode>): Likewise.
3987 (neon_vld3_dup<mode>): Likewise.
3988 (neon_vst3<mode>): Likewise.
3989 (neon_vst3qa<mode>): Likewise.
3990 (neon_vst3qb<mode>): Likewise.
3991 (neon_vst3_lane<mode>): Likewise.
3992 (neon_vld4<mode>): Likewise.
3993 (neon_vld4qa<mode>): Likewise.
3994 (neon_vld4qb<mode>): Likewise.
3995 (neon_vld4_lane<mode>): Likewise.
3996 (neon_vld4_dup<mode>): Likewise.
3997 (neon_vst4<mode>): Likewise.
3998 (neon_vst4qa<mode>): Likewise.
3999 (neon_vst4qb<mode>): Likewise.
4000 (neon_vst4_lane<mode>): Likewise.
4001 (neon_vec_unpack<US>_lo_<mode>): Likewise.
4002 (neon_vec_unpack<US>_hi_<mode>): Likewise.
4003 (neon_vec_<US>mult_lo_<mode>): Likewise.
4004 (neon_vec_<US>mult_hi_<mode>): Likewise.
4005 (neon_vec_<US>shiftl_<mode>): Likewise.
4006 (neon_unpack<US>_<mode>): Likewise.
4007 (neon_vec_<US>mult_<mode>): Likewise.
4008 (vec_pack_trunc_<mode>): Likewise.
4009 (neon_vec_pack_trunk_<mode>): Likewise.
4010 (neon_vabd<mode>_2): Likewise.
4011 (neon_vabd<mode>_3): Likewise.
4012 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
4013 (thumb2_movsi_vfp): Likewise.
4014 (movdi_vfp): Likewise.
4015 (movdi_vfp_cortexa8): Likewise.
4016 (movhf_vfp_neon): Likewise.
4017 (movhf_vfp): Likewiwse.
4018 (movsf_vfp): Likewiwse.
4019 (thumb2_movsf_vfp): Likewiwse.
4020 (movdf_vfp): Likewise.
4021 (thumb2_movdf_vfp): Likewise.
4022 (movsfcc_vfp): Likewise.
4023 (thumb2_movsfcc_vfp): Likewise.
4024 (movdfcc_vfp): Likewise.
4025 (thumb2_movdfcc_vfp): Likewise.
4026 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
4027 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
4028 (v10_v2c): Likewise.
4029 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
4031 (cortex_a15_neon_int_2): Likewise.
4032 (cortex_a15_neon_int_3): Likewise.
4033 (cortex_a15_neon_int_4): Likewise.
4034 (cortex_a15_neon_int_5): Likewise.
4035 (cortex_a15_neon_vqneg_vqabs): Likewise.
4036 (cortex_a15_neon_vmov): Likewise.
4037 (cortex_a15_neon_vaba): Likewise.
4038 (cortex_a15_neon_vaba_qqq): Likewise.
4039 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4040 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4041 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
4042 scalar_64_32_long_scalar): Likewise.
4043 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4044 (cortex_a15_neon_mla_qqq_8_16): Likewise.
4045 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
4046 lotype_qdd_64_32_long): Likewise.
4047 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4048 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4049 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
4050 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4051 (cortex_a15_neon_shift_1): Likewise.
4052 (cortex_a15_neon_shift_2): Likewise.
4053 (cortex_a15_neon_shift_3): Likewise.
4054 (cortex_a15_neon_vshl_ddd): Likewise.
4055 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4056 (cortex_a15_neon_vsra_vrsra): Likewise.
4057 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
4058 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
4059 (cortex_a15_neon_fp_vmul_ddd): Likewise.
4060 (cortex_a15_neon_fp_vmul_qqd): Likewise.
4061 (cortex_a15_neon_fp_vmla_ddd): Likewise.
4062 (cortex_a15_neon_fp_vmla_qqq): Likewise.
4063 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
4064 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
4065 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4066 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4067 (cortex_a15_neon_bp_simple): Likewise.
4068 (cortex_a15_neon_bp_2cycle): Likewise.
4069 (cortex_a15_neon_bp_3cycle): Likewise.
4070 (cortex_a15_neon_vld1_1_2_regs): Likewise.
4071 (cortex_a15_neon_vld1_3_4_regs): Likewise.
4072 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4073 (cortex_a15_neon_vld2_4_regs): Likewise.
4074 (cortex_a15_neon_vld3_vld4): Likewise.
4075 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4076 (cortex_a15_neon_vst1_3_4_regs): Likewise.
4077 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
4078 (cortex_a15_neon_vst3_vst4): Likewise.
4079 (cortex_a15_neon_vld1_vld2_lane): Likewise.
4080 (cortex_a15_neon_vld3_vld4_lane" 10
4081 (cortex_a15_neon_vst1_vst2_lane): Likewise.
4082 (cortex_a15_neon_vst3_vst4_lane): Likewise.
4083 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
4084 (cortex_a15_neon_ldm_2): Likewise.0
4085 (cortex_a15_neon_stm_2): Likewise.
4086 (cortex_a15_neon_mcr): Likewise.
4087 (cortex_a15_neon_mcr_2_mcrr): Likewise.
4088 (cortex_a15_neon_mrc): Likewise.
4089 (cortex_a15_neon_mrrc): Likewise.
4090 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
4092 (cortex_a15_alu_shift): Likewise.
4093 (cortex_a15_alu_shift_reg): Likewise.
4094 (cortex_a15_mult32): Likewise.
4095 (cortex_a15_mult64): Likewise.
4096 (cortex_a15_block): Likewise.
4097 (cortex_a15_branch): Likewise.
4098 (cortex_a15_load1): Likewise.
4099 (cortex_a15_load3): Likewise.
4100 (cortex_a15_store1): Likewise.
4101 (cortex_a15_store3): Likewise.
4102 (cortex_a15_call): Likewise.
4103 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
4104 (cortex_a5_f2r): Likewise.
4105 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
4107 (cortex_a53_f2r): Likewise.
4108 * config/arm/cortex-a7.md
4109 (cortex_a7_branch): Update for attribute change.
4110 (cortex_a7_call): Likewise.
4111 (cortex_a7_alu_imm): Likewise.
4112 (cortex_a7_alu_reg): Likewise.
4113 (cortex_a7_alu_shift): Likewise.
4114 (cortex_a7_mul): Likewise.
4115 (cortex_a7_load1): Likewise.
4116 (cortex_a7_store1): Likewise.
4117 (cortex_a7_load2): Likewise.
4118 (cortex_a7_store2): Likewise.
4119 (cortex_a7_load3): Likewise.
4120 (cortex_a7_store3): Likewise.
4121 (cortex_a7_load4): Likewise.
4122 (cortex_a7_store4): Likewise.
4123 (cortex_a7_fpalu): Likewise.
4124 (cortex_a7_fconst): Likewise.
4125 (cortex_a7_fpmuls): Likewise.
4126 (cortex_a7_neon_mul): Likewise.
4127 (cortex_a7_fpmacs): Likewise.
4128 (cortex_a7_neon_mla: Likewise.
4129 (cortex_a7_fpmuld: Likewise.
4130 (cortex_a7_fpmacd: Likewise.
4131 (cortex_a7_fpfmad: Likewise.
4132 (cortex_a7_fdivs: Likewise.
4133 (cortex_a7_fdivd: Likewise.
4134 (cortex_a7_r2f: Likewise.
4135 (cortex_a7_f2r: Likewise.
4136 (cortex_a7_f_flags: Likewise.
4137 (cortex_a7_f_loads: Likewise.
4138 (cortex_a7_f_loadd: Likewise.
4139 (cortex_a7_f_stores: Likewise.
4140 (cortex_a7_f_stored: Likewise.
4141 (cortex_a7_neon): Likewise.
4142 * config/arm/cortex-a8-neon.md
4143 (cortex_a8_neon_mrc): Update for attribute change.
4144 (cortex_a8_neon_mrrc): Likewise.
4145 (cortex_a8_neon_int_1): Likewise.
4146 (cortex_a8_neon_int_2): Likewise.
4147 (cortex_a8_neon_int_3): Likewise.
4148 (cortex_a8_neon_int_4): Likewise.
4149 (cortex_a8_neon_int_5): Likewise.
4150 (cortex_a8_neon_vqneg_vqabs): Likewise.
4151 (cortex_a8_neon_vmov): Likewise.
4152 (cortex_a8_neon_vaba): Likewise.
4153 (cortex_a8_neon_vaba_qqq): Likewise.
4154 (cortex_a8_neon_vsma): Likewise.
4155 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4156 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4157 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
4158 long_scalar): Likewise.
4159 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4160 (cortex_a8_neon_mla_qqq_8_16): Likewise.
4161 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
4162 long_scalar_qdd_64_32_long): Likewise.
4163 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4164 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4165 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
4166 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4167 (cortex_a8_neon_shift_1): Likewise.
4168 (cortex_a8_neon_shift_2): Likewise.
4169 (cortex_a8_neon_shift_3): Likewise.
4170 (cortex_a8_neon_vshl_ddd): Likewise.
4171 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4172 (cortex_a8_neon_vsra_vrsra): Likewise.
4173 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
4174 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
4175 (cortex_a8_neon_fp_vsum): Likewise.
4176 (cortex_a8_neon_fp_vmul_ddd): Likewise.
4177 (cortex_a8_neon_fp_vmul_qqd): Likewise.
4178 (cortex_a8_neon_fp_vmla_ddd): Likewise.
4179 (cortex_a8_neon_fp_vmla_qqq): Likewise.
4180 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
4181 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
4182 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4183 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4184 (cortex_a8_neon_bp_simple): Likewise.
4185 (cortex_a8_neon_bp_2cycle): Likewise.
4186 (cortex_a8_neon_bp_3cycle): Likewise.
4187 (cortex_a8_neon_ldr): Likewise.
4188 (cortex_a8_neon_str): Likewise.
4189 (cortex_a8_neon_vld1_1_2_regs): Likewise.
4190 (cortex_a8_neon_vld1_3_4_regs): Likewise.
4191 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4192 (cortex_a8_neon_vld2_4_regs): Likewise.
4193 (cortex_a8_neon_vld3_vld4): Likewise.
4194 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4195 (cortex_a8_neon_vst1_3_4_regs): Likewise.
4196 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
4197 (cortex_a8_neon_vst3_vst4): Likewise.
4198 (cortex_a8_neon_vld1_vld2_lane): Likewise.
4199 (cortex_a8_neon_vld3_vld4_lane): Likewise.
4200 (cortex_a8_neon_vst1_vst2_lane): Likewise.
4201 (cortex_a8_neon_vst3_vst4_lane): Likewise.
4202 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
4203 (cortex_a8_neon_mcr): Likewise.
4204 (cortex_a8_neon_mcr_2_mcrr): Likewise.
4205 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4206 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
4208 (ca9_neon_mrrc): Likewise.
4209 (cortex_a9_neon_int_1): Likewise.
4210 (cortex_a9_neon_int_2): Likewise.
4211 (cortex_a9_neon_int_3): Likewise.
4212 (cortex_a9_neon_int_4): Likewise.
4213 (cortex_a9_neon_int_5): Likewise.
4214 (cortex_a9_neon_vqneg_vqabs): Likewise.
4215 (cortex_a9_neon_vmov): Likewise.
4216 (cortex_a9_neon_vaba): Likewise.
4217 (cortex_a9_neon_vaba_qqq): Likewise.
4218 (cortex_a9_neon_vsma): Likewise.
4219 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4220 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4221 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
4222 long_scalar): Likewise.
4223 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4224 (cortex_a9_neon_mla_qqq_8_16): Likewise.
4225 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
4226 long_scalar_qdd_64_32_long): Likewise.
4227 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4228 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4229 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
4230 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4231 (cortex_a9_neon_shift_1): Likewise.
4232 (cortex_a9_neon_shift_2): Likewise.
4233 (cortex_a9_neon_shift_3): Likewise.
4234 (cortex_a9_neon_vshl_ddd): Likewise.
4235 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4236 (cortex_a9_neon_vsra_vrsra): Likewise.
4237 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
4238 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
4239 (cortex_a9_neon_fp_vsum): Likewise.
4240 (cortex_a9_neon_fp_vmul_ddd): Likewise.
4241 (cortex_a9_neon_fp_vmul_qqd): Likewise.
4242 (cortex_a9_neon_fp_vmla_ddd): Likewise.
4243 (cortex_a9_neon_fp_vmla_qqq): Likewise.
4244 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
4245 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
4246 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4247 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4248 (cortex_a9_neon_bp_simple): Likewise.
4249 (cortex_a9_neon_bp_2cycle): Likewise.
4250 (cortex_a9_neon_bp_3cycle): Likewise.
4251 (cortex_a9_neon_ldr): Likewise.
4252 (cortex_a9_neon_str): Likewise.
4253 (cortex_a9_neon_vld1_1_2_regs): Likewise.
4254 (cortex_a9_neon_vld1_3_4_regs): Likewise.
4255 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4256 (cortex_a9_neon_vld2_4_regs): Likewise.
4257 (cortex_a9_neon_vld3_vld4): Likewise.
4258 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4259 (cortex_a9_neon_vst1_3_4_regs): Likewise.
4260 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
4261 (cortex_a9_neon_vst3_vst4): Likewise.
4262 (cortex_a9_neon_vld1_vld2_lane): Likewise.
4263 (cortex_a9_neon_vld3_vld4_lane): Likewise.
4264 (cortex_a9_neon_vst1_vst2_lane): Likewise.
4265 (cortex_a9_neon_vst3_vst4_lane): Likewise.
4266 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4267 (cortex_a9_neon_mcr): Likewise.
4268 (cortex_a9_neon_mcr_2_mcrr): Likewise.
4269 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
4270 (cortex_a9_fps): Likewise.
4271 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
4273 (cortex_m4_fmuls): Likewise.
4274 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
4276 (cortex_r4_mrc): Likewise.
4277 * config/arm/iterators.md: Update comment referring to neon_type.
4278 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
4279 (iwmmxt_movsi_insn): Likewise.
4280 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
4282 (pj4_core_to_vfp): Likewise.
4283 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
4285 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
4286 (vfp_fstore): Likewise.
4287 * doc/md.texi: Change references to neon_type to refer to type.
4289 2013-09-04 Dodji Seketeli <dodji@redhat.com>
4291 * tree.h (DECL_BUILT_IN): Fix typo in comment.
4293 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
4295 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
4298 2013-09-04 Easwaran Raman <eraman@google.com>
4301 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
4302 (build_and_add_sum): Use it.
4303 (appears_later_in_bb): Simplify code.
4305 2013-09-04 Teresa Johnson <tejohnson@google.com>
4307 * dumpfile.c (dump_finish): Don't close stderr/stdout.
4309 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
4311 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
4313 2013-09-04 Jan Hubicka <jh@suse.cz>
4315 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
4316 * ipa-devirt.c: Include diganostic.h
4317 (odr_type_d): Add types and types_set.
4318 (hash_type_name): Work for types with vtables during LTO.
4319 (odr_hasher::remove): Fix comment; destroy types_set.
4320 (add_type_duplicate): New function,
4321 (get_odr_type): Use it.
4322 (dump_type_inheritance_graph): Dump type duplicates.
4323 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
4325 * tree.c (types_same_for_odr): Give exact answers on types with
4328 2013-09-04 Dodji Seketeli <dodji@redhat.com>
4330 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
4331 explaining their differences.
4333 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
4335 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
4337 2013-09-03 Jeff Law <law@redhat.com>
4339 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
4340 when not threading through a joiner block. Pass joiner/no joiner
4341 state to register_jump_thread.
4342 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
4343 state from argument rather than implying on path length.
4344 Dump the entire jump thread path into debugging dump.
4345 * tree-flow.h (register_jump_thread): Update prototype.
4347 2013-08-29 Xinliang David Li <davidxl@google.com>
4349 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
4350 Remove a trivial gcc_assert.
4352 2013-08-29 Xinliang David Li <davidxl@google.com>
4354 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
4355 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
4356 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
4357 Delay base decl alignment adjustment.
4358 * tree-vectorizer.c (vect_destroy_datarefs): New function.
4359 * tree-vectorizer.h: New data structure.
4360 (set_dr_misalignment): New function.
4361 (dr_misalignment): Ditto.
4362 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
4363 (vectorizable_load): Ditto.
4364 (ensure_base_align): New function.
4365 (vectorize_loops): Add dbg_cnt support.
4366 (execute_vect_slp): Ditto.
4367 * dbgcnt.def: New debug counter.
4368 * Makefile: New dependency.
4370 2013-09-03 Meador Inge <meadori@codesourcery.com>
4374 2013-08-30 Meador Inge <meadori@codesourcery.com>
4376 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
4378 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
4380 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
4381 function descriptor.
4383 2013-09-03 Richard Biener <rguenther@suse.de>
4385 * tree-affine.c (add_elt_to_tree): Fix association issue,
4386 avoid useless converts and make sure to always return a
4387 properly typed result.
4389 2013-09-03 Richard Biener <rguenther@suse.de>
4392 * fold-const.c (negate_expr_p): Fix division case.
4393 (negate_expr): Likewise.
4395 2013-09-03 Richard Biener <rguenther@suse.de>
4398 * tree-streamer-out.c: Include tm.h.
4399 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
4401 2013-09-03 Jan Hubicka <jh@suse.cz>
4403 * tree-profile.c (tree_profiling): Cleanup CFG when done.
4405 2013-09-03 Alan Modra <amodra@gmail.com>
4407 * config.gcc (powerpc*-*-linux*): Add support for little-endian
4408 multilibs to big-endian target and vice versa.
4409 * config/rs6000/t-linux64: Use := assignment on all vars.
4410 (MULTILIB_EXTRA_OPTS): Remove fPIC.
4411 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
4412 * config/rs6000/t-linux64le: New file.
4413 * config/rs6000/t-linux64bele: New file.
4414 * config/rs6000/t-linux64lebe: New file.
4416 2013-09-02 Jan Hubicka <jh@suse.cz>
4418 * ipa-inline-transform.c (inline_transform): Do not
4419 optimize_inline_calls when not optimizing.
4421 2013-09-02 Jan Hubicka <jh@suse.cz>
4423 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
4424 duplicated nodes for assembler names.
4425 * symtab.c (symtab_unregister_node): Do not attempt to unlink
4426 hard registers from assembler name hash.
4428 2013-09-02 Jan Hubicka <jh@suse.cz>
4430 * ipa-split.c (execute_split_functions): Split externally visible
4431 functions called once.
4433 2013-09-02 Martin Jambor <mjambor@suse.cz>
4436 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
4437 linked list. When finding the correct duplicate, also consider also
4438 the caller in additon to its inlined_to node.
4440 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
4442 * config/aarch64/aarch64-simd-builtins.def
4443 (dup_lane_scalar): Remove.
4444 * config/aarch64/aarch64-simd.md
4445 (aarch64_simd_dup): Add 'w->w' alternative.
4446 (aarch64_dup_lane<mode>): Allow for VALL.
4447 (aarch64_dup_lane_scalar<mode>): Remove.
4448 (aarch64_dup_lane_<vswap_width_name><mode>): New.
4449 (aarch64_get_lane_signed<mode>): Add w->w altenative.
4450 (aarch64_get_lane_unsigned<mode>): Likewise.
4451 (aarch64_get_lane<mode>): Likewise.
4452 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
4453 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
4454 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
4455 (VCON): Change container of V2SF.
4456 (vswap_width_name): Likewise.
4457 * config/aarch64/arm_neon.h
4458 (__aarch64_vdup_lane_any): New.
4459 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
4460 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
4461 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
4463 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
4466 * expr.c (emit_move_complex): Do not move complex FP values as parts if
4467 the source or the destination is a single hard register.
4469 2013-09-02 Richard Biener <rguenther@suse.de>
4472 * tree-scalar-evolution.c (instantiate_scev_name): Allow
4475 2013-09-02 Richard Biener <rguenther@suse.de>
4477 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
4478 arithmetic to sizetype.
4480 2013-09-02 Bin Cheng <bin.cheng@arm.com>
4482 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
4483 Find auto-increment use both before and after candidate.
4485 2013-09-02 Marek Polacek <polacek@redhat.com>
4487 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
4489 2013-09-01 Jan Hubicka <jh@suse.cz>
4491 * Makefile.in: Add ipa-profile.o
4492 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
4493 * cgraph.c (struct cgraph_propagate_frequency_data,
4494 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
4495 ipa-profile.c; replace cgraph_ by ipa_ prefix.
4496 * cgraph.h (cgraph_propagate_frequency): Remove.
4497 * ipa-inline-analysis.c: Include ipa-utils.h;
4498 drop duplicated cfgloop.h.
4499 (inline_update_callee_summaries): Update.
4500 * ipa-profile.c: New file.
4501 * ipa-utils.h (ipa_propagate_frequency): Declare.
4502 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
4503 data-streamer.h, value-prof.h.
4504 (symtab_remove_unreachable_nodes): Update profile.
4505 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
4506 account_time_size, cmp_counts, dump_histogram,
4507 ipa_profile_generate_summary, ipa_profile_write_summary,
4508 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
4509 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
4510 Move to ipa-profile.c.
4512 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
4514 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
4516 2013-09-01 Jan Hubicka <jh@suse.cz>
4518 * common.opt (fdevirtualize-speculatively): New function.
4519 * invoke.texi (fdevirtualize-speculatively): Document.
4520 * ipa-devirt.c: Include ipa-inline.h
4521 (likely_target_p): New function.
4522 (ipa_devirt): New function.
4523 (gate_ipa_devirt): New function.
4524 (pass_data_ipa_devirt): New static var.
4525 (pass_ipa_devirt): Likewise.
4526 (make_pass_ipa_devirt): New function.
4527 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
4528 (common_handle_option): Disable devirtualization when
4529 value range profiling is available.
4530 * passes.def (pass_ipa_devirt): Add.
4531 * timever.def (TV_IPA_DEVIRT): New timevar.
4532 * tree-pass.h (make_pass_ipa_devirt):
4534 2013-09-01 Iain Sandoe <iain@codesourcery.com>
4536 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
4537 include sanitize(undefined).
4539 2013-08-31 Diego Novillo <dnovillo@google.com>
4541 * Makefile.in (TREE_CORE_H): Define.
4543 (GTFILES): Add tree-core.h.
4544 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
4546 * tree-core.h: New file.
4547 Move all data structures, enum, typedefs, global
4548 declarations and constants from ...
4551 2013-08-31 Jan Hubicka <jh@suse.cz>
4553 * bulitins.c (expand_builtin): Do not early exit for gcov
4554 instrumented functions.
4556 2013-08-31 Marek Polacek <polacek@redhat.com>
4558 * ubsan.c: Include tm_p.h.
4560 2013-08-31 Jan Hubicka <jh@suse.cz>
4562 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
4565 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
4566 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
4567 * tree-cfg.c (verify_gimple_label): ... here.
4568 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
4569 (ipa_merge_profiles): New function.
4570 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
4571 (lto_input_function_body): Likewise.
4572 * ipa-utils.h (ipa_merge_profiles): Declare.
4573 * lto-streamer.h (lto_input_function_body): Update prototype.
4574 (emit_label_in_global_context_p): Remove.
4575 * lto-symtab.c: Include ipa-utils.h
4576 (lto_cgraph_replace_node): Use ipa_merge_profiles.
4578 2013-08-31 Jan Hubicka <jh@suse.cz>
4580 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
4582 2013-08-31 Jan Hubicka <jh@suse.cz>
4584 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
4586 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
4588 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
4589 "cmp" RTX before signed_comparison_operator check to account
4592 2013-08-30 Jan Hubicka <jh@suse.cz>
4594 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
4595 (detect_type_change_1): Rename to ...
4596 (detect_type_change): ... this one; early return on non-polymorphic
4598 (detect_type_change_ssa): Add comp_type parameter; update
4599 use of detect_type_change.
4600 (compute_complex_assign_jump_func): Add param_type parameter;
4601 update use of detect_type_change_ssa.
4602 (compute_complex_ancestor_jump_func): Likewise.
4603 (ipa_get_callee_param_type): New function.
4604 (ipa_compute_jump_functions_for_edge): Compute parameter type;
4605 update calls to the jump function computation functions.
4607 2013-08-30 Teresa Johnson <tejohnson@google.com>
4608 Steven Bosscher <steven@gcc.gnu.org>
4610 * cfgrtl.c (fixup_new_cold_bb): New routine.
4611 (commit_edge_insertions): Invoke fixup_partitions.
4612 (find_partition_fixes): New routine.
4613 (fixup_partitions): Ditto.
4614 (verify_hot_cold_block_grouping): Update comments.
4615 (rtl_verify_edges): Invoke find_partition_fixes.
4616 (rtl_verify_bb_pointers): Update comments.
4617 (rtl_verify_bb_layout): Ditto.
4618 * basic-block.h (probably_never_executed_edge_p): Declare.
4619 (fixup_partitions): Ditto.
4620 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
4621 * bb-reorder.c (sanitize_hot_paths): New function.
4622 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
4624 * predict.c (probably_never_executed_edge_p): New routine.
4625 * cfg.c (check_bb_profile): Add partition insanity warnings.
4627 2013-08-30 Meador Inge <meadori@codesourcery.com>
4629 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
4631 2013-08-30 Marek Polacek <polacek@redhat.com>
4633 * Makefile.in (ubsan.o): Add.
4634 (c-family/c-ubsan.o): Add.
4635 (builtins.o): Add ubsan.h dependency.
4636 * ubsan.h: New file.
4637 * ubsan.c: New file.
4638 * common.opt: Add -fsanitize=undefined option.
4639 (flag_sanitize): Add variable.
4640 (fsanitize=): Add option. Add Driver.
4641 (fsanitize=thread): Remove option.
4642 (fsanitize=address): Likewise.
4643 (static-libubsan): New option.
4644 * doc/invoke.texi: Document the new flag and -static-libubsan.
4645 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
4646 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
4647 * builtin-attrs.def (ATTR_COLD): Define.
4648 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
4649 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
4650 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
4651 * flag-types.h (sanitize_code): New enum.
4652 * opts.c (common_handle_option): Parse command line arguments
4653 of -fsanitize=. Add -fsanitize=unreachable option.
4654 * varasm.c (get_variable_section): Adjust.
4655 (assemble_noswitch_variable): Likewise.
4656 (assemble_variable): Likewise.
4657 (output_constant_def_contents): Likewise.
4658 (categorize_decl_for_section): Likewise.
4659 (place_block_symbol): Likewise.
4660 (output_object_block): Likewise.
4661 * builtins.def: Likewise.
4662 * toplev.c (compile_file): Likewise.
4663 (process_options): Likewise.
4664 * cppbuiltin.c: Likewise.
4665 * tsan.c (tsan_pass): Likewise.
4666 (tsan_gate): Likewise.
4667 (tsan_gate_O0): Likewise.
4668 * cfgexpand.c (partition_stack_vars): Likewise.
4669 (expand_stack_vars): Likewise.
4670 (defer_stack_allocation): Likewise.
4671 (expand_used_vars): Likewise.
4672 * cfgcleanup.c (old_insns_match_p): Likewise.
4673 * asan.c (asan_finish_file): Likewise.
4674 (asan_instrument): Likewise.
4675 (gate_asan): Likewise.
4676 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
4677 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
4678 (asan_global_struct): Use pointer_sized_int_node instead
4679 calling build_nonstandard_integer_type.
4680 (initialize_sanitizer_builtins): Likewise.
4681 (asan_finish_file): Likewise.
4682 * gcc.c: Document %{%:function(args):X}.
4683 (static_spec_functions): Add sanitize.
4684 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
4685 store funcval != NULL there.
4686 (do_spec_1): Adjust handle_spec_function caller.
4687 (handle_braces): Allow %:function(args) as condition.
4688 (sanitize_spec_function): New function.
4689 (ADD_STATIC_LIBUBSAN_LIBS): Define.
4690 (LIBUBSAN_SPEC): Likewise.
4691 (LIBUBSAN_EARLY_SPEC): Likewise.
4692 (SANITIZER_SPEC): Handle libubsan.
4693 (SANITIZER_EARLY_SPEC): Likewise.
4694 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
4695 instead of fsanitize=address.
4696 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
4697 instead of fsanitize=address*.
4698 * builtins.c: Include ubsan.h.
4699 (fold_builtin_0): Instrument __builtin_unreachable.
4700 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
4701 instead of flag_asan.
4702 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
4703 (pointer_sized_int_node): Define.
4704 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
4706 2013-08-30 Mike Stump <mikestump@comcast.net>
4708 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
4711 2013-08-29 Jan Hubicka <jh@suse.cz>
4713 * cgraph.c (cgraph_function_body_availability): Handle weakref
4715 * passes.def: Remove pass_fixup_cfg.
4716 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
4717 track when we need to remove functions.
4718 (gate_ipa_inline): Execute inlining always; add comment why.
4719 (pass_data_ipa_inline): Remove TODO_remove_functions.
4720 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
4721 do not produce summaries.
4722 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
4723 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
4724 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
4725 constant pool and vtable.
4727 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
4729 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
4730 New arm_neon.h's internal macros to specify 64-bit constants.
4731 Avoid using stdint.h's macros.
4733 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
4735 * recog.c (verify_changes): Verify that changes[i].old is non-zero
4736 before applying REG_P.
4738 2013-08-30 Jakub Jelinek <jakub@redhat.com>
4740 PR tree-optimization/58277
4741 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
4742 after seeing too many stmts with vdef in between dombb and current
4743 bb, invalidate everything.
4745 2013-08-30 Richard Biener <rguenther@suse.de>
4747 * fold-const.c (fold_single_bit_test): Fix overflow test.
4749 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
4751 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
4752 and which can live in a register, always retrieve the value on entry.
4753 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
4754 passed by invisible reference specially.
4755 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
4756 (vt_add_function_parameter): Correctly deal with a parameter passed by
4757 invisible reference.
4759 2013-08-30 Jan Hubicka <jh@suse.cz>
4761 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
4763 2013-08-30 Richard Biener <rguenther@suse.de>
4765 PR tree-optimization/58228
4766 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
4767 allow invariant loads in nested loop vectorization.
4769 2013-08-30 Richard Biener <rguenther@suse.de>
4771 PR tree-optimization/58223
4772 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
4773 (has_anti_or_output_dependence): ... this and adjust to also
4774 look for output dependences.
4775 (mark_nodes_having_upstream_mem_writes): Adjust.
4776 (rdg_flag_uses): Likewise.
4778 2013-08-30 Richard Biener <rguenther@suse.de>
4780 PR tree-optimization/58010
4781 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
4782 assert that we have a loop-closed PHI.
4784 2013-08-29 Jan Hubicka <jh@suse.cz>
4786 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
4787 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
4788 * lto-section-in.c (lto_free_function_in_decl_state): New function.
4789 (lto_free_function_in_decl_state_for_node): New function.
4791 2013-08-29 Xinliang David Li <davidxl@google.com>
4793 * loop-unroll.c (report_unroll_peel): Minor message change.
4794 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
4795 Emit alignment peeling message with default -fopt-info.
4796 (vect_loop_versioning): Emit loop version info message.
4797 * tree-vectorizer.c (vectorize_loops): Minor message change.
4798 (execute_vect_slp): Ditto.
4800 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
4802 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
4803 of the clone from the DECL_NAME of the original function.
4805 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
4807 * passes.c (register_pass): Add overload.
4808 * tree-pass.h (register_pass): Forward declare it. Add comment.
4810 2013-08-29 Jan Hubicka <jh@suse.cz>
4812 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
4813 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
4814 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
4815 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
4816 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
4817 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
4818 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
4819 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
4821 2013-08-29 Teresa Johnson <tejohnson@google.com>
4823 * dumpfile.c (dump_loc): Output column number.
4824 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
4825 * doc/invoke.texi: Document optall -fopt-info flag.
4826 * profile.c (read_profile_edge_counts): Use new dump framework.
4827 (compute_branch_probabilities): Ditto.
4828 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
4829 when pass not in any opt group.
4830 * pass_manager.h (pass_manager::get_pass_profile): New method.
4831 * value-prof.c (check_counter): Use new dump framework.
4832 (check_ic_target): Ditto.
4833 * coverage.c (get_coverage_counts): Ditto.
4834 (coverage_init): Setup new dump framework.
4836 2013-08-29 Richard Biener <rguenther@suse.de>
4838 PR tree-optimization/58246
4839 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
4840 handle the dominance check inside a basic-block.
4842 2013-08-29 Richard Biener <rguenther@suse.de>
4845 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
4846 of default defs that appear in abnormal PHI nodes.
4848 2013-08-29 Richard Biener <rguenther@suse.de>
4850 PR tree-optimization/57685
4851 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
4852 single-use operands to avoid exponential complexity.
4854 2013-08-28 Dehao Chen <dehao@google.com>
4856 * ipa-inline.c (edge_badness): Fix integer underflow.
4858 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
4860 * gtm-builtins.def (_ITM_free): Declare leaf.
4862 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4865 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
4866 (*tls_local_dynamic_base_64_largepic): Likewise.
4867 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
4868 Remove predicate from call operand.
4869 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
4870 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
4872 2013-08-28 Jeff Law <law@redhat.com>
4874 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
4875 checks for the number of predecessors and successors allowed.
4876 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
4877 which require copying a joiner block if there is a request which
4878 is a subpath that requires no joiner block copying.
4880 2013-08-28 Jan Hubicka <jh@suse.cz>
4882 * lto-streamer-out.c (DFS_write_tree_body): Drop
4883 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
4884 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
4885 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
4886 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
4887 Do not read DECL_ERROR_ISSUED.
4888 (unpack_ts_decl_with_vis_value_fields): Do not read
4890 (lto_input_ts_binfo_tree_pointers): Do not read
4891 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
4892 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
4893 write DECL_ERROR_ISSUED..
4894 (pack_ts_decl_with_vis_value_fields): Do not write
4896 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
4897 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
4898 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
4899 * tree.h (tree_decl_common): Update comment.
4900 (DECL_ERROR_ISSUED): Remove.
4902 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4905 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
4907 2013-08-28 Jan Hubicka <jh@suse.cz>
4909 * builtins.def (free): Declare leaf.
4911 2013-08-27 David Malcolm <dmalcolm@redhat.com>
4914 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
4915 * configure: Regenerate.
4917 2013-08-27 Martin Jambor <mjambor@suse.cz>
4919 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
4920 (ipa_ancestor_jf_data): Likewise.
4921 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
4922 (ipa_get_jf_pass_through_type_preserved): New function.
4923 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
4924 (ipa_get_jf_ancestor_type_preserved): New function.
4925 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
4926 (ipa_get_jf_ancestor_result): Likewise.
4927 (propagate_vals_accross_pass_through): Use
4928 ipa_get_jf_pass_through_result to do all the value mappings.
4929 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
4930 type_preserved flag.
4931 (ipa_set_jf_cst_copy): New function.
4932 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
4933 (ipa_set_jf_arith_pass_through): Likewise.
4934 (ipa_set_ancestor_jf): Likewise.
4935 (compute_complex_assign_jump_func): Set type_preserved instead of
4937 (ipa_compute_jump_functions_for_edge): Likewise.
4938 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
4939 (update_jump_functions_after_inlining): Update type_preserved.
4940 Explicitely create jump functions when combining one with pass_through.
4941 (ipa_write_jump_function): Stream the type_preserved flags.
4942 (ipa_read_jump_function): Likewise.
4944 2013-08-27 Jakub Jelinek <jakub@redhat.com>
4945 Aldy Hernandez <aldyh@redhat.com>
4947 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
4948 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
4949 * function.h (struct function): Add has_force_vect_loops and
4951 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
4952 * gimple.c (gimple_build_omp_critical): Add KIND argument and
4954 * gimple.def: Update CLAUSES comments.
4955 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
4956 (gimple_build_omp_for): Add argument to prototype.
4957 (gimple_omp_for_kind): New.
4958 (gimple_omp_for_set_kind): New.
4959 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
4960 GOVD_DATA_SHARE_CLASS.
4961 (enum omp_region_type): Add ORT_SIMD.
4962 (gimple_add_tmp_var): Handle ORT_SIMD.
4963 (gimplify_var_or_parm_decl): Same.
4964 (is_gimple_stmt): Same.
4965 (omp_firstprivatize_variable): Same.
4966 (omp_add_variable): Only use splay_tree_insert if lookup failed.
4967 (omp_notice_variable): Handle ORT_SIMD.
4968 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
4969 (omp_check_private): Handle ORT_SIMD.
4970 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4972 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
4973 Handle OMP_CLAUSE_LASTPRIVATE.
4974 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4976 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
4977 (gimplify_expr): Handle OMP_SIMD.
4978 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
4979 (expand_GOMP_SIMD_VF): New.
4980 (expand_GOMP_SIMD_LAST_LANE): New.
4981 * internal-fn.def (GOMP_SIMD_LANE): New.
4982 (GOMP_SIMD_VF): New.
4983 (GOMP_SIMD_LAST_LANE): New.
4984 * omp-low.c: Include target.h.
4985 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
4987 (check_omp_nesting_restrictions): Same.
4989 (lower_rec_simd_input_clauses): New.
4990 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
4992 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
4993 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
4994 (expand_omp_build_assign): New.
4995 (expand_omp_for_init_counts): New.
4996 (expand_omp_for_init_vars): New.
4997 (extract_omp_for_update_vars): New.
4998 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
4999 and rewrite accordingly.
5000 (expand_omp_simd): New.
5001 (expand_omp_for): Use expand_omp_simd.
5002 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
5003 (lower_omp_for): Do not lower the body.
5004 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
5006 * tree-flow.h (find_omp_clause): Remove prototype.
5007 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
5008 forcing vectorization of the loop, or if flag_tree_vectorize.
5009 (gate_tree_if_conversion): Similarly.
5010 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
5011 gimple_build_omp_for.
5012 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
5013 * tree-parloops (create_parallel_loop): Pass kind argument to
5014 gimple_build_omp_for.
5015 * tree-pretty-print.c (dump_omp_clause): Add cases for
5016 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
5017 OMP_CLAUSE__SIMDUID_.
5018 (dump_generic_node): Handle OMP_SIMD.
5019 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
5020 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
5021 unroll OMP_SIMD loops here.
5022 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
5023 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
5025 (vect_analyze_data_refs): Handle simd loops.
5026 * tree-vect-loop.c (vectorizable_live_operation): Handle
5028 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
5029 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
5030 (vectorizable_load): Same.
5031 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
5032 (struct simduid_to_vf): New.
5033 (simduid_to_vf::hash): New.
5034 (simduid_to-vf::equal): New.
5035 (struct simd_array_to_simduid): New.
5036 (simd_array_to_simduid::hash): New.
5037 (simd_array_to_simduid::equal): New.
5038 (adjust_simduid_builtins): New.
5039 (struct note_simd_array_uses_struct): New.
5040 (note_simd_array_uses_cb): New.
5041 (note_simd_array_uses): New.
5042 (vectorize_loops): Handle simd hints and adjust simd builtins
5044 * tree-vectorizer.h (struct _stmt_vec_info): Add
5045 simd_lane_access_p field.
5046 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
5047 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
5048 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
5049 (omp_clause_code_name): Same.
5050 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
5051 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
5052 * tree.def (OMP_SIMD): New entry.
5053 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
5054 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
5055 (OMP_CLAUSE_DECL): Adjust range for new clauses.
5056 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
5057 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
5058 (OMP_CLAUSE_LINEAR_STEP): New.
5059 (OMP_CLAUSE_SAFELEN_EXPR): New.
5060 (OMP_CLAUSE__SIMDUID__DECL): New.
5061 (find_omp_clause): New prototype.
5063 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
5065 * config/i386/driver-i386.c (host_detect_local_cpu): Update
5066 Haswell processor detection.
5068 2013-08-27 Christian Widmer <shadow@umbrox.de>
5071 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
5072 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
5073 AVX2 capable processors.
5075 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
5077 * config/aarch64/arm_neon.h: Replace all inline asm implementations
5078 of vget_low_* with implementations in terms of other intrinsics.
5080 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
5083 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
5084 values to -1, 0 and 1.
5086 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
5088 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
5089 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
5090 (<optab><mode>3_insn): Remove.
5091 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
5092 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
5093 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
5094 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
5095 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
5096 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
5097 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
5098 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
5099 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
5100 (ror<mode>3_insn): Likewise.
5101 * config/aarch64/predicates.md (aarch64_simd_register): New.
5103 2013-08-27 Richard Biener <rguenther@suse.de>
5105 PR tree-optimization/57521
5106 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
5107 one edge is non-critical.
5108 (find_phi_replacement_condition): Make sure to use a non-critical
5109 edge. Cleanup and remove old bug workarounds.
5110 (bb_postdominates_preds): Remove.
5111 (if_convertible_loop_p_1): Do not compute post-dominators.
5112 (combine_blocks): Do not free post-dominators.
5113 (main_tree_if_conversion): Likewise.
5114 (pass_data_if_conversion): Add TODO_verify_ssa.
5116 2013-08-27 DJ Delorie <dj@redhat.com>
5118 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
5120 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
5122 * function.c (assign_parm_find_data_types): Set passed_mode and
5123 nominal_mode to the TYPE_MODE of nominal_type for the built
5124 pointer type in case of the struct-pass-by-reference.
5126 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
5128 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
5129 (UINT16_TYPE): Change default to "unsigned int".
5131 * config/avr/avr.opt (mfract-convert-truncate): New option.
5132 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
5133 is set, round negative fractional integers according to n1169
5134 when converting to integer types.
5136 2013-08-26 Jan Hubicka <jh@suse.cz>
5138 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
5139 methods can not be called indirectly when their address is not taken.
5141 2013-08-26 Jan Hubicka <jh@suse.cz>
5143 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
5146 2013-08-26 Jan Hubicka <jh@suse.cz>
5148 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
5151 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
5153 * reload.c (find_valid_class): Allow classes that do not include
5154 FIRST_PSEUDO_REGISTER - 1.
5156 2013-08-26 Jan Hubicka <jh@suse.cz>
5158 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
5159 fix edge count/frequency when speculation failed; fix type check
5160 for the direct call.
5162 2013-08-26 Jan Hubicka <jh@suse.cz>
5164 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
5166 2013-08-26 Jan Hubicka <jh@suse.cz>
5168 * ipa-inline-transform.c (inline_transform): Be ready for basic block
5169 to be changed by edge redirection.
5171 2013-08-26 Jan Hubicka <jh@suse.cz>
5173 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
5174 formating; add sanity check.
5175 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
5176 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
5177 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
5178 (resolve_noninline_speculation): Update callee keys, too.
5180 2013-08-26 Jan Hubicka <jh@suse.cz>
5182 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
5183 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
5185 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
5187 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
5190 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
5192 * config/i386/i386.c (ix86_debug_options): Remove prototype.
5193 (x86_64_elf_select_section): Ditto.
5194 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
5196 (ix86_pass_by_reference): Ditto.
5197 (output_set_got): Ditto.
5198 (ix86_unary_operator_ok): Ditto.
5199 (ix86_expand_builtin): Ditto.
5201 2013-08-23 Jan Hubicka <jh@suse.cz>
5203 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
5205 2013-08-23 Jan Hubicka <jh@suse.cz>
5207 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
5208 (tree_decl_with_vis): Add FINAL field.
5210 2013-08-23 Jeff Law <law@redhat.com>
5212 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
5213 the debugging dump when the expression is fully redundant.
5215 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
5217 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
5218 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
5219 * pretty-print.c (pp_formatted_text_data): Likewise.
5220 (pp_write_text_to_stream): Likewise.
5221 (pp_write_text_as_dot_label_to_stream): Likewise.
5222 (pp_append_r): Likewise.
5223 (pp_format): Likewise.
5224 (pp_flush): Likewise.
5225 (pp_clear_output_area): Likewise.
5226 (pp_append_text): Likewise.
5227 (pp_formatted_text): Likewise.
5228 (pp_remaining_character_count_for_line): Likewise.
5229 (pp_newline): Likewise.
5230 (pp_character): Likewise.
5231 (output_buffer::~output_buffer): Define.
5232 (pretty_printer::~pretty_printer): Destruct output buffer.
5233 * pretty-print.h (output_buffer::~output_buffer): Declare.
5234 (pretty_printer::~pretty_printer): Declare virtual.
5236 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
5239 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
5240 HOST_WIDE_INT_M1U): New macros.
5241 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
5242 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
5243 unsigned -1 for lshift.
5244 * cse.c (cse_insn): Likewise.
5245 * double-int.c (rshift_double, lshift_double): Likewise.
5246 * builtins.c (fold_builtin_bitop): Likewise.
5247 * combine.c (force_to_mode): Likewise.
5248 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
5249 * simplify-rtx.c (simplify_const_unary_operation,
5250 simplify_const_binary_operation): Likewise.
5251 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
5252 check_va_list_escapes): Likewise.
5253 * rtlanal.c (nonzero_bits1): Likewise.
5254 * expmed.c (expand_smod_pow2): Likewise.
5255 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
5257 2013-08-23 Jan Hubicka <jh@suse.cz>
5259 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
5260 as having address taken.
5262 2013-08-23 Jan Hubicka <jh@suse.cz>
5264 * ipa-utils.h (method_class_type): Declare.
5265 * ipa-devirt.c (method_class_type): Export.
5267 * cgraphunit.c (analyze_functions): Do basic devirtualization;
5268 do not walk base classes of anonymous types.
5270 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
5272 PR rtl-optimization/58220
5274 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
5275 handle SEQUENCE insns properly.
5277 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
5279 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
5281 (pp_newline_and_indent): Likewise.
5282 (pp_separate_with): Likewise.
5283 * pretty-print.c (pp_newline_and_flush): Define.
5284 (pp_newline_and_indent): Likewise.
5285 (pp_separate_with): Likewise.
5287 2013-08-23 Jakub Jelinek <jakub@redhat.com>
5290 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
5291 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
5293 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
5295 * config/i386/predicates.md (ext_sse_reg_operand): New.
5296 * config/i386/i386.md (*movti_internal): Use
5297 predicate to determine if EVEX is needed.
5298 (*movsi_internal): Ditto.
5299 (*movdf_internal): Ditto.
5300 (*movsf_internal): Ditto.
5301 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
5303 2013-08-23 Jakub Jelinek <jakub@redhat.com>
5305 PR tree-optimization/58209
5306 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
5307 (find_tail_calls): Give up for pointer result types if m is non-NULL.
5308 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
5309 emit POINTER_PLUS_EXPR.
5310 (create_tailcall_accumulator): For pointer result type accumulate in
5313 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
5315 * configure.ac: Add backslashes missing from the last change.
5316 * configure: Regenerate.
5318 2013-08-22 Jan Hubicka <jh@susue.cz>
5320 * ipa.c (function_and_variable_visibility): First remember function
5321 was global and then make it local.
5323 2013-08-22 Julian Brown <julian@codesourcery.com>
5325 * configure.ac: Add aarch64 to list of arches which use "nop" in
5327 * configure: Regenerate.
5329 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5331 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
5332 gnu_libc_has_function.
5333 * config/s390/tpf.h: Likewise.
5335 2013-08-22 Jan Hubicka <jh@susue.cz>
5337 * timevar.c (validate_phases): Add cast.
5339 2013-08-22 Jan Hubicka <jh@susue.cz>
5341 * timevar.c (validate_phases): Use size_t for memory.
5342 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
5344 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
5346 * pretty-print.h (output_buffer::output_buffer): Declare.
5347 (pretty_printer::pretty_printer): Likewise.
5348 (pp_construct): Remove.
5349 * pretty-print.c (output_buffer::output_buffer): Define.
5350 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
5351 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
5352 (print_gimple_expr): Likewise.
5353 (print_gimple_seq): Likewise.
5354 (gimple_dump_bb): Likewise.
5355 * sched-vis.c (dump_value_slim): Likewise.
5356 (dump_insn_slim): Likewise.
5357 (dump_rtl_slim): Likewise.
5358 (str_pattern_slim): Likewise.
5359 * tree-mudflap.c (mf_varname_tree): Likewise.
5360 * graph.c (print_graph_cfg): Likewise.
5361 (start_graph_dump): Likewise.
5362 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
5364 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
5365 pretty printer initialization.
5366 * coretypes.h (diagnostic_context): Remove superflous type alias
5368 (pretty_printer): Likewise. Declare directly as a class.
5369 (pretty_print_info): Remove declaration as class.
5370 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
5371 and pp_clear_output_area.
5372 (asan_add_global): Likewise.
5374 2013-08-22 Jan Hubicka <jh@suse.cz>
5376 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
5377 * ipa-utils.h (update_type_inheritance_graph): Declare.
5378 (possible_polymorphic_call_target_p): Declare.
5379 (possible_polymorphic_call_target_p): New.
5380 * ipa-devirt.c: Update toplevel comments.
5381 (cached_polymorphic_call_targets): Move up.
5382 (odr_type_d): Move ID down.
5383 (polymorphic_type_binfo_p): Update comment.
5384 (odr_hasher::remove): Likewise;
5385 (get_odr_type): Set anonymous_namespace.
5386 (dump_odr_type): Dump it.
5387 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
5388 (maybe_record_node): Record node in cached_polymorphic_call_targets.
5389 (record_binfo): Add comment.
5390 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
5392 (devirt_node_removal_hook): Do not iCE when cache is freed.
5393 (possible_polymorphic_call_target_p): New predicate.
5394 (update_type_inheritance_graph): New function.
5396 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
5397 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5398 Sergey Lega <sergey.s.lega@intel.com>
5399 Anna Tikhonova <anna.tikhonova@intel.com>
5400 Ilya Tocar <ilya.tocar@intel.com>
5401 Andrey Turetskiy <andrey.turetskiy@intel.com>
5402 Ilya Verbin <ilya.verbin@intel.com>
5403 Kirill Yukhin <kirill.yukhin@intel.com>
5404 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5406 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
5407 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
5408 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
5409 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
5410 (OPTION_MASK_ISA_AVX2_UNSET): Update.
5411 (OPTION_MASK_ISA_AVX512F_UNSET): New.
5412 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
5413 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
5414 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
5415 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
5416 OPT_mavx512pf, OPT_mavx512er cases.
5417 * config/i386/constraints.md (v): New constraint.
5418 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
5419 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
5420 (bit_AVX512CD): New.
5421 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
5422 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
5423 * config/i386/i386-c.c (ix86_target_macros_internal):
5424 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
5426 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
5427 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
5428 * config/i386/i386.c (regclass_map, dbx_register_map)
5429 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
5430 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
5431 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
5433 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
5434 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
5435 -mavx512pf options. Fix formatting.
5436 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
5437 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
5438 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
5439 -mavx512cd, -mavx512pf options.
5440 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
5441 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
5442 (ix86_preferred_output_reload_class): Replace SSE_REGS with
5444 (ix86_hard_regno_mode_ok): Support 512-bit registers.
5445 (ix86_set_reg_reg_cost): Ditto.
5446 (x86_order_regs_for_local_alloc): Ditto.
5447 (MAX_VECT_LEN): Extend to 64-byte.
5448 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
5449 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
5450 (TARGET_AVX512ER, TARGET_AVX512CD): New.
5451 (BIGGEST_ALIGNMENT): Extend to 512-bits.
5452 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
5453 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
5454 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
5455 (SSE_REG_MODE_P): Support new modes.
5456 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
5457 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
5458 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
5459 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
5460 (REG_CLASS_CONTENTS): Add new registers.
5461 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
5462 (EXT_REX_SSE_REGNO_P): New.
5463 (HI_REGISTER_NAMES): Add new registers.
5464 * config/i386/i386.md: Define constants for new registers.
5465 (mode): Add new 512-bit modes.
5466 (prefix): Support evex prefix.
5467 (isa): Support avx512f, noavx512f, fma_avx512f.
5468 (ssemodesuffix): Add new 512-bit modes.
5470 (*movxi_internal_avx512f): Ditto.
5471 (*movdi_internal): Replace constraint "x" with the new constraint "v".
5473 (*movsi_internal): Likewise.
5474 (*movdf_internal): Likewise.
5475 (*movsf_internal): Likewise.
5476 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
5477 (<code><mode>3): Likewise.
5478 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
5480 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
5481 with the new constraint "v".
5482 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
5484 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
5485 with the new constraint "v".
5486 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
5487 (<sse2>_storedqu<avxsizesuffix>): Likewise.
5488 (*<plusminus_insn><mode>3): Likewise.
5489 (<sse>_vm<plusminus_insn><mode>3): Likewise.
5490 (*mul<mode>3): Likewise.
5491 (<sse>_vmmul<mode>3): Likewise.
5492 (<sse>_div<mode>3): Likewise.
5493 (<sse>_vmdiv<mode>3): Likewise.
5494 (<sse>_sqrt<mode>2): Likewise.
5495 (<sse>_vmsqrt<mode>2): Likewise.
5496 (*<code><mode>3_finite): Likewise.
5497 (*<code><mode>3) <smaxmin>: Likewise.
5498 (<sse>_vm<code><mode>3): Likewise.
5499 (*<code><mode>3) <any_logic>: Likewise.
5500 (*fma_fmadd_<mode>): Likewise.
5501 (*fma_fmsub_<mode>): Likewise.
5502 (*fma_fnmadd_<mode>): Likewise.
5503 (*fma_fnmsub_<mode>): Likewise.
5504 (*fma_fmaddsub_<mode>): Likewise.
5505 (*fma_fmsubadd_<mode>): Likewise.
5506 (*fmai_fmadd_<mode>): Likewise.
5507 (*fmai_fmsub_<mode>): Likewise.
5508 (*fmai_fnmadd_<mode>): Likewise.
5509 (*fmai_fnmsub_<mode>): Likewise.
5510 (sse_cvtsi2ss): Likewise.
5511 (sse_cvtsi2ssq): Likewise.
5512 (sse_cvtss2si): Likewise.
5513 (sse_cvtss2si_2): Likewise.
5514 (sse_cvtss2siq): Likewise.
5515 (sse_cvtss2siq_2): Likewise.
5516 (sse_cvttss2si): Likewise.
5517 (sse_cvtss2siq_2): Likewise.
5518 (float<sseintvecmodelower><mode>2): Likewise.
5519 (sse2_cvtsd2si_2): Likewise.
5520 (sse2_cvtsd2siq_2): Likewise.
5521 (*<plusminus_insn><mode>3): Likewise.
5522 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
5523 (*<sse4_1_avx2>_mul<mode>3): Likewise.
5524 (ashr<mode>3): Likewise.
5525 (<shift_insn><mode>3): Likewise.
5526 (avx2_<code><mode>3): Likewise.
5527 (*avx2_<code><mode>3): Likewise.
5528 (*andnot<mode>3): Likewise.
5529 (*<code><mode>3) <any_logic>: Likewise.
5530 (abs<mode>2): Likewise.
5531 (avx2_permvar<mode>): Likewise.
5532 (avx2_perm<mode>_1): Likewise.
5533 (*avx_vpermilp<mode>): Likewise.
5534 (avx_vpermilvar<mode>3): Likewise.
5535 (avx2_ashrv<mode>): Likewise.
5536 (avx2_<shift_insn>v<mode>): Likewise.
5537 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
5539 * doc/rtl.texi: Document XImode.
5541 2013-08-21 Jeff Law <law@redhat.com>
5543 * tree-flow.h (register_jump_thread): Pass vector of edges
5544 instead of each important edge.
5545 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
5546 thread path into a vector and pass that to register_jump_thread.
5547 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
5548 passed in edge vector to the current 3-edge form.
5551 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
5553 * dce.c (fini_dce): Call df_analyze again just in case
5554 delete_unmarked_insns removed anything.
5556 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
5558 * reload.h (struct reg_equivs): Rename to ..
5559 (struct reg_equivs_s): .. this.
5561 2013-08-20 Martin Liska <marxin.liska@gmail.com>
5563 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
5565 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5567 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
5569 2013-08-21 Jeff Law <law@redhat.com>
5571 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
5572 simplify assignments too. If the RHS collapses to a singleton
5573 range, then return the value for the range.
5575 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
5577 * config/i386/sse.md (V16): Rename to...
5579 (mov<mode>): Update iterator name.
5580 (*mov<mode>_internal): Ditto.
5581 (push<mode>1): Ditto.
5582 (movmisalign<mode>): Ditto.
5584 2013-08-20 Jan Hubicka <jh@suse.cz>
5587 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
5588 entry for direct edges.
5589 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
5591 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5593 Revert my last two changes, r201865 and r201864:
5596 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5598 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
5599 instances can own GC refs.
5601 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
5602 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
5603 (gcc::context::gt_pch_nx): Likewise.
5604 (gcc::context::gt_pch_nx): Likewise.
5605 * ggc.h (gt_ggc_mx <T>): New.
5606 (gt_pch_nx_with_op <T>): New.
5607 (gt_pch_nx <T>): New.
5608 * passes.c (opt_pass::gt_ggc_mx): New.
5609 (opt_pass::gt_pch_nx): New.
5610 (opt_pass::gt_pch_nx_with_op): New.
5611 (pass_manager::gt_ggc_mx): New.
5612 (pass_manager::gt_pch_nx): New.
5613 (pass_manager::gt_pch_nx_with_op): New.
5614 (pass_manager::operator new): Use
5615 ggc_internal_cleared_alloc_stat rather than xcalloc.
5616 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
5617 (pass_manager::gt_ggc_mx): New.
5618 (pass_manager::gt_pch_nx): New.
5619 (pass_manager::gt_pch_nx_with_op): New.
5620 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
5621 (opt_pass::operator new): New.
5622 (opt_pass::gt_ggc_mx): New.
5623 (opt_pass::gt_pch_nx): New.
5624 (opt_pass::gt_pch_nx_with_op): New.
5627 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5629 * Makefile.in (GTFILES): Add context.h.
5630 * context.c (gcc::context::operator new): New.
5631 (gcc::context::gt_ggc_mx): New.
5632 (gcc::context::gt_pch_nx): New.
5633 (gcc::context::gt_pch_nx): New.
5634 * context.h (gcc::context): Add GTY((user)) marking.
5635 (gcc::context::operator new): New.
5636 (gcc::context::gt_ggc_mx): New.
5637 (gcc::context::gt_pch_nx): New.
5638 (gcc::context::gt_pch_nx): New.
5639 (g): Add GTY marking.
5640 (gt_ggc_mx (gcc::context *)): New.
5641 (gt_pch_nx (gcc::context *)): New.
5642 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
5643 void *cookie)): New.
5644 * gengtype.c (open_base_files) <ifiles>: Add context.h.
5646 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
5648 * dce.c (fini_dce): Call df_analyze again just in case
5649 delete_unmarked_insns removed anything.
5651 2013-08-20 Teresa Johnson <tejohnson@google.com>
5653 PR rtl-optimizations/57451
5654 * final.c (reemit_insn_block_notes): Prevent lexical blocks
5655 from crossing split section boundaries.
5657 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
5659 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
5660 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
5661 with MULTLIB_DEFAULTS.
5663 2013-08-20 Nick Clifton <nickc@redhat.com>
5665 * target.def (narrow_volatile_bitfield): Note that the default
5666 value is false, not !TARGET_STRICT_ALIGN.
5667 * doc/tm.texi: Regenerate.
5669 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
5671 Fix LIB_SPEC for systems without libpthread.
5673 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
5674 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
5676 * config/i386/linux-common.h: Likewise.
5677 * config/mips/linux-common.h: Likewise.
5679 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
5681 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
5684 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5686 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
5687 instances can own GC refs.
5689 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
5690 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
5691 (gcc::context::gt_pch_nx): Likewise.
5692 (gcc::context::gt_pch_nx): Likewise.
5693 * ggc.h (gt_ggc_mx <T>): New.
5694 (gt_pch_nx_with_op <T>): New.
5695 (gt_pch_nx <T>): New.
5696 * passes.c (opt_pass::gt_ggc_mx): New.
5697 (opt_pass::gt_pch_nx): New.
5698 (opt_pass::gt_pch_nx_with_op): New.
5699 (pass_manager::gt_ggc_mx): New.
5700 (pass_manager::gt_pch_nx): New.
5701 (pass_manager::gt_pch_nx_with_op): New.
5702 (pass_manager::operator new): Use
5703 ggc_internal_cleared_alloc_stat rather than xcalloc.
5704 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
5705 (pass_manager::gt_ggc_mx): New.
5706 (pass_manager::gt_pch_nx): New.
5707 (pass_manager::gt_pch_nx_with_op): New.
5708 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
5709 (opt_pass::operator new): New.
5710 (opt_pass::gt_ggc_mx): New.
5711 (opt_pass::gt_pch_nx): New.
5712 (opt_pass::gt_pch_nx_with_op): New.
5714 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5716 * Makefile.in (GTFILES): Add context.h.
5717 * context.c (gcc::context::operator new): New.
5718 (gcc::context::gt_ggc_mx): New.
5719 (gcc::context::gt_pch_nx): New.
5720 (gcc::context::gt_pch_nx): New.
5721 * context.h (gcc::context): Add GTY((user)) marking.
5722 (gcc::context::operator new): New.
5723 (gcc::context::gt_ggc_mx): New.
5724 (gcc::context::gt_pch_nx): New.
5725 (gcc::context::gt_pch_nx): New.
5726 (g): Add GTY marking.
5727 (gt_ggc_mx (gcc::context *)): New.
5728 (gt_pch_nx (gcc::context *)): New.
5729 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
5730 void *cookie)): New.
5731 * gengtype.c (open_base_files) <ifiles>: Add context.h.
5733 2013-08-20 Alan Modra <amodra@gmail.com>
5736 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
5737 (rs6000_emit_epilogue): Likewise.
5739 2013-08-19 Dehao Chen <dehao@google.com>
5741 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
5743 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
5744 Jakub Jelinek <jakub@redhat.com>
5746 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
5747 (BUILT_IN_FABSD64): Likewise.
5748 (BUILT_IN_FABSD128): Likewise.
5749 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
5750 (fold_builtin_1): Likewise.
5751 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
5752 destination and source operands.
5753 (*abstd2_fpr): Likewise.
5754 (*nabstd2_fpr): Likewise.
5756 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
5758 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
5761 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
5763 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
5766 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
5768 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
5769 * builtins.c (default_libc_has_function): New.
5770 (gnu_libc_has_function): Ditto.
5771 (no_c99_libc_has_function): Ditto.
5772 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
5773 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
5774 (fold_builtin_sincos): Likewise.
5775 (fold_builtin_cexp): Likewise.
5776 * builtins.def (DEF_C94_BUILTIN): Likewise.
5777 (DEF_C99_BUILTIN): Likewise.
5778 (DEF_C99_C90RES_BUILTIN): Likewise.
5779 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
5780 definitions to using this define.
5781 * config/darwin-protos.h (darwin_libc_has_function): New.
5782 * config/darwin.c (darwin_libc_has_function): Ditto.
5783 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
5784 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
5785 * config/darwin.h: Ditto.
5786 * config/elfos.h: Ditto.
5787 * config/freebsd.h: Ditto.
5788 * config/i386/cygming.h: Ditto.
5789 * config/i386/djgpp.h: Ditto.
5790 * config/i386/i386-interix.h: Ditto.
5791 * config/microblaze/microblaze.h: Ditto.
5792 * config/mmix/mmix.h: Ditto.
5793 * config/gnu-user.h: Ditto.
5794 * config/ia64/hpux.h: Ditto.
5795 * config/pa/pa-hpux.h: Ditto.
5796 * config/pdp11/pdp11.h: Ditto.
5797 * config/picochip/picochip.h: Ditto.
5798 * config/linux.h: Ditto.
5799 * config/netbsd.h: Ditto.
5800 * config/openbsd.h: Ditto.
5801 * config/rs6000/aix43.h: Ditto.
5802 * config/rs6000/aix51.h: Ditto.
5803 * config/rs6000/aix52.h: Ditto.
5804 * config/rs6000/aix53.h: Ditto.
5805 * config/rs6000/aix61.h: Ditto.
5806 * config/rs6000/darwin.h: Ditto.
5807 * config/rs6000/linux.h: Ditto.
5808 * config/rs6000/linux64.h: Ditto.
5809 * config/s390/tpf.h: Ditto.
5810 * config/sol2-10.h: Ditto.
5811 * config/sol2.h: Ditto.
5812 * config/vms/vms.h: Ditto.
5813 * config/vxworks.h: Ditto.
5814 * config/linux-android.c (linux_android_libc_has_function):
5815 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
5816 * config/linux-protos.h (linux_android_libc_has_function):
5818 * config/i386/i386.c (ix86_libc_has_function): New.
5819 * config/i386/i386-protos.h
5820 (ix86_libc_has_function): New declaration.
5821 * config/i386/i386.md
5822 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
5823 ("isinf<mode>2): Likewise.
5824 * convert.c (convert_to_integer): Using new target hook
5825 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
5826 TARGET_C99_FUNCTIONS.
5827 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
5828 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
5829 * coretypes.h (function_class): New enum for different
5830 classes of functions.
5831 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
5832 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
5833 (TARGET_HAS_SINCOS): Likewise.
5834 (TARGET_LIBC_HAS_FUNCTION): New.
5835 * doc/tm.texi: Regenerated.
5836 * targhooks.h (default_libc_has_function): New declaration.
5837 (no_c99_libc_has_function): Ditto.
5838 (gnu_libc_has_function): Ditto.
5839 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
5840 and TARGET_HAS_SINCOS.
5842 2013-08-18 Jan Hubicka <jh@suse.cz>
5844 * Makeifle-in (ipa-devirt.o): New.
5845 (GTFILES): Add ipa-utils.h and ipa-devirt.c
5846 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
5847 (analyze_functions): Look into possible targets of polymorphic call.
5848 * dumpfile.c (dump_files): Add type-inheritance dump.
5849 * dumpfile.h (TDI_inheritance): New.
5850 * ipa-devirt.c: New file.
5851 * ipa-utils.h (odr_type_d): Forward declare.
5852 (odr_type): New type.
5853 (build_type_inheritance_graph): Declare.
5854 (possible_polymorphic_call_targets): Declare and introduce inline
5855 variant when only edge is pased.
5856 (dump_possible_polymorphic_call_targets): Likewise.
5857 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
5858 * tree.c (type_in_anonymous_namespace_p): Break out from ...
5859 (types_same_for_odr): ... here.
5860 * tree.h (type_in_anonymous_namespace_p): Declare.
5862 2013-08-18 Jakub Jelinek <jakub@redhat.com>
5864 PR tree-optimization/58006
5865 * tree-parloops.c (take_address_of): Don't ICE if get_name
5867 (eliminate_local_variables_stmt): Remove clobber stmts.
5869 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
5871 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
5872 error message is issued for an alias to undefined symbol.
5874 2013-08-18 Jan Hubicka <jh@suse.cz>
5876 * cgraph.c (cgraph_create_indirect_edge): Discover
5877 polymorphic calls and record basic info into indirect_info.
5878 * gimple-fold.c (gimple_fold_call): When doing BINFO based
5879 devirtualization, ignore objc function calls.
5880 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
5881 call with no parm index info.
5882 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
5883 * tree.c (virtual_method_call_p): New function.
5884 * tree.h (virtual_method_call_p): Declare.
5886 2013-08-16 Jan Hubicka <jh@suse.cz>
5889 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
5891 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
5893 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
5896 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5898 * gengtype.c (type_for_name): Add special-case support for
5899 locating types within the "gcc::" namespace.
5900 (open_base_files): Emit a "using namespace gcc" directive.
5902 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
5905 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
5906 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
5908 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
5909 array instead of each individual operand as a separate argument.
5910 (emit_fusion_gpr_load): Likewise.
5911 (expand_fusion_gpr_load): Add new function declaration.
5913 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
5914 signature to have the operands passed as an array, instead of as
5915 separate arguments. Allow ZERO_EXTEND to be in the memory
5916 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
5917 depend on the register live/dead flags when peepholes are run.
5918 (expand_fusion_gpr_load): New function to be called from the
5919 peephole2 pass, to change the register that addis sets to be the
5921 (emit_fusion_gpr_load): Change the calling signature to have the
5922 operands passed as an array, instead of as separate arguments.
5923 Allow ZERO_EXTEND to be in the memory address, and also
5924 SIGN_EXTEND if -mpower8-fusion-sign.
5926 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
5928 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
5929 adjust the register addis loads up in the peephole2 pass. Do not
5930 depend on the register live/dead state when the peephole pass is done.
5932 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5934 * gengtype.c (create_user_defined_type): Ensure that the kind
5935 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
5936 declaration is seen before the GTY((user)) marking.
5938 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
5941 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
5943 2013-08-16 Jan Hubicka <jh@suse.cz>
5945 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
5946 arugment expected_type.
5947 (gimple_fold_call): Use it.
5948 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
5949 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
5950 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
5951 (try_make_edge_direct_virtual_call): Likewise.
5952 * tree.c (obj_type_ref_class): New.
5953 * tree.h (obj_type_ref_class): Use it.
5955 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
5957 * sched-vis.c (rtl_slim_pp_initialized): Remove.
5958 (rtl_slim_pp): Likewise.
5959 (init_rtl_slim_pretty_print): Likewise.
5960 (dump_value_slim): Don't call it. Use local pretty printer.
5961 (dump_insn_slim): Likewise.
5962 (dump_rtl_slim): Likewise.
5963 (str_pattern_slim): Likewise.
5964 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
5967 2013-08-16 Jakub Jelinek <jakub@redhat.com>
5969 PR tree-optimization/58164
5970 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
5971 walk gimple_goto_dest of GIMPLE_GOTO.
5973 PR tree-optimization/58165
5974 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
5975 bi_call must be the last stmt in a bb, don't split_block, instead
5976 use fallthru edge from it and give up if there is none.
5977 Release conds vector when returning early.
5979 2013-08-14 Xinliang David Li <davidxl@google.com>
5981 * config/i386/i386.c (ix86_option_override_internal):
5982 Remove unused variable and field.
5984 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5987 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
5988 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
5989 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
5990 and Linux, correct BLKmode alignment when 128-bit alignment is
5991 required and compatibility flag is not set.
5992 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
5993 for zero-size arguments when compatibility flag is not set.
5995 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5997 PR tree-optimization/58145
5998 * tree-sra.c (build_ref_for_offset): If prev_base has
5999 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
6001 2013-08-14 Xinliang David Li <davidxl@google.com>
6003 * config/i386/i386.c (ix86_option_override_internal):
6004 Fix uninitialized variable error.
6006 2013-08-14 Xinliang David Li <davidxl@google.com>
6008 * config/i386/i386.opt: Define two new options.
6009 * config/i386/x86-tune.def: Add arch selector field in macros.
6010 * config/i386/i386.h: Adjust macro definition.
6011 * config/i386/i386.c (ix86_option_override_internal):
6013 (parse_mtune_ctrl_str): New function.
6014 (set_ix86_tune_features): New function.
6015 (ix86_function_specific_restore): Call the new helper function.
6017 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
6019 PR rtl-optimization/57662
6020 * sel-sched.c (code_motion_process_successors): When the current insn
6021 is removed after the recursive traversal, break from the loop.
6022 Add comments and debug printouts.
6024 2013-08-14 Jakub Jelinek <jakub@redhat.com>
6025 Alexandre Oliva <aoliva@redhat.com>
6028 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
6029 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
6030 there also UNSPEC_PLTOFF.
6032 2013-08-14 Marek Polacek <polacek@redhat.com>
6034 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
6035 NUM_CONDITIONS bit positions.
6037 2013-08-13 Cary Coutant <ccoutant@google.com>
6039 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
6040 (attr_checksum): Hash vector contents instead of pointer.
6041 (attr_checksum_ordered): Likewise.
6043 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
6045 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
6046 when Pmode != word_mode. Add length_address attribute.
6047 (sse3_monitor_<mode>): Merge from sse3_monitor and
6048 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
6049 Pmode != word_mode. Update insn length attribute.
6050 * config/i386/i386.c (ix86_option_override_internal): Update
6051 ix86_gen_monitor selection for merged sse3_monitor insn.
6053 2013-08-13 Julian Brown <julian@codesourcery.com>
6055 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
6056 perform invalid legitimization on greater-than-word-size modes for
6059 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
6061 * ira.c (setup_class_translate_array): Use aclass instead of cl
6062 for classes not fully covered by allocno classes.
6064 2013-08-13 Jakub Jelinek <jakub@redhat.com>
6066 PR tree-optimization/57661
6067 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
6068 * tree-inline.c (tree_function_versioning): Initialize it.
6069 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
6070 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
6071 that is not being copied.
6074 * asan.c (instrument_strlen_call): Fix typo in comment.
6075 Use char * type even for the lhs of POINTER_PLUS_EXPR.
6077 2013-08-13 Steve Ellcey <sellcey@mips.com>
6079 * config/mips/mips.md (prefetch): Use lw instead of ld on
6080 loongson in 32bit mode.
6082 2013-08-13 Nick Clifton <nickc@redhat.com>
6084 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
6086 2013-08-13 Jan Hubicka <jh@suse.cz>
6088 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
6089 introduced edge; fix typo in sanity check.
6090 (cgraph_resolve_speculation): Export; improve diagnostic.
6091 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
6092 speculation at type mismatch.
6093 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
6094 (cgraph_resolve_speculation): Declare.
6095 (symtab_can_be_discarded): New function.
6096 * value-prof.c (gimple_ic_transform): Remove actual transform code.
6097 * ipa-inline-transform.c (speculation_removed): New global var.
6098 (clone_inlined_nodes): See if speculation can be removed.
6099 (inline_call): If speculations was removed, we growths may not match.
6100 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
6101 (speculation_useful_p): New function.
6102 (resolve_noninline_speculation): New function.
6103 (inline_small_functions): Resolve useless speculations.
6104 * ipa-inline.h (speculation_useful_p): Declare
6105 * ipa.c (can_replace_by_local_alias): Simplify.
6106 (ipa_profile): Produce speculative calls in non-lto, too;
6107 add simple cost model; produce local aliases.
6109 2013-08-13 David Malcolm <dmalcolm@redhat.com>
6111 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
6114 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
6116 * config/i386/i386.c (ix86_function_versions): Use error + inform.
6118 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
6120 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
6121 iterator instead of X87MODEF.
6123 2013-08-12 Perez Read <netfirewall@gmail.com>
6126 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
6127 operand 0 for intel asm alternative.
6128 (*movabs<mode>_2): Ditto for operand 1.
6130 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
6132 * config/aarch64/arm_none.h
6133 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
6135 2013-08-12 Nick Clifton <nickc@redhat.com>
6137 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
6139 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
6141 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
6142 expand for QI/HImode operand to produce more effictive code for
6143 unsigned char(short) --> float(double) conversion.
6145 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
6147 * doc/invoke.texi: Mention that -ftls-model does not force the final
6150 2013-08-12 Marek Polacek <polacek@redhat.com>
6151 Marc Glisse <marc.glisse@inria.fr>
6153 PR tree-optimization/57980
6154 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
6155 when creating -1 constant.
6157 2013-08-10 Jan Hubicka <jh@suse.cz>
6159 Workaround binutils PR14342.
6160 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
6161 (gimple_init_edge_profiler): Likewise.
6162 (gimple_gen_ic_func_profiler): Likewise.
6164 2013-08-09 Jan Hubicka <jh@suse.cz>
6166 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
6168 2013-08-09 Xinliang David Li <davidxl@google.com>
6170 * config/i386/stringop.def: New file.
6171 * config/i386/stringop.opt: New file.
6172 * config/i386/i386-opts.h: Include stringopt.def.
6173 * config/i386/i386.opt: Include stringopt.opt.
6174 * config/i386/i386.c (ix86_option_override_internal):
6175 Override default size based stringop inline strategies with options.
6176 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
6179 2013-08-09 Jan Hubicka <jh@suse.cz>
6181 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
6183 2013-08-09 Jan Hubicka <jh@suse.cz>
6185 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
6187 (dump_cgraph_node): Dump profile-id.
6188 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
6189 and common_target_probability.
6190 * lto-cgraph.c (lto_output_edge): Stream common targets.
6191 (lto_output_node): Stream profile ids.
6192 (input_node): Stream profile ids.
6193 (input_edge): Stream common targets.
6194 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
6195 * ipa.c: Include value-prof.h
6196 (ipa_profile_generate_summary): Turn indirect call statement histograms
6197 into common targets.
6198 (ipa_profile): Turn common targets into speculative edges.
6200 2013-08-09 Jan Hubicka <jh@suse.cz>
6202 * cgraph.h (cgraph_node): Add profile_id.
6203 * value-prof.c (cgraph_node_map): Turn into pointer_map.
6204 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
6205 (del_node_map): Update.
6206 (find_func_by_funcdef_no): Replace by ...
6207 (find_func_by_profile_id): ... this one.
6208 (gimple_ic_transform): Do not remove useful histograms when
6209 speculation is not done; dump info when indirect call removal
6211 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
6212 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
6213 (__gcov_indirect_call_profiler_v2): .. this one.
6214 * profile.h (init_node_map): Update.
6215 * coverage.c (coverage_compute_profile_id): New function.
6216 * coverage.h (coverage_compute_profile_id): Declare.
6217 * tree-profile.c (init_ic_make_global_vars): Make
6218 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
6219 (gimple_init_edge_profiler): Update prototype of
6220 __gcov_indirect_call_profiler.
6221 (gimple_gen_ic_func_profiler): Simplify.
6222 (tree_profiling): Use init_node_map
6224 2013-08-09 Jan Hubicka <jh@suse.cz>
6226 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
6227 non-speculative refs.
6228 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
6229 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
6230 (cgraph_set_call_stmt): Likewise.
6231 (cgraph_create_edge_1): Fix release checking compilatoin;
6233 (cgraph_free_edge): Free indirect info.
6234 (cgraph_turn_edge_to_speculative): New function.
6235 (cgraph_speculative_call_info): New function.
6236 (cgraph_make_edge_direct): Return direct edge; handle speculation.
6237 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
6238 (dump_cgraph_node): Dump speculation.
6239 (verify_edge_count_and_frequency): Accept speculative edges.
6240 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
6241 (verify_cgraph_node): Handle speculation.
6242 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
6243 (cgraph_set_call_stmt): Update prototype.
6244 (cgraph_make_edge_direct): Update prototype.
6245 (cgraph_speculative_call_info): Declare.
6246 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
6247 to change; update call of ipa_find_references.
6248 * ipa-ref.c (ipa_record_reference): Fix return value; clear
6249 lto_stmt_uid and speculative flags.
6250 (ipa_dump_references): Dump speculation.
6251 (ipa_clone_references): Clone speculative flag.
6252 (ipa_clone_referring): Likewise.
6253 (ipa_clone_ref): New function.
6254 (ipa_find_reference): Look into lto_stmt_uids
6255 (ipa_clear_stmts_in_references): Do not clear speculative calls.
6256 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
6257 (ipa_find_reference): Update declaration.
6258 (ipa_clone_ref): Declare.
6259 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
6260 stream speculative flag.
6261 (lto_output_ref): Stream statements uids and speculation.
6262 (input_ref): Likewise.
6263 (input_edge): Stream speuclation.
6264 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
6265 (cgraph_set_call_stmt_including_clones): Handle speculation.
6266 * ipa-inline.c (heap_edge_removal_hook): New function.
6267 (inline_small_functions): Register it.
6268 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
6269 also initialize refs.
6270 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
6272 (try_make_edge_direct_simple_call): Likewise.
6273 (try_make_edge_direct_simple_call): Likewise.
6274 (update_indirect_edges_after_inlining): Likewise.
6275 (remove_described_reference): Look proper lto_stmt_uid.
6276 (propagate_controlled_uses): Likewise.
6277 (propagate_controlled_uses): Liekwise.
6278 * tree-inline.c (copy_bb): Copy speculative edges.
6279 (redirect_all_calls): New function.
6280 (copy_cfg_body): Do redirection after loop info is updated.
6281 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
6283 2013-08-09 Jan Hubicka <jh@suse.cz>
6285 * lto-streamer-out.c (output_function): Renumber PHIs.
6286 * lto-streamer-in.c (input_function): Likewise.
6288 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
6290 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
6291 (get_lane_unsigned): Likewise.
6292 (dup_lane_scalar): Likewise.
6293 (get_lane): enable for VALL.
6294 * config/aarch64/aarch64-simd.md
6295 (aarch64_dup_lane_scalar<mode>): Remove.
6296 (aarch64_get_lane_signed<mode>): Likewise.
6297 (aarch64_get_lane_unsigned<mode>): Likewise.
6298 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
6299 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
6300 (aarch64_get_lane<mode>): Enable for all vector modes.
6301 (aarch64_get_lanedi): Remove misleading constraints.
6302 * config/aarch64/arm_neon.h
6303 (__aarch64_vget_lane_any): Define.
6304 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
6305 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
6306 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
6307 * config/aarch64/iterators.md (VDQQH): New.
6311 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
6313 * configure.ac: Add GAS check for LEON instructions on SPARC.
6314 * configure: Regenerate.
6315 * config.in: Likewise.
6316 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
6318 * config/sparc/sparc.opt (LEON, LEON3): New masks.
6319 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
6321 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
6322 (AS_LEON_FLAG): New macro.
6323 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
6324 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
6325 Deal with LEON and LEON3 for the memory model.
6326 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
6327 (atomic_compare_and_swap<mode>_1): Likewise.
6328 (*atomic_compare_and_swap<mode>_1): Likewise.
6330 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6332 * config/arm/neon.md (vcond): Fix floating-point vector
6333 comparisons against 0.
6335 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
6337 * lra-constraints.c (emit_spill_move): Remove assert.
6338 (process_alt_operands): Add more debugging
6339 output. Increase reject for spilling into memory. Decrease
6340 reject for reloading scratch.
6341 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
6343 2013-08-08 Steve Ellcey <sellcey@mips.com>
6345 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
6346 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
6347 micromips incompatible. Add nan2008.
6348 (MULTILIB_DIRNAMES): Add nan2008.
6349 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
6350 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
6351 and micromips incompatible. Add nan2008.
6352 (MULTILIB_DIRNAMES): Add nan2008.
6353 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
6355 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
6357 PR rtl-optimization/58079
6358 * combine.c (combine_simplify_rtx): Avoid using SUBST if
6359 simplify_comparison has widened a comparison with an integer.
6361 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6363 * config/arm/neon.md (movmisalign<mode>): Disable when we
6364 don't allow unaligned accesses.
6365 (*movmisalign<mode>_neon_store): Likewise.
6366 (*movmisalign<mode>_neon_load): Likewise.
6367 (*movmisalign<mode>_neon_store): Likewise.
6368 (*movmisalign<mode>_neon_load): Likewise.
6370 2013-08-08 Jan Hubicka <jh@suse.cz>
6372 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
6373 (make_pass_rebuild_cgraph_edges): Also clear references.
6374 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
6375 * ipa-inline-transform.c (inline_transform): Remove all references
6377 * cgraphunit.c (expand_function): Remove all references after
6379 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
6380 (ipa_find_reference): Rewrite to iterator.
6381 (remove_stmt_references): Likewise.
6382 (ipa_clear_stmts_in_references): New function.
6383 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
6384 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
6386 * ipa-split.c (split_function): Remove references in split function.
6388 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
6391 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
6392 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
6394 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
6397 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
6398 suggested mode for the assignment isn't compatible with the
6401 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
6404 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
6406 2013-08-07 Xinliang David Li <davidxl@google.com>
6408 * config/i386/i386.opt: New option -mtune-ctrl=.
6409 * config/i386/x86-tune.def: New file.
6410 * config/i386/i386.h: include x86-tune.def.
6411 * config/i386/i386.c (ix86_option_override_internal):
6412 Parsing -mtune-ctrl= option and set tune features.
6414 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
6417 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
6418 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
6421 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
6423 * rtl.h (update_alignments): Declare.
6424 * final.c (grow_label_align): New function extracted from...
6425 (shorten_branches): ...here. Call it.
6426 (update_alignments): New function.
6427 * reorg.c (sibling_labels): New variable.
6428 (get_label_before): Add SIBLING parameter. If it is non-zero, push
6429 the new label along with it onto the sibling_labels vector.
6430 (fill_simple_delay_slots): Adjust call to get_label_before.
6431 (fill_slots_from_thread): Likewise.
6432 (relax_delay_slots): Likewise.
6433 (make_return_insns): Likewise.
6434 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
6436 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
6438 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
6439 document its semantics.
6440 (diagnostic_report_diagnostic): Adjust accordingly.
6442 2013-08-07 David Malcolm <dmalcolm@redhat.com>
6444 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
6445 (sparc_option_override): ...and port to new C++ pass API.
6446 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
6448 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
6450 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
6452 2013-08-06 Caroline Tice <cmtice@google.com>
6454 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
6455 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
6456 * tree-pass.h: Add pass_vtable_verify.
6457 * varasm.c (assemble_variable): Add code to properly set the comdat
6458 section and name for the .vtable_map_vars section.
6459 (assemble_vtyv_preinit_initializer): New function.
6460 (default_sectin_type_flags): Make sure .vtable_map_vars section has
6462 * output.h: Add function decl for assemble_vtv_preinit_initializer.
6463 * vtable-verify.c: New file.
6464 * vtable-verify.h: New file.
6465 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
6466 initialiation levels.
6467 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
6468 * passes.def: Insert pass_vtable_verify.
6469 * aclocal.m4: Reorder includes.
6470 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
6471 -fvtv-counts options.
6472 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
6473 as appropriate, if -fvtable-verify=... is used.
6474 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
6475 -fvtable-verify=... is used.
6476 * Makefile.in (OBJS): Add vtable-verify.o to list.
6477 (vtable-verify.o): Add new build rule.
6478 (GTFILES): Add vtable-verify.c to list.
6479 * common.opt (fvtable-verify=): New flag.
6480 (vtv_priority): Values for fvtable-verify= flag.
6481 (fvtv-counts): New flag.
6482 (fvtv-debug): New flag.
6483 * tree.h (save_vtable_map_decl): New extern function decl.
6485 2013-08-07 David Malcolm <dmalcolm@redhat.com>
6487 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
6488 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
6489 (pass_data_rl78_devirt): ...new pass_data instance and...
6490 (make_pass_rl78_devirt): ...new function.
6491 (rl78_asm_file_start): Port pass registration to new C++ API.
6493 2013-08-07 David Malcolm <dmalcolm@redhat.com>
6495 * coretypes.h (rtl_opt_pass): Add.
6496 (gcc::context): Add.
6497 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
6498 (epiphany_init): Port to new C++ pass API.
6499 (epiphany_optimize_mode_switching): Likewise.
6500 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
6501 (pass_manager::get_pass_mode_switching): New.
6502 (pass_manager::get_pass_peephole2): New.
6503 * mode-switching.c (pass_mode_switching): Add clone method.
6504 * recog.c (pass_peephole2): Add clone method.
6505 (pass_split_all_insns): Add clone method.
6507 2013-08-06 David Malcolm <dmalcolm@redhat.com>
6509 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
6510 (mips_option_override): ...here, porting to new C++ API for passes.
6512 2013-08-06 Jan Hubicka <jh@suse.cz>
6514 * cgraph.c (cgraph_get_body): New function based on lto.c
6516 * cgraph.h (cgraph_get_body): Declare.
6517 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
6519 * cgraphunit.c (expand_function): Get body prior expanding.
6520 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
6521 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
6523 * passes.c (do_per_function_toporder): Get body.
6524 * tree-inline.c (expand_call_inline): Get body prior inlining it.
6525 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
6527 2013-08-06 Martin Jambor <mjambor@suse.cz>
6530 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
6531 is not re-finalized. Rename second parameter to no_collect.
6533 2013-08-06 Martin Jambor <mjambor@suse.cz>
6536 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
6537 MEM_REF has proper alignment information.
6539 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
6542 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
6544 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
6545 rtx (*) (rtx, ...) with insn_gen_fn.
6546 * genoutput.c (output_insn_data): Cast gen_? function pointers to
6547 insn_gen_fn::stored_funcptr. Add initializer braces.
6549 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6551 Rewrite how instances of passes are cloned to remove assumptions
6552 about their sizes (thus allowing pass subclasses to have
6553 additional data fields, albeit non-GC-managed ones at this point).
6555 * passes.c (make_pass_instance): Now that passes have clone
6556 methods, rewrite this function to eliminate XNEW and memcpy
6557 calls that used hardcoded sizes. Since this function no longer
6558 creates pass instances, rename it to...
6559 (add_pass_instance): ...this. Document the old way that passes were
6560 numbered and flagged, and rework this function to continue using it.
6561 (next_pass_1): Add an initial_pass argument for use by
6563 (position_pass): When adding multiple instances of a pass, use
6564 the pass's clone method, rather than relying on the XNEW/memcpy
6565 within the former make_pass_instance (now add_pass_instance).
6566 (pass_manager::pass_manager): When invoking next_pass_1, also supply
6567 the initial instance of the current pass within the pass manager.
6569 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6571 This is the automated part of the conversion of passes from C
6572 structs to C++ classes.
6574 Patch autogenerated by refactor_passes.py from
6575 https://github.com/davidmalcolm/gcc-refactoring-scripts
6576 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
6578 * asan.c (pass_asan): Convert from a global struct to a subclass of
6579 gimple_opt_pass along with...
6580 (pass_data_asan): ...new pass_data instance and...
6581 (make_pass_asan): ...new function.
6582 (pass_asan_O0): Convert from a global struct to a subclass of
6583 gimple_opt_pass along with...
6584 (pass_data_asan_O0): ...new pass_data instance and...
6585 (make_pass_asan_O0): ...new function.
6586 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
6587 subclass of rtl_opt_pass along with...
6588 (pass_data_inc_dec): ...new pass_data instance and...
6589 (make_pass_inc_dec): ...new function.
6590 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
6591 a subclass of rtl_opt_pass along with...
6592 (pass_data_reorder_blocks): ...new pass_data instance and...
6593 (make_pass_reorder_blocks): ...new function.
6594 (pass_duplicate_computed_gotos): Convert from a global struct to a
6595 subclass of rtl_opt_pass along with...
6596 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
6597 (make_pass_duplicate_computed_gotos): ...new function.
6598 (pass_partition_blocks): Convert from a global struct to a subclass of
6599 rtl_opt_pass along with...
6600 (pass_data_partition_blocks): ...new pass_data instance and...
6601 (make_pass_partition_blocks): ...new function.
6602 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
6603 struct to a subclass of rtl_opt_pass along with...
6604 (pass_data_branch_target_load_optimize1): ...new pass_data instance
6606 (make_pass_branch_target_load_optimize1): ...new function.
6607 (pass_branch_target_load_optimize2): Convert from a global struct to a
6608 subclass of rtl_opt_pass along with...
6609 (pass_data_branch_target_load_optimize2): ...new pass_data instance
6611 (make_pass_branch_target_load_optimize2): ...new function.
6612 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
6613 of rtl_opt_pass along with...
6614 (pass_data_jump): ...new pass_data instance and...
6615 (make_pass_jump): ...new function.
6616 (pass_jump2): Convert from a global struct to a subclass of
6617 rtl_opt_pass along with...
6618 (pass_data_jump2): ...new pass_data instance and...
6619 (make_pass_jump2): ...new function.
6620 * cfgexpand.c (pass_expand): Convert from a global struct to a
6621 subclass of rtl_opt_pass along with...
6622 (pass_data_expand): ...new pass_data instance and...
6623 (make_pass_expand): ...new function.
6624 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
6625 of rtl_opt_pass along with...
6626 (pass_data_free_cfg): ...new pass_data instance and...
6627 (make_pass_free_cfg): ...new function.
6628 (pass_into_cfg_layout_mode): Convert from a global struct to a
6629 subclass of rtl_opt_pass along with...
6630 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
6631 (make_pass_into_cfg_layout_mode): ...new function.
6632 (pass_outof_cfg_layout_mode): Convert from a global struct to a
6633 subclass of rtl_opt_pass along with...
6634 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
6635 (make_pass_outof_cfg_layout_mode): ...new function.
6636 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
6637 struct to a subclass of gimple_opt_pass along with...
6638 (pass_data_build_cgraph_edges): ...new pass_data instance and...
6639 (make_pass_build_cgraph_edges): ...new function.
6640 (pass_rebuild_cgraph_edges): Convert from a global struct to a
6641 subclass of gimple_opt_pass along with...
6642 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
6643 (make_pass_rebuild_cgraph_edges): ...new function.
6644 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
6645 subclass of gimple_opt_pass along with...
6646 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
6648 (make_pass_remove_cgraph_callee_edges): ...new function.
6649 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
6650 struct to a subclass of rtl_opt_pass along with...
6651 (pass_data_stack_adjustments): ...new pass_data instance and...
6652 (make_pass_stack_adjustments): ...new function.
6653 * combine.c (pass_combine): Convert from a global struct to a subclass
6654 of rtl_opt_pass along with...
6655 (pass_data_combine): ...new pass_data instance and...
6656 (make_pass_combine): ...new function.
6657 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
6658 global struct to a subclass of rtl_opt_pass along with...
6659 (pass_data_compare_elim_after_reload): ...new pass_data instance
6661 (make_pass_compare_elim_after_reload): ...new function.
6662 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
6663 of rtl_opt_pass along with...
6664 (pass_data_rtl_cprop): ...new pass_data instance and...
6665 (make_pass_rtl_cprop): ...new function.
6666 * cse.c (pass_cse): Convert from a global struct to a subclass of
6667 rtl_opt_pass along with...
6668 (pass_data_cse): ...new pass_data instance and...
6669 (make_pass_cse): ...new function.
6670 (pass_cse2): Convert from a global struct to a subclass of
6671 rtl_opt_pass along with...
6672 (pass_data_cse2): ...new pass_data instance and...
6673 (make_pass_cse2): ...new function.
6674 (pass_cse_after_global_opts): Convert from a global struct to a
6675 subclass of rtl_opt_pass along with...
6676 (pass_data_cse_after_global_opts): ...new pass_data instance and...
6677 (make_pass_cse_after_global_opts): ...new function.
6678 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
6679 of rtl_opt_pass along with...
6680 (pass_data_ud_rtl_dce): ...new pass_data instance and...
6681 (make_pass_ud_rtl_dce): ...new function.
6682 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
6683 rtl_opt_pass along with...
6684 (pass_data_fast_rtl_dce): ...new pass_data instance and...
6685 (make_pass_fast_rtl_dce): ...new function.
6686 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
6687 a subclass of rtl_opt_pass along with...
6688 (pass_data_df_initialize_opt): ...new pass_data instance and...
6689 (make_pass_df_initialize_opt): ...new function.
6690 (pass_df_initialize_no_opt): Convert from a global struct to a
6691 subclass of rtl_opt_pass along with...
6692 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
6693 (make_pass_df_initialize_no_opt): ...new function.
6694 (pass_df_finish): Convert from a global struct to a subclass of
6695 rtl_opt_pass along with...
6696 (pass_data_df_finish): ...new pass_data instance and...
6697 (make_pass_df_finish): ...new function.
6698 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
6699 rtl_opt_pass along with...
6700 (pass_data_rtl_dse1): ...new pass_data instance and...
6701 (make_pass_rtl_dse1): ...new function.
6702 (pass_rtl_dse2): Convert from a global struct to a subclass of
6703 rtl_opt_pass along with...
6704 (pass_data_rtl_dse2): ...new pass_data instance and...
6705 (make_pass_rtl_dse2): ...new function.
6706 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
6707 subclass of rtl_opt_pass along with...
6708 (pass_data_dwarf2_frame): ...new pass_data instance and...
6709 (make_pass_dwarf2_frame): ...new function.
6710 * except.c (pass_set_nothrow_function_flags): Convert from a global
6711 struct to a subclass of rtl_opt_pass along with...
6712 (pass_data_set_nothrow_function_flags): ...new pass_data instance
6714 (make_pass_set_nothrow_function_flags): ...new function.
6715 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
6716 subclass of rtl_opt_pass along with...
6717 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
6719 (make_pass_convert_to_eh_region_ranges): ...new function.
6720 * final.c (pass_compute_alignments): Convert from a global struct to a
6721 subclass of rtl_opt_pass along with...
6722 (pass_data_compute_alignments): ...new pass_data instance and...
6723 (make_pass_compute_alignments): ...new function.
6724 (pass_final): Convert from a global struct to a subclass of
6725 rtl_opt_pass along with...
6726 (pass_data_final): ...new pass_data instance and...
6727 (make_pass_final): ...new function.
6728 (pass_shorten_branches): Convert from a global struct to a subclass of
6729 rtl_opt_pass along with...
6730 (pass_data_shorten_branches): ...new pass_data instance and...
6731 (make_pass_shorten_branches): ...new function.
6732 (pass_clean_state): Convert from a global struct to a subclass of
6733 rtl_opt_pass along with...
6734 (pass_data_clean_state): ...new pass_data instance and...
6735 (make_pass_clean_state): ...new function.
6736 * function.c (pass_instantiate_virtual_regs): Convert from a global
6737 struct to a subclass of rtl_opt_pass along with...
6738 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
6739 (make_pass_instantiate_virtual_regs): ...new function.
6740 (pass_leaf_regs): Convert from a global struct to a subclass of
6741 rtl_opt_pass along with...
6742 (pass_data_leaf_regs): ...new pass_data instance and...
6743 (make_pass_leaf_regs): ...new function.
6744 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
6745 subclass of rtl_opt_pass along with...
6746 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
6748 (make_pass_thread_prologue_and_epilogue): ...new function.
6749 (pass_match_asm_constraints): Convert from a global struct to a
6750 subclass of rtl_opt_pass along with...
6751 (pass_data_match_asm_constraints): ...new pass_data instance and...
6752 (make_pass_match_asm_constraints): ...new function.
6753 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
6754 subclass of rtl_opt_pass along with...
6755 (pass_data_rtl_fwprop): ...new pass_data instance and...
6756 (make_pass_rtl_fwprop): ...new function.
6757 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
6758 rtl_opt_pass along with...
6759 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
6760 (make_pass_rtl_fwprop_addr): ...new function.
6761 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
6762 rtl_opt_pass along with...
6763 (pass_data_rtl_pre): ...new pass_data instance and...
6764 (make_pass_rtl_pre): ...new function.
6765 (pass_rtl_hoist): Convert from a global struct to a subclass of
6766 rtl_opt_pass along with...
6767 (pass_data_rtl_hoist): ...new pass_data instance and...
6768 (make_pass_rtl_hoist): ...new function.
6769 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
6770 subclass of gimple_opt_pass along with...
6771 (pass_data_lower_cf): ...new pass_data instance and...
6772 (make_pass_lower_cf): ...new function.
6773 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
6774 from a global struct to a subclass of gimple_opt_pass along with...
6775 (pass_data_strength_reduction): ...new pass_data instance and...
6776 (make_pass_strength_reduction): ...new function.
6777 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
6778 of rtl_opt_pass along with...
6779 (pass_data_rtl_ifcvt): ...new pass_data instance and...
6780 (make_pass_rtl_ifcvt): ...new function.
6781 (pass_if_after_combine): Convert from a global struct to a subclass of
6782 rtl_opt_pass along with...
6783 (pass_data_if_after_combine): ...new pass_data instance and...
6784 (make_pass_if_after_combine): ...new function.
6785 (pass_if_after_reload): Convert from a global struct to a subclass of
6786 rtl_opt_pass along with...
6787 (pass_data_if_after_reload): ...new pass_data instance and...
6788 (make_pass_if_after_reload): ...new function.
6789 * init-regs.c (pass_initialize_regs): Convert from a global struct to
6790 a subclass of rtl_opt_pass along with...
6791 (pass_data_initialize_regs): ...new pass_data instance and...
6792 (make_pass_initialize_regs): ...new function.
6793 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
6794 of ipa_opt_pass_d along with...
6795 (pass_data_ipa_cp): ...new pass_data instance and...
6796 (make_pass_ipa_cp): ...new function.
6797 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
6798 global struct to a subclass of gimple_opt_pass along with...
6799 (pass_data_inline_parameters): ...new pass_data instance and...
6800 (make_pass_inline_parameters): ...new function.
6801 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
6802 subclass of gimple_opt_pass along with...
6803 (pass_data_early_inline): ...new pass_data instance and...
6804 (make_pass_early_inline): ...new function.
6805 (pass_ipa_inline): Convert from a global struct to a subclass of
6806 ipa_opt_pass_d along with...
6807 (pass_data_ipa_inline): ...new pass_data instance and...
6808 (make_pass_ipa_inline): ...new function.
6809 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
6810 struct to a subclass of gimple_opt_pass along with...
6811 (pass_data_local_pure_const): ...new pass_data instance and...
6812 (make_pass_local_pure_const): ...new function.
6813 (pass_ipa_pure_const): Convert from a global struct to a subclass of
6814 ipa_opt_pass_d along with...
6815 (pass_data_ipa_pure_const): ...new pass_data instance and...
6816 (make_pass_ipa_pure_const): ...new function.
6817 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
6818 to a subclass of ipa_opt_pass_d along with...
6819 (pass_data_ipa_reference): ...new pass_data instance and...
6820 (make_pass_ipa_reference): ...new function.
6821 * ipa-split.c (pass_split_functions): Convert from a global struct to
6822 a subclass of gimple_opt_pass along with...
6823 (pass_data_split_functions): ...new pass_data instance and...
6824 (make_pass_split_functions): ...new function.
6825 (pass_feedback_split_functions): Convert from a global struct to a
6826 subclass of gimple_opt_pass along with...
6827 (pass_data_feedback_split_functions): ...new pass_data instance and...
6828 (make_pass_feedback_split_functions): ...new function.
6829 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
6830 global struct to a subclass of simple_ipa_opt_pass along with...
6831 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
6833 (make_pass_ipa_function_and_variable_visibility): ...new function.
6834 (pass_ipa_free_inline_summary): Convert from a global struct to a
6835 subclass of simple_ipa_opt_pass along with...
6836 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
6837 (make_pass_ipa_free_inline_summary): ...new function.
6838 (pass_ipa_whole_program_visibility): Convert from a global struct to a
6839 subclass of ipa_opt_pass_d along with...
6840 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
6842 (make_pass_ipa_whole_program_visibility): ...new function.
6843 (pass_ipa_profile): Convert from a global struct to a subclass of
6844 ipa_opt_pass_d along with...
6845 (pass_data_ipa_profile): ...new pass_data instance and...
6846 (make_pass_ipa_profile): ...new function.
6847 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
6848 ipa_opt_pass_d along with...
6849 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
6850 (make_pass_ipa_cdtor_merge): ...new function.
6851 * ira.c (pass_ira): Convert from a global struct to a subclass of
6852 rtl_opt_pass along with...
6853 (pass_data_ira): ...new pass_data instance and...
6854 (make_pass_ira): ...new function.
6855 (pass_reload): Convert from a global struct to a subclass of
6856 rtl_opt_pass along with...
6857 (pass_data_reload): ...new pass_data instance and...
6858 (make_pass_reload): ...new function.
6859 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
6860 subclass of rtl_opt_pass along with...
6861 (pass_data_cleanup_barriers): ...new pass_data instance and...
6862 (make_pass_cleanup_barriers): ...new function.
6863 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
6864 of rtl_opt_pass along with...
6865 (pass_data_loop2): ...new pass_data instance and...
6866 (make_pass_loop2): ...new function.
6867 (pass_rtl_loop_init): Convert from a global struct to a subclass of
6868 rtl_opt_pass along with...
6869 (pass_data_rtl_loop_init): ...new pass_data instance and...
6870 (make_pass_rtl_loop_init): ...new function.
6871 (pass_rtl_loop_done): Convert from a global struct to a subclass of
6872 rtl_opt_pass along with...
6873 (pass_data_rtl_loop_done): ...new pass_data instance and...
6874 (make_pass_rtl_loop_done): ...new function.
6875 (pass_rtl_move_loop_invariants): Convert from a global struct to a
6876 subclass of rtl_opt_pass along with...
6877 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
6878 (make_pass_rtl_move_loop_invariants): ...new function.
6879 (pass_rtl_unswitch): Convert from a global struct to a subclass of
6880 rtl_opt_pass along with...
6881 (pass_data_rtl_unswitch): ...new pass_data instance and...
6882 (make_pass_rtl_unswitch): ...new function.
6883 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
6884 subclass of rtl_opt_pass along with...
6885 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
6887 (make_pass_rtl_unroll_and_peel_loops): ...new function.
6888 (pass_rtl_doloop): Convert from a global struct to a subclass of
6889 rtl_opt_pass along with...
6890 (pass_data_rtl_doloop): ...new pass_data instance and...
6891 (make_pass_rtl_doloop): ...new function.
6892 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
6893 a subclass of rtl_opt_pass along with...
6894 (pass_data_lower_subreg): ...new pass_data instance and...
6895 (make_pass_lower_subreg): ...new function.
6896 (pass_lower_subreg2): Convert from a global struct to a subclass of
6897 rtl_opt_pass along with...
6898 (pass_data_lower_subreg2): ...new pass_data instance and...
6899 (make_pass_lower_subreg2): ...new function.
6900 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
6901 struct to a subclass of ipa_opt_pass_d along with...
6902 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
6903 (make_pass_ipa_lto_gimple_out): ...new function.
6904 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
6905 of ipa_opt_pass_d along with...
6906 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
6907 (make_pass_ipa_lto_finish_out): ...new function.
6908 * mode-switching.c (pass_mode_switching): Convert from a global struct
6909 to a subclass of rtl_opt_pass along with...
6910 (pass_data_mode_switching): ...new pass_data instance and...
6911 (make_pass_mode_switching): ...new function.
6912 * modulo-sched.c (pass_sms): Convert from a global struct to a
6913 subclass of rtl_opt_pass along with...
6914 (pass_data_sms): ...new pass_data instance and...
6915 (make_pass_sms): ...new function.
6916 * omp-low.c (pass_expand_omp): Convert from a global struct to a
6917 subclass of gimple_opt_pass along with...
6918 (pass_data_expand_omp): ...new pass_data instance and...
6919 (make_pass_expand_omp): ...new function.
6920 (pass_lower_omp): Convert from a global struct to a subclass of
6921 gimple_opt_pass along with...
6922 (pass_data_lower_omp): ...new pass_data instance and...
6923 (make_pass_lower_omp): ...new function.
6924 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
6925 of gimple_opt_pass along with...
6926 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
6927 (make_pass_diagnose_omp_blocks): ...new function.
6928 * passes.c (pass_early_local_passes): Convert from a global struct to
6929 a subclass of simple_ipa_opt_pass along with...
6930 (pass_data_early_local_passes): ...new pass_data instance and...
6931 (make_pass_early_local_passes): ...new function.
6932 (pass_all_early_optimizations): Convert from a global struct to a
6933 subclass of gimple_opt_pass along with...
6934 (pass_data_all_early_optimizations): ...new pass_data instance and...
6935 (make_pass_all_early_optimizations): ...new function.
6936 (pass_all_optimizations): Convert from a global struct to a subclass
6937 of gimple_opt_pass along with...
6938 (pass_data_all_optimizations): ...new pass_data instance and...
6939 (make_pass_all_optimizations): ...new function.
6940 (pass_all_optimizations_g): Convert from a global struct to a subclass
6941 of gimple_opt_pass along with...
6942 (pass_data_all_optimizations_g): ...new pass_data instance and...
6943 (make_pass_all_optimizations_g): ...new function.
6944 (pass_rest_of_compilation): Convert from a global struct to a subclass
6945 of rtl_opt_pass along with...
6946 (pass_data_rest_of_compilation): ...new pass_data instance and...
6947 (make_pass_rest_of_compilation): ...new function.
6948 (pass_postreload): Convert from a global struct to a subclass of
6949 rtl_opt_pass along with...
6950 (pass_data_postreload): ...new pass_data instance and...
6951 (make_pass_postreload): ...new function.
6952 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
6953 subclass of rtl_opt_pass along with...
6954 (pass_data_gcse2): ...new pass_data instance and...
6955 (make_pass_gcse2): ...new function.
6956 * postreload.c (pass_postreload_cse): Convert from a global struct to
6957 a subclass of rtl_opt_pass along with...
6958 (pass_data_postreload_cse): ...new pass_data instance and...
6959 (make_pass_postreload_cse): ...new function.
6960 * predict.c (pass_profile): Convert from a global struct to a subclass
6961 of gimple_opt_pass along with...
6962 (pass_data_profile): ...new pass_data instance and...
6963 (make_pass_profile): ...new function.
6964 (pass_strip_predict_hints): Convert from a global struct to a subclass
6965 of gimple_opt_pass along with...
6966 (pass_data_strip_predict_hints): ...new pass_data instance and...
6967 (make_pass_strip_predict_hints): ...new function.
6968 * recog.c (pass_peephole2): Convert from a global struct to a subclass
6969 of rtl_opt_pass along with...
6970 (pass_data_peephole2): ...new pass_data instance and...
6971 (make_pass_peephole2): ...new function.
6972 (pass_split_all_insns): Convert from a global struct to a subclass of
6973 rtl_opt_pass along with...
6974 (pass_data_split_all_insns): ...new pass_data instance and...
6975 (make_pass_split_all_insns): ...new function.
6976 (pass_split_after_reload): Convert from a global struct to a subclass
6977 of rtl_opt_pass along with...
6978 (pass_data_split_after_reload): ...new pass_data instance and...
6979 (make_pass_split_after_reload): ...new function.
6980 (pass_split_before_regstack): Convert from a global struct to a
6981 subclass of rtl_opt_pass along with...
6982 (pass_data_split_before_regstack): ...new pass_data instance and...
6983 (make_pass_split_before_regstack): ...new function.
6984 (pass_split_before_sched2): Convert from a global struct to a subclass
6985 of rtl_opt_pass along with...
6986 (pass_data_split_before_sched2): ...new pass_data instance and...
6987 (make_pass_split_before_sched2): ...new function.
6988 (pass_split_for_shorten_branches): Convert from a global struct to a
6989 subclass of rtl_opt_pass along with...
6990 (pass_data_split_for_shorten_branches): ...new pass_data instance
6992 (make_pass_split_for_shorten_branches): ...new function.
6993 * ree.c (pass_ree): Convert from a global struct to a subclass of
6994 rtl_opt_pass along with...
6995 (pass_data_ree): ...new pass_data instance and...
6996 (make_pass_ree): ...new function.
6997 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
6998 subclass of rtl_opt_pass along with...
6999 (pass_data_stack_regs): ...new pass_data instance and...
7000 (make_pass_stack_regs): ...new function.
7001 (pass_stack_regs_run): Convert from a global struct to a subclass of
7002 rtl_opt_pass along with...
7003 (pass_data_stack_regs_run): ...new pass_data instance and...
7004 (make_pass_stack_regs_run): ...new function.
7005 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
7006 subclass of rtl_opt_pass along with...
7007 (pass_data_cprop_hardreg): ...new pass_data instance and...
7008 (make_pass_cprop_hardreg): ...new function.
7009 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
7010 subclass of rtl_opt_pass along with...
7011 (pass_data_reginfo_init): ...new pass_data instance and...
7012 (make_pass_reginfo_init): ...new function.
7013 * regmove.c (pass_regmove): Convert from a global struct to a subclass
7014 of rtl_opt_pass along with...
7015 (pass_data_regmove): ...new pass_data instance and...
7016 (make_pass_regmove): ...new function.
7017 * regrename.c (pass_regrename): Convert from a global struct to a
7018 subclass of rtl_opt_pass along with...
7019 (pass_data_regrename): ...new pass_data instance and...
7020 (make_pass_regrename): ...new function.
7021 * reorg.c (pass_delay_slots): Convert from a global struct to a
7022 subclass of rtl_opt_pass along with...
7023 (pass_data_delay_slots): ...new pass_data instance and...
7024 (make_pass_delay_slots): ...new function.
7025 (pass_machine_reorg): Convert from a global struct to a subclass of
7026 rtl_opt_pass along with...
7027 (pass_data_machine_reorg): ...new pass_data instance and...
7028 (make_pass_machine_reorg): ...new function.
7029 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
7030 of rtl_opt_pass along with...
7031 (pass_data_sched): ...new pass_data instance and...
7032 (make_pass_sched): ...new function.
7033 (pass_sched2): Convert from a global struct to a subclass of
7034 rtl_opt_pass along with...
7035 (pass_data_sched2): ...new pass_data instance and...
7036 (make_pass_sched2): ...new function.
7037 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
7038 to a subclass of rtl_opt_pass along with...
7039 (pass_data_stack_ptr_mod): ...new pass_data instance and...
7040 (make_pass_stack_ptr_mod): ...new function.
7041 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
7042 to a subclass of rtl_opt_pass along with...
7043 (pass_data_rtl_store_motion): ...new pass_data instance and...
7044 (make_pass_rtl_store_motion): ...new function.
7045 * tracer.c (pass_tracer): Convert from a global struct to a subclass
7046 of gimple_opt_pass along with...
7047 (pass_data_tracer): ...new pass_data instance and...
7048 (make_pass_tracer): ...new function.
7049 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
7050 to a subclass of gimple_opt_pass along with...
7051 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
7052 (make_pass_diagnose_tm_blocks): ...new function.
7053 (pass_lower_tm): Convert from a global struct to a subclass of
7054 gimple_opt_pass along with...
7055 (pass_data_lower_tm): ...new pass_data instance and...
7056 (make_pass_lower_tm): ...new function.
7057 (pass_tm_init): Convert from a global struct to a subclass of
7058 gimple_opt_pass along with...
7059 (pass_data_tm_init): ...new pass_data instance and...
7060 (make_pass_tm_init): ...new function.
7061 (pass_tm_mark): Convert from a global struct to a subclass of
7062 gimple_opt_pass along with...
7063 (pass_data_tm_mark): ...new pass_data instance and...
7064 (make_pass_tm_mark): ...new function.
7065 (pass_tm_edges): Convert from a global struct to a subclass of
7066 gimple_opt_pass along with...
7067 (pass_data_tm_edges): ...new pass_data instance and...
7068 (make_pass_tm_edges): ...new function.
7069 (pass_tm_memopt): Convert from a global struct to a subclass of
7070 gimple_opt_pass along with...
7071 (pass_data_tm_memopt): ...new pass_data instance and...
7072 (make_pass_tm_memopt): ...new function.
7073 (pass_ipa_tm): Convert from a global struct to a subclass of
7074 simple_ipa_opt_pass along with...
7075 (pass_data_ipa_tm): ...new pass_data instance and...
7076 (make_pass_ipa_tm): ...new function.
7077 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
7078 subclass of gimple_opt_pass along with...
7079 (pass_data_call_cdce): ...new pass_data instance and...
7080 (make_pass_call_cdce): ...new function.
7081 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
7082 subclass of gimple_opt_pass along with...
7083 (pass_data_build_cfg): ...new pass_data instance and...
7084 (make_pass_build_cfg): ...new function.
7085 (pass_split_crit_edges): Convert from a global struct to a subclass of
7086 gimple_opt_pass along with...
7087 (pass_data_split_crit_edges): ...new pass_data instance and...
7088 (make_pass_split_crit_edges): ...new function.
7089 (pass_warn_function_return): Convert from a global struct to a
7090 subclass of gimple_opt_pass along with...
7091 (pass_data_warn_function_return): ...new pass_data instance and...
7092 (make_pass_warn_function_return): ...new function.
7093 (pass_warn_function_noreturn): Convert from a global struct to a
7094 subclass of gimple_opt_pass along with...
7095 (pass_data_warn_function_noreturn): ...new pass_data instance and...
7096 (make_pass_warn_function_noreturn): ...new function.
7097 (pass_warn_unused_result): Convert from a global struct to a subclass
7098 of gimple_opt_pass along with...
7099 (pass_data_warn_unused_result): ...new pass_data instance and...
7100 (make_pass_warn_unused_result): ...new function.
7101 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
7102 a subclass of gimple_opt_pass along with...
7103 (pass_data_merge_phi): ...new pass_data instance and...
7104 (make_pass_merge_phi): ...new function.
7105 * tree-complex.c (pass_lower_complex): Convert from a global struct to
7106 a subclass of gimple_opt_pass along with...
7107 (pass_data_lower_complex): ...new pass_data instance and...
7108 (make_pass_lower_complex): ...new function.
7109 (pass_lower_complex_O0): Convert from a global struct to a subclass of
7110 gimple_opt_pass along with...
7111 (pass_data_lower_complex_O0): ...new pass_data instance and...
7112 (make_pass_lower_complex_O0): ...new function.
7113 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
7114 subclass of gimple_opt_pass along with...
7115 (pass_data_lower_eh): ...new pass_data instance and...
7116 (make_pass_lower_eh): ...new function.
7117 (pass_refactor_eh): Convert from a global struct to a subclass of
7118 gimple_opt_pass along with...
7119 (pass_data_refactor_eh): ...new pass_data instance and...
7120 (make_pass_refactor_eh): ...new function.
7121 (pass_lower_resx): Convert from a global struct to a subclass of
7122 gimple_opt_pass along with...
7123 (pass_data_lower_resx): ...new pass_data instance and...
7124 (make_pass_lower_resx): ...new function.
7125 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
7126 of gimple_opt_pass along with...
7127 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
7128 (make_pass_lower_eh_dispatch): ...new function.
7129 (pass_cleanup_eh): Convert from a global struct to a subclass of
7130 gimple_opt_pass along with...
7131 (pass_data_cleanup_eh): ...new pass_data instance and...
7132 (make_pass_cleanup_eh): ...new function.
7133 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
7134 to a subclass of simple_ipa_opt_pass along with...
7135 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
7136 (make_pass_ipa_lower_emutls): ...new function.
7137 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
7138 a subclass of gimple_opt_pass along with...
7139 (pass_data_if_conversion): ...new pass_data instance and...
7140 (make_pass_if_conversion): ...new function.
7141 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
7142 subclass of gimple_opt_pass along with...
7143 (pass_data_build_ssa): ...new pass_data instance and...
7144 (make_pass_build_ssa): ...new function.
7145 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
7146 global struct to a subclass of gimple_opt_pass along with...
7147 (pass_data_loop_distribution): ...new pass_data instance and...
7148 (make_pass_loop_distribution): ...new function.
7149 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
7150 subclass of gimple_opt_pass along with...
7151 (pass_data_mudflap_1): ...new pass_data instance and...
7152 (make_pass_mudflap_1): ...new function.
7153 (pass_mudflap_2): Convert from a global struct to a subclass of
7154 gimple_opt_pass along with...
7155 (pass_data_mudflap_2): ...new pass_data instance and...
7156 (make_pass_mudflap_2): ...new function.
7157 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
7158 subclass of gimple_opt_pass along with...
7159 (pass_data_mudflap_1): ...new pass_data instance and...
7160 (make_pass_mudflap_1): ...new function.
7161 (pass_mudflap_2): Convert from a global struct to a subclass of
7162 gimple_opt_pass along with...
7163 (pass_data_mudflap_2): ...new pass_data instance and...
7164 (make_pass_mudflap_2): ...new function.
7165 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
7166 gimple_opt_pass along with...
7167 (pass_data_nrv): ...new pass_data instance and...
7168 (make_pass_nrv): ...new function.
7169 (pass_return_slot): Convert from a global struct to a subclass of
7170 gimple_opt_pass along with...
7171 (pass_data_return_slot): ...new pass_data instance and...
7172 (make_pass_return_slot): ...new function.
7173 * tree-object-size.c (pass_object_sizes): Convert from a global struct
7174 to a subclass of gimple_opt_pass along with...
7175 (pass_data_object_sizes): ...new pass_data instance and...
7176 (make_pass_object_sizes): ...new function.
7177 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
7178 global struct to a subclass of gimple_opt_pass along with...
7179 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
7181 (make_pass_cleanup_cfg_post_optimizing): ...new function.
7182 (pass_fixup_cfg): Convert from a global struct to a subclass of
7183 gimple_opt_pass along with...
7184 (pass_data_fixup_cfg): ...new pass_data instance and...
7185 (make_pass_fixup_cfg): ...new function.
7186 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
7187 (make_pass_mudflap_1): ...new function.
7188 (pass_mudflap_2): Replace declaration with that of...
7189 (make_pass_mudflap_2): ...new function.
7190 (pass_asan): Replace declaration with that of...
7191 (make_pass_asan): ...new function.
7192 (pass_asan_O0): Replace declaration with that of...
7193 (make_pass_asan_O0): ...new function.
7194 (pass_tsan): Replace declaration with that of...
7195 (make_pass_tsan): ...new function.
7196 (pass_tsan_O0): Replace declaration with that of...
7197 (make_pass_tsan_O0): ...new function.
7198 (pass_lower_cf): Replace declaration with that of...
7199 (make_pass_lower_cf): ...new function.
7200 (pass_refactor_eh): Replace declaration with that of...
7201 (make_pass_refactor_eh): ...new function.
7202 (pass_lower_eh): Replace declaration with that of...
7203 (make_pass_lower_eh): ...new function.
7204 (pass_lower_eh_dispatch): Replace declaration with that of...
7205 (make_pass_lower_eh_dispatch): ...new function.
7206 (pass_lower_resx): Replace declaration with that of...
7207 (make_pass_lower_resx): ...new function.
7208 (pass_build_cfg): Replace declaration with that of...
7209 (make_pass_build_cfg): ...new function.
7210 (pass_early_tree_profile): Replace declaration with that of...
7211 (make_pass_early_tree_profile): ...new function.
7212 (pass_cleanup_eh): Replace declaration with that of...
7213 (make_pass_cleanup_eh): ...new function.
7214 (pass_sra): Replace declaration with that of...
7215 (make_pass_sra): ...new function.
7216 (pass_sra_early): Replace declaration with that of...
7217 (make_pass_sra_early): ...new function.
7218 (pass_early_ipa_sra): Replace declaration with that of...
7219 (make_pass_early_ipa_sra): ...new function.
7220 (pass_tail_recursion): Replace declaration with that of...
7221 (make_pass_tail_recursion): ...new function.
7222 (pass_tail_calls): Replace declaration with that of...
7223 (make_pass_tail_calls): ...new function.
7224 (pass_tree_loop): Replace declaration with that of...
7225 (make_pass_tree_loop): ...new function.
7226 (pass_tree_loop_init): Replace declaration with that of...
7227 (make_pass_tree_loop_init): ...new function.
7228 (pass_lim): Replace declaration with that of...
7229 (make_pass_lim): ...new function.
7230 (pass_tree_unswitch): Replace declaration with that of...
7231 (make_pass_tree_unswitch): ...new function.
7232 (pass_predcom): Replace declaration with that of...
7233 (make_pass_predcom): ...new function.
7234 (pass_iv_canon): Replace declaration with that of...
7235 (make_pass_iv_canon): ...new function.
7236 (pass_scev_cprop): Replace declaration with that of...
7237 (make_pass_scev_cprop): ...new function.
7238 (pass_empty_loop): Replace declaration with that of...
7239 (make_pass_empty_loop): ...new function.
7240 (pass_record_bounds): Replace declaration with that of...
7241 (make_pass_record_bounds): ...new function.
7242 (pass_graphite): Replace declaration with that of...
7243 (make_pass_graphite): ...new function.
7244 (pass_graphite_transforms): Replace declaration with that of...
7245 (make_pass_graphite_transforms): ...new function.
7246 (pass_if_conversion): Replace declaration with that of...
7247 (make_pass_if_conversion): ...new function.
7248 (pass_loop_distribution): Replace declaration with that of...
7249 (make_pass_loop_distribution): ...new function.
7250 (pass_vectorize): Replace declaration with that of...
7251 (make_pass_vectorize): ...new function.
7252 (pass_slp_vectorize): Replace declaration with that of...
7253 (make_pass_slp_vectorize): ...new function.
7254 (pass_complete_unroll): Replace declaration with that of...
7255 (make_pass_complete_unroll): ...new function.
7256 (pass_complete_unrolli): Replace declaration with that of...
7257 (make_pass_complete_unrolli): ...new function.
7258 (pass_parallelize_loops): Replace declaration with that of...
7259 (make_pass_parallelize_loops): ...new function.
7260 (pass_loop_prefetch): Replace declaration with that of...
7261 (make_pass_loop_prefetch): ...new function.
7262 (pass_iv_optimize): Replace declaration with that of...
7263 (make_pass_iv_optimize): ...new function.
7264 (pass_tree_loop_done): Replace declaration with that of...
7265 (make_pass_tree_loop_done): ...new function.
7266 (pass_ch): Replace declaration with that of...
7267 (make_pass_ch): ...new function.
7268 (pass_ccp): Replace declaration with that of...
7269 (make_pass_ccp): ...new function.
7270 (pass_phi_only_cprop): Replace declaration with that of...
7271 (make_pass_phi_only_cprop): ...new function.
7272 (pass_build_ssa): Replace declaration with that of...
7273 (make_pass_build_ssa): ...new function.
7274 (pass_build_alias): Replace declaration with that of...
7275 (make_pass_build_alias): ...new function.
7276 (pass_build_ealias): Replace declaration with that of...
7277 (make_pass_build_ealias): ...new function.
7278 (pass_dominator): Replace declaration with that of...
7279 (make_pass_dominator): ...new function.
7280 (pass_dce): Replace declaration with that of...
7281 (make_pass_dce): ...new function.
7282 (pass_dce_loop): Replace declaration with that of...
7283 (make_pass_dce_loop): ...new function.
7284 (pass_cd_dce): Replace declaration with that of...
7285 (make_pass_cd_dce): ...new function.
7286 (pass_call_cdce): Replace declaration with that of...
7287 (make_pass_call_cdce): ...new function.
7288 (pass_merge_phi): Replace declaration with that of...
7289 (make_pass_merge_phi): ...new function.
7290 (pass_split_crit_edges): Replace declaration with that of...
7291 (make_pass_split_crit_edges): ...new function.
7292 (pass_pre): Replace declaration with that of...
7293 (make_pass_pre): ...new function.
7294 (pass_profile): Replace declaration with that of...
7295 (make_pass_profile): ...new function.
7296 (pass_strip_predict_hints): Replace declaration with that of...
7297 (make_pass_strip_predict_hints): ...new function.
7298 (pass_lower_complex_O0): Replace declaration with that of...
7299 (make_pass_lower_complex_O0): ...new function.
7300 (pass_lower_complex): Replace declaration with that of...
7301 (make_pass_lower_complex): ...new function.
7302 (pass_lower_vector): Replace declaration with that of...
7303 (make_pass_lower_vector): ...new function.
7304 (pass_lower_vector_ssa): Replace declaration with that of...
7305 (make_pass_lower_vector_ssa): ...new function.
7306 (pass_lower_omp): Replace declaration with that of...
7307 (make_pass_lower_omp): ...new function.
7308 (pass_diagnose_omp_blocks): Replace declaration with that of...
7309 (make_pass_diagnose_omp_blocks): ...new function.
7310 (pass_expand_omp): Replace declaration with that of...
7311 (make_pass_expand_omp): ...new function.
7312 (pass_expand_omp_ssa): Replace declaration with that of...
7313 (make_pass_expand_omp_ssa): ...new function.
7314 (pass_object_sizes): Replace declaration with that of...
7315 (make_pass_object_sizes): ...new function.
7316 (pass_strlen): Replace declaration with that of...
7317 (make_pass_strlen): ...new function.
7318 (pass_fold_builtins): Replace declaration with that of...
7319 (make_pass_fold_builtins): ...new function.
7320 (pass_stdarg): Replace declaration with that of...
7321 (make_pass_stdarg): ...new function.
7322 (pass_early_warn_uninitialized): Replace declaration with that of...
7323 (make_pass_early_warn_uninitialized): ...new function.
7324 (pass_late_warn_uninitialized): Replace declaration with that of...
7325 (make_pass_late_warn_uninitialized): ...new function.
7326 (pass_cse_reciprocals): Replace declaration with that of...
7327 (make_pass_cse_reciprocals): ...new function.
7328 (pass_cse_sincos): Replace declaration with that of...
7329 (make_pass_cse_sincos): ...new function.
7330 (pass_optimize_bswap): Replace declaration with that of...
7331 (make_pass_optimize_bswap): ...new function.
7332 (pass_optimize_widening_mul): Replace declaration with that of...
7333 (make_pass_optimize_widening_mul): ...new function.
7334 (pass_warn_function_return): Replace declaration with that of...
7335 (make_pass_warn_function_return): ...new function.
7336 (pass_warn_function_noreturn): Replace declaration with that of...
7337 (make_pass_warn_function_noreturn): ...new function.
7338 (pass_cselim): Replace declaration with that of...
7339 (make_pass_cselim): ...new function.
7340 (pass_phiopt): Replace declaration with that of...
7341 (make_pass_phiopt): ...new function.
7342 (pass_forwprop): Replace declaration with that of...
7343 (make_pass_forwprop): ...new function.
7344 (pass_phiprop): Replace declaration with that of...
7345 (make_pass_phiprop): ...new function.
7346 (pass_tree_ifcombine): Replace declaration with that of...
7347 (make_pass_tree_ifcombine): ...new function.
7348 (pass_dse): Replace declaration with that of...
7349 (make_pass_dse): ...new function.
7350 (pass_nrv): Replace declaration with that of...
7351 (make_pass_nrv): ...new function.
7352 (pass_rename_ssa_copies): Replace declaration with that of...
7353 (make_pass_rename_ssa_copies): ...new function.
7354 (pass_sink_code): Replace declaration with that of...
7355 (make_pass_sink_code): ...new function.
7356 (pass_fre): Replace declaration with that of...
7357 (make_pass_fre): ...new function.
7358 (pass_check_data_deps): Replace declaration with that of...
7359 (make_pass_check_data_deps): ...new function.
7360 (pass_copy_prop): Replace declaration with that of...
7361 (make_pass_copy_prop): ...new function.
7362 (pass_vrp): Replace declaration with that of...
7363 (make_pass_vrp): ...new function.
7364 (pass_uncprop): Replace declaration with that of...
7365 (make_pass_uncprop): ...new function.
7366 (pass_return_slot): Replace declaration with that of...
7367 (make_pass_return_slot): ...new function.
7368 (pass_reassoc): Replace declaration with that of...
7369 (make_pass_reassoc): ...new function.
7370 (pass_rebuild_cgraph_edges): Replace declaration with that of...
7371 (make_pass_rebuild_cgraph_edges): ...new function.
7372 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
7373 (make_pass_remove_cgraph_callee_edges): ...new function.
7374 (pass_build_cgraph_edges): Replace declaration with that of...
7375 (make_pass_build_cgraph_edges): ...new function.
7376 (pass_local_pure_const): Replace declaration with that of...
7377 (make_pass_local_pure_const): ...new function.
7378 (pass_tracer): Replace declaration with that of...
7379 (make_pass_tracer): ...new function.
7380 (pass_warn_unused_result): Replace declaration with that of...
7381 (make_pass_warn_unused_result): ...new function.
7382 (pass_diagnose_tm_blocks): Replace declaration with that of...
7383 (make_pass_diagnose_tm_blocks): ...new function.
7384 (pass_lower_tm): Replace declaration with that of...
7385 (make_pass_lower_tm): ...new function.
7386 (pass_tm_init): Replace declaration with that of...
7387 (make_pass_tm_init): ...new function.
7388 (pass_tm_mark): Replace declaration with that of...
7389 (make_pass_tm_mark): ...new function.
7390 (pass_tm_memopt): Replace declaration with that of...
7391 (make_pass_tm_memopt): ...new function.
7392 (pass_tm_edges): Replace declaration with that of...
7393 (make_pass_tm_edges): ...new function.
7394 (pass_split_functions): Replace declaration with that of...
7395 (make_pass_split_functions): ...new function.
7396 (pass_feedback_split_functions): Replace declaration with that of...
7397 (make_pass_feedback_split_functions): ...new function.
7398 (pass_strength_reduction): Replace declaration with that of...
7399 (make_pass_strength_reduction): ...new function.
7400 (pass_ipa_lower_emutls): Replace declaration with that of...
7401 (make_pass_ipa_lower_emutls): ...new function.
7402 (pass_ipa_function_and_variable_visibility): Replace declaration with
7404 (make_pass_ipa_function_and_variable_visibility): ...new function.
7405 (pass_ipa_tree_profile): Replace declaration with that of...
7406 (make_pass_ipa_tree_profile): ...new function.
7407 (pass_early_local_passes): Replace declaration with that of...
7408 (make_pass_early_local_passes): ...new function.
7409 (pass_ipa_whole_program_visibility): Replace declaration with that
7411 (make_pass_ipa_whole_program_visibility): ...new function.
7412 (pass_ipa_lto_gimple_out): Replace declaration with that of...
7413 (make_pass_ipa_lto_gimple_out): ...new function.
7414 (pass_ipa_increase_alignment): Replace declaration with that of...
7415 (make_pass_ipa_increase_alignment): ...new function.
7416 (pass_ipa_inline): Replace declaration with that of...
7417 (make_pass_ipa_inline): ...new function.
7418 (pass_ipa_free_lang_data): Replace declaration with that of...
7419 (make_pass_ipa_free_lang_data): ...new function.
7420 (pass_ipa_free_inline_summary): Replace declaration with that of...
7421 (make_pass_ipa_free_inline_summary): ...new function.
7422 (pass_ipa_cp): Replace declaration with that of...
7423 (make_pass_ipa_cp): ...new function.
7424 (pass_ipa_reference): Replace declaration with that of...
7425 (make_pass_ipa_reference): ...new function.
7426 (pass_ipa_pure_const): Replace declaration with that of...
7427 (make_pass_ipa_pure_const): ...new function.
7428 (pass_ipa_pta): Replace declaration with that of...
7429 (make_pass_ipa_pta): ...new function.
7430 (pass_ipa_lto_finish_out): Replace declaration with that of...
7431 (make_pass_ipa_lto_finish_out): ...new function.
7432 (pass_ipa_tm): Replace declaration with that of...
7433 (make_pass_ipa_tm): ...new function.
7434 (pass_ipa_profile): Replace declaration with that of...
7435 (make_pass_ipa_profile): ...new function.
7436 (pass_ipa_cdtor_merge): Replace declaration with that of...
7437 (make_pass_ipa_cdtor_merge): ...new function.
7438 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
7440 (make_pass_cleanup_cfg_post_optimizing): ...new function.
7441 (pass_init_datastructures): Replace declaration with that of...
7442 (make_pass_init_datastructures): ...new function.
7443 (pass_fixup_cfg): Replace declaration with that of...
7444 (make_pass_fixup_cfg): ...new function.
7445 (pass_expand): Replace declaration with that of...
7446 (make_pass_expand): ...new function.
7447 (pass_instantiate_virtual_regs): Replace declaration with that of...
7448 (make_pass_instantiate_virtual_regs): ...new function.
7449 (pass_rtl_fwprop): Replace declaration with that of...
7450 (make_pass_rtl_fwprop): ...new function.
7451 (pass_rtl_fwprop_addr): Replace declaration with that of...
7452 (make_pass_rtl_fwprop_addr): ...new function.
7453 (pass_jump): Replace declaration with that of...
7454 (make_pass_jump): ...new function.
7455 (pass_jump2): Replace declaration with that of...
7456 (make_pass_jump2): ...new function.
7457 (pass_lower_subreg): Replace declaration with that of...
7458 (make_pass_lower_subreg): ...new function.
7459 (pass_cse): Replace declaration with that of...
7460 (make_pass_cse): ...new function.
7461 (pass_fast_rtl_dce): Replace declaration with that of...
7462 (make_pass_fast_rtl_dce): ...new function.
7463 (pass_ud_rtl_dce): Replace declaration with that of...
7464 (make_pass_ud_rtl_dce): ...new function.
7465 (pass_rtl_dce): Replace declaration with that of...
7466 (make_pass_rtl_dce): ...new function.
7467 (pass_rtl_dse1): Replace declaration with that of...
7468 (make_pass_rtl_dse1): ...new function.
7469 (pass_rtl_dse2): Replace declaration with that of...
7470 (make_pass_rtl_dse2): ...new function.
7471 (pass_rtl_dse3): Replace declaration with that of...
7472 (make_pass_rtl_dse3): ...new function.
7473 (pass_rtl_cprop): Replace declaration with that of...
7474 (make_pass_rtl_cprop): ...new function.
7475 (pass_rtl_pre): Replace declaration with that of...
7476 (make_pass_rtl_pre): ...new function.
7477 (pass_rtl_hoist): Replace declaration with that of...
7478 (make_pass_rtl_hoist): ...new function.
7479 (pass_rtl_store_motion): Replace declaration with that of...
7480 (make_pass_rtl_store_motion): ...new function.
7481 (pass_cse_after_global_opts): Replace declaration with that of...
7482 (make_pass_cse_after_global_opts): ...new function.
7483 (pass_rtl_ifcvt): Replace declaration with that of...
7484 (make_pass_rtl_ifcvt): ...new function.
7485 (pass_into_cfg_layout_mode): Replace declaration with that of...
7486 (make_pass_into_cfg_layout_mode): ...new function.
7487 (pass_outof_cfg_layout_mode): Replace declaration with that of...
7488 (make_pass_outof_cfg_layout_mode): ...new function.
7489 (pass_loop2): Replace declaration with that of...
7490 (make_pass_loop2): ...new function.
7491 (pass_rtl_loop_init): Replace declaration with that of...
7492 (make_pass_rtl_loop_init): ...new function.
7493 (pass_rtl_move_loop_invariants): Replace declaration with that of...
7494 (make_pass_rtl_move_loop_invariants): ...new function.
7495 (pass_rtl_unswitch): Replace declaration with that of...
7496 (make_pass_rtl_unswitch): ...new function.
7497 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
7498 (make_pass_rtl_unroll_and_peel_loops): ...new function.
7499 (pass_rtl_doloop): Replace declaration with that of...
7500 (make_pass_rtl_doloop): ...new function.
7501 (pass_rtl_loop_done): Replace declaration with that of...
7502 (make_pass_rtl_loop_done): ...new function.
7503 (pass_web): Replace declaration with that of...
7504 (make_pass_web): ...new function.
7505 (pass_cse2): Replace declaration with that of...
7506 (make_pass_cse2): ...new function.
7507 (pass_df_initialize_opt): Replace declaration with that of...
7508 (make_pass_df_initialize_opt): ...new function.
7509 (pass_df_initialize_no_opt): Replace declaration with that of...
7510 (make_pass_df_initialize_no_opt): ...new function.
7511 (pass_reginfo_init): Replace declaration with that of...
7512 (make_pass_reginfo_init): ...new function.
7513 (pass_inc_dec): Replace declaration with that of...
7514 (make_pass_inc_dec): ...new function.
7515 (pass_stack_ptr_mod): Replace declaration with that of...
7516 (make_pass_stack_ptr_mod): ...new function.
7517 (pass_initialize_regs): Replace declaration with that of...
7518 (make_pass_initialize_regs): ...new function.
7519 (pass_combine): Replace declaration with that of...
7520 (make_pass_combine): ...new function.
7521 (pass_if_after_combine): Replace declaration with that of...
7522 (make_pass_if_after_combine): ...new function.
7523 (pass_ree): Replace declaration with that of...
7524 (make_pass_ree): ...new function.
7525 (pass_partition_blocks): Replace declaration with that of...
7526 (make_pass_partition_blocks): ...new function.
7527 (pass_match_asm_constraints): Replace declaration with that of...
7528 (make_pass_match_asm_constraints): ...new function.
7529 (pass_regmove): Replace declaration with that of...
7530 (make_pass_regmove): ...new function.
7531 (pass_split_all_insns): Replace declaration with that of...
7532 (make_pass_split_all_insns): ...new function.
7533 (pass_fast_rtl_byte_dce): Replace declaration with that of...
7534 (make_pass_fast_rtl_byte_dce): ...new function.
7535 (pass_lower_subreg2): Replace declaration with that of...
7536 (make_pass_lower_subreg2): ...new function.
7537 (pass_mode_switching): Replace declaration with that of...
7538 (make_pass_mode_switching): ...new function.
7539 (pass_sms): Replace declaration with that of...
7540 (make_pass_sms): ...new function.
7541 (pass_sched): Replace declaration with that of...
7542 (make_pass_sched): ...new function.
7543 (pass_ira): Replace declaration with that of...
7544 (make_pass_ira): ...new function.
7545 (pass_reload): Replace declaration with that of...
7546 (make_pass_reload): ...new function.
7547 (pass_clean_state): Replace declaration with that of...
7548 (make_pass_clean_state): ...new function.
7549 (pass_branch_prob): Replace declaration with that of...
7550 (make_pass_branch_prob): ...new function.
7551 (pass_value_profile_transformations): Replace declaration with that
7553 (make_pass_value_profile_transformations): ...new function.
7554 (pass_postreload_cse): Replace declaration with that of...
7555 (make_pass_postreload_cse): ...new function.
7556 (pass_gcse2): Replace declaration with that of...
7557 (make_pass_gcse2): ...new function.
7558 (pass_split_after_reload): Replace declaration with that of...
7559 (make_pass_split_after_reload): ...new function.
7560 (pass_branch_target_load_optimize1): Replace declaration with that
7562 (make_pass_branch_target_load_optimize1): ...new function.
7563 (pass_thread_prologue_and_epilogue): Replace declaration with that
7565 (make_pass_thread_prologue_and_epilogue): ...new function.
7566 (pass_stack_adjustments): Replace declaration with that of...
7567 (make_pass_stack_adjustments): ...new function.
7568 (pass_peephole2): Replace declaration with that of...
7569 (make_pass_peephole2): ...new function.
7570 (pass_if_after_reload): Replace declaration with that of...
7571 (make_pass_if_after_reload): ...new function.
7572 (pass_regrename): Replace declaration with that of...
7573 (make_pass_regrename): ...new function.
7574 (pass_cprop_hardreg): Replace declaration with that of...
7575 (make_pass_cprop_hardreg): ...new function.
7576 (pass_reorder_blocks): Replace declaration with that of...
7577 (make_pass_reorder_blocks): ...new function.
7578 (pass_branch_target_load_optimize2): Replace declaration with that
7580 (make_pass_branch_target_load_optimize2): ...new function.
7581 (pass_leaf_regs): Replace declaration with that of...
7582 (make_pass_leaf_regs): ...new function.
7583 (pass_split_before_sched2): Replace declaration with that of...
7584 (make_pass_split_before_sched2): ...new function.
7585 (pass_compare_elim_after_reload): Replace declaration with that of...
7586 (make_pass_compare_elim_after_reload): ...new function.
7587 (pass_sched2): Replace declaration with that of...
7588 (make_pass_sched2): ...new function.
7589 (pass_stack_regs): Replace declaration with that of...
7590 (make_pass_stack_regs): ...new function.
7591 (pass_stack_regs_run): Replace declaration with that of...
7592 (make_pass_stack_regs_run): ...new function.
7593 (pass_df_finish): Replace declaration with that of...
7594 (make_pass_df_finish): ...new function.
7595 (pass_compute_alignments): Replace declaration with that of...
7596 (make_pass_compute_alignments): ...new function.
7597 (pass_duplicate_computed_gotos): Replace declaration with that of...
7598 (make_pass_duplicate_computed_gotos): ...new function.
7599 (pass_variable_tracking): Replace declaration with that of...
7600 (make_pass_variable_tracking): ...new function.
7601 (pass_free_cfg): Replace declaration with that of...
7602 (make_pass_free_cfg): ...new function.
7603 (pass_machine_reorg): Replace declaration with that of...
7604 (make_pass_machine_reorg): ...new function.
7605 (pass_cleanup_barriers): Replace declaration with that of...
7606 (make_pass_cleanup_barriers): ...new function.
7607 (pass_delay_slots): Replace declaration with that of...
7608 (make_pass_delay_slots): ...new function.
7609 (pass_split_for_shorten_branches): Replace declaration with that of...
7610 (make_pass_split_for_shorten_branches): ...new function.
7611 (pass_split_before_regstack): Replace declaration with that of...
7612 (make_pass_split_before_regstack): ...new function.
7613 (pass_convert_to_eh_region_ranges): Replace declaration with that
7615 (make_pass_convert_to_eh_region_ranges): ...new function.
7616 (pass_shorten_branches): Replace declaration with that of...
7617 (make_pass_shorten_branches): ...new function.
7618 (pass_set_nothrow_function_flags): Replace declaration with that of...
7619 (make_pass_set_nothrow_function_flags): ...new function.
7620 (pass_dwarf2_frame): Replace declaration with that of...
7621 (make_pass_dwarf2_frame): ...new function.
7622 (pass_final): Replace declaration with that of...
7623 (make_pass_final): ...new function.
7624 (pass_rtl_seqabstr): Replace declaration with that of...
7625 (make_pass_rtl_seqabstr): ...new function.
7626 (pass_release_ssa_names): Replace declaration with that of...
7627 (make_pass_release_ssa_names): ...new function.
7628 (pass_early_inline): Replace declaration with that of...
7629 (make_pass_early_inline): ...new function.
7630 (pass_inline_parameters): Replace declaration with that of...
7631 (make_pass_inline_parameters): ...new function.
7632 (pass_update_address_taken): Replace declaration with that of...
7633 (make_pass_update_address_taken): ...new function.
7634 (pass_convert_switch): Replace declaration with that of...
7635 (make_pass_convert_switch): ...new function.
7636 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
7637 to a subclass of simple_ipa_opt_pass along with...
7638 (pass_data_ipa_tree_profile): ...new pass_data instance and...
7639 (make_pass_ipa_tree_profile): ...new function.
7640 * tree-sra.c (pass_sra_early): Convert from a global struct to a
7641 subclass of gimple_opt_pass along with...
7642 (pass_data_sra_early): ...new pass_data instance and...
7643 (make_pass_sra_early): ...new function.
7644 (pass_sra): Convert from a global struct to a subclass of
7645 gimple_opt_pass along with...
7646 (pass_data_sra): ...new pass_data instance and...
7647 (make_pass_sra): ...new function.
7648 (pass_early_ipa_sra): Convert from a global struct to a subclass of
7649 gimple_opt_pass along with...
7650 (pass_data_early_ipa_sra): ...new pass_data instance and...
7651 (make_pass_early_ipa_sra): ...new function.
7652 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
7653 subclass of gimple_opt_pass along with...
7654 (pass_data_ccp): ...new pass_data instance and...
7655 (make_pass_ccp): ...new function.
7656 (pass_fold_builtins): Convert from a global struct to a subclass of
7657 gimple_opt_pass along with...
7658 (pass_data_fold_builtins): ...new pass_data instance and...
7659 (make_pass_fold_builtins): ...new function.
7660 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
7661 subclass of gimple_opt_pass along with...
7662 (pass_data_copy_prop): ...new pass_data instance and...
7663 (make_pass_copy_prop): ...new function.
7664 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
7665 global struct to a subclass of gimple_opt_pass along with...
7666 (pass_data_rename_ssa_copies): ...new pass_data instance and...
7667 (make_pass_rename_ssa_copies): ...new function.
7668 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
7669 subclass of gimple_opt_pass along with...
7670 (pass_data_dce): ...new pass_data instance and...
7671 (make_pass_dce): ...new function.
7672 (pass_dce_loop): Convert from a global struct to a subclass of
7673 gimple_opt_pass along with...
7674 (pass_data_dce_loop): ...new pass_data instance and...
7675 (make_pass_dce_loop): ...new function.
7676 (pass_cd_dce): Convert from a global struct to a subclass of
7677 gimple_opt_pass along with...
7678 (pass_data_cd_dce): ...new pass_data instance and...
7679 (make_pass_cd_dce): ...new function.
7680 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
7681 subclass of gimple_opt_pass along with...
7682 (pass_data_dominator): ...new pass_data instance and...
7683 (make_pass_dominator): ...new function.
7684 (pass_phi_only_cprop): Convert from a global struct to a subclass of
7685 gimple_opt_pass along with...
7686 (pass_data_phi_only_cprop): ...new pass_data instance and...
7687 (make_pass_phi_only_cprop): ...new function.
7688 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
7689 subclass of gimple_opt_pass along with...
7690 (pass_data_dse): ...new pass_data instance and...
7691 (make_pass_dse): ...new function.
7692 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
7693 a subclass of gimple_opt_pass along with...
7694 (pass_data_forwprop): ...new pass_data instance and...
7695 (make_pass_forwprop): ...new function.
7696 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
7697 struct to a subclass of gimple_opt_pass along with...
7698 (pass_data_tree_ifcombine): ...new pass_data instance and...
7699 (make_pass_tree_ifcombine): ...new function.
7700 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
7701 subclass of gimple_opt_pass along with...
7702 (pass_data_ch): ...new pass_data instance and...
7703 (make_pass_ch): ...new function.
7704 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
7705 subclass of gimple_opt_pass along with...
7706 (pass_data_tree_loop): ...new pass_data instance and...
7707 (make_pass_tree_loop): ...new function.
7708 (pass_tree_loop_init): Convert from a global struct to a subclass of
7709 gimple_opt_pass along with...
7710 (pass_data_tree_loop_init): ...new pass_data instance and...
7711 (make_pass_tree_loop_init): ...new function.
7712 (pass_lim): Convert from a global struct to a subclass of
7713 gimple_opt_pass along with...
7714 (pass_data_lim): ...new pass_data instance and...
7715 (make_pass_lim): ...new function.
7716 (pass_tree_unswitch): Convert from a global struct to a subclass of
7717 gimple_opt_pass along with...
7718 (pass_data_tree_unswitch): ...new pass_data instance and...
7719 (make_pass_tree_unswitch): ...new function.
7720 (pass_predcom): Convert from a global struct to a subclass of
7721 gimple_opt_pass along with...
7722 (pass_data_predcom): ...new pass_data instance and...
7723 (make_pass_predcom): ...new function.
7724 (pass_vectorize): Convert from a global struct to a subclass of
7725 gimple_opt_pass along with...
7726 (pass_data_vectorize): ...new pass_data instance and...
7727 (make_pass_vectorize): ...new function.
7728 (pass_graphite): Convert from a global struct to a subclass of
7729 gimple_opt_pass along with...
7730 (pass_data_graphite): ...new pass_data instance and...
7731 (make_pass_graphite): ...new function.
7732 (pass_graphite_transforms): Convert from a global struct to a subclass
7733 of gimple_opt_pass along with...
7734 (pass_data_graphite_transforms): ...new pass_data instance and...
7735 (make_pass_graphite_transforms): ...new function.
7736 (pass_check_data_deps): Convert from a global struct to a subclass of
7737 gimple_opt_pass along with...
7738 (pass_data_check_data_deps): ...new pass_data instance and...
7739 (make_pass_check_data_deps): ...new function.
7740 (pass_iv_canon): Convert from a global struct to a subclass of
7741 gimple_opt_pass along with...
7742 (pass_data_iv_canon): ...new pass_data instance and...
7743 (make_pass_iv_canon): ...new function.
7744 (pass_scev_cprop): Convert from a global struct to a subclass of
7745 gimple_opt_pass along with...
7746 (pass_data_scev_cprop): ...new pass_data instance and...
7747 (make_pass_scev_cprop): ...new function.
7748 (pass_record_bounds): Convert from a global struct to a subclass of
7749 gimple_opt_pass along with...
7750 (pass_data_record_bounds): ...new pass_data instance and...
7751 (make_pass_record_bounds): ...new function.
7752 (pass_complete_unroll): Convert from a global struct to a subclass of
7753 gimple_opt_pass along with...
7754 (pass_data_complete_unroll): ...new pass_data instance and...
7755 (make_pass_complete_unroll): ...new function.
7756 (pass_complete_unrolli): Convert from a global struct to a subclass of
7757 gimple_opt_pass along with...
7758 (pass_data_complete_unrolli): ...new pass_data instance and...
7759 (make_pass_complete_unrolli): ...new function.
7760 (pass_parallelize_loops): Convert from a global struct to a subclass
7761 of gimple_opt_pass along with...
7762 (pass_data_parallelize_loops): ...new pass_data instance and...
7763 (make_pass_parallelize_loops): ...new function.
7764 (pass_loop_prefetch): Convert from a global struct to a subclass of
7765 gimple_opt_pass along with...
7766 (pass_data_loop_prefetch): ...new pass_data instance and...
7767 (make_pass_loop_prefetch): ...new function.
7768 (pass_iv_optimize): Convert from a global struct to a subclass of
7769 gimple_opt_pass along with...
7770 (pass_data_iv_optimize): ...new pass_data instance and...
7771 (make_pass_iv_optimize): ...new function.
7772 (pass_tree_loop_done): Convert from a global struct to a subclass of
7773 gimple_opt_pass along with...
7774 (pass_data_tree_loop_done): ...new pass_data instance and...
7775 (make_pass_tree_loop_done): ...new function.
7776 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
7777 struct to a subclass of gimple_opt_pass along with...
7778 (pass_data_cse_reciprocals): ...new pass_data instance and...
7779 (make_pass_cse_reciprocals): ...new function.
7780 (pass_cse_sincos): Convert from a global struct to a subclass of
7781 gimple_opt_pass along with...
7782 (pass_data_cse_sincos): ...new pass_data instance and...
7783 (make_pass_cse_sincos): ...new function.
7784 (pass_optimize_bswap): Convert from a global struct to a subclass of
7785 gimple_opt_pass along with...
7786 (pass_data_optimize_bswap): ...new pass_data instance and...
7787 (make_pass_optimize_bswap): ...new function.
7788 (pass_optimize_widening_mul): Convert from a global struct to a
7789 subclass of gimple_opt_pass along with...
7790 (pass_data_optimize_widening_mul): ...new pass_data instance and...
7791 (make_pass_optimize_widening_mul): ...new function.
7792 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
7793 subclass of gimple_opt_pass along with...
7794 (pass_data_phiopt): ...new pass_data instance and...
7795 (make_pass_phiopt): ...new function.
7796 (pass_cselim): Convert from a global struct to a subclass of
7797 gimple_opt_pass along with...
7798 (pass_data_cselim): ...new pass_data instance and...
7799 (make_pass_cselim): ...new function.
7800 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
7801 subclass of gimple_opt_pass along with...
7802 (pass_data_phiprop): ...new pass_data instance and...
7803 (make_pass_phiprop): ...new function.
7804 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
7805 subclass of gimple_opt_pass along with...
7806 (pass_data_pre): ...new pass_data instance and...
7807 (make_pass_pre): ...new function.
7808 (pass_fre): Convert from a global struct to a subclass of
7809 gimple_opt_pass along with...
7810 (pass_data_fre): ...new pass_data instance and...
7811 (make_pass_fre): ...new function.
7812 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
7813 subclass of gimple_opt_pass along with...
7814 (pass_data_reassoc): ...new pass_data instance and...
7815 (make_pass_reassoc): ...new function.
7816 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
7817 subclass of gimple_opt_pass along with...
7818 (pass_data_sink_code): ...new pass_data instance and...
7819 (make_pass_sink_code): ...new function.
7820 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
7821 subclass of gimple_opt_pass along with...
7822 (pass_data_strlen): ...new pass_data instance and...
7823 (make_pass_strlen): ...new function.
7824 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
7825 struct to a subclass of gimple_opt_pass along with...
7826 (pass_data_build_alias): ...new pass_data instance and...
7827 (make_pass_build_alias): ...new function.
7828 (pass_build_ealias): Convert from a global struct to a subclass of
7829 gimple_opt_pass along with...
7830 (pass_data_build_ealias): ...new pass_data instance and...
7831 (make_pass_build_ealias): ...new function.
7832 (pass_ipa_pta): Convert from a global struct to a subclass of
7833 simple_ipa_opt_pass along with...
7834 (pass_data_ipa_pta): ...new pass_data instance and...
7835 (make_pass_ipa_pta): ...new function.
7836 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
7837 subclass of gimple_opt_pass along with...
7838 (pass_data_uncprop): ...new pass_data instance and...
7839 (make_pass_uncprop): ...new function.
7840 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
7841 global struct to a subclass of gimple_opt_pass along with...
7842 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
7843 (make_pass_late_warn_uninitialized): ...new function.
7844 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
7845 to a subclass of gimple_opt_pass along with...
7846 (pass_data_init_datastructures): ...new pass_data instance and...
7847 (make_pass_init_datastructures): ...new function.
7848 (pass_early_warn_uninitialized): Convert from a global struct to a
7849 subclass of gimple_opt_pass along with...
7850 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
7851 (make_pass_early_warn_uninitialized): ...new function.
7852 (pass_update_address_taken): Convert from a global struct to a
7853 subclass of gimple_opt_pass along with...
7854 (pass_data_update_address_taken): ...new pass_data instance and...
7855 (make_pass_update_address_taken): ...new function.
7856 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
7857 struct to a subclass of gimple_opt_pass along with...
7858 (pass_data_release_ssa_names): ...new pass_data instance and...
7859 (make_pass_release_ssa_names): ...new function.
7860 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
7861 subclass of gimple_opt_pass along with...
7862 (pass_data_stdarg): ...new pass_data instance and...
7863 (make_pass_stdarg): ...new function.
7864 * tree-switch-conversion.c (pass_convert_switch): Convert from a
7865 global struct to a subclass of gimple_opt_pass along with...
7866 (pass_data_convert_switch): ...new pass_data instance and...
7867 (make_pass_convert_switch): ...new function.
7868 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
7869 to a subclass of gimple_opt_pass along with...
7870 (pass_data_tail_recursion): ...new pass_data instance and...
7871 (make_pass_tail_recursion): ...new function.
7872 (pass_tail_calls): Convert from a global struct to a subclass of
7873 gimple_opt_pass along with...
7874 (pass_data_tail_calls): ...new pass_data instance and...
7875 (make_pass_tail_calls): ...new function.
7876 * tree-vect-generic.c (pass_lower_vector): Convert from a global
7877 struct to a subclass of gimple_opt_pass along with...
7878 (pass_data_lower_vector): ...new pass_data instance and...
7879 (make_pass_lower_vector): ...new function.
7880 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
7881 gimple_opt_pass along with...
7882 (pass_data_lower_vector_ssa): ...new pass_data instance and...
7883 (make_pass_lower_vector_ssa): ...new function.
7884 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
7885 to a subclass of gimple_opt_pass along with...
7886 (pass_data_slp_vectorize): ...new pass_data instance and...
7887 (make_pass_slp_vectorize): ...new function.
7888 (pass_ipa_increase_alignment): Convert from a global struct to a
7889 subclass of simple_ipa_opt_pass along with...
7890 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
7891 (make_pass_ipa_increase_alignment): ...new function.
7892 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
7893 gimple_opt_pass along with...
7894 (pass_data_vrp): ...new pass_data instance and...
7895 (make_pass_vrp): ...new function.
7896 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
7897 subclass of simple_ipa_opt_pass along with...
7898 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
7899 (make_pass_ipa_free_lang_data): ...new function.
7900 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
7901 gimple_opt_pass along with...
7902 (pass_data_tsan): ...new pass_data instance and...
7903 (make_pass_tsan): ...new function.
7904 (pass_tsan_O0): Convert from a global struct to a subclass of
7905 gimple_opt_pass along with...
7906 (pass_data_tsan_O0): ...new pass_data instance and...
7907 (make_pass_tsan_O0): ...new function.
7908 * var-tracking.c (pass_variable_tracking): Convert from a global
7909 struct to a subclass of rtl_opt_pass along with...
7910 (pass_data_variable_tracking): ...new pass_data instance and...
7911 (make_pass_variable_tracking): ...new function.
7912 * web.c (pass_web): Convert from a global struct to a subclass of
7913 rtl_opt_pass along with...
7914 (pass_data_web): ...new pass_data instance and...
7915 (make_pass_web): ...new function.
7916 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
7917 declaration with that of...
7918 (make_pass_mode_switch_use): ...new function.
7919 (pass_resolve_sw_modes): Replace declaration with that of...
7920 (make_pass_resolve_sw_modes): ...new function.
7921 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
7922 from a global struct to a subclass of rtl_opt_pass along with...
7923 (pass_data_mode_switch_use): ...new pass_data instance and...
7924 (make_pass_mode_switch_use): ...new function.
7925 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
7926 from a global struct to a subclass of rtl_opt_pass along with...
7927 (pass_data_resolve_sw_modes): ...new pass_data instance and...
7928 (make_pass_resolve_sw_modes): ...new function.
7929 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
7930 struct to a subclass of rtl_opt_pass along with...
7931 (pass_data_insert_vzeroupper): ...new pass_data instance and...
7932 (make_pass_insert_vzeroupper): ...new function.
7933 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
7934 global struct to a subclass of rtl_opt_pass along with...
7935 (pass_data_work_around_errata): ...new pass_data instance and...
7936 (make_pass_work_around_errata): ...new function.
7937 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
7938 struct to a subclass of rtl_opt_pass along with...
7939 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
7940 (make_pass_mips_machine_reorg2): ...new function.
7942 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7944 * passes.c (pass_manager::operator new): New.
7946 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7948 Handwritten part of conversion of passes to C++ classes.
7950 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
7951 (toplev.o): Add dep on PASS_MANAGER_H.
7952 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
7953 of early local pases to reflect this moving from a global to a
7954 member of gcc::pass_manager.
7955 (cgraph_add_new_function): Likewise.
7956 * lto-cgraph.c (lto_output_node): Update for conversion of
7957 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
7958 * passes.c (opt_pass::clone): New.
7959 (opt_pass::gate): New.
7960 (opt_pass::execute): New.
7961 (opt_pass::opt_pass): New.
7962 (pass_manager::execute_early_local_passes): New.
7963 (pass_manager::execute_pass_mode_switching): new.
7964 (finish_optimization_passes): Convert to...
7965 (pass_manager::finish_optimization_passes): ...this.
7966 (finish_optimization_passes): Update for conversion of passes to
7968 (register_dump_files_1): Use has_gate since we cannot portably
7969 check a vtable entry against NULL.
7970 (dump_one_pass): Likewise.
7971 (ipa_write_summaries_2): Likewise.
7972 (ipa_write_optimization_summaries_1): Likewise.
7973 (ipa_read_summaries_1): Likewise.
7974 (ipa_read_optimization_summaries_1): Likewise.
7975 (execute_ipa_stmt_fixups): Likewise.
7976 (pass_manager::pass_manager): Rewrite pass-creation, invoking
7977 pass-creation functions rather than wiring up globals, and
7978 storing the results in fields of pass_manager generated using
7980 (pass_manager::dump_profile_report): Update for conversion of
7981 passes to C++ classes.
7982 (pass_manager::execute_ipa_summary_passes): Likewise.
7983 (execute_one_ipa_transform_pass): Likewise.
7984 (execute_one_pass): Use has_gate and has_execute since we cannot
7985 portably check a vtable entry against NULL.
7986 * pass_manager.h (pass_manager::finish_optimization_passes): New.
7987 (pass_manager): Use pass-instances.def to add fields for the
7988 various pass instances.
7989 * toplev.c (finalize): Update for move of
7990 finish_optimization_passes to a method of gcc::pass_manager.
7991 * toplev.h (finish_optimization_passes): Move to method of class
7993 * tree-pass.h (struct pass_data): New.
7994 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
7995 (opt_pass::gate): Convert to virtual function.
7996 (opt_pass::~opt_pass): New.
7997 (opt_pass::clone): New.
7998 (opt_pass::execute): Convert to virtual function.
7999 (opt_pass::opt_pass): New.
8000 (opt_pass::ctxt_): new.
8001 (gimple_opt_pass): Convert to subclass of opt_pass.
8002 (gimple_opt_pass::gimple_opt_pass): New.
8003 (rtl_opt_pass): Convert to subclass of opt_pass.
8004 (rtl_opt_pass::rtl_opt_pass): New.
8005 (ipa_opt_pass_d): Convert to subclass of opt_pass.
8006 (ipa_opt_pass_d::ipa_opt_pass_d): New.
8007 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
8008 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
8009 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
8010 invocation of pass_mode_switching to reflect this moving from a
8011 global to a member of gcc::pass_manager.
8012 (ix86_option_override): Rework how pass_insert_vzeroupper is
8013 added to the pass_manager to reflect autogenerated changes.
8014 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
8016 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
8018 PR rtl-optimization/57708
8019 * recog.c (peep2_find_free_register): Validate all regs in a
8022 2013-08-05 Jan Hubicka <jh@suse.cz>
8025 * cgraph.c (verify_cgraph_node): Accept local flags from other
8027 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
8028 (function_and_variable_visibility): Likewise.
8029 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
8031 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
8033 * graph.c (init_graph_slim_pretty_print): Remove.
8034 (print_graph_cfg): Do not call it. Use local pretty printer.
8035 (start_graph_dump): Likewise.
8037 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
8039 * gimple-pretty-print.c (buffer): Remove.
8040 (initialized): Likewise.
8041 (maybe_init_pretty_print): Likewise.
8042 (print_gimple_stmt): Do not call it. Use non-static local
8043 pretty_printer variable.
8044 (print_gimple_expr): Likewise.
8045 (print_gimple_seq): Likewise.
8046 (gimple_dump_bb): Likewise.
8048 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
8050 * asan.c (asan_pp): Remove.
8051 (asan_pp_initialized): Likewise.
8052 (asan_pp_initialize): Likewise.
8053 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
8054 (asan_emit_stack_protection): Tidy. Use local pretty printer.
8055 (asan_add_global): Likewise.
8057 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
8059 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
8060 * diagnostic.h (diagnostic_flush_buffer): Adjust.
8061 * pretty-print.c (pp_formatted_text_data): Likewise.
8062 (pp_indent): Rename from pp_base_indent.
8063 (pp_format): Rename from pp_base_format.
8064 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
8065 (pp_format_verbatim): Rename from pp_base_format_verbatim.
8066 (pp_flush): Rename from pp_base_flush.
8067 (pp_set_line_maximum_length): Rename from
8068 pp_base_set_line_maximum_length.
8069 (pp_clear_output_area): Rename from pp_base_clear_output_area.
8070 (pp_set_prefix): Rename from pp_base_set_prefix.
8071 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
8072 (pp_emit_prefix): Rename from pp_base_emit_prefix.
8073 (pp_append_text): Rename from pp_base_append_text.
8074 (pp_formatted_text): Rename from pp_base_formatted_text.
8075 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
8076 (pp_remaining_character_count_for_line): Rename from
8077 pp_base_remaining_character_count_for_line.
8078 (pp_newline): Rename from pp_base_newline.
8079 (pp_character): Rename from pp_base_character.
8080 (pp_string): Rename from pp_base_string.
8081 (pp_maybe_space): Rename from pp_base_maybe_space.
8082 * asan.c (asan_pp_string): Adjust.
8083 (asan_emit_stack_protection): Likewise.
8084 (asan_add_global): Likewise.
8085 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
8086 * tree-mudflap.c (mf_varname_tree): Likewise.
8087 * tree-pretty-print.c (pp_tree_identifier): Rename from
8088 pp_base_tree_identifier.
8089 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
8090 Declare as function.
8092 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
8094 * pretty-print.h (pp_bar_bar): New.
8095 (pp_ampersand_ampersand): Likewise.
8096 (pp_less_equal): Likewise.
8097 (pp_greater_equal): Likewise.
8098 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
8099 printer functions instead of pp_string or operators and punctuators.
8100 (dump_gimple_call): Likewise.
8101 (dump_gimple_omp_for): Likewise.
8102 (dump_gimple_transaction): Likewise.
8103 (dump_gimple_phi): Likewise.
8104 (pp_gimple_stmt_1): Likewise.
8105 * sched-vis.c (print_insn): Likewise.
8106 * tree-mudflap.c (mf_varname_tree): Likewise.
8107 * tree-pretty-print.c (dump_block_node): Likewise.
8108 (dump_generic_node): Likewise.
8110 2013-08-02 Jan Hubicka <jh@suse.cz>
8112 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
8114 * lto-streamer-out.c (tree_is_indexable): Results decls and
8115 parm decls are not indexable.
8116 (DFS_write_tree_body): Do not follow args and results.
8117 (hash_tree): Likewise.
8118 (output_functions): Rearrange so struct function is needed
8119 only when real body is output; be able to also ouptut abstract
8120 functions; output DECL_ARGUMENTS and DECL_RESULT.
8121 (lto_output): When not in WPA, ale store abstract functions.
8122 (write_symbol): Do not care about RESULT_DECL.
8123 (output_symbol_p): Handle correctly sbtract decls.
8124 * lto-streamer-in.c (input_function): Rearrange so struct
8125 function can be NULL at entry; allow streaming of
8126 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
8127 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
8128 sanity check during LTO.
8129 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
8130 RESULT_DECl and DECL_ARGUMENTS.
8131 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
8134 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
8136 * pretty-print.h (pp_underscore): New.
8138 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
8139 printer functions instead of pp_character.
8140 (dump_binary_rhs): Likewise.
8141 (dump_ternary_rhs): Likewise.
8142 (dump_gimple_call_args): Likewise.
8143 (pp_points_to_solution): Likewise.
8144 (dump_gimple_call): Likewise.
8145 (dump_gimple_switch): Likewise.
8146 (dump_gimple_cond): Likewise.
8147 (dump_gimple_bind): Likewise.
8148 (dump_gimple_try): Likewise.
8149 (dump_gimple_omp_for): Likewise.
8150 (dump_gimple_omp_continue): Likewise.
8151 (dump_gimple_omp_single): Likewise.
8152 (dump_gimple_omp_sections): Likewise.
8153 (dump_gimple_omp_block): Likewise.
8154 (dump_gimple_omp_critical): Likewise.
8155 (dump_gimple_transaction): Likewise.
8156 (dump_gimple_asm): Likewise.
8157 (dump_gimple_phi): Likewise.
8158 (dump_gimple_omp_parallel): Likewise.
8159 (dump_gimple_omp_task): Likewise.
8160 (dump_gimple_omp_atomic_load): Likewise.
8161 (dump_gimple_omp_atomic_store): Likewise.
8162 (dump_gimple_mem_ops): Likewise.
8163 (pp_gimple_stmt_1): Likewise.
8164 (pp_cfg_jump): Likewise.
8165 (dump_implicit_edges): Likewise.
8166 (gimple_dump_bb_for_graph): Likewise.
8167 * graph.c (draw_cfg_node): Likewise.
8168 * langhooks.c (lhd_print_error_function): Likewise.
8169 * sched-vis.c (print_exp): Likewise.
8170 (print_value): Likewise.
8171 (print_pattern): Likewise.
8172 (print_insn): Likewise.
8173 (rtl_dump_bb_for_graph): Likewise.
8174 * tree-pretty-print.c (dump_function_declaration): Likewise.
8175 (dump_array_domain): Likewise.
8176 (dump_omp_clause): Likewise.
8177 (dump_location): Likewise.
8178 (dump_generic_node): Likewise.
8179 (print_struct_decl): Likewise.
8180 * diagnostic.c (diagnostic_show_locus): Use pp_space.
8182 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8184 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
8185 candidate table when replacing a candidate statement.
8186 (replace_rhs_if_not_dup): Likewise.
8187 (replace_one_candidate): Likewise.
8189 2013-08-02 Jan Hubicka <jh@suse.cz>
8190 Martin Liska <marxin.liska@gmail.com>
8192 * cgraphunit.c (add_new_function): Fix logic when adding from
8194 (assemble_thunk): Rename to ...
8195 (expand_thunk); .. this one; export; get it working with
8196 general functions; make produced gimple valid.
8197 * cgraph.h (expand_thunk): Declare.
8199 2013-08-02 Jan Hubicka <jh@suse.cz>
8201 * ipa-cp.c (gather_context_independent_values): Use
8202 ipa_get_param_move_cost.
8203 (get_replacement_map): Remove PARAM; move parameter folding
8205 (create_specialized_node): Update.
8206 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
8207 assert that we have gimple body; update move_cost.
8208 (count_formal_params): Assert that we have gimple body.
8209 (ipa_dump_param): New function.
8210 (ipa_alloc_node_params): Break out from ...
8211 (ipa_initialize_node_params): ... here.
8212 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
8213 (ipa_write_node_info): Stream move costs.
8214 (ipa_read_node_info): Read move costs.
8215 (ipa_update_after_lto_read): Do not recompute node params.
8216 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
8217 (ipa_get_param): Check we are not in WPA.
8218 (ipa_get_param_move_cost): New.
8219 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
8220 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
8221 parm numbers to be present.
8223 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
8225 PR rtl-optimization/58048
8226 * lra-constraints.c (process_alt_operands): Don't check asm
8227 operand on register.
8229 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
8231 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
8232 the implied StoreLoad barrier for atomic operations if before.
8234 2013-08-02 Jan Hubicka <jh@suse.cz>
8235 Martin Liska <marxin.liska@gmail.com>
8237 * cgraph.c (cgraph_function_body_availability): Do not check
8239 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
8240 symtab_node_availability): Declare.
8241 * ipa.c (can_replace_by_local_alias): New.
8242 (function_and_variable_visibility): Use it.
8243 * symtab.c (symtab_for_node_and_aliases,
8244 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
8246 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
8248 PR rtl-optimization/57963
8249 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
8250 (lra_constraints): Use them.
8252 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
8254 * config/arm/types.md (define_attr "type"): Add "load_acq"
8256 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
8258 (cortex_a53_store1): Likewise.
8260 2013-08-01 Jan Hubicka <jh@suse.cz>
8262 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
8263 partitions are not needed.
8265 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
8267 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
8268 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
8269 MAYBE_NON_Q_CLASS_P where appropriate.
8271 2013-08-01 Jan Hubicka <jh@suse.cz>
8273 * cgraph.h (release_function_body): Declare.
8274 * tree.c (free_lang_data_in_decl): Free, parameters and return values
8275 of unused delcarations.
8277 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8279 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
8280 RTL form when subtracting a constant.
8282 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8284 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
8285 Generate canonical plus rtx with negated immediate instead of minus
8287 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
8289 2013-08-01 Jan Hubicka <jh@suse.cz>
8291 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
8292 (cgraph_release_function_body): Likewise.
8293 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
8294 * cgraph.h (cgrpah_node): Rename abstract_and_needed
8295 to used_as_abstract_origin.
8296 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
8297 symbols used as abstract origins.
8298 * cgraphunit.c (analyze_functions): Update.
8299 * ipa.c (symtab_remove_unreachable_nodes): Recompute
8300 used_as_abstract_origin.
8301 * tree-inline.c (tree_function_versioning): Update
8302 used_as_abstract_origin; be ready for DECL_RESULT and
8303 DECL_ARGUMENTS to be NULL.
8305 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
8306 for abstract functions.
8307 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
8310 2013-08-01 Jan Hubicka <jh@suse.cz>
8312 * profile.c (compute_value_histograms): Fix thinko.
8314 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
8316 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
8317 aarch-common-protos.h to extra_headers.
8318 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
8319 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
8320 * config/aarch64/t-aarch64 (aarch-common.o): Define.
8322 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
8324 * config/aarch64/aarch64.md (define_attr "type"): Delete.
8325 Include "../arm/types.md". Define "type" attribute for all patterns.
8326 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
8329 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
8331 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
8332 to support power8 load fusion.
8333 (fusion_gpr_mem_load): Likewise.
8335 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
8337 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
8338 declarations for power8 load fusion.
8339 (emit_fusion_gpr_load): Likewise.
8341 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
8342 tuning for power8, turn on fusion mode by default. Turn on sign
8343 extending fusion mode if normal fusion mode is on, and we are at
8345 (fusion_gpr_load_p): New function, return true if we can fuse an
8346 addis instruction with a dependent load to a GPR.
8347 (emit_fusion_gpr_load): Emit the instructions for power8 load
8350 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
8351 (VSX load fusion peepholes): New peepholes to fuse together an
8352 addi instruction with a VSX load instruction.
8354 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
8355 peepholes to fuse an addis instruction with a load to a GPR base
8356 register. If we are supporting sign extending fusions, convert
8357 sign extending loads to zero extending loads and add an explicit
8360 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
8362 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
8363 aarch-common-protos.h to extra_headers.
8364 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
8365 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
8366 (arm_early_store_addr_dep): Likewise.
8367 (arm_no_early_alu_shift_dep): Likewise.
8368 (arm_no_early_alu_shift_value_dep): Likewise.
8369 (arm_no_early_mul_dep): Likewise.
8370 (arm_no_early_store_addr_dep): Likewise.
8371 (arm_mac_accumulator_is_mul_result): Likewise.
8372 (arm_mac_accumulator_is_result): Likewise.
8373 * config/arm/aarch-common.c: ... here. New file.
8374 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
8376 (arm_early_store_addr_dep): Likewise.
8377 (arm_no_early_alu_shift_dep): Likewise.
8378 (arm_no_early_alu_shift_value_dep): Likewise.
8379 (arm_no_early_mul_dep): Likewise.
8380 (arm_no_early_store_addr_dep): Likewise.
8381 (arm_mac_accumulator_is_mul_result): Likewise.
8382 (arm_mac_accumulator_is_result): Likewise.
8383 * config/arm/aarch-common-protos.h: ... here. New file.
8384 * config/arm/t-arm (aarch-common.o): Define.
8386 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
8388 * config/arm/arm.md: Include new file "types.md".
8389 (define_attr "type"): Move from here to ...
8390 (define_attr "mul32"): Likewise.
8391 (define_attr "mul64"): Likewise.
8392 * config/arm/types.md: ... here. New file.
8394 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
8396 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
8397 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
8399 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8401 * gen-pass-instances.awk: Fix offset of substr().
8403 2013-07-31 David Malcolm <dmalcolm@redhat.com>
8405 * Makefile.in (pass-instances.def): New.
8406 (passes.o): Replace dependency on passes.def with one on
8409 * gen-pass-instances.awk: New.
8411 * passes.c (pass_manager::pass_manager): Use pass-instances.def
8412 rather than passes.def, updating local definition of NEXT_PASS
8413 macro to add an extra NUM parameter (currently unused).
8415 2013-07-30 David Malcolm <dmalcolm@redhat.com>
8417 * Makefile.in (PASS_MANAGER_H): New.
8418 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
8419 (passes.o): Likewise.
8420 (statistics.o): Likewise.
8421 (cgraphunit.o): Likewise.
8422 (context.o): Depend on PASS_MANAGER_H.
8424 * pass_manager.h: New.
8426 * cgraphunit.c (cgraph_add_new_function): Update for moves
8427 of globals to fields of pass_manager.
8428 (analyze_function): Likewise.
8429 (expand_function): Likewise.
8430 (ipa_passes): Likewise.
8431 (compile): Likewise.
8433 * context.c (context::context): New.
8434 * context.h (context::context): New.
8435 (context::get_passes): New.
8436 (context::passes_): New.
8438 * lto-cgraph.c (input_node): Update for moves of globals to
8439 fields of pass_manager.
8441 * passes.c (all_passes): Remove, in favor of a field of the
8442 same name within the new class pass_manager.
8443 (all_small_ipa_passes): Likewise.
8444 (all_lowering_passes): Likewise.
8445 (all_regular_ipa_passes): Likewise.
8446 (all_late_ipa_passes): Likewise.
8447 (all_lto_gen_passes): Likewise.
8448 (passes_by_id): Likewise.
8449 (passes_by_id_size): Likewise.
8450 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
8451 the new class pass_manager.
8452 (set_pass_for_id): Convert to...
8453 (pass_manager::set_pass_for_id): ...method.
8454 (get_pass_for_id): Convert to...
8455 (pass_manager::get_pass_for_id): ...method.
8456 (register_one_dump_file): Move body of implementation into...
8457 (pass_manager::register_one_dump_file): ...here.
8458 (register_dump_files_1): Convert to...
8459 (pass_manager::register_dump_files_1): ...method.
8460 (register_dump_files): Convert to...
8461 (pass_manager::register_dump_files): ...method.
8462 (create_pass_tab): Update for moves of globals to fields of
8464 (dump_passes): Move body of implementation into...
8465 (pass_manager::dump_passes): ...here.
8466 (register_pass): Move body of implementation into...
8467 (pass_manager::register_pass): ...here.
8468 (init_optimization_passes): Convert into...
8469 (pass_manager::pass_manager): ...constructor for new
8470 pass_manager class, and initialize the pass_lists array.
8471 (check_profile_consistency): Update for moves of globals to
8472 fields of pass_manager.
8473 (dump_profile_report): Move body of implementation into...
8474 (pass_manager::dump_profile_report): ...here.
8475 (ipa_write_summaries_1): Update for moves of pass lists from
8476 being globals to fields of pass_manager.
8477 (ipa_write_optimization_summaries): Likewise.
8478 (ipa_read_summaries): Likewise.
8479 (ipa_read_optimization_summaries): Likewise.
8480 (execute_all_ipa_stmt_fixups): Likewise.
8482 * statistics.c (statistics_fini): Update for moves of globals to
8483 fields of pass_manager.
8485 * toplev.c (general_init): Replace call to
8486 init_optimization_passes with construction of the pass_manager
8489 * tree-pass.h (all_passes): Remove, in favor of a field of the
8490 same name within the new class pass_manager.
8491 (all_small_ipa_passes): Likewise.
8492 (all_lowering_passes): Likewise.
8493 (all_regular_ipa_passes): Likewise.
8494 (all_lto_gen_passes): Likewise.
8495 (all_late_ipa_passes): Likewise.
8496 (passes_by_id): Likewise.
8497 (passes_by_id_size): Likewise.
8498 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
8499 the new class pass_manager.
8500 (get_pass_for_id): Remove.
8502 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
8504 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
8507 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
8509 * arm.md (mulhi3): New expand pattern.
8511 2013-07-30 Jan Hubicka <jh@suse.cz>
8512 Martin Liska <marxin.liska@gmail.com>
8514 * profile.c (compute_value_histograms): Do not ICE when
8515 there is mismatch only on some counters.
8517 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8519 PR rtl-optimization/57637
8520 * function.c (move_insn_for_shrink_wrap): Also check the
8521 GEN set of the LIVE problem for the liveness analysis
8522 if it exists, otherwise give up.
8524 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8526 PR tree-optimization/57993
8527 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
8528 replaced statement in the candidate table.
8529 (phi_add_costs): Return infinite cost when the hidden basis does
8530 not dominate all phis on which the candidate is dependent.
8531 (replace_one_candidate): Record replaced statement in the
8534 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
8536 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
8537 (ashlv2si3): New expander.
8538 (*ashlv2si3_i): New define_insn_and_split.
8539 * predicates.md (float_operation): Allow patterns with three
8542 PR rtl-optimization/58021
8543 * mode-switching.c (create_pre_exit): Always split off preceding
8544 insns if we are not at the basic block head.
8546 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
8548 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
8549 (UCLIBC_DYNAMIC_LINKER): New macro.
8550 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
8552 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
8553 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
8555 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
8556 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
8557 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
8558 for SF and DF modes. Use ieee_quad_format for TF mode.
8559 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
8560 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
8561 (mips_option_override): Handle `-mnan=legacy'.
8562 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
8563 `-mabs=2008' and `-mnan=2008'.
8564 (OPTION_DEFAULT_SPECS): Add "nan" default.
8565 (ASM_SPEC): Handle `-mnan='.
8566 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
8567 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
8568 comment accordingly.
8569 (neg<mode>2): Likewise.
8570 * config/mips/mips.opt (mabs, mnan): New options.
8571 * doc/install.texi (Configuration): Document `--with-nan=' option.
8572 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
8574 (MIPS Options): Document them.
8575 * config.gcc <mips*-*-*>: Handle `--with-nan='.
8576 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
8577 * configure: Regenerate.
8578 * config.in: Regenerate.
8580 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
8582 * config/i386/i386.md (float post-reload splitters): Do not check
8583 for subregs of SSE registers.
8585 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
8586 H.J. Lu <hongjiu.lu@intel.com>
8590 * config/i386/i386.md (post-reload splitter
8591 to avoid partial SSE reg dependency stalls): New pattern.
8593 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
8595 * config/s390/s390.md ("movcc"): Swap load and store instructions.
8597 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
8599 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
8600 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
8602 2013-07-26 Cary Coutant <ccoutant@google.com>
8604 * dwarf2out.c (die_checksum_ordered): Don't include template
8605 instantiations in signature.
8606 (is_template_parameter): New function.
8607 (is_template_instantiation): New function.
8608 (generate_skeleton_bottom_up): Don't include template instantiations
8610 (generate_skeleton): Likewise.
8611 (break_out_comdat_types): Move recursive call to break out nested
8613 (prune_unused_types_mark_generic_parms_dies): Call
8614 is_template_parameter.
8616 2013-07-26 Ian Bolton <ian.bolton@arm.com>
8618 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
8619 uses vector registers.
8620 * config/aarch64/iterators.md: Add attributes rtn and vas.
8622 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8623 Richard Earnshaw <richard.earnshaw@arm.com>
8625 * combine.c (simplify_comparison): Re-canonicalize operands
8627 * config/arm/arm.md (movcond_addsi): New splitter.
8629 2013-07-25 Sterling Augustine <saugustine@google.com>
8631 * dwarf2out.c (size_of_pubnames): Move code to...
8632 (include_pubname_in_output): ...here. New.
8633 (want_pubnames): Rearrange.
8634 (output_pubnames): Call include_pubname_in_output. Move assertion.
8636 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
8638 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
8640 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
8643 * doc/extend.texi: Remove obsolete builtins. Fix
8644 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
8646 2013-07-25 Jan Hubicka <jh@suse.cz>
8648 * cgraph.c (release_function_body): Break out from ...
8649 (cgraph_release_function_body): ... this one; also release DECL_RESULT
8651 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
8652 old_tree in the map.
8653 (create_specialized_node): Update.
8654 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
8656 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
8657 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
8658 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
8659 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
8662 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8664 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
8665 addsi3_carryin_alt2_<optab>): Correct output template.
8667 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8669 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
8670 Adjust for arm_restrict_it.
8671 Remove trailing whitespace.
8673 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
8675 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
8676 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
8678 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
8680 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
8682 PR rtl-optimization/57960
8683 * lra-constraints.c (process_alt_operands): Use the right mode
8684 when checking strict_low.
8686 2013-07-25 Jan Hubicka <jh@suse.cz>
8688 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
8689 * cgraph.c (cgraph_remove_node): Do not release function body
8690 when in cgraph streaming.
8691 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
8692 in other partitions are not considered reachable; fix handling of
8695 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8697 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
8699 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8704 * config/arm/arm.md ("*sibcall_insn): Replace use of
8705 Ss with US. Adjust output for v5 and v4t.
8706 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
8708 * config/arm/constraints.md ("Ss"): Rename to US.
8710 2013-07-25 Terry Guo <terry.guo@arm.com>
8712 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
8713 shift_add/shift_sub0/shift_sub1 RTXs.
8715 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
8716 Anton Blanchard <anton@au1.ibm.com>
8718 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
8719 (altivec_vpks<VI_char>ss): Likewise.
8720 (altivec_vpks<VI_char>us): Likewise.
8721 (altivec_vpku<VI_char>us): Likewise.
8722 (altivec_vpku<VI_char>um): Likewise.
8724 2013-07-24 David Malcolm <dmalcolm@redhat.com>
8726 Introduce context class.
8728 * Makefile.in (CONTEXT_H): New.
8729 (OBJS): Add context.o.
8730 (toplev.o): Add CONTEXT_H to dependencies.
8733 * toplev.c (general_init): Create the singleton gcc::context instance.
8739 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
8741 PR rtl-optimization/57968
8742 * mode-switching.c (create_pre_exit): Allow instructions that
8743 don't set a return register to need a non-exit mode.
8745 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8746 Anton Blanchard <anton@au1.ibm.com>
8748 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
8749 operands to vperm for little endian.
8750 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
8751 of lvsl to create the control mask for a vperm for little endian.
8753 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8754 Anton Blanchard <anton@au1.ibm.com>
8756 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
8757 two operands for little-endian.
8759 2013-07-23 Steve Ellcey <sellcey@mips.com>
8761 * config/mips/mips.c (mips_case_values_threshold): New.
8762 (TARGET_CASE_VALUES_THRESHOLD): Define.
8764 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8765 Anton Blanchard <anton@au1.ibm.com>
8767 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
8768 selection of field for vector splat in little endian mode.
8770 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8772 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
8773 expanders to rs6000.md.
8774 (ior<mode>3): Likewise.
8775 (and<mode>3): Likewise.
8776 (one_cmpl<mode>2): Likewise.
8777 (nor<mode>3): Likewise.
8778 (andc<mode>3): Likewise.
8779 (eqv<mode>3): Likewise.
8780 (nand<mode>3): Likewise.
8781 (orc<mode>3): Likewise.
8783 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
8786 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
8787 to split multi-word logical operations.
8788 (rs6000_split_logical_di): Likewise.
8789 (rs6000_split_logical): Likewise.
8791 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
8792 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
8793 and allow TImode operations in 32-bit.
8794 (vsx_and<mode>3_64bit): Likewise.
8795 (vsx_ior<mode>3_32bit): Likewise.
8796 (vsx_ior<mode>3_64bit): Likewise.
8797 (vsx_xor<mode>3_32bit): Likewise.
8798 (vsx_xor<mode>3_64bit): Likewise.
8799 (vsx_one_cmpl<mode>2_32bit): Likewise.
8800 (vsx_one_cmpl<mode>2_64bit): Likewise.
8801 (vsx_nor<mode>3_32bit): Likewise.
8802 (vsx_nor<mode>3_64bit): Likewise.
8803 (vsx_andc<mode>3_32bit): Likewise.
8804 (vsx_andc<mode>3_64bit): Likewise.
8805 (vsx_eqv<mode>3_32bit): Likewise.
8806 (vsx_eqv<mode>3_64bit): Likewise.
8807 (vsx_nand<mode>3_32bit): Likewise.
8808 (vsx_nand<mode>3_64bit): Likewise.
8809 (vsx_orc<mode>3_32bit): Likewise.
8810 (vsx_orc<mode>3_64bit): Likewise.
8812 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
8813 logical types in GPRs.
8815 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
8816 logical insns to rs6000.md, and allow TImode operations in
8818 (altivec_ior<mode>3): Likewise.
8819 (altivec_xor<mode>3): Likewise.
8820 (altivec_one_cmpl<mode>2): Likewise.
8821 (altivec_nor<mode>3): Likewise.
8822 (altivec_andc<mode>3): Likewise.
8824 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
8825 attributes for moving the 128-bit logical operations into
8827 (BOOL_REGS_OUTPUT): Likewise.
8828 (BOOL_REGS_OP1): Likewise.
8829 (BOOL_REGS_OP2): Likewise.
8830 (BOOL_REGS_UNARY): Likewise.
8831 (BOOL_REGS_AND_CR0): Likewise.
8832 (one_cmpl<mode>2): Add support for DI logical operations on
8833 32-bit, splitting the operations to 32-bit.
8837 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
8838 changes to combine the 32/64-bit code, allow logical operations on
8839 TI mode in 32-bit, and to use similar match_operator patterns like
8840 scalar mode uses. Combine the Altivec and VSX code for logical
8841 operations, and move it here.
8842 (ior<mode>3, 128-bit types): Likewise.
8843 (xor<mode>3, 128-bit types): Likewise.
8844 (one_cmpl<mode>3, 128-bit types): Likewise.
8845 (nor<mode>3, 128-bit types): Likewise.
8846 (andc<mode>3, 128-bit types): Likewise.
8847 (eqv<mode>3, 128-bit types): Likewise.
8848 (nand<mode>3, 128-bit types): Likewise.
8849 (orc<mode>3, 128-bit types): Likewise.
8850 (and<mode>3_internal): Likewise.
8851 (bool<mode>3_internal): Likewise.
8852 (boolc<mode>3_internal1): Likewise.
8853 (boolc<mode>3_internal2): Likewise.
8854 (boolcc<mode>3_internal1): Likewise.
8855 (boolcc<mode>3_internal2): Likewise.
8856 (eqv<mode>3_internal1): Likewise.
8857 (eqv<mode>3_internal2): Likewise.
8858 (one_cmpl1<mode>3_internal): Likewise.
8860 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8862 * config/microblaze/microblaze.c (microblaze_expand_prologue):
8863 Rename flag_stack_usage to flag_stack_usage_info.
8865 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8867 * config/microblaze/sync.md: New file.
8868 * config/microblaze/microblaze.md: Include sync.md
8869 * config/microblaze/microblaze.c: Add print_operand 'y'.
8870 * config/microblaze/constraints.md: Add memory_contraint
8871 'Q' which is a single register.
8873 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
8875 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
8877 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8879 * reload.c (find_reloads): Exit loop once we find this operand
8880 cannot be reloaded somehow for this alternative.
8882 * reload.c (find_reloads): Exit loop once we find a hard register.
8884 * rtlanal.c (computed_jump_p): Exit loop once we find label
8887 * i386.c (ix86_pad_returns): Exit loop after setting replace.
8889 * cfgloopmanip.c (remove_path): Exit loop after setting
8892 * gensupport.c (subst_dup): Avoid loop if code is not
8893 MATCH_DUP nor MATCH_OP_DUP.
8895 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
8897 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
8899 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8901 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
8902 true for SP_REGNUM if mode == ptr_mode.
8903 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
8904 with value R0_REGNUM + 31.
8906 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8908 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
8909 pad pointer-typed argument downward.
8911 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8913 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
8914 and __ILP32__ when the ILP32 model is in use.
8916 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8918 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
8919 (aarch64_load_symref_appropriately): In the case of
8920 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
8921 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
8922 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
8923 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
8924 if 'mode' doesn't equal to 'ptr_mode'.
8925 (aarch64_output_mi_thunk): Add an assertion on the alignment of
8926 'vcall_offset'; change to call aarch64_emit_move differently depending
8927 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
8928 to calculate the upper bound of 'vcall_offset'.
8929 (aarch64_cannot_force_const_mem): Change to also return true if
8931 (aarch64_legitimize_reload_address): In the case of large
8932 displacements, add new local variable 'xmode' and an assertion
8933 based on it; change to use 'xmode' to generate the new rtx and
8935 (aarch64_asm_trampoline_template): Change to generate the template
8936 differently depending on TARGET_ILP32 or not; change to use
8937 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
8938 (aarch64_trampoline_size): Removed.
8939 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
8940 and replace immediate literals with it. Change to use 'ptr_mode'
8941 instead of 'DImode' and call convert_memory_address if the mode
8942 of 'fnaddr' doesn't equal to 'ptr_mode'.
8943 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
8945 (aarch64_elf_asm_destructor): Likewise.
8946 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
8947 on TARGET_ILP32 instead of aarch64_trampoline_size.
8948 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
8949 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
8950 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
8951 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
8952 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
8953 (storewb_pair<GPI:mode>_<P:mode>): ... this.
8954 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
8955 depending on the value of 'mode'.
8956 (add_losym_<mode>): New.
8957 (ldr_got_small_<mode>): New, based on ldr_got_small.
8958 (ldr_got_small): Remove.
8959 (ldr_got_small_sidi): New.
8960 * config/aarch64/iterators.md (P): New.
8961 (PTR): Change to 'ptr_mode' in the condition.
8963 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8965 * config.gcc (aarch64*-*-*): Support --with-abi.
8966 (aarch64*-*-elf): Support --with-multilib-list.
8967 (aarch64*-*-linux*): Likewise.
8968 (supported_defaults): Add abi to aarch64*-*-*.
8969 * configure.ac: Mention AArch64 for --with-multilib-list.
8970 * configure: Re-generated.
8971 * config/aarch64/biarchilp32.h: New file.
8972 * config/aarch64/biarchlp64.h: New file.
8973 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
8975 (MULTILIB_DEFAULTS): Ditto.
8976 (DRIVER_SELF_SPECS): Ditto.
8977 (ASM_SPEC): Update to also substitute -mabi.
8978 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
8979 file whose name depends on -mabi= and -mbig-endian.
8980 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
8982 (POINTER_SIZE): New define.
8983 (POINTERS_EXTEND_UNSIGNED): Ditto.
8984 (enum aarch64_abi_type): New enumeration tag.
8985 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
8986 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
8987 (TARGET_ILP32): New define.
8988 * config/aarch64/aarch64.opt (mabi): New.
8990 (ilp32, lp64): New values for -mabi.
8991 * config/aarch64/t-aarch64 (comma): New define.
8992 (MULTILIB_OPTIONS): Ditto.
8993 (MULTILIB_DIRNAMES): Ditto.
8994 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
8995 * doc/invoke.texi: Document -mabi for AArch64.
8997 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
8999 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
9001 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9002 Anton Blanchard <anton@au1.ibm.com>
9004 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
9005 endianness when selecting field to splat.
9007 2013-07-22 Eric Christopher <echristo@gmail.com>
9009 * dwarf2out.c (die_odr_checksum): New function to use
9010 CHECKSUM_ macros and ULEB128 for DIE tag.
9011 (generate_type_signature): Use.
9013 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
9015 * config.gcc (sparc*-*-*): Accept leon3 processor.
9016 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
9017 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
9018 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
9019 * config/sparc/sparc.opt (enum processor_type): Add leon3.
9020 (mfix-ut699): Adjust comment.
9021 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
9022 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
9023 (CPP_CPU_SPEC): Likewise.
9024 (ASM_CPU_SPEC): Likewise.
9025 * config/sparc/sparc.c (leon3_cost): New constant.
9026 (sparc_option_override): Add leon3 support.
9027 (mem_ref): New function.
9028 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
9029 (sparc_do_work_around_errata): Look into the instruction in the delay
9030 slot and adjust accordingly. Add fix for the data cache nullify issues
9031 of the UT699. Change insertion position for the NOP.
9032 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
9033 (leon3_load): New reservation.
9034 (leon_store): Bump latency to 2.
9035 (grfpu): New automaton.
9036 (grfpu_alu): New unit.
9037 (grfpu_ds): Likewise.
9038 (leon_fp_alu): Adjust.
9039 (leon_fp_mult): Delete.
9040 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
9041 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
9042 * config/sparc/sparc.md (cpu): Add leon3.
9043 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
9045 (atomic_test_and_set): Likewise.
9048 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
9050 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
9051 default for R5900 targets.
9052 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
9053 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
9054 * config/mips/mips.c (mips_option_override): Report an error for
9055 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
9056 for -march=r5900 -mhard-float.
9058 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
9060 * df-problems.c (can_move_insns_across): Exit loop once we
9061 find a non-fixed, non-global register.
9063 * ipa-pure-const.c (propagate_nothrow): Exit loop after
9066 * omega.c (omega_eliminate_red): Break after setting red_found.
9067 (omega_problem_has_red_equations): Similarly after setting found.
9068 (omega_query_variable): Similarly after setting coupled.
9070 2013-07-22 Marek Polacek <polacek@redhat.com>
9072 * gimplify.c: Don't include gimple.h twice.
9074 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9076 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
9077 instead of TARGET_THUMB1.
9078 (Pz): New constraint.
9079 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
9081 (compare_negsi_si): Likewise.
9082 (compare_addsi2_op0): Likewise.
9083 (compare_addsi2_op1): Likewise.
9084 (addsi3_carryin_<optab>): Likewise.
9085 (addsi3_carryin_alt2_<optab>): Likewise.
9086 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
9087 for arm_restrict_it.
9088 (subsi3_carryin): Likewise.
9089 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
9090 (minmax_arithsi): Disable for arm_restrict_it.
9091 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
9092 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
9093 (satsi_<SAT:code>_shift): Likewise.
9094 (arm_shiftsi3): Add alternative for 16-bit encoding.
9095 (arm32_movhf): Disable for arm_restrict_it.
9096 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
9097 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
9099 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
9101 * config/arm/arm.md (attribute "insn"): Delete.
9102 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
9103 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
9104 (not_shiftsi): Update for attribute change.
9105 (not_shiftsi_compare0): Likewise.
9106 (not_shiftsi_compare0_scratch): Likewise.
9107 (arm_one_cmplsi2): Likewise.
9108 (thumb1_one_cmplsi2): Likewise.
9109 (notsi_compare0): Likewise.
9110 (notsi_compare0_scratch): Likewise.
9111 (thumb1_movdi_insn): Likewise.
9112 (arm_movsi_insn): Likewise.
9113 (movhi_insn_arch4): Likewise.
9114 (movhi_bytes): Likewise.
9115 (arm_movqi_insn): Likewise.
9116 (thumb1_movqi_insn): Likewise.
9117 (arm32_movhf): Likewise.
9118 (thumb1_movhf): Likewise.
9119 (arm_movsf_soft_insn): Likewise.
9120 (thumb1_movsf_insn): Likewise.
9121 (thumb_movdf_insn): Likewise.
9122 (movsicc_insn): Likewise.
9123 (movsfcc_soft_insn): Likewise.
9124 (and_scc): Likewise.
9125 (cond_move): Likewise.
9126 (if_move_not): Likewise.
9127 (if_not_move): Likewise.
9128 (if_shift_move): Likewise.
9129 (if_move_shift): Likewise.
9130 (if_shift_shift): Likewise.
9131 (if_not_arith): Likewise.
9132 (if_arith_not): Likewise.
9133 (cond_move_not): Likewise.
9134 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
9135 (neon_mov<mode>): Likewise.
9136 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
9137 (thumb2_movsi_vfp): Likewise.
9138 (movsf_vfp): Likewise.
9139 (thumb2_movsf_vfp): Likewise.
9140 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
9142 (cortexa7_older_only): Likewise.
9143 (cortexa7_younger): Likewise.
9144 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
9145 (1020alu_shift_op): Likewise.
9146 (1020alu_shift_reg_op): Likewise.
9147 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
9148 (alu_shift_op): Likewise.
9149 (alu_shift_reg_op): Likewise.
9150 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
9151 (11_alu_shift_op): Likewise.
9152 (11_alu_shift_reg_op): Likewise.
9153 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
9154 (9_alu_shift_reg_op): Likewise.
9155 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9157 (cortex_a15_alu_shift): Likewise.
9158 (cortex_a15_alu_shift_reg): Likewise.
9159 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
9160 (cortex_a5_alu_shift): Likewise.
9161 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
9163 (cortex_a53_alu_shift): Likewise.
9164 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
9166 (cortex_a7_alu_reg): Likewise.
9167 (cortex_a7_alu_shift): Likewise.
9168 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
9169 (cortex_a8_alu_shift): Likewise.
9170 (cortex_a8_alu_shift_reg): Likewise.
9171 (cortex_a8_mov): Likewise.
9172 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
9173 (cortex_a9_dp_shift): Likewise.
9174 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
9175 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
9176 (cortex_r4_mov): Likewise.
9177 (cortex_r4_alu_shift): Likewise.
9178 (cortex_r4_alu_shift_reg): Likewise.
9179 * config/arm/fa526.md (526_alu_op): Update for attribute change.
9180 (526_alu_shift_op): Likewise.
9181 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
9182 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
9183 (626te_alu_shift_op): Likewise.
9184 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
9185 (726te_alu_op): Likewise.
9186 (726te_alu_shift_op): Likewise.
9187 (726te_alu_shift_reg_op): Likewise.
9188 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
9189 (mp626_alu_shift_op): Likewise.
9190 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
9191 (pj4_alu_e1_conds): Likewise.
9192 (pj4_alu): Likewise.
9193 (pj4_alu_conds): Likewise.
9194 (pj4_shift): Likewise.
9195 (pj4_shift_conds): Likewise.
9196 (pj4_alu_shift): Likewise.
9197 (pj4_alu_shift_conds): Likewise.
9199 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9201 * config/arm/predicates.md (shiftable_operator_strict_it):
9203 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
9204 Disable cond_exec version for arm_restrict_it.
9205 (thumb2_smaxsi3): Convert to generate cond_exec.
9206 (thumb2_sminsi3): Likewise.
9207 (thumb32_umaxsi3): Likewise.
9208 (thumb2_uminsi3): Likewise.
9209 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
9210 (thumb2_neg_abssi2): Likewise.
9211 (thumb2_mov_scc): Add alternative for 16-bit encoding.
9212 (thumb2_movsicc_insn): Adjust alternatives.
9213 (thumb2_mov_negscc): Disable for arm_restrict_it.
9214 (thumb2_mov_negscc_strict_it): New pattern.
9215 (thumb2_mov_notscc_strict_it): New pattern.
9216 (thumb2_mov_notscc): Disable for arm_restrict_it.
9217 (thumb2_ior_scc): Likewise.
9218 (thumb2_ior_scc_strict_it): New pattern.
9219 (thumb2_cond_move): Adjust for arm_restrict_it.
9220 (thumb2_cond_arith): Disable for arm_restrict_it.
9221 (thumb2_cond_arith_strict_it): New pattern.
9222 (thumb2_cond_sub): Adjust for arm_restrict_it.
9223 (thumb2_movcond): Likewise.
9224 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
9225 (thumb2_zero_extendhisi2_v6): Likewise.
9226 (thumb2_zero_extendqisi2_v6): Likewise.
9227 (orsi_notsi_si): Likewise.
9228 (orsi_not_shiftsi_si): Likewise.
9230 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
9232 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
9233 instruction sequence is 1 byte shorter.
9235 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
9237 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
9238 it is not needed after split.
9240 2013-07-20 Iain Sandoe <iain@codesourcery.com>
9243 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
9244 second label for nonlocal goto receivers. Don't output pic base labels
9245 unless we're producing PIC; mark that action unreachable().
9246 (ix86_save_reg): If the function contains a nonlocal label, save the
9248 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
9249 * config/darwin.c (emitted_pic_label_num): New GTY.
9250 (update_pic_label_number_if_needed): New.
9251 (machopic_output_function_base_name): Adjust for nonlocal receiver
9253 (machopic_should_output_picbase_label): New.
9254 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
9255 (nonlocal_goto_receiver): New insn and split.
9257 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
9259 * config/aarch64/aarch64-builtins.c
9260 (aarch64_fold_builtin): Fold abs in all modes.
9261 * config/aarch64/aarch64-simd-builtins.def
9262 (abs): Enable for all modes.
9263 * config/aarch64/arm_neon.h
9264 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
9265 (vabs_f64): Add missing intrinsic.
9267 2013-07-19 Ian Bolton <ian.bolton@arm.com>
9269 * config/aarch64/arm_neon.h (vabs_s64): New function
9271 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
9274 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
9275 * config/avr/avr.md (adjust_len): Add `round'.
9276 * config/avr/avr-protos.h (avr_out_round): New prototype.
9277 (avr_out_plus): Add `out_label' argument.
9278 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
9279 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
9280 Handle the case where `insn' is just a pattern.
9281 (avr_out_bitop): Handle the case where `insn' is just a pattern.
9282 (avr_out_round): New function.
9283 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
9285 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
9287 * config/microblaze/microblaze.c (microblaze_expand_prologue):
9288 Add check for flag_stack_usage to handle -fstack-usage support
9290 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
9292 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
9293 interaction for new Power8 flags and VSX.
9295 2013-07-18 Sriraman Tallam <tmsriram@google.com>
9298 * tree-inline.c (expand_call_inline): Emit errors during
9299 early_inlining only if optimization is not turned on.
9301 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9305 * passes.c (init_optimization_passes): Move the construction of
9306 the pass hierarchy into a new passes.def file.
9308 * Makefile.in (passes.o): Add dependency on passes.def.
9310 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9312 * passes.c (init_optimization_passes): Introduce macros for
9313 constructing the tree of passes (INSERT_PASSES_AFTER,
9314 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
9315 TERMINATE_PASS_LIST).
9317 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
9318 Wei Mi <wmi@google.com>
9320 PR rtl-optimization/57878
9321 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
9323 (reload_pseudo_compare_func): Check nregs first for reload
9326 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9328 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
9330 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
9332 * read-rtl.c (validate_const_int): Once an invalid character is
9333 seen, quit the loop.
9335 * gengtype.c (write_roots): Similarly once we find the "deletable"
9336 or "if_marked" option.
9338 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
9340 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
9341 "xtab" and "sat". Move value "clz" from here to ...
9342 (attriubte "type"): ... here.
9343 (satsi_<SAT:code>): Delete "insn" attribute.
9344 (satsi_<SAT:code>_shift): Likewise.
9345 (arm_zero_extendqisi2addsi): Likewise.
9346 (arm_extendqisi2addsi): Likewise.
9347 (clzsi2): Update for attribute changes.
9348 (rbitsi2): Likewise.
9349 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
9351 (arm_usatsihi): Likewise.
9352 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
9354 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
9356 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
9357 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
9358 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
9359 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
9360 in alphabetical order.
9361 (attribute "core_cycles"): Update for attribute changes.
9362 (arm_addsi3): Likewise.
9363 (addsi3_compare0): Likewise.
9364 (addsi3_compare0_scratch): Likewise.
9365 (addsi3_compare_op1): Likewise.
9366 (addsi3_compare_op2): Likewise.
9367 (compare_addsi2_op0): Likewise.
9368 (compare_addsi2_op1): Likewise.
9369 (addsi3_carryin_shift_<optab>): Likewise.
9370 (subsi3_carryin_shift): Likewise.
9371 (rsbsi3_carryin_shift): Likewise.
9372 (arm_subsi3_insn): Likewise.
9373 (subsi3_compare0): Likewise.
9374 (subsi3_compare): Likewise.
9375 (arm_andsi3_insn): Likewise.
9376 (thumb1_andsi3_insn): Likewise.
9377 (andsi3_compare0): Likewise.
9378 (andsi3_compare0_scratch): Likewise.
9379 (zeroextractsi_compare0_scratch
9380 (andsi_not_shiftsi_si): Likewise.
9381 (iorsi3_insn): Likewise.
9382 (iorsi3_compare0): Likewise.
9383 (iorsi3_compare0_scratch): Likewise.
9384 (arm_xorsi3): Likewise.
9385 (thumb1_xorsi3_insn): Likewise.
9386 (xorsi3_compare0): Likewise.
9387 (xorsi3_compare0_scratch): Likewise.
9388 (satsi_<SAT:code>_shift): Likewise.
9390 (arm_shiftsi3): Likewise.
9391 (shiftsi3_compare0): Likewise.
9392 (not_shiftsi): Likewise.
9393 (not_shiftsi_compare0): Likewise.
9394 (not_shiftsi_compare0_scratch): Likewise.
9395 (arm_one_cmplsi2): Likewise.
9396 (thumb_one_complsi2): Likewise.
9397 (notsi_compare0): Likewise.
9398 (notsi_compare0_scratch): Likewise.
9399 (thumb1_zero_extendhisi2): Likewise.
9400 (arm_zero_extendhisi2): Likewise.
9401 (arm_zero_extendhisi2_v6): Likewise.
9402 (arm_zero_extendhisi2addsi): Likewise.
9403 (thumb1_zero_extendqisi2): Likewise.
9404 (thumb1_zero_extendqisi2_v6): Likewise.
9405 (arm_zero_extendqisi2): Likewise.
9406 (arm_zero_extendqisi2_v6): Likewise.
9407 (arm_zero_extendqisi2addsi): Likewise.
9408 (thumb1_extendhisi2): Likewise.
9409 (arm_extendhisi2): Likewise.
9410 (arm_extendhisi2_v6): Likewise.
9411 (arm_extendqisi): Likewise.
9412 (arm_extendqisi_v6): Likewise.
9413 (arm_extendqisi2addsi): Likewise.
9414 (thumb1_extendqisi2): Likewise.
9415 (thumb1_movdi_insn): Likewise.
9416 (arm_movsi_insn): Likewise.
9417 (movsi_compare0): Likewise.
9418 (movhi_insn_arch4): Likewise.
9419 (movhi_bytes): Likewise.
9420 (arm_movqi_insn): Likewise.
9421 (thumb1_movqi_insn): Likewise.
9422 (arm32_movhf): Likewise.
9423 (thumb1_movhf): Likewise.
9424 (arm_movsf_soft_insn): Likewise.
9425 (thumb1_movsf_insn): Likewise.
9426 (movdf_soft_insn): Likewise.
9427 (thumb_movdf_insn): Likewise.
9428 (arm_cmpsi_insn): Likewise.
9429 (cmpsi_shiftsi): Likewise.
9430 (cmpsi_shiftsi_swp): Likewise.
9431 (arm_cmpsi_negshiftsi_si): Likewise.
9432 (movsicc_insn): Likewise.
9433 (movsfcc_soft_insn): Likewise.
9434 (arith_shiftsi): Likewise.
9435 (arith_shiftsi_compare0
9436 (arith_shiftsi_compare0_scratch
9437 (sub_shiftsi): Likewise.
9438 (sub_shiftsi_compare0
9439 (sub_shiftsi_compare0_scratch
9440 (and_scc): Likewise.
9441 (cond_move): Likewise.
9442 (if_plus_move): Likewise.
9443 (if_move_plus): Likewise.
9444 (if_move_not): Likewise.
9445 (if_not_move): Likewise.
9446 (if_shift_move): Likewise.
9447 (if_move_shift): Likewise.
9448 (if_shift_shift): Likewise.
9449 (if_not_arith): Likewise.
9450 (if_arith_not): Likewise.
9451 (cond_move_not): Likewise.
9452 (thumb1_ashlsi3): Set type attribute.
9453 (thumb1_ashrsi3): Likewise.
9454 (thumb1_lshrsi3): Likewise.
9455 (thumb1_rotrsi3): Likewise.
9456 (shiftsi3_compare0_scratch): Likewise.
9457 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
9458 (neon_mov<mode>): Likewise.
9459 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
9461 (thumb2_movsi_insn): Likewise.
9462 (thumb2_cmpsi_neg_shiftsi): Likewise.
9463 (thumb2_extendqisi_v6): Likewise.
9464 (thumb2_zero_extendhisi2_v6): Likewise.
9465 (thumb2_zero_extendqisi2_v6): Likewise.
9466 (thumb2_shiftsi3_short): Likewise.
9467 (thumb2_addsi3_compare0_scratch): Likewise.
9468 (orsi_not_shiftsi_si): Likewise.
9469 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
9470 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
9472 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9473 (1020alu_shift_op): Likewise.
9474 (1020alu_shift_reg_op): Likewise.
9475 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9476 (alu_shift_op): Likewise.
9477 (alu_shift_reg_op): Likewise.
9478 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9479 (11_alu_shift_op): Likewise.
9480 (11_alu_shift_reg_op): Likewise.
9481 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9482 (9_alu_shift_reg_op): Likewise.
9483 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9485 (cortex_a15_alu_shift): Likewise.
9486 (cortex_a15_alu_shift_reg): Likewise.
9487 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
9489 (cortex_a5_alu_shift): Likewise.
9490 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
9492 (cortex_a53_alu_shift): Likewise.
9493 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
9495 (cortex_a7_alu_reg): Likewise.
9496 (cortex_a7_alu_shift): Likewise.
9497 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
9499 (cortex_a8_alu_shift): Likewise.
9500 (cortex_a8_alu_shift_reg): Likewise.
9501 (cortex_a8_mov): Likewise.
9502 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9503 (cortex_a9_dp_shift): Likewise.
9504 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
9506 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
9508 (cortex_r4_mov): Likewise.
9509 (cortex_r4_alu_shift): Likewise.
9510 (cortex_r4_alu_shift_reg): Likewise.
9511 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9512 (526_alu_shift_op): Likewise.
9513 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9514 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9515 (626te_alu_shift_op): Likewise.
9516 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
9517 (726te_alu_op): Likewise.
9518 (726te_alu_shift_op): Likewise.
9519 (726te_alu_shift_reg_op): Likewise.
9520 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9521 (mp626_alu_shift_op): Likewise.
9522 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
9523 (pj4_alu_e1_conds): Likewise.
9524 (pj4_alu): Likewise.
9525 (pj4_alu_conds): Likewise.
9526 (pj4_shift): Likewise.
9527 (pj4_shift_conds): Likewise.
9528 (pj4_alu_shift): Likewise.
9529 (pj4_alu_shift_conds): Likewise.
9530 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
9532 (cortexa7_older_only): Likewise.
9533 (cortexa7_younger): Likewise.
9535 2013-07-18 David Malcolm <dmalcolm@redhat.com>
9537 * ipa-pure-const.c (generate_summary): Rename to...
9538 (pure_const_generate_summary): ... this.
9540 2013-07-17 Iain Sandoe <iain@codesourcery.com>
9542 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
9544 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
9547 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
9550 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9552 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
9553 enabled without -march=zEC12.
9554 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
9557 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
9559 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
9560 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
9561 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
9562 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
9563 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
9564 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
9565 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
9567 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
9568 and ISA_HAS_NMADD3_NMSUB3.
9569 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
9570 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
9571 (nmsub4<mode>, nmsub3<mode>): Likewise.
9572 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
9574 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
9576 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
9577 TARGET_MIPS5400 checking.
9579 2013-07-16 Jakub Jelinek <jakub@redhat.com>
9580 Peter Bergner <bergner@vnet.ibm.com>
9582 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
9583 registers in the comment.
9584 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
9585 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
9586 rather than FIRST_PSEUDO_REGISTERS.
9588 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
9590 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
9591 enable extra ISA flags with TARGET_HTM.
9593 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
9595 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
9598 2013-07-15 Cong Hou <congh@google.com>
9600 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
9601 in compare function for sorting.
9603 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
9605 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
9606 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
9607 * config/rs6000/rs6000.opt: Add -mhtm option.
9608 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
9609 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
9610 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
9611 __HTM__ if the HTM instructions are available.
9612 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
9613 htm_spr_reg_operand): New define_predicates.
9614 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
9615 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
9617 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
9618 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
9619 HTM builtin functions.
9620 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
9621 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
9622 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
9623 (rs6000_builtin_mask_calculate): Likewise.
9624 (rs6000_option_override_internal): Likewise.
9625 (bdesc_htm): Add new HTM builtin support.
9626 (htm_spr_num): New function.
9627 (htm_spr_regno): Likewise.
9628 (rs6000_htm_spr_icode): Likewise.
9629 (htm_expand_builtin): Likewise.
9630 (htm_init_builtins): Likewise.
9631 (rs6000_expand_builtin): Add support for HTM builtin functions.
9632 (rs6000_init_builtins): Likewise.
9633 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
9635 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
9636 (TARGET_HTM, MASK_HTM): Define macros.
9637 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
9638 (FIXED_REGISTERS): Likewise.
9639 (CALL_USED_REGISTERS): Likewise.
9640 (CALL_REALLY_USED_REGISTERS): Likewise.
9641 (REG_ALLOC_ORDER): Likewise.
9642 (enum reg_class): Likewise.
9643 (REG_CLASS_NAMES): Likewise.
9644 (REG_CLASS_CONTENTS): Likewise.
9645 (REGISTER_NAMES): Likewise.
9646 (ADDITIONAL_REGISTER_NAMES): Likewise.
9647 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
9648 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
9649 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
9650 * config/rs6000/htm.md: New file.
9651 * config/rs6000/htmintrin.h: New file.
9652 * config/rs6000/htmxlintrin.h: New file.
9654 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
9656 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
9657 Define SYMBOL_TINY_GOT, update comment.
9658 * config/aarch64/aarch64.c
9659 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
9660 (aarch64_expand_mov_immediate): Likewise.
9661 (aarch64_print_operand): Likewise.
9662 (aarch64_classify_symbol): Likewise.
9663 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
9664 (ldr_got_tiny): Define.
9666 2013-07-13 Tobias Grosser <tobias@grosser.es>
9668 PR tree-optimization/54094
9669 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
9670 scheduling dimension for the parallelism check from the polyhedral
9671 information in the AST.
9672 * graphite-dependences.c (carries_deps): Do not assume the schedule is
9675 2013-07-13 Jason Merrill <jason@redhat.com>
9677 * print-tree.c (debug_vec_tree): Use debug_raw.
9678 (debug_raw (vec<tree, va_gc> &)): New.
9679 (debug_raw (vec<tree, va_gc> *)): New.
9680 * tree.h: Declare them.
9682 2013-07-13 Bin Cheng <bin.cheng@arm.com>
9684 * ifcvt.c (ifcvt_after_combine): New static variable.
9685 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
9687 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
9688 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
9689 rest_of_handle_if_after_reload): Pass new argument for if_convert.
9691 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
9693 * config/mips/mips.c (mips_expand_call): Remove empty statement.
9695 2013-07-12 Michael Matz <matz@suse.de>
9698 * convert.c (convert_to_real): Reject non-float inner types.
9700 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
9702 * config/aarch64/aarch64-protos.h
9703 (aarch64_simd_immediate_valid_for_move): Remove.
9704 * config/aarch64/aarch64.c (simd_immediate_info): New member.
9705 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
9707 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
9709 2013-07-11 Steve Ellcey <sellcey@mips.com>
9711 * config/mips/mips.c (mips_conditional_register_usage): Do not
9712 use t[0-7] registers in MIPS16 mode when optimizing for size.
9714 2013-07-11 Sriraman Tallam <tmsriram@google.com>
9716 * config/i386/i386.c (dispatch_function_versions): Fix array
9717 indexing of function_version_info to match actual_versions.
9719 2013-07-11 Teresa Johnson <tejohnson@google.com>
9721 * vec.h (struct va_gc): Move release out-of-line.
9722 (va_gc::release): Call ggc_free on released vec.
9724 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
9726 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
9727 Require GOT register as additional operand in UNSPEC.
9728 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
9729 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
9730 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
9731 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
9732 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
9733 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
9734 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
9736 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
9739 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
9740 name seen by assembler/linker rather if available.
9742 2013-07-11 Andreas Schwab <schwab@suse.de>
9744 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
9746 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
9748 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
9750 2013-07-10 Joseph Myers <joseph@codesourcery.com>
9752 * doc/tm.texi.in: Move hook documentation to ....
9753 * target.def: ... here.
9755 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
9757 * doc/tm.texi: Regenerate.
9759 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
9762 * doc/invoke.texi: Document -Wconditionally-supported.
9764 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9767 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
9770 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9773 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
9774 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
9775 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
9776 Remove duplicate devices.
9777 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
9778 * config/avr/t-multilib: Regenerate.
9779 * config/avr/avr-tables.opt: Regenerate.
9780 * doc/avr-mmcu.texi: Regenerate.
9782 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9785 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
9787 2013-07-10 Graham Stott <graham.stott@btinternet.com>
9789 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
9790 the cost of MULT when optimizing for size.
9792 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9794 * config/cr16/cr16-protos.h: Don't include target.h.
9796 2013-07-09 Joseph Myers <joseph@codesourcery.com>
9798 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
9799 adjust register size for TDmode and TFmode for VSX registers.
9801 2013-07-08 Kai Tietz <ktietz@redhat.com>
9804 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
9805 hook_bool_const_tree_true.
9807 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9809 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
9810 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
9811 * config/s390/s390.md: Define FPR*_REGNUM constants.
9812 Fix FPR2_REGNUM constant (18 -> 17).
9813 ("*trunc<BFP:mode><DFP_ALL:mode>2")
9814 ("*trunc<DFP_ALL:mode><BFP:mode>2")
9815 ("trunc<BFP:mode><DFP_ALL:mode>2")
9816 ("trunc<DFP_ALL:mode><BFP:mode>2")
9817 ("*extend<BFP:mode><DFP_ALL:mode>2")
9818 ("*extend<DFP_ALL:mode><BFP:mode>2")
9819 ("extend<BFP:mode><DFP_ALL:mode>2")
9820 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
9823 2013-07-08 Graham Stott <graham.stott@btinternet.com>
9825 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
9827 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9829 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
9830 and cfun_fpr_bit_p to cfun_fpr_save_p.
9831 (s390_frame_area, s390_register_info, s390_frame_info)
9832 (s390_emit_prologue, s390_emit_epilogue)
9833 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
9835 * config/s390/s390.h: Define *_REGNUM macros for floating point
9838 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
9840 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
9842 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
9844 PR rtl-optimization/57786
9845 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
9846 and break out of the loop when it is set to false.
9848 2013-07-08 Jakub Jelinek <jakub@redhat.com>
9851 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
9852 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
9853 (const_int 63)) 0)).
9854 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
9855 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
9856 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
9858 PR rtl-optimization/57829
9859 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
9860 mask bits outside of mode are just sign-extension from mode to HWI.
9862 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
9864 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
9865 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
9866 adjust_address instead of change_address to keep info about alignment.
9867 (emit_strmov): Remove.
9868 (emit_memmov): New function.
9869 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
9870 (expand_movmem_epilogue): Likewise and return updated rtx for
9872 (expand_constant_movmem_prologue): Likewise and return updated rtx for
9873 destination and source.
9874 (decide_alignment): Refactor, handle vector_loop.
9875 (ix86_expand_movmem): Likewise.
9876 (ix86_expand_setmem): Likewise.
9877 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
9879 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
9881 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
9882 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
9884 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9886 * config/i386/sse.md (sse_movlhps): Change alternative 3
9887 of operand 2 to "m".
9889 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9892 * config/i386/sse.md (iptr): New mode attribute.
9893 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
9894 (<sse>_vm<plusminus_insn><mode>3): Ditto.
9895 (<sse>_vmmul<mode>3): Ditto.
9896 (<sse>_vmdiv<mode>3): Ditto.
9897 (sse_vmrcpv4sf2): Ditto.
9898 (<sse>_vmsqrt<mode>2): Ditto.
9899 (sse_vmrsqrtv4sf2): Ditto.
9900 (<sse>_vm<code><mode>3): Ditto.
9901 (avx_vmcmp<mode>3): Ditto.
9902 (<sse>_vmmaskcmp<mode>3): Ditto.
9903 (<sse>_comi): Ditto.
9904 (<sse>_ucomi): Ditto.
9905 (*xop_vmfrcz_<mode>): Ditto.
9906 (*fmai_fmadd_<mode>): Ditto.
9907 (*fmai_fmsub_<mode>): Ditto.
9908 (*fmai_fnmadd_<mode>): Ditto.
9909 (*fmai_fnmsub_<mode>): Ditto.
9910 (*fma4i_vmfmadd_<mode>): Ditto.
9911 (*fma4i_vmfmsub_<mode>): Ditto.
9912 (*fma4i_vmfnmadd_<mode>): Ditto.
9913 (*fma4i_vmfnmsub_<mode>): Ditto.
9914 (*xop_vmfrcz_<mode>): Ditto.
9915 (sse_cvtps2pi): Ditto.
9916 (sse_cvttps2pi): Ditto.
9917 (sse_cvtss2si): Ditto.
9918 (sse_cvtss2si_2): Ditto.
9919 (sse_cvtss2siq_2): Ditto.
9920 (sse_cvttss2si): Ditto.
9921 (sse_cvttss2siq): Ditto.
9922 (sse_cvtsd2si): Ditto.
9923 (sse_cvtsd2si_2): Ditto.
9924 (sse_cvtsd2siq_2): Ditto.
9925 (sse_cvttsd2si): Ditto.
9926 (sse_cvttsd2siq): Ditto.
9927 (sse_cvtsd2ss): Ditto.
9928 (sse_cvtss2sd): Ditto.
9929 (avx2_pbroadcast<mode>): Ditto.
9930 (avx2_pbroadcast<mode>_1): Ditto.
9931 (*avx_vperm_broadcast_v4sf): Ditto.
9933 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
9934 (sse_movlhps): Ditto.
9935 (sse_storehps): Ditto.
9936 (sse_loadhps): Ditto.
9937 (sse_storelps): Ditto.
9938 (sse_loadlps): Ditto.
9939 (*vec_concatv4sf): Ditto.
9940 (*vec_interleave_highv2df): Ditto.
9941 (*vec_interleave_lowv2df): Ditto.
9942 (*vec_extractv2df_1_sse): Ditto.
9943 (*vec_extractv2df_0_sse): Ditto.
9944 (sse2_storelpd): Ditto.
9945 (sse2_loadlpd): Ditto.
9946 (sse2_movsd): Ditto.
9947 (*vec_concatv4si): Ditto.
9948 (vec_concatv2di): Ditto.
9950 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
9951 for Intel asm dialect.
9952 (mmx_punpcklwd): Ditto.
9953 (mmx_punpckldq): Ditto.
9955 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
9956 for intel assembler dialect.
9958 2013-07-06 Jakub Jelinek <jakub@redhat.com>
9961 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
9962 for BUILT_IN_C{LZ,LRSB}*.
9963 * tree.h (CASE_INT_FN): Add FN##IMAX case.
9964 * tree-vrp.c (extract_range_basic): Handle
9965 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
9966 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
9967 fall thru to code calling set_value*.
9968 * builtins.c (expand_builtin): Remove *IMAX cases.
9969 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
9970 if width is bigger than 2*HWI.
9972 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
9974 PR rtl-optimization/55342
9975 * lra-int.h (lra_subreg_reload_pseudos): New.
9976 * lra.c: Add undoing optional reloads to the block diagram.
9977 (lra_subreg_reload_pseudos): New.
9978 (lra_optional_reload_pseudos): Change comments.
9979 (lra): Init and clear lra_subreg_reload_pseudos. Clear
9980 lra_optional_reload_pseudos after undo transformations.
9981 * lra-assigns.c (pseudo_prefix_title): New.
9982 (lra_setup_reg_renumber): Use it.
9983 (spill_for): Ditto. Check subreg reload pseudos too.
9984 (assign_by_spills): Consider subreg reload pseudos too.
9985 * lra-constraints.c (simplify_operand_subreg): Use
9986 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
9987 (curr_insn_transform): Recognize and do optional reloads.
9988 (undo_optional_reloads): New.
9989 (lra_undo_inheritance): Call undo_optional_reloads.
9991 2013-07-05 Thomas Quinot <quinot@adacore.com>
9993 * tree-complex.c (expand_complex_operations_1): Fix typo.
9995 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
9997 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
9998 (tune_params): New member 'const vec_costs'.
9999 * config/aarch64/aarch64.c (generic_vector_cost): New.
10000 (generic_tunings): New member 'generic_vector_cost'.
10001 (aarch64_builtin_vectorization_cost): New.
10002 (aarch64_add_stmt_cost): New.
10003 (TARGET_VECTORIZE_ADD_STMT_COST): New.
10004 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
10006 2013-07-03 Jakub Jelinek <jakub@redhat.com>
10009 * config/i386/predicates.md (vsib_address_operand): Disallow
10010 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
10012 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
10014 PR middle-end/55030
10015 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
10016 expand_builtin_setjmp_receiver.
10017 (expand_label): Adjust, call expand_builtin_setjmp_receiver
10018 with NULL for the label parameter.
10019 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
10020 the frame-pointer. Adjust comments.
10021 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
10022 only if LABEL is non-NULL.
10024 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
10026 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
10027 (ARM_ABI_AAPCS64): Ditto.
10029 (ARM_DEFAULT_ABI): Ditto.
10031 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
10033 * config/aarch64/aarch64-builtins.c
10034 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
10035 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
10037 * config/aarch64/aarch64-simd.md
10038 (aarch64_ld1<VALL:mode>): New.
10039 (aarch64_st1<VALL:mode>): Likewise.
10040 * config/aarch64/arm_neon.h
10041 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
10043 2013-07-02 Sriraman Tallam <tmsriram@google.com>
10045 * config/i386/i386.c (gate_insert_vzeroupper): Check if
10047 (ix86_option_override_internal):Turn on all -mavx target flags by
10048 default as they are dependent on AVX anyway.
10050 2013-07-02 Cary Coutant <ccoutant@google.com>
10052 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
10053 deterministic hash.
10054 (loc_checksum_ordered): Likewise.
10055 (hash_loc_operands): Remove inline keyword.
10057 2013-07-02 Jakub Jelinek <jakub@redhat.com>
10059 PR tree-optimization/57741
10060 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
10061 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
10062 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
10063 Allow REAL_CST step_exprs if flag_associative_math.
10064 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
10066 2013-07-02 Ian Bolton <ian.bolton@arm.com>
10068 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
10070 2013-07-02 Ian Bolton <ian.bolton@arm.com>
10072 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
10074 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10076 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
10078 (iorsi3_insn): Likewise.
10079 (arm_xorsi3): Likewise.
10081 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
10083 * arm.md (attribute "wtype"): Delete. Move attribute values from here
10085 (attribute "type"): ... here, and prefix with "wmmx_".
10086 (attribute "core_cycles"): Update for attribute changes.
10087 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
10088 (tbcstv4hi): Likewise.
10089 (tbcstv2si): Likewise.
10090 (iwmmxt_iordi3): Likewise.
10091 (iwmmxt_xordi3): Likewise.
10092 (iwmmxt_anddi3): Likewise.
10093 (iwmmxt_nanddi3): Likewise.
10094 (iwmmxt_arm_movdi): Likewise.
10095 (iwmmxt_movsi_insn): Likewise.
10096 (mov<mode>_internal): Likewise.
10097 (and<mode>3_iwmmxt): Likewise.
10098 (ior<mode>3_iwmmxt): Likewise.
10099 (xor<mode>3_iwmmxt): Likewise.
10100 (add<mode>3_iwmmxt): Likewise.
10101 (ssaddv8qi3): Likewise.
10102 (ssaddv4hi3): Likewise.
10103 (ssaddv2si3): Likewise.
10104 (usaddv8qi3): Likewise.
10105 (usaddv4hi3): Likewise.
10106 (usaddv2si3): Likewise.
10107 (sub<mode>3_iwmmxt): Likewise.
10108 (sssubv8qi3): Likewise.
10109 (sssubv4hi3): Likewise.
10110 (sssubv2si3): Likewise.
10111 (ussubv8qi3): Likewise.
10112 (ussubv4hi3): Likewise.
10113 (ussubv2si3): Likewise.
10114 (mulv4hi3_iwmmxt): Likewise.
10115 (smulv4hi3_highpart): Likewise.
10116 (umulv4hi3_highpart): Likewise.
10117 (iwmmxt_wmacs): Likewise.
10118 (iwmmxt_wmacsz): Likewise.
10119 (iwmmxt_wmacu): Likewise.
10120 (iwmmxt_wmacuz): Likewise.
10121 (iwmmxt_clrdi): Likewise.
10122 (iwmmxt_clrv8qi): Likewise.
10123 (iwmmxt_clr4hi): Likewise.
10124 (iwmmxt_clr2si): Likewise.
10125 (iwmmxt_uavgrndv8qi3): Likewise.
10126 (iwmmxt_uavgrndv4hi3): Likewise.
10127 (iwmmxt_uavgv8qi3): Likewise.
10128 (iwmmxt_uavgv4hi3): Likewise.
10129 (iwmmxt_tinsrb): Likewise.
10130 (iwmmxt_tinsrh): Likewise.
10131 (iwmmxt_tinsrw): Likewise.
10132 (iwmmxt_textrmub): Likewise.
10133 (iwmmxt_textrmsb): Likewise.
10134 (iwmmxt_textrmuh): Likewise.
10135 (iwmmxt_textrmsh): Likewise.
10136 (iwmmxt_textrmw): Likewise.
10137 (iwmxxt_wshufh): Likewise.
10138 (eqv8qi3): Likewise.
10139 (eqv4hi3): Likewise.
10140 (eqv2si3): Likewise.
10141 (gtuv8qi3): Likewise.
10142 (gtuv4hi3): Likewise.
10143 (gtuv2si3): Likewise.
10144 (gtv8qi3): Likewise.
10145 (gtv4hi3): Likewise.
10146 (gtv2si3): Likewise.
10147 (smax<mode>3_iwmmxt): Likewise.
10148 (umax<mode>3_iwmmxt): Likewise.
10149 (smin<mode>3_iwmmxt): Likewise.
10150 (umin<mode>3_iwmmxt): Likewise.
10151 (iwmmxt_wpackhss): Likewise.
10152 (iwmmxt_wpackwss): Likewise.
10153 (iwmmxt_wpackdss): Likewise.
10154 (iwmmxt_wpackhus): Likewise.
10155 (iwmmxt_wpackwus): Likewise.
10156 (iwmmxt_wpackdus): Likewise.
10157 (iwmmxt_wunpckihb): Likewise.
10158 (iwmmxt_wunpckihh): Likewise.
10159 (iwmmxt_wunpckihw): Likewise.
10160 (iwmmxt_wunpckilb): Likewise.
10161 (iwmmxt_wunpckilh): Likewise.
10162 (iwmmxt_wunpckilw): Likewise.
10163 (iwmmxt_wunpckehub): Likewise.
10164 (iwmmxt_wunpckehuh): Likewise.
10165 (iwmmxt_wunpckehuw): Likewise.
10166 (iwmmxt_wunpckehsb): Likewise.
10167 (iwmmxt_wunpckehsh): Likewise.
10168 (iwmmxt_wunpckehsw): Likewise.
10169 (iwmmxt_wunpckelub): Likewise.
10170 (iwmmxt_wunpckeluh): Likewise.
10171 (iwmmxt_wunpckeluw): Likewise.
10172 (iwmmxt_wunpckelsb): Likewise.
10173 (iwmmxt_wunpckelsh): Likewise.
10174 (iwmmxt_wunpckelsw): Likewise.
10175 (ror<mode>3): Likewise.
10176 (ashr<mode>3_iwmmxt): Likewise.
10177 (lshr<mode>3_iwmmxt): Likewise.
10178 (ashl<mode>3_iwmmxt): Likewise.
10179 (ror<mode>3_di): Likewise.
10180 (ashr<mode>3_di): Likewise.
10181 (lshr<mode>3_di): Likewise.
10182 (ashl<mode>3_di): Likewise.
10183 (iwmmxt_wmadds): Likewise.
10184 (iwmmxt_wmaddu): Likewise.
10185 (iwmmxt_tmia): Likewise.
10186 (iwmmxt_tmiaph): Likewise.
10187 (iwmmxt_tmiabb): Likewise.
10188 (iwmmxt_tmiatb): Likewise.
10189 (iwmmxt_tmiabt): Likewise.
10190 (iwmmxt_tmiatt): Likewise.
10191 (iwmmxt_tmovmskb): Likewise.
10192 (iwmmxt_tmovmskh): Likewise.
10193 (iwmmxt_tmovmskw): Likewise.
10194 (iwmmxt_waccb): Likewise.
10195 (iwmmxt_wacch): Likewise.
10196 (iwmmxt_waccw): Likewise.
10197 (iwmmxt_waligni): Likewise.
10198 (iwmmxt_walignr): Likewise.
10199 (iwmmxt_walignr0): Likewise.
10200 (iwmmxt_walignr1): Likewise.
10201 (iwmmxt_walignr2): Likewise.
10202 (iwmmxt_walignr3): Likewise.
10203 (iwmmxt_wsadb): Likewise.
10204 (iwmmxt_wsadh): Likewise.
10205 (iwmmxt_wsadbz): Likewise.
10206 (iwmmxt_wsadhz): Likewise.
10207 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
10208 (iwmmxt_wabsdiffb): Likewise.
10209 (iwmmxt_wabsdiffh): Likewise.
10210 (iwmmxt_wabsdiffw): Likewise.
10211 (iwmmxt_waddsubhx): Likewise
10212 (iwmmxt_wsubaddhx): Likewise.
10213 (addc<mode>3): Likewise.
10214 (iwmmxt_avg4): Likewise.
10215 (iwmmxt_avg4r): Likewise.
10216 (iwmmxt_wmaddsx): Likewise.
10217 (iwmmxt_wmaddux): Likewise.
10218 (iwmmxt_wmaddsn): Likewise.
10219 (iwmmxt_wmaddun): Likewise.
10220 (iwmmxt_wmulwsm): Likewise.
10221 (iwmmxt_wmulwum): Likewise.
10222 (iwmmxt_wmulsmr): Likewise.
10223 (iwmmxt_wmulumr): Likewise.
10224 (iwmmxt_wmulwsmr): Likewise.
10225 (iwmmxt_wmulwumr): Likewise.
10226 (iwmmxt_wmulwl): Likewise.
10227 (iwmmxt_wqmulm): Likewise.
10228 (iwmmxt_wqmulwm): Likewise.
10229 (iwmmxt_wqmulmr): Likewise.
10230 (iwmmxt_wqmulwmr): Likewise.
10231 (iwmmxt_waddbhusm): Likewise.
10232 (iwmmxt_waddbhusl): Likewise.
10233 (iwmmxt_wqmiabb): Likewise.
10234 (iwmmxt_wqmiabt): Likewise.
10235 (iwmmxt_wqmiatb): Likewise.
10236 (iwmmxt_wqmiatt): Likewise.
10237 (iwmmxt_wqmiabbn): Likewise.
10238 (iwmmxt_wqmiabtn): Likewise.
10239 (iwmmxt_wqmiatbn): Likewise.
10240 (iwmmxt_wqmiattn): Likewise.
10241 (iwmmxt_wmiabb): Likewise.
10242 (iwmmxt_wmiabt): Likewise.
10243 (iwmmxt_wmiatb): Likewise.
10244 (iwmmxt_wmiatt): Likewise.
10245 (iwmmxt_wmiabbn): Likewise.
10246 (iwmmxt_wmiabtn): Likewise.
10247 (iwmmxt_wmiatbn): Likewise.
10248 (iwmmxt_wmiattn): Likewise.
10249 (iwmmxt_wmiawbb): Likewise.
10250 (iwmmxt_wmiawbt): Likewise.
10251 (iwmmxt_wmiawtb): Likewise.
10252 (iwmmxt_wmiawtt): Likewise.
10253 (iwmmxt_wmiawbbn): Likewise.
10254 (iwmmxt_wmiawbtn): Likewise.
10255 (iwmmxt_wmiawtbn): Likewise.
10256 (iwmmxt_wmiawttn): Likewise.
10257 (iwmmxt_wmerge): Likewise.
10258 (iwmmxt_tandc<mode>3): Likewise.
10259 (iwmmxt_torc<mode>3): Likewise.
10260 (iwmmxt_torvsc<mode>3): Likewise.
10261 (iwmmxt_textrc<mode>3): Likewise.
10262 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
10263 (wmmxt_pack): Likewise.
10264 (wmmxt_mult_c1): Likewise.
10265 (wmmxt_mult_c2): Likewise.
10266 (wmmxt_alu_c1): Likewise.
10267 (wmmxt_alu_c2): Likewise.
10268 (wmmxt_alu_c3): Likewise.
10269 (wmmxt_transfer_c1): Likewise.
10270 (wmmxt_transfer_c2): Likewise.
10271 (wmmxt_transfer_c3): Likewise.
10272 (marvell_f_iwmmxt_wstr): Likewise.
10273 (marvell_f_iwmmxt_wldr): Likewise.
10275 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
10277 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
10279 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
10282 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
10283 * lra-constraints.c (need_for_split_p): Check call used hard regs
10284 living through calls.
10286 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
10287 call used regs for call insn.
10289 2013-06-28 Jakub Jelinek <jakub@redhat.com>
10292 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
10293 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
10295 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
10297 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
10299 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
10301 * lra-constraints.c (need_for_split_p): Check call used hard regs
10302 living through calls.
10304 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
10307 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
10308 to tie with any other modes. Eliminate Altivec vector mode tests,
10309 since these are a subset of ALTIVEC or VSX vector modes. Simplify
10310 code, to return 0 if testing MODE2 for a condition, if we've
10311 already tested MODE1 for the same condition.
10313 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
10315 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
10318 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
10320 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
10321 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
10323 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
10325 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
10327 (aarch64_symbolic_constant_p): Remove.
10328 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
10329 static. Fix line length and white space.
10330 (aarch64_symbolic_constant_p): Remove.
10331 * config/aarch64/predicates.md (aarch64_valid_symref):
10332 Use aarch64_classify_symbol_expression.
10334 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10336 * config/arm/constraints.md (Ts): New constraint.
10337 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
10339 (compare_scc): Use "Ts" constraint for operand 0.
10340 (ior_scc_scc): Likewise.
10341 (and_scc_scc): Likewise.
10342 (and_scc_scc_nodom): Likewise.
10343 (ior_scc_scc_cmp): Likewise for operand 7.
10344 (and_scc_scc_cmp): Likewise.
10345 * config/arm/thumb2.md (thumb2_movsi_insn):
10346 Add alternatives for 16-bit encodings.
10347 (thumb2_movhi_insn): Likewise.
10348 (thumb2_movsicc_insn): Likewise.
10349 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
10350 (thumb2_negscc): Use "Ts" constraint.
10351 Move mvn instruction outside cond_exec block.
10352 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
10353 for 16-bit encodings.
10355 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10357 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
10359 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
10360 (mulsi3subsi): Likewise.
10361 (mulsidi3adddi): Likewise.
10362 (mulsidi3_v6): Likewise.
10363 (umulsidi3_v6): Likewise.
10364 (umulsidi3adddi_v6): Likewise.
10365 (smulsi3_highpart_v6): Likewise.
10366 (umulsi3_highpart_v6): Likewise.
10367 (mulhisi3tb): Likewise.
10368 (mulhisi3bt): Likewise.
10369 (mulhisi3tt): Likewise.
10370 (maddhisi4): Likewise.
10371 (maddhisi4tb): Likewise.
10372 (maddhisi4tt): Likewise.
10373 (maddhidi4): Likewise.
10374 (maddhidi4tb): Likewise.
10375 (maddhidi4tt): Likewise.
10376 (zeroextractsi_compare0_scratch): Likewise.
10377 (insv_zero): Likewise.
10378 (insv_t2): Likewise.
10379 (anddi_notzesidi_di): Likewise.
10380 (anddi_notsesidi_di): Likewise.
10381 (andsi_notsi_si): Likewise.
10382 (iordi_zesidi_di): Likewise.
10383 (xordi_zesidi_di): Likewise.
10384 (andsi_iorsi3_notsi): Likewise.
10385 (smax_0): Likewise.
10386 (smax_m1): Likewise.
10387 (smin_0): Likewise.
10388 (not_shiftsi): Likewise.
10389 (unaligned_loadsi): Likewise.
10390 (unaligned_loadhis): Likewise.
10391 (unaligned_loadhiu): Likewise.
10392 (unaligned_storesi): Likewise.
10393 (unaligned_storehi): Likewise.
10394 (extv_reg): Likewise.
10395 (extzv_t2): Likewise.
10396 (divsi3): Likewise.
10397 (udivsi3): Likewise.
10398 (arm_zero_extendhisi2addsi): Likewise.
10399 (arm_zero_extendqisi2addsi): Likewise.
10400 (compareqi_eq0): Likewise.
10401 (arm_extendhisi2_v6): Likewise.
10402 (arm_extendqisi2addsi): Likewise.
10403 (arm_movt): Likewise.
10404 (thumb2_ldrd): Likewise.
10405 (thumb2_ldrd_base): Likewise.
10406 (thumb2_ldrd_base_neg): Likewise.
10407 (thumb2_strd): Likewise.
10408 (thumb2_strd_base): Likewise.
10409 (thumb2_strd_base_neg): Likewise.
10410 (arm_negsi2): Add alternative for 16-bit encoding.
10411 (arm_one_cmplsi2): Likewise.
10413 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10415 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
10416 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
10417 (movdfcc): Likewise.
10418 * config/arm/vfp.md (*thumb2_movsf_vfp):
10419 Disable predication for arm_restrict_it.
10420 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
10421 (*thumb2_movdfcc_vfp): Likewise.
10422 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
10423 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
10424 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
10425 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
10426 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
10427 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
10428 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
10429 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
10430 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
10431 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
10432 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
10433 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
10434 Disable predication for arm_restrict_it.
10436 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
10438 * config/i386/bmiintrin.h (_bextr_u32): New.
10439 (_bextr_u64): Ditto.
10441 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
10443 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
10444 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
10445 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
10446 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
10447 * config/mips/n32-elf.h: ...this new file.
10449 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
10452 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
10453 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
10455 2013-06-27 Catherine Moore <clm@codesourcery.com>
10457 * config/mips/mips-tables.opt: Regenerate.
10458 * config/mips/mips-cpus.def: Add m14ke and m14kec.
10459 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
10460 * doc/invoke.texi: Add -m14kc.
10462 2013-06-27 Jakub Jelinek <jakub@redhat.com>
10465 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
10466 constraints of operand 1 and 2.
10469 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
10470 to match RTL canonicalization. Swap predicates and
10471 constraints of operand 1 and 2.
10473 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
10475 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
10476 Process OP_INOUT regs for splitting too.
10478 2013-06-27 Jakub Jelinek <jakub@redhat.com>
10480 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
10481 decl before the loop, initialize to NULL.
10482 (vectorizable_load): Initialize ptr_incr to NULL.
10484 2013-06-27 Martin Jambor <mjambor@suse.cz>
10487 * ipa-ref.h (ipa_maybe_record_reference): Declare.
10488 * ipa-ref.c (ipa_maybe_record_reference): New function.
10489 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
10490 * ipa-cp.c (create_specialized_node): Record potential references from
10492 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
10494 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
10496 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
10497 parameter 'mode' of type 'enum machine_mode mode'; change to pass
10498 'mode' to force_reg.
10499 (aarch64_add_offset): Update calls to aarch64_force_temporary.
10500 (aarch64_expand_mov_immediate): Likewise.
10502 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
10504 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
10505 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
10507 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10509 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
10510 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
10511 (struct machine_function): Add tbegin_p.
10512 (s390_canonicalize_comparison): Fold CC mode compares to
10513 conditional jump if possible.
10514 (s390_emit_jump): Return the emitted jump.
10515 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
10516 Handle CCRAWmode compares.
10517 (s390_option_override): Default to -mhtm if available.
10518 (s390_reg_clobbered_rtx): Handle floating point regs as well.
10519 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
10520 FPRs instead of df_regs_ever_live_p.
10521 (s390_optimize_nonescaping_tx): New function.
10522 (s390_init_frame_layout): Extend clobbered_regs array to cover
10524 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
10525 (s390_expand_tbegin): New function.
10526 (enum s390_builtin): New enum definition.
10527 (code_for_builtin): New array definition.
10528 (s390_init_builtins): New function.
10529 (s390_expand_builtin): New function.
10530 (TARGET_INIT_BUILTINS): Define.
10531 (TARGET_EXPAND_BUILTIN): Define.
10532 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
10533 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
10534 (s390_alc_comparison): Likewise.
10535 * config/s390/s390-modes.def: Add CCRAWmode.
10536 * config/s390/s390.h (processor_flags): Add PF_TX.
10537 (TARGET_CPU_HTM): Define macro.
10538 (TARGET_HTM): Define macro.
10539 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
10540 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
10541 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
10542 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
10543 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
10544 (TBEGIN_MASK, TBEGINC_MASK): New constants.
10545 ("*cc_to_int"): Move up.
10546 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
10547 constants other than 0.
10548 ("*ccraw_to_int"): New insn and splitter definition.
10549 ("tbegin", "tbegin_nofloat", "tbegin_retry")
10550 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
10551 ("tx_assist"): New expander.
10552 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
10553 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
10554 * config/s390/s390.opt: Add -mhtm option.
10555 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
10556 * config/s390/htmxlintrin.h: New file.
10557 * config/s390/htmintrin.h: New file.
10558 * config/s390/s390intrin.h: New file.
10559 * doc/extend.texi: Document htm builtins.
10560 * config.gcc: Add the new header files to extra_headers.
10562 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
10564 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
10565 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
10567 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
10568 Pat Haugen <pthaugen@us.ibm.com>
10569 Peter Bergner <bergner@vnet.ibm.com>
10571 * config/rs6000/power8.md: New.
10572 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
10573 setting for power8 entry.
10574 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
10575 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
10576 test for Power4/Power5 only.
10577 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
10579 (force_new_group): Adjust comment.
10580 * config/rs6000/rs6000.md: Include power8.md.
10582 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
10584 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
10585 * config/arm/arm-protos.h (arm_max_conditional_execute): New
10587 (tune_params): Update comment.
10588 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
10589 (arm_max_conditional_execute): New function.
10590 (thumb2_final_prescan_insn): Use max_insn_skipped and
10591 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
10593 2013-06-25 Jakub Jelinek <jakub@redhat.com>
10595 PR tree-optimization/57705
10596 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
10597 SSA_NAME step, provided that it is not defined inside the loop.
10598 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
10599 (get_initial_def_for_induction): Handle SSA_NAME IV step.
10601 2013-06-25 Martin Jambor <mjambor@suse.cz>
10603 PR middle-end/57670
10604 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
10605 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
10607 (ipa_note_param_call): Initialize member_ptr flag.
10608 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
10609 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
10610 (ipa_write_indirect_edge_info): Stream member_ptr flag.
10611 (ipa_read_indirect_edge_info): Likewise.
10613 2013-06-25 Richard Biener <rguenther@suse.de>
10615 PR middle-end/56977
10616 * passes.c (init_optimization_passes): Move pass_fold_builtins
10617 and pass_dce earlier with -Og.
10619 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
10621 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
10622 <BIT_FIELD_REF>: Remove trailing TAB.
10623 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
10626 2013-06-24 Martin Jambor <mjambor@suse.cz>
10628 PR tree-optimization/57358
10629 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
10630 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
10631 (ipa_analyze_params_uses): Generate pessimistic info when true.
10633 2013-06-24 Martin Jambor <mjambor@suse.cz>
10635 PR tree-optimization/57539
10636 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
10637 global.inlined_to of the new node to it. All callers changed.
10638 * ipa-inline-transform.c (clone_inlined_nodes): New variable
10639 inlining_into, pass it to cgraph_clone_node.
10640 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
10641 ipa_free_edge_args_substructures.
10642 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
10643 rdesc linked list. Do not assert rdesc edges have inlined caller.
10644 Assert we have found an rdesc in the rdesc list.
10646 2013-06-24 Richard Biener <rguenther@suse.de>
10648 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
10649 (pointer_set_lookup): Declare.
10650 (class pointer_map): New template class implementing a
10651 generic pointer to T map.
10652 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
10653 pointer_map<T>::contains, pointer_map<T>::insert,
10654 pointer_map<T>::traverse): New functions.
10655 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
10656 (pointer_set_lookup): New function.
10657 (pointer_set_contains): Use pointer_set_lookup.
10658 (pointer_set_insert): Likewise.
10659 (insert_aux): Remove.
10660 (struct pointer_map_t): Embed a pointer_set_t.
10661 (pointer_map_create): Adjust.
10662 (pointer_map_destroy): Likewise.
10663 (pointer_map_contains): Likewise.
10664 (pointer_map_insert): Likewise.
10665 (pointer_map_traverse): Likewise.
10666 * tree-streamer.h (struct streamer_tree_cache_d): Use a
10667 pointer_map<unsigned> instead of a pointer_map_t.
10668 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
10669 (streamer_tree_cache_lookup): Likewise.
10670 (streamer_tree_cache_create): Likewise.
10671 (streamer_tree_cache_delete): Likewise.
10672 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
10673 pointer_map<unsigned> instead of a pointer_map_t.
10674 (lto_init_tree_ref_encoder): Adjust.
10675 (lto_destroy_tree_ref_encoder): Likewise.
10676 * lto-section-out.c (lto_output_decl_index): Likewise.
10677 (lto_record_function_out_decl_state): Likewise.
10678 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
10680 2013-06-24 Richard Biener <rguenther@suse.de>
10682 PR tree-optimization/57488
10683 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
10685 2013-06-24 Alan Modra <amodra@gmail.com>
10687 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
10688 (gen_easy_altivec_constant): Likewise.
10689 * config/rs6000/predicates.md (easy_vector_constant_add_self,
10690 easy_vector_constant_msb): Likewise.
10692 2013-06-23 Jakub Jelinek <jakub@redhat.com>
10695 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
10696 add missing return true.
10698 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
10701 * config/sh/predicates.md (general_extend_operand): Invoke
10702 general_movsrc_operand for memory operands.
10703 (general_movsrc_operand): Allow reg+reg addressing, do not use
10704 general_operand for memory operands.
10706 2013-06-23 Sriraman Tallam <tmsriram@google.com>
10708 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
10709 when current target options does not apply.
10710 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
10711 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
10712 * config/i386/bmiintrin.h: Pass appropriate target
10713 attributes to header.
10714 * config/i386/mmintrin.h: Ditto.
10715 * config/i386/nmmintrin.h: Ditto.
10716 * config/i386/avx2intrin.h: Ditto.
10717 * config/i386/fxsrintrin.h: Ditto.
10718 * config/i386/tbmintrin.h: Ditto.
10719 * config/i386/xsaveintrin.h: Ditto.
10720 * config/i386/f16cintrin.h: Ditto.
10721 * config/i386/xtestintrin.h: Ditto.
10722 * config/i386/xsaveoptintrin.h: Ditto.
10723 * config/i386/bmi2intrin.h: Ditto.
10724 * config/i386/lzcntintrin.h: Ditto.
10725 * config/i386/smmintrin.h: Ditto.
10726 * config/i386/wmmintrin.h: Ditto.
10727 * config/i386/x86intrin.h: Remove all header include guards.
10728 * config/i386/prfchwintrin.h: Ditto.
10729 * config/i386/pmmintrin.h: Ditto.
10730 * config/i386/tmmintrin.h: Ditto.
10731 * config/i386/xmmintrin.h: Ditto.
10732 * config/i386/popcntintrin.h: Ditto.
10733 * config/i386/rdseedintrin.h: Ditto.
10734 * config/i386/ammintrin.h: Ditto.
10735 * config/i386/emmintrin.h: Ditto.
10736 * config/i386/immintrin.h: Remove all header include guards.
10737 * config/i386/fma4intrin.h: Ditto.
10738 * config/i386/lwpintrin.h: Ditto.
10739 * config/i386/xopintrin.h: Ditto.
10740 * config/i386/ia32intrin.h: Ditto.
10741 * config/i386/avxintrin.h: Ditto.
10742 * config/i386/rtmintrin.h: Ditto.
10743 * config/i386/fmaintrin.h: Ditto.
10744 * config/i386/mm3dnow.h: Ditto.
10746 2013-06-22 Sriraman Tallam <tmsriram@google.com>
10748 * common/config/i386/i386-common.c: Handle LZCNT.
10750 2013-06-22 Andi Kleen <ak@linux.intel.com>
10752 * doc/extend.texi: Use __atomic_store_n instead of
10753 __atomic_store in HLE example.
10755 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
10757 * config/sh/sh.c: Remove <cstdlib> workaround.
10759 2013-06-21 Andi Kleen <ak@linux.intel.com>
10761 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
10763 2013-06-21 Andi Kleen <ak@linux.intel.com>
10765 * doc/extend.texi: Document that __atomic_clear and
10766 __atomic_test_and_set should only be used with bool.
10768 2013-06-20 Jan Hubicka <jh@suse.cz>
10770 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
10771 types_same_for_odr.
10772 * tree.c (decls_same_for_odr): New function.
10773 (same_for_edr): New function.
10774 (types_same_for_odr): New function.
10775 (get_binfo_at_offset): Use it.
10776 * tree.h (types_same_for_odr): Declare.
10778 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
10779 Jason Merrill <jason@redhat.com>
10781 * system.h: Include <cstdlib> as well as <stdlib.h>.
10783 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
10786 * config/i386/i386.c (construct_container): Report error if
10787 long double is used with disabled x87 float returns.
10789 2013-06-20 Jan Hubicka <jh@suse.cz>
10791 * lto-cgraph.c (input_symtab): Do not set cgraph state.
10793 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
10795 PR rtl-optimization/57425
10796 PR rtl-optimization/57569
10797 * alias.c (write_dependence_p): Remove parameters mem_mode and
10798 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
10799 Changed all callers.
10800 (canon_anti_dependence): Get comments and semantics in sync.
10801 Add parameter mem_canonicalized. Changed all callers.
10802 * rtl.h (canon_anti_dependence): Update prototype.
10804 2013-06-20 Richard Biener <rguenther@suse.de>
10806 * data-streamer-in.c (streamer_read_uhwi): Optimize single
10807 byte case, inline streamer_read_uchar and defer section
10810 2013-06-20 Richard Biener <rguenther@suse.de>
10812 PR tree-optimization/57584
10813 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
10814 SSA names into the expanded expression that take part in
10815 abnormal coalescing.
10817 2013-06-19 Sharad Singhai <singhai@google.com>
10819 * gcov.c (print_usage): Handle new option.
10820 (process_args): Ditto.
10821 (get_gcov_intermediate_filename): New function.
10822 (output_intermediate_file): New function.
10823 (output_gcov_file): New function
10824 (generate_results): Handle new option.
10825 (release_function): Relase demangled name.
10826 (read_graph_file): Handle demangled name.
10827 (output_lines): Ditto.
10828 * doc/gcov.texi: Document gcov intermediate format.
10830 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
10833 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
10834 (lra_emit_add): Use the functions. Add comment about Y as an
10837 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
10840 * collect2.c (collect_atexit): New.
10841 (collect_exit): Delete.
10842 (main): Register collect_atexit with atexit.
10843 (collect_wait): Change collect_exit to exit.
10845 * collect2.h (collect_exit): Delete.
10846 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
10848 2013-06-19 Wei Mi <wmi@google.com>
10850 PR rtl-optimization/57518
10851 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
10852 if regno is used in paradoxical subreg.
10853 (update_equiv_regs): Check pdx_subregs[regno] before
10854 set a reg to be equivalent with a mem.
10856 2013-06-19 Matthias Klose <doko@ubuntu.com>
10859 * file-find.h (find_a_file): Add a mode parameter.
10860 * file-find.c (find_a_file): Likewise.
10861 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
10862 with X_OK for the executables.
10863 * collect2.c (main): Call find_a_file with X_OK.
10865 2013-06-19 Steve Ellcey <sellcey@mips.com>
10868 * config/mips/mips.md (casesi_internal_mips16_<mode>):
10869 Use NEXT_INSN instead of next_real_insn.
10871 2013-06-19 Jan Hubicka <jh@suse.cz>
10873 * cgraph.h (const_value_known_p): Replace by ...
10874 (ctor_for_folding): .. this one.
10875 * cgraphunit.c (process_function_and_variable_attributes): Use it.
10876 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
10877 * expr.c (expand_expr_real_1): Likewise.
10878 (string_constant): Likewise.
10879 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
10880 * ipa.c (process_references): Likewise.
10881 (symtab_remove_unreachable_nodes): Likewise.
10882 * ipa-inline-analysis.c (param_change_prob): Likewise.
10883 * gimple-fold.c (canonicalize_constructor_val): Likewise.
10884 (get_base_constructor): Likwise.
10885 * varpool.c (varpool_remove_node): Likewise.
10886 (varpool_remove_initializer): LIkewise.
10887 (dump_varpool_node): LIkwise.
10888 (const_value_known_p): Rewrite to ...
10889 (ctor_for_folding): ... this one.
10891 2013-06-19 Jakub Jelinek <jakub@redhat.com>
10894 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
10895 PERSONALITY in $PATH derived prefixes.
10897 2013-06-19 Jeff Law <law@redhat.com>
10899 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
10902 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
10903 (simplify_bitwise_binary): Use it to simpify certain binary ops on
10906 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
10908 * config/arm/vfp.md: Move VFP instruction classification documentation
10910 * config/arm/arm.md: ... here. Update instruction classification
10913 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
10915 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
10916 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
10917 pattern. Use more efficient sequences on ARMv5 and Thumb2.
10919 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
10922 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10923 with NEXT_INSN. Use tablejump_p to check for jump table data
10926 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
10929 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
10930 that now in C++ the value is correct per the C++ standards.
10932 2013-06-19 Richard Biener <rguenther@suse.de>
10934 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
10935 for global context.
10937 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10940 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10943 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10944 with next_active_insn.
10946 2013-06-18 Sriraman Tallam <tmsriram@google.com>
10948 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
10949 functions are inlined during failures to flag an error.
10950 * tree-inline.c (expand_call_inline): Allow the error to be flagged
10951 in early inline pass.
10953 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
10955 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
10958 2013-06-18 Julian Brown <julian@codesourcery.com>
10960 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
10961 Permit virtual register pre-reload if !strict.
10962 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
10964 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
10966 * config/arm/neon.md (movmisalign<mode>): Use
10967 neon_perm_struct_or_reg_operand instead of
10968 neon_struct_or_register_operand.
10969 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
10970 neon_permissive_struct_operand instead of neon_struct_operand.
10971 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
10972 neon_vector_mem_operand.
10973 * config/arm/predicates.md (neon_struct_operand): Adjust call to
10974 neon_vector_mem_operand.
10975 (neon_permissive_struct_operand): New.
10976 (neon_struct_or_register_operand): Rename to...
10977 (neon_perm_struct_or_reg_operand): This. Adjust call to
10978 neon_vector_mem_operand.
10980 2013-06-18 Richard Biener <rguenther@suse.de>
10982 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
10983 * lto-streamer.h: Include pointer-set.h.
10984 (struct lto_decl_slot): Remove.
10985 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
10986 Remove next_index entry.
10987 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10988 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10989 (lto_init_tree_ref_encoder): Adjust.
10990 (lto_destroy_tree_ref_encoder): Likewise.
10991 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10992 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10993 (lto_output_decl_index): Adjust.
10994 (lto_new_out_decl_state): Likewise.
10995 (lto_record_function_out_decl_state): Likewise.
10996 * lto-streamer-out.c (copy_function): Likewise.
10998 2013-06-18 Richard Biener <rguenther@suse.de>
11000 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
11001 * cgraphunit.c: Include cfgloop.h.
11002 (init_lowered_empty_function): Initialize the loop tree.
11003 (assemble_thunk): Insert new BBs into loops.
11005 2013-06-18 Richard Biener <rguenther@suse.de>
11007 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
11008 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
11009 the map from cache entry to cache index optional.
11010 (streamer_tree_cache_replace_tree): Adjust accordingly.
11011 (streamer_tree_cache_append): Likewise.
11012 (streamer_tree_cache_delete): Likewise.
11013 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
11014 streamer cache map from cache entry to cache index.
11015 * lto-streamer-out.c (create_output_block): Adjust.
11017 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
11019 * config/arm/arm.md (attribute "insn"): Move multiplication and
11020 division attributes to...
11021 (attribute "type"): ... here. Remove mult.
11022 (attribute "mul32"): New attribute.
11023 (attribute "mul64"): Add umaal.
11024 (*arm_mulsi3): Update attributes.
11025 (*arm_mulsi3_v6): Likewise.
11026 (*thumb_mulsi3): Likewise.
11027 (*thumb_mulsi3_v6): Likewise.
11028 (*mulsi3_compare0): Likewise.
11029 (*mulsi3_compare0_v6): Likewise.
11030 (*mulsi_compare0_scratch): Likewise.
11031 (*mulsi_compare0_scratch_v6): Likewise.
11032 (*mulsi3addsi): Likewise.
11033 (*mulsi3addsi_v6): Likewise.
11034 (*mulsi3addsi_compare0): Likewise.
11035 (*mulsi3addsi_compare0_v6): Likewise.
11036 (*mulsi3addsi_compare0_scratch): Likewise.
11037 (*mulsi3addsi_compare0_scratch_v6): Likewise.
11038 (*mulsi3subsi): Likewise.
11039 (*mulsidi3adddi): Likewise.
11040 (*mulsi3addsi_v6): Likewise.
11041 (*mulsidi3adddi_v6): Likewise.
11042 (*mulsidi3_nov6): Likewise.
11043 (*mulsidi3_v6): Likewise.
11044 (*umulsidi3_nov6): Likewise.
11045 (*umulsidi3_v6): Likewise.
11046 (*umulsidi3adddi): Likewise.
11047 (*umulsidi3adddi_v6): Likewise.
11048 (*smulsi3_highpart_nov6): Likewise.
11049 (*smulsi3_highpart_v6): Likewise.
11050 (*umulsi3_highpart_nov6): Likewise.
11051 (*umulsi3_highpart_v6): Likewise.
11052 (mulhisi3): Likewise.
11053 (*mulhisi3tb): Likewise.
11054 (*mulhisi3bt): Likewise.
11055 (*mulhisi3tt): Likewise.
11056 (maddhisi4): Likewise.
11057 (*maddhisi4tb): Likewise.
11058 (*maddhisi4tt): Likewise.
11059 (maddhidi4): Likewise.
11060 (*maddhidi4tb): Likewise.
11061 (*maddhidi4tt): Likewise.
11062 (divsi3): Likewise.
11063 (udivsi3): Likewise.
11064 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
11065 (thumb2_mulsi_short_compare0): Likewise.
11066 (thumb2_mulsi_short_compare0_scratch): Likewise.
11067 * config/arm/arm1020e.md (1020mult1): Update attribute change.
11068 (1020mult2): Likewise.
11069 (1020mult3): Likewise.
11070 (1020mult4): Likewise.
11071 (1020mult5): Likewise.
11072 (1020mult6): Likewise.
11073 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
11075 (cortex_a15_mult64): Likewise.
11076 (cortex_a15_sdiv): Likewise.
11077 (cortex_a15_udiv): Likewise.
11078 * config/arm/arm1026ejs.md (mult1): Update attribute change.
11084 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
11085 (pj4_ir_div): Likewise.
11086 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
11087 (11_mult2): Likewise.
11088 (11_mult3): Likewise.
11089 (11_mult4): Likewise.
11090 (11_mult5): Likewise.
11091 (11_mult6): Likewise.
11092 (11_mult7): Likewise.
11093 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
11094 (cortex_a8_mla): Likewise.
11095 (cortex_a8_mull): Likewise.
11096 (cortex_a8_smulwy): Likewise.
11097 (cortex_a8_smlald): Likewise.
11098 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
11099 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
11100 (cortex_r4_mul_3): Likewise.
11101 (cortex_r4_mla_4): Likewise.
11102 (cortex_r4_mla_3): Likewise.
11103 (cortex_r4_smlald): Likewise.
11104 (cortex_r4_mull): Likewise.
11105 (cortex_r4_sdiv): Likewise.
11106 (cortex_r4_udiv): Likewise.
11107 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
11108 (cortex_a7_idiv): Likewise.
11109 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
11110 (9_mult2): Likewise.
11111 (9_mult3): Likewise.
11112 (9_mult4): Likewise.
11113 (9_mult5): Likewise.
11114 (9_mult6): Likewise.
11115 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
11116 (cortex_a53_sdiv): Likewise.
11117 (cortex_a53_udiv): Likewise.
11118 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
11119 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
11120 (mp626_mult2): Likewise.
11121 (mp626_mult3): Likewise.
11122 (mp626_mult4): Likewise.
11123 * config/arm/fa526.md (526_mult1): Update attribute change.
11124 (526_mult2): Likewise.
11125 * config/arm/arm-generic.md (mult): Update attribute change.
11126 (mult_ldsched_strongarm): Likewise.
11127 (mult_ldsched): Likewise.
11128 (multi_cycle): Likewise.
11129 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
11130 * config/arm/fa606te.md (606te_mult1): Update attribute change.
11131 (606te_mult2): Likewise.
11132 (606te_mult3): Likewise.
11133 (606te_mult4): Likewise.
11134 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
11135 (cortex_a9_mac16): Likewise.
11136 (cortex_a9_multiply): Likewise.
11137 (cortex_a9_mac): Likewise.
11138 (cortex_a9_multiply_long): Likewise.
11139 * config/arm/fa626te.md (626te_mult1): Update attribute change.
11140 (626te_mult2): Likewise.
11141 (626te_mult3): Likewise.
11142 (626te_mult4): Likewise.
11144 2013-06-18 Richard Biener <rguenther@suse.de>
11147 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
11149 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11152 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
11153 with next_active_insn.
11155 2013-06-18 Alan Modra <amodra@gmail.com>
11157 * config/rs6000/rs6000.h (enum data_align): New.
11158 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
11159 (DATA_ABI_ALIGNMENT): Define.
11160 (CONSTANT_ALIGNMENT): Correct comment.
11161 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
11162 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
11164 2013-06-17 David Malcolm <dmalcolm@redhat.com>
11166 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
11167 ATTRIBUTE_UNUSED marking.
11169 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
11171 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
11172 alternative and update.
11173 (aarch64_dup_lanedi): Delete.
11174 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
11175 * config/aarch64/aarch64-simd-builtins.def: Update.
11177 2013-06-17 Richard Biener <rguenther@suse.de>
11179 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
11180 (lto_input_scc): Declare.
11181 (lto_input_tree_1): Likewise.
11182 (struct lto_stats_d): Add num_tree_bodies_output and
11183 num_pickle_refs_output.
11184 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
11185 (lto_read_tree_1): Split out from ...
11186 (lto_read_tree): ... this.
11187 (lto_input_scc): New function.
11188 (lto_input_tree_1): Split out from ...
11189 (lto_input_tree): ... this. Handle LTO_tree_scc.
11190 (lto_data_in_create): Create the streamer cache without hashes.
11191 * lto-streamer-out.c (create_output_block): Create the streamer
11192 cache with hashes when not doing WPA.
11193 (lto_write_tree_1): Split out from ...
11194 (lto_write_tree): ... this.
11195 (get_symbol_initial_value): New function.
11196 (lto_output_tree_1): Split out from ...
11197 (lto_output_tree): ... this. Write trees as series of SCCs
11198 using a DFS walk via DFS_write_tree.
11199 (struct sccs, struct scc_entry): New types.
11200 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
11201 (DFS_write_tree_body): New function.
11202 (DFS_write_tree): Likewise.
11203 (hash_tree): Likewise.
11204 (scc_entry_compare): Likewise.
11205 (hash_scc): Likewise.
11206 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
11207 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
11208 TREE_CHAIN as regular reference.
11209 (streamer_read_integer_cst): Remove.
11210 (streamer_get_pickled_tree): Adjust.
11211 * tree-streamer-out.c (streamer_write_chain): Disable streaming
11212 of DECL_EXTERNALs in BLOCK_VARS for now.
11213 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
11215 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
11216 Add hash value argument and record that if hashes are recorded
11218 (streamer_tree_cache_insert_1): Adjust.
11219 (streamer_tree_cache_insert): Likewise.
11220 (streamer_tree_cache_insert_at): Rename to ...
11221 (streamer_tree_cache_replace_tree): ... this and adjust.
11222 (streamer_tree_cache_append): Adjust.
11223 (record_common_node): Likewise.
11224 (streamer_tree_cache_create): Add argument whether to
11225 record hash values together with trees.
11226 (streamer_tree_cache_delete): Adjust.
11227 * tree-streamer.h (struct streamer_tree_cache_d): Add
11229 (streamer_read_integer_cst): Remove.
11230 (streamer_tree_cache_insert): Adjust.
11231 (streamer_tree_cache_append): Likewise.
11232 (streamer_tree_cache_insert_at): Rename to ...
11233 (streamer_tree_cache_replace_tree): ... this and adjust.
11234 (streamer_tree_cache_create): Add argument whether to record hashes.
11235 (streamer_tree_cache_get): Rename to ...
11236 (streamer_tree_cache_get_tree): ... this.
11237 (streamer_tree_cache_get_hash): New function.
11238 * tree.c (cache_integer_cst): New function.
11239 * tree.h (cache_integer_cst): Declare.
11240 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
11241 * lto-symtab.c (lto_varpool_replace_node): Only release
11242 DECL_INITIAL of non-prevailing decls.
11243 * varpool.c (varpool_remove_initializer): Do not release
11244 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
11246 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
11248 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
11249 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
11250 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
11251 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
11252 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
11253 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
11254 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
11255 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
11256 instead of TARGET_64BIT.
11257 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
11258 Require ISA_HAS_<D>DIV.
11260 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
11262 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
11263 (mips*-*-linux*): Move default with_llsc setting to where other
11265 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
11267 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
11268 Likewise. Remove default with_tune setting. Move default float
11269 setting to its own block. Handle with_llsc in the same block as above.
11271 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
11273 PR rtl-optimization/57425
11274 PR rtl-optimization/57569
11275 * alias.c (write_dependence_p): Add new parameters mem_mode,
11276 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
11277 Changed all callers.
11278 (canon_anti_dependence): New function.
11279 * cse.c (check_dependence): Use canon_anti_dependence.
11280 * cselib.c (cselib_invalidate_mem): Likewise.
11281 * rtl.h (canon_anti_dependence): Declare.
11283 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
11285 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
11286 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
11287 ".set mips3" for 64-bit targets.
11289 2013-06-15 Dehao Chen <dehao@google.com>
11291 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
11292 * gimple-low.c (gimple_check_call_matching_types): Likewise.
11293 (gimple_check_call_args): Likewise.
11294 * value-prof.c (check_ic_target): Likewise.
11295 * ipa-inline.c (early_inliner): Likewise.
11296 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
11297 * cgraph.c (cgraph_create_edge_1): Likewise.
11298 (cgraph_make_edge_direct): Likewise.
11300 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
11303 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
11304 rs6000_output_move_128bit to handle emitting quad memory
11305 operations. Set attribute length to 8 bytes.
11307 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
11309 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
11311 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
11312 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
11313 (aarch64_<su>mlsl<mode>): Likewise.
11315 2013-06-14 Mike Stump <mikestump@comcast.net>
11317 * Makefile.in (TARGET_H): Add insn-codes.h.
11319 2013-06-14 Alan Modra <amodra@gmail.com>
11321 PR middle-end/57134
11322 PR middle-end/57586
11323 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
11324 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
11325 bitfield expansion when EXPAND_MEMORY.
11326 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
11328 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
11330 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
11331 test for clearing quad memory on 32-bit later.
11333 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
11335 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
11336 (fold_negate_expr): Likewise.
11337 (fold_real_zero_addition_p): Handle vectors.
11338 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
11340 2013-06-14 Alan Modra <amodra@gmail.com>
11342 * varasm.c (force_const_mem): Revert 2013-06-07 change.
11344 2013-06-13 Jan Hubicka <jh@suse.cz>
11346 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
11347 Local comdats are not externally visible.
11348 * symtab.c (dump_symtab_base): Dump externally visible.
11349 (verify_symtab_base): Verify back links in the symtab hash.
11351 2013-06-13 Bin Cheng <bin.cheng@arm.com>
11353 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
11354 CONVERT_EXPR as equal nodes.
11356 2013-06-13 Bin Cheng <bin.cheng@arm.com>
11358 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
11360 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
11362 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
11363 Generalize to complex and vector.
11364 * tree.c (build_all_ones_cst): New function.
11365 * tree.h (build_all_ones_cst): Declare it.
11367 2013-06-13 Alan Modra <amodra@gmail.com>
11369 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
11370 * config/rs6000/rs6000.md (signbittf2): New insn.
11371 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
11372 (abstf2_internal, cmptf_internal2): Likewise.
11373 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
11375 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
11376 Pat Haugen <pthaugen@us.ibm.com>
11377 Peter Bergner <bergner@vnet.ibm.com>
11379 * config/rs6000/rs6000.c (emit_load_locked): Add support for
11380 power8 byte, half-word, and quad-word atomic instructions.
11381 (emit_store_conditional): Likewise.
11382 (rs6000_expand_atomic_compare_and_swap): Likewise.
11383 (rs6000_expand_atomic_op): Likewise.
11385 * config/rs6000/sync.md (larx): Add new modes for power8.
11387 (AINT): New mode iterator to include TImode as well as normal
11388 integer modes on power8.
11389 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
11390 that VSX registers are not considered. Use AINT mode iterator
11391 instead of INT1 to allow inclusion of quad word atomic operations
11393 (load_locked<mode>): Likewise.
11394 (store_conditional<mode>): Likewise.
11395 (atomic_compare_and_swap<mode>): Likewise.
11396 (atomic_exchange<mode>): Likewise.
11397 (atomic_nand<mode>): Likewise.
11398 (atomic_fetch_<fetchop_name><mode>): Likewise.
11399 (atomic_nand_fetch<mode>): Likewise.
11400 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
11402 (ATOMIC): On power8, add QImode, HImode modes.
11403 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
11404 modes that promote to SImode.
11405 (load_lockedti): Convert TImode arguments to PTImode, so that we
11406 get a guaranteed even/odd register pair.
11407 (load_lockedpti): Likewise.
11408 (store_conditionalti): Likewise.
11409 (store_conditionalpti): Likewise.
11411 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
11412 atomic load/store instructions.
11415 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
11417 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
11419 (insn_count): New attribute, with most cases extracted from...
11420 (length): ...here. Redefine most cases in terms of insn_count.
11421 (single_insn): Delete.
11422 (can_delay): Use insn_count to check for single instructions.
11423 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
11424 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
11425 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
11426 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
11427 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
11428 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
11429 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
11430 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
11431 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
11432 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
11433 rather than "length".
11434 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
11435 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
11436 Use "insn_count" rather than "length".
11437 * config/mips/mips-dsp.md
11438 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
11439 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
11442 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
11444 PR tree-optimization/57361
11445 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
11447 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
11449 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
11451 (aarch64_simd_combine<mode>): New instruction expansion.
11452 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
11453 function prototype.
11454 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
11455 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
11457 2013-06-12 Jan Hubicka <jh@suse.cz>
11459 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
11460 decl has when in streaming stage.
11461 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
11462 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
11464 2013-06-12 Roland Stigge <stigge@antcom.de>
11467 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
11469 2013-06-12 Jakub Jelinek <jakub@redhat.com>
11471 PR tree-optimization/57537
11472 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
11473 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
11475 2013-06-12 Richard Biener <rguenther@suse.de>
11477 * data-streamer.h (streamer_write_char_stream): CSE
11478 obs->current_pointer.
11479 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
11480 streamer_write_char_stream manually and optimize the resulting loop.
11481 (streamer_write_hwi_stream): Likewise.
11483 2013-06-12 Jan Hubicka <jh@suse.cz>
11485 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
11486 * cgraph.h (varpool_create_empty_node): Declare.
11487 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
11489 * symtab.c (symtab_unregister_node): Be lax about missin entries
11491 (symtab_get_node): Update comment.
11492 * varpool.c (varpool_create_empty_node): Break out from ...
11493 (varpool_node_for_decl): ... here.
11494 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
11496 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
11498 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
11499 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
11500 part. Use straight-line flow at the end.
11501 <COMPONENT_REF>: Remove superfluous else.
11502 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
11504 2013-06-12 Jakub Jelinek <jakub@redhat.com>
11507 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
11508 target hook even for !TREE_PUBLIC decls. If no resolution info
11509 is available, return false for common and external decls.
11511 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11513 * config/rl78/constraints.md (U): New constraint.
11514 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
11517 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
11520 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
11521 to allow returning address to AT_PLATFORM name.
11523 2013-06-11 Jan Hubicka <jh@suse.cz>
11525 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
11526 * cgraph.h (symtab_node_base): Add weakref flag.
11527 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
11528 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
11529 (output_weakrefs): Use weakref flag.
11530 * fold-const.c (simple_operand_p): Handle WEAK.
11531 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
11532 * ipa.c (varpool_externally_visible_p): Drop weakref.
11533 (function_and_variable_visibility): Update comment; fix weakref
11534 sanity checks; do not clear DECL_WEAK on them.
11535 * lto-cgraph.c (lto_output_node): update.
11536 (lto_output_varpool_node): Update.
11537 (input_overwrite_node): Update.
11538 (input_node): Update.
11539 (input_varpool_node): Update.
11540 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
11541 (lto_symtab_merge_symbols): Add sanity check.
11542 (lto_symtab_prevailing_decl): Do not special case weakrefs.
11543 * passes.c (rest_of_decl_compilation): Set static flag, too.
11544 * symtab.c (dump_symtab_base): Dump weakref.
11545 (verify_symtab_base): Sanity check weakrefs.
11546 (symtab_make_decl_local): Remove duplicated code.
11547 (symtab_alias_ultimate_target): Simplify.
11548 * varpool.c (varpool_create_variable_alias): Set weakref flag.
11550 2013-06-11 Tom de Vries <tom@codesourcery.com>
11552 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
11553 sequence_vect == NULL.
11555 2013-06-11 DJ Delorie <dj@redhat.com>
11557 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
11558 (rl78_unwind_word_mode): New.
11560 2013-06-11 David Malcolm <dmalcolm@redhat.com>
11562 * final.c (debug_prefix_maps): Make static.
11564 2013-06-11 David Malcolm <dmalcolm@redhat.com>
11566 * function.c (initial_trampoline): Remove stray copy.
11568 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
11570 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
11572 2013-06-11 Martin Jambor <mjambor@suse.cz>
11574 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
11575 within bounds at the beginning of the function.
11577 2013-06-11 Alan Modra <amodra@gmail.com>
11579 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
11581 (get_named_section): Don't NULL !DECL_P decl.
11583 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
11585 * doc/invoke.texi (core-avx2): Document.
11587 (atom): Updated with MOVBE.
11589 2013-06-11 Richard Biener <rguenther@suse.de>
11591 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
11593 2013-06-11 Anton Blanchard <anton@samba.org>
11595 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
11596 correct shift value in little-endian mode.
11598 2013-06-11 Jakub Jelinek <jakub@redhat.com>
11601 * varasm.c (get_variable_align): Move #endif to the right place.
11603 2013-06-10 Cary Coutant <ccoutant@google.com>
11605 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
11606 for hash so that hash table traversal order is deterministic.
11608 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
11609 Pat Haugen <pthaugen@us.ibm.com>
11610 Peter Bergner <bergner@vnet.ibm.com>
11612 * config/rs6000/vector.md (GPR move splitter): Do not split moves
11613 of vectors in GPRS if they are direct moves or quad word load or
11616 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
11618 (direct_move_p): Likewise.
11619 (quad_load_store_p): Likewise.
11621 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
11622 classes into bins based on the physical register type.
11623 (reg_class_to_reg_type): Likewise.
11624 (IS_STD_REG_TYPE): Likewise.
11625 (IS_FP_VECT_REG_TYPE): Likewise.
11626 (reload_fpr_gpr): Arrays to determine what insn to use if we can
11627 use direct move instructions.
11628 (reload_gpr_vsx): Likewise.
11629 (reload_vsx_gpr): Likewise.
11630 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
11631 information that is a simplification of register classes. Also
11632 precalculate direct move reload helpers.
11633 (direct_move_p): New function to return true if the operation can
11634 be done as a direct move instruciton.
11635 (quad_load_store_p): New function to return true if the operation
11636 is a quad memory operation.
11637 (rs6000_legitimize_address): If quad memory, only allow register
11638 indirect for TImode addresses.
11639 (rs6000_legitimate_address_p): Likewise.
11640 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
11641 (rs6000_reload_register_type): Likewise.
11642 (register_to_reg_type): Return register type.
11643 (rs6000_secondary_reload_simple_move): New helper function for
11644 secondary reload and secondary memory needed to identify anything
11645 that is a simple move, and does not need reloading.
11646 (rs6000_secondary_reload_direct_move): New helper function for
11647 secondary reload to identify cases that can be done with several
11648 instructions via the direct move instructions.
11649 (rs6000_secondary_reload_move): New helper function for secondary
11650 reload to identify moves between register types that can be done.
11651 (rs6000_secondary_reload): Add support for quad memory operations
11652 and for direct move.
11653 (rs6000_secondary_memory_needed): Likewise.
11654 (rs6000_debug_secondary_memory_needed): Change argument names.
11655 (rs6000_output_move_128bit): New function to return the move to
11656 use for 128-bit moves, including knowing about the various
11657 limitations of quad memory operations.
11659 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
11660 memory operations. call rs6000_output_move_128bit for the actual
11661 instruciton(s) to generate.
11662 (vsx_movti_64bit): Likewise.
11664 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
11665 (UNSPEC_P8V_MTVSRWZ): Likewise.
11666 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
11667 (UNSPEC_P8V_MTVSRD): Likewise.
11668 (UNSPEC_P8V_XXPERMDI): Likewise.
11669 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
11670 (UNSPEC_FUSION_GPR): Likewise.
11671 (FMOVE128_GPR): New iterator for direct move.
11672 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
11673 (f32_sv): Likewise.
11674 (f32_dm): Likewise.
11675 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
11676 loads and direct move instructions.
11677 (zero_extendsidi2_lfiwzx): Likewise.
11678 (extendsidi2_lfiwax): Likewise.
11679 (extendsidi2_nocell): Likewise.
11680 (floatsi<mode>2_lfiwax): Likewise.
11681 (lfiwax): Likewise.
11682 (floatunssi<mode>2_lfiwzx): Likewise.
11683 (lfiwzx): Likewise.
11684 (fix_trunc<mode>_stfiwx): Likewise.
11685 (fixuns_trunc<mode>_stfiwx): Likewise.
11686 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
11687 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
11688 (parity<mode>2_cmpb): Set length/type attr.
11689 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
11690 for 'mr.' to fast_compare.
11691 (bpermd_<mode>): Change type attr to popcnt.
11692 (p8_fmrgow_<mode>): New insns for power8 direct move support.
11693 (p8_mtvsrwz_1): Likewise.
11694 (p8_mtvsrwz_2): Likewise.
11695 (reload_fpr_from_gpr<mode>): Likewise.
11696 (p8_mtvsrd_1): Likewise.
11697 (p8_mtvsrd_2): Likewise.
11698 (p8_xxpermdi_<mode>): Likewise.
11699 (reload_vsx_from_gpr<mode>): Likewise.
11700 (reload_vsx_from_gprsf): Likewise.
11701 (p8_mfvsrd_3_<mode>): LIkewise.
11702 (reload_gpr_from_vsx<mode>): Likewise.
11703 (reload_gpr_from_vsxsf): Likewise.
11704 (p8_mfvsrd_4_disf): Likewise.
11705 (multi-word GPR splits): Do not split direct moves or quad memory
11708 2013-06-10 David Malcolm <dmalcolm@redhat.com>
11710 * tree-into-ssa.c (interesting_blocks): Make static.
11712 2013-06-10 Jakub Jelinek <jakub@redhat.com>
11715 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
11716 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
11717 Use DATA_ABI_ALIGNMENT for that case instead if defined.
11718 (get_variable_align): New function.
11719 (get_variable_section, emit_bss, emit_common,
11720 assemble_variable_contents, place_block_symbol): Use
11721 get_variable_align instead of DECL_ALIGN.
11722 (assemble_noswitch_variable): Add align argument, use it
11723 instead of DECL_ALIGN.
11724 (assemble_variable): Adjust caller. Use get_variable_align
11725 instead of DECL_ALIGN.
11726 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
11728 (DATA_ABI_ALIGNMENT): Define.
11729 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
11730 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
11731 opt is false, only return the psABI mandated alignment increase.
11732 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
11733 (DATA_ABI_ALIGNMENT): ... this.
11734 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
11735 (DATA_ABI_ALIGNMENT): ... this.
11736 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
11737 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
11738 (DATA_ABI_ALIGNMENT): ... this.
11739 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
11740 * doc/tm.texi: Regenerated.
11742 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
11744 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
11745 cmp_code to construct REG_EQUAL note.
11747 2013-06-09 Jakub Jelinek <jakub@redhat.com>
11750 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
11751 that operands[2] doesn't overlap with operands[0].
11753 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
11754 Jan Hubicka <jh@suse.cz>
11756 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
11757 hack to mark symbols as used.
11759 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
11761 PR rtl-optimization/57559
11762 * lra-constraints.c (process_alt_operands): Don't discourage
11763 memory with known offset for offsetable memory constraint.
11764 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
11766 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
11768 * varasm.c (struct oc_local_state): Reorder fields.
11769 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
11770 and adjust accordingly.
11771 (output_constructor): Reorder initialization code and adjust call to
11772 output_constructor_bitfield.
11774 2013-06-07 Jan Hubicka <jh@suse.cz>
11776 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
11778 2013-06-07 David Malcolm <dmalcolm@redhat.com>
11780 * tree-object-size.c (unknown): Make const.
11782 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11784 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
11785 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
11786 for last alternative in the cpu_facility attribute.
11788 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11791 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
11792 (xordi3): Change operand 2 constraint to arm_xordi_operand.
11793 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
11794 * config/arm/constraints.md (Dg): New constraint.
11795 * config/arm/neon.md (xordi3_neon): Remove.
11796 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
11797 * config/arm/predicates.md (arm_xordi_operand): New predicate.
11799 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11801 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
11802 Clean up alternatives.
11804 2013-06-07 Alan Modra <amodra@gmail.com>
11806 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
11809 2013-06-07 Alan Modra <amodra@gmail.com>
11811 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
11813 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11815 * config/arm/constraints.md (Df): New constraint.
11816 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
11817 Correct length attribute for last two alternatives.
11819 2013-06-07 Alan Modra <amodra@gmail.com>
11821 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11822 override user -mfp-in-toc.
11823 (offsettable_ok_by_alignment): Consider just the current access
11824 rather than the whole object, unless BLKmode. Handle
11825 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
11826 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
11827 for -mcmodel=medium.
11828 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
11829 override user -mfp-in-toc or -msum-in-toc. Default to
11830 -mno-fp-in-toc for -mcmodel=medium.
11832 2013-06-06 DJ Delorie <dj@redhat.com>
11834 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
11835 TARGET_VALID_POINTER_MODE.
11837 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
11838 Pat Haugen <pthaugen@us.ibm.com>
11839 Peter Bergner <bergner@vnet.ibm.com>
11841 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
11842 Document new power8 builtins.
11844 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
11845 condition code register, to allow 128-bit logical operations to be
11846 done in the VSX or GPR registers.
11847 (nor<mode>3): Use the canonical form for nor.
11848 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
11849 vclz*, and vpopcnt* vector instructions.
11850 (nand<mode>3): Likewise.
11851 (orc<mode>3): Likewise.
11852 (clz<mode>2): LIkewise.
11853 (popcount<mode>2): Likewise.
11855 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
11856 that only the GPRs are recognized.
11858 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11859 support for new power8 builtins.
11861 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
11863 (xscvdpspn): Likewise.
11869 (vpopcnt): Likewise.
11870 (vpopcntb): Likewise.
11871 (vpopcnth): Likewise.
11872 (vpopcntw): Likewise.
11873 (vpopcntd): Likewise.
11875 (vmrgew): Likewise.
11876 (vmrgow): Likewise.
11878 (eqv_v16qi3): Likewise.
11879 (eqv_v8hi3): Likewise.
11880 (eqv_v4si3): Likewise.
11881 (eqv_v2di3): Likewise.
11882 (eqv_v4sf3): Likewise.
11883 (eqv_v2df3): Likewise.
11885 (nand_v16qi3): Likewise.
11886 (nand_v8hi3): Likewise.
11887 (nand_v4si3): Likewise.
11888 (nand_v2di3): Likewise.
11889 (nand_v4sf3): Likewise.
11890 (nand_v2df3): Likewise.
11892 (orc_v16qi3): Likewise.
11893 (orc_v8hi3): Likewise.
11894 (orc_v4si3): Likewise.
11895 (orc_v2di3): Likewise.
11896 (orc_v4sf3): Likewise.
11897 (orc_v2df3): Likewise.
11899 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
11900 allow power8 quad mode in 64-bit.
11901 (rs6000_builtin_vectorized_function): Add support to vectorize
11902 ISA 2.07 count leading zeros, population count builtins.
11903 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
11904 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
11905 (builtin_function_type): Add vgbbd builtin function which takes an
11907 (altivec_expand_vec_perm_const): Add support for new power8 merge
11910 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
11911 that does not include TImdoe for use with 32-bit.
11912 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
11914 (UNSPEC_VSX_CVDPSPN): Likewise.
11915 (vsx_xscvdpspn): Likewise.
11916 (vsx_xscvspdpn): Likewise.
11917 (vsx_xscvdpspn_scalar): Likewise.
11918 (vsx_xscvspdpn_directmove): Likewise.
11919 (vsx_and<mode>3): Split logical operations into 32-bit and
11920 64-bit. Add support to do logical operations on TImode as well as
11921 VSX vector types. Allow logical operations to be done in either
11922 VSX registers or in general purpose registers in 64-bit mode. Add
11923 splitters if GPRs were used. For AND, add clobber of CCmode to
11924 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
11926 (vsx_and<mode>3_32bit): Likewise.
11927 (vsx_and<mode>3_64bit): Likewise.
11928 (vsx_ior<mode>3): Likewise.
11929 (vsx_ior<mode>3_32bit): Likewise.
11930 (vsx_ior<mode>3_64bit): Likewise.
11931 (vsx_xor<mode>3): Likewise.
11932 (vsx_xor<mode>3_32bit): Likewise.
11933 (vsx_xor<mode>3_64bit): Likewise.
11934 (vsx_one_cmpl<mode>2): Likewise.
11935 (vsx_one_cmpl<mode>2_32bit): Likewise.
11936 (vsx_one_cmpl<mode>2_64bit): Likewise.
11937 (vsx_nor<mode>3): Likewise.
11938 (vsx_nor<mode>3_32bit): Likewise.
11939 (vsx_nor<mode>3_64bit): Likewise.
11940 (vsx_andc<mode>3): Likewise.
11941 (vsx_andc<mode>3_32bit): Likewise.
11942 (vsx_andc<mode>3_64bit): Likewise.
11943 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
11944 and xxlorc instructions.
11945 (vsx_eqv<mode>3_64bit): Likewise.
11946 (vsx_nand<mode>3_32bit): Likewise.
11947 (vsx_nand<mode>3_64bit): Likewise.
11948 (vsx_orc<mode>3_32bit): Likewise.
11949 (vsx_orc<mode>3_64bit): Likewise.
11951 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
11953 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
11955 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
11956 (p8_vmrgow): Likewise.
11957 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
11958 GPRs to be split under VSX.
11959 (p8v_clz<mode>2): Add power8 count leading zero support.
11960 (p8v_popcount<mode>2): Add power8 population count support.
11961 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
11964 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
11967 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
11969 (vec_nand): Likewise.
11970 (vec_vclz): Likewise.
11971 (vec_vclzb): Likewise.
11972 (vec_vclzd): Likewise.
11973 (vec_vclzh): Likewise.
11974 (vec_vclzw): Likewise.
11975 (vec_vgbbd): Likewise.
11976 (vec_vmrgew): Likewise.
11977 (vec_vmrgow): Likewise.
11978 (vec_vpopcnt): Likewise.
11979 (vec_vpopcntb): Likewise.
11980 (vec_vpopcntd): Likewise.
11981 (vec_vpopcnth): Likewise.
11982 (vec_vpopcntw): Likewise.
11984 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11986 PR rtl-optimization/57468
11987 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
11990 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11992 PR rtl-optimization/57459
11993 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
11994 type when setting live regs.
11996 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11998 * config/s390/s390.opt (mlra): New option.
11999 * config/s390/s390.c (s390_decompose_address): Check displacement
12000 for all registers for LRA.
12001 (s390_secondary_reload): Don't used secondary reloads for LRA.
12002 (s390_lra_p): New function.
12003 (TARGET_LRA_P): Define.
12004 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
12005 of attribute cpu_facility to zarch for the last alternative.
12006 (*cmpmem_short): Ditto.
12008 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
12010 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
12011 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
12012 (arm_expand_prologue): Likewise.
12014 2013-06-06 Teresa Johnson <tejohnson@google.com>
12017 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
12018 as this is now done by redirect_edge_and_branch_force.
12019 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
12020 barriers, and fix interaction with splitting.
12021 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
12022 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
12023 reflect changes made in the routine.
12024 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
12025 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
12026 since this is called in cfglayout mode, and replace partition fixup
12027 with assert as that is now done by force_nonfallthru_and_redirect.
12028 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
12029 already be marked with region crossing note.
12030 (insert_section_boundary_note): Make non-static, gate on flag
12031 has_bb_partition, rewrite to also check for multiple partitions.
12032 (rest_of_handle_reorder_blocks): Remove call to
12033 insert_section_boundary_note, now done later during free_cfg.
12034 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
12035 * bb-reorder.h (insert_section_boundary_note): Declare.
12036 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
12037 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
12038 invoke insert_section_boundary_note.
12039 (try_redirect_by_replacing_jump): Remove unnecessary
12040 check for region crossing note.
12041 (fixup_partition_crossing): New function.
12042 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
12043 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
12044 in non-cfglayout mode.
12045 (force_nonfallthru_and_redirect): Fixup partition boundaries,
12046 remove old code that tried to do this. Emit barrier correctly
12047 when we are in cfglayout mode.
12048 (last_bb_in_partition): New function.
12049 (rtl_split_edge): Correctly fixup partition boundaries.
12050 (commit_one_edge_insertion): Remove old code that tried to
12051 fixup region crossing edge since this is now handled in
12052 split_block, and set up insertion point correctly since
12053 block may now end in a jump.
12054 (verify_hot_cold_block_grouping): Guard against checking when not in
12055 linearized RTL mode.
12056 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
12058 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
12059 rtl_verify_flow_info, so not called in cfglayout mode.
12060 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
12061 (fixup_reorder_chain): Remove old code that attempted to fixup region
12062 crossing note as this is now handled in force_nonfallthru_and_redirect.
12063 (duplicate_insn_chain): Don't duplicate switch section notes.
12064 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
12066 * basic-block.h (emit_barrier_after_bb): Declare.
12068 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12070 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
12071 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
12072 arm_usatsihi): Adjust alternatives for arm_restrict_it.
12074 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12076 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
12078 * config/arm/ldmstm.md: Regenerate.
12080 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12082 * config/arm/sync.md (atomic_loaddi_1):
12083 Disable predication for arm_restrict_it.
12084 (arm_load_exclusive<mode>): Likewise.
12085 (arm_load_exclusivesi): Likewise.
12086 (arm_load_exclusivedi): Likewise.
12087 (arm_load_acquire_exclusive<mode>): Likewise.
12088 (arm_load_acquire_exclusivesi): Likewise.
12089 (arm_load_acquire_exclusivedi): Likewise.
12090 (arm_store_exclusive<mode>): Likewise.
12091 (arm_store_exclusive<mode>): Likewise.
12092 (arm_store_release_exclusivedi): Likewise.
12093 (arm_store_release_exclusive<mode>): Likewise.
12095 2013-06-06 Richard Biener <rguenther@suse.de>
12097 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
12099 (lto_tag_is_tree_code_p): Adjust.
12100 (lto_tag_is_gimple_code_p): Likewise.
12101 (lto_gimple_code_to_tag): Likewise.
12102 (lto_tag_to_gimple_code): Likewise.
12103 (lto_tree_code_to_tag): Likewise.
12104 (lto_tag_to_tree_code): Likewise.
12105 * data-streamer.h (streamer_write_hwi_in_range): Use
12106 uhwi streaming to stream the normalized range.
12107 (streamer_read_hwi_in_range): Likewise.
12109 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12111 * config/arm/arm.md (enabled_for_depr_it): New attribute.
12112 (predicable_short_it): Likewise.
12113 (predicated): Likewise.
12114 (enabled): Handle above.
12115 (define_cond_exec): Set predicated attribute to yes.
12117 2013-06-05 Mike Stump <mikestump@comcast.net>
12119 * gdbinit.in (__FUNCTION__): Add.
12121 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
12123 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
12124 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
12126 2013-06-05 Jan Hubicka <jh@suse.cz>
12128 * varasm.c (mark_decl_referenced): Revert the removal until targets
12131 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
12133 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
12134 instead of mark_decl_referenced.
12136 2013-06-05 Jan Hubicka <jh@suse.cz>
12138 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
12139 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
12140 and symtab_used_from_object_file_p.
12141 (cgraph_make_node_local_1): Clear forced_by_abi.
12142 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
12143 * cgraph.h (symtab_node_base): Add forced_by_abi.
12144 (decide_is_variable_needed): Remove.
12145 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
12146 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
12147 (decide_is_symbol_needed): ... this one; handle symbols in general;
12148 always analyze virtuals; honnor forced_by_abi.
12149 (cgraph_finalize_function): Update.
12150 (varpool_finalize_decl): Update.
12151 (symbol_defined_and_needed): Remove.
12152 (analyze_functions): Update.
12153 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
12154 output_refs, input_overwrite_node): Handle forced_by_abi.
12155 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
12156 (address_taken_from_non_vtable_p): ... this one.
12157 (comdat_can_be_unshared_p_1): New function.
12158 (cgraph_comdat_can_be_unshared_p): Rename to ...
12159 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
12160 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
12161 (function_and_variable_visibility): Clear forced_by_abi as needed.
12162 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
12163 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
12164 * symtab.c (dump_symtab_base): Dump forced_by_abi.
12165 * varpool.c (decide_is_variable_needed): Remove.
12167 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12169 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
12170 (arm_option_override): Override arm_restrict_it where appropriate.
12171 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
12172 * config/arm/arm.opt (mrestrict-it): New command-line option.
12173 * doc/invoke.texi: Document -mrestrict-it.
12175 2013-06-05 David Malcolm <dmalcolm@redhat.com>
12177 * tsan.c (tsan_atomic_table): Make const.
12179 2013-06-05 Richard Biener <rguenther@suse.de>
12181 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
12182 index associated with the tree we are supposed to replace.
12183 * tree-streamer-out.c (pack_ts_base_value_fields): Output
12184 TREE_ASM_WRITTEN as zero for everything but SSA names.
12186 2013-06-05 David Malcolm <dmalcolm@redhat.com>
12188 * tree-ssa-structalias.c (call_stmt_vars): Make static.
12190 2013-06-04 Jan Hubicka <jh@suse.cz>
12192 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
12193 (input_node, input_varpool_node): Handle correctly external same
12195 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
12196 nodes at ltrans stage.
12198 2013-06-04 Jan Hubicka <jh@suse.cz>
12200 * ipa-inline.c (update_caller_keys): Fix availability test.
12201 (update_callee_keys): Likewise.
12202 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
12203 to follow ELF standard.
12205 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
12207 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
12208 (mips64r5900el-*-elf*): New configurations.
12209 * config/mips/mips-cpus.def (r5900): New processor.
12210 * config/mips/mips-tables.opt: Regenerate.
12211 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
12212 (mips_issue_rate): Handle PROCESSOR_R5900.
12213 (mips_reorg_process_insns): Force reorder mode for the R5900.
12214 * config/mips/mips.h (TARGET_MIPS5900): Define.
12215 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
12217 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
12219 * config/mips/mips.md (processor): Add r5900.
12220 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
12222 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12224 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
12225 into function to generate MOVI instruction.
12226 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
12227 (aarch64_preferred_simd_mode): Turn into wrapper.
12228 (aarch64_output_scalar_simd_mov_immediate): New function.
12229 * config/aarch64/aarch64-protos.h: Add prototype for above.
12231 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12233 * config/aarch64/aarch64.c (simd_immediate_info): Remove
12234 element_char member.
12235 (sizetochar): Return signed char.
12236 (aarch64_simd_valid_immediate): Remove elchar and other
12237 unnecessary variables.
12238 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
12239 Calculate element_char as required.
12240 * config/aarch64/aarch64-protos.h: Update and move prototype
12241 for aarch64_output_simd_mov_immediate.
12242 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
12245 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12247 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
12248 information completed by aarch64_simd_valid_immediate.
12249 (aarch64_legitimate_constant_p): Update arguments.
12250 (aarch64_simd_valid_immediate): Work with struct rather than many
12252 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
12253 (aarch64_simd_make_constant): Update arguments.
12254 (aarch64_output_simd_mov_immediate): Work with struct rather than
12255 many pointers. Output immediate directly rather than as operand.
12256 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
12258 * config/aarch64/constraints.md (Dn): Update arguments.
12260 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12262 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
12264 (aarch64_simd_immediate_valid_for_move): Remove.
12265 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
12266 (aarch64_simd_make_constant): Update call.
12267 (aarch64_output_simd_mov_immediate): Update call.
12268 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
12270 * config/aarch64/constraints.md (Dn): Update call.
12272 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12274 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
12275 return type to bool for prototype.
12276 (aarch64_legitimate_constant_p): Check for true instead of not -1.
12277 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
12278 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
12280 2013-06-04 Catherine Moore <clm@codesourcery.com>
12282 * config/mips/mips.opt (meva): New.
12283 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
12284 (ASM_SPEC): Handle -meva.
12285 * doc/invoke.texi (meva): Document.
12287 2013-06-04 Alan Modra <amodra@gmail.com>
12289 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
12292 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12294 * rtl.def: Add extra fourth optional field to define_cond_exec.
12295 * gensupport.c (process_one_cond_exec): Process attributes from
12297 * doc/md.texi: Document fourth field in define_cond_exec.
12299 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
12301 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
12302 out the processing order as in store_bit_field_1.
12304 2013-06-04 Jan Hubicka <jh@suse.cz>
12306 PR middle-end/57500
12307 * cgraphunit.c (cgraph_process_same_body_aliases): Create
12308 non-VAR_DECL node if it does not exist yet.
12310 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
12312 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
12313 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
12314 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
12315 target_cpu_default setting.
12317 2013-06-03 Teresa Johnson <tejohnson@google.com>
12319 * dumpfile.c (opt_info_switch_p): Change -fopt-info
12320 default to -fopt-info=optimized instead of all.
12321 * doc/invoke.texi: Ditto.
12322 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
12323 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
12324 (execute_vect_slp): Emit BB vectorization success under
12325 MSG_OPTIMIZED_LOCATIONS.
12326 * tree-vect-slp.c (vect_slp_transform_bb): Change
12327 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
12328 * tree-vect-loop.c (vect_transform_loop): Ditto.
12330 2013-06-03 Jason Merrill <jason@redhat.com>
12333 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
12334 Use TARGET_EXPR for C++.
12336 2013-06-03 Jakub Jelinek <jakub@redhat.com>
12338 PR rtl-optimization/57268
12339 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
12340 if DEBUG_INSN_P (insn).
12343 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
12345 PR rtl-optimization/57268
12346 * sched-deps.c (sched_analyze_2): Flush dependence lists if
12347 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
12349 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
12351 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
12352 (ix86_avoid_lea_for_addr): Likewise.
12353 (exact_dependency_1): Likewise.
12354 (ix86_adjust_cost): Likewise.
12355 (swap_top_of_ready_list): Fix formatting and !reload_completed check
12357 (do_reorder_for_imul): Fix typo, formatting and
12358 !reload_completed check removed.
12359 (ix86_sched_reorder): Fix typo and formatting.
12360 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
12363 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
12365 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
12367 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
12369 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
12370 <STRING_CST>: Likewise.
12371 <VECTOR_CST>: Likewise.
12373 2013-06-01 Janus Weil <janus@gcc.gnu.org>
12374 Mikael Morin <mikael@gcc.gnu.org>
12376 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
12377 * config.in: Regenerated.
12378 * configure: Regenerated.
12380 2013-06-01 Jan Hubicka <jh@suse.cz>
12382 PR middle-end/57366
12383 * cgraphunit.c (compile): When weakref is not supported,
12384 set up transparent aliases before final output pass.
12385 * varasm.c (assemble_alias): Do not try to do it here.
12387 2013-06-01 Jan Hubicka <jh@suse.cz>
12389 PR middle-end/57467
12390 * passes.c (for_per_function): Skip unanalyzed functions.
12392 2013-06-01 Jan Hubicka <jh@suse.cz>
12394 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
12395 (lto_symtab_merge_symbols_1): ... this one.
12396 (lto_symtab_merge_cgraph_nodes): Rename to ...
12397 (lto_symtab_merge_symbols): ... this one; simplify.
12398 * cgraph.c (same_body_aliases_done): Rename to ...
12399 (cpp_implicit_aliases_done): ... this one.
12400 (cgraph_create_function_alias): Update.
12401 (cgraph_same_body_alias): Update.
12402 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
12403 (verify_edge_corresponds_to_fndecl): Simplify.
12404 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
12405 (cgraph_node): Remove same_body_alias.
12406 (varpool_node): Remove alias_of and extra_name_alias.
12407 (same_body_aliases_done): Rename to ..
12408 (cpp_implicit_aliases_done): ... this one.
12409 (symtab_alias_ultimate_target): Add default parameter.
12410 (symtab_resolve_alias): New function.
12411 (fixup_same_cpp_alias_visibility): Declare.
12412 (cgraph_function_node): Add default parameter.
12413 (cgraph_node_asm_name): Likewise.
12414 (cgraph_function_or_thunk_node): Add default parameter; do
12415 not ICE when it is NULL.
12416 (varpool_variable_node): Likewise.
12417 * tree-emutls.c (create_emultls_var): Update.
12418 (ipa_lower_emutls): Update.
12419 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
12420 (cgraph_reset_node): Reset alias info.
12421 (cgraph_finalize_function): Update.
12422 (fixup_same_cpp_alias_visibility): Move to symtab.c.
12423 (analyze_function): Simplify.
12424 (cgraph_process_same_body_aliases): Simplify.
12425 (analyze_functions): Fixup same body aliases.
12426 (handle_alias_pairs): Simplify.
12427 (assemble_thunk): Update.
12428 (assemble_thunks_and_aliases): Update.
12429 (output_weakrefs): Rewrite.
12430 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
12431 (lto_output_varpool_node): Likewise.
12432 (compute_ltrans_boundary): Remve assert.
12433 (get_alias_symbol): New functoin.
12434 (input_node): Rewrite alias handling.
12435 (input_varpool_node): Likewise.
12436 * ipa-pure-const.c (propagate_pure_const): Fix formating.
12437 * ipa.c (process_references): Handle weakrefs correctly.
12438 (symtab_remove_unreachable_nodes): Likewise.
12439 * trans-mem.c (get_cg_data): Update.
12440 (ipa_tm_create_version_alias): Update.
12441 (ipa_tm_execute): Update.
12442 * symtab.c (dump_symtab_base): Dump aliases.
12443 (verify_symtab_base): Verify aliases.
12444 (symtab_node_availability): New function.
12445 (symtab_alias_ultimate_target): Simplify.
12446 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
12447 handle all the fixup cases.
12448 (symtab_resolve_alias): New function.
12449 * passes.c (ipa_write_summaries): Handle weakrefs.
12450 * varpool.c (varpool_analyze_node): Simplify.
12451 (assemble_aliases): Update.
12452 (varpool_create_variable_alias): Simplify.
12453 (varpool_extra_name_alias): Simplify.
12454 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
12455 (lto_symtab_merge_symbols): ... this one.
12457 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
12460 PR rtl-optimization/57268
12461 * sched-deps.c (sched_analyze_2): Flush dependence lists if
12462 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
12464 2013-06-01 Tobias Burnus <burnus@net-b.de>
12466 Partially reverted:
12467 2013-05-31 Tobias Burnus <burnus@net-b.de>
12469 PR middle-end/57073
12470 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
12473 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
12475 PR rtl-optimization/57268
12476 * sched-deps.c (sched_analyze_2): Flush dependence lists if
12477 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
12479 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
12481 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
12482 unordered comparison operators when -fno-trapping-math is in effect
12484 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
12485 and implement unordered comparison operators properly on the e500.
12487 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
12489 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
12490 for constant scalar integers.
12491 (simplify_relational_operation_1): Likewise.
12493 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
12495 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
12496 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
12499 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
12500 Igor Zamyatin <igor.zamyatin@intel.com>
12502 Silvermont (SLM) architecture performance tuning.
12503 * config/i386/i386.h (enum ix86_tune_indices): Add
12504 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
12505 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
12507 * config/i386/i386.c (initial_ix86_tune_features)
12508 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
12509 (ix86_lea_outperforms): Handle Silvermont tuning.
12510 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
12512 (ix86_use_lea_for_mov): Likewise.
12513 (ix86_avoid_lea_for_addr): Likewise.
12514 (ix86_lea_for_add_ok): Likewise.
12515 (exact_dependency_1): New function.
12516 (exact_store_load_dependency): Likewise.
12517 (ix86_adjust_cost): Handle Silvermont tuning.
12518 (do_reoder_for_imul): Likewise.
12519 (swap_top_of_ready_list): New function.
12520 (ix86_sched_reorder): Changed to handle Silvermont tuning.
12522 * config/i386/i386.md (peepholes that split memory operand in fp
12525 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
12527 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
12528 Remove un-necessary braces.
12530 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
12532 * config/aarch64/aarch64.c (aarch64_classify_symbol):
12533 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
12535 2013-05-31 Tobias Burnus <burnus@net-b.de>
12537 PR middle-end/57073
12538 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
12540 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12543 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
12544 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
12545 * config/arm/neon.md (iordi3_neon): Remove.
12546 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
12547 * config/arm/predicates.md (imm_for_neon_logic_operand):
12548 Move to earlier in the file.
12549 (neon_logic_op2): Likewise.
12550 (arm_iordi_operand_neon): New predicate.
12552 2013-05-31 Richard Biener <rguenther@suse.de>
12554 PR tree-optimization/57478
12555 PR tree-optimization/57453
12556 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
12559 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
12561 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
12562 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
12564 2013-05-30 Tobias Burnus <burnus@net-b.de>
12565 Thomas Koenig <tkoenig@gcc.gnu.org>
12567 PR middle-end/57073
12568 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
12569 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
12571 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
12573 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
12575 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
12577 * target.def (register_usage_leveling_p): New hook.
12578 * targhooks.c (default_register_usage_leveling_p): New.
12579 * targhooks.h (default_register_usage_leveling_p): New prototype.
12580 * lra-assigns.c (register_usage_leveling_p): Use the hook.
12581 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
12582 * doc/tm.texi: Update.
12583 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
12585 2013-05-30 Ian Bolton <ian.bolton@arm.com>
12587 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
12588 (*insv_reg<mode>): New define_insn.
12590 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
12592 PR rtl-optimization/57439
12593 * postreload.c (move2add_valid_value_p): Check that we have
12594 a zero subreg_regno_offset when accessing the register in
12595 the requested mode.
12597 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
12598 Igor Zamyatin <igor.zamyatin@intel.com>
12600 Silvermont (SLM) architecture pipeline model, tuning and
12602 * config.gcc: Add slm config options and target.
12604 * config/i386/slm.md: New.
12606 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
12608 * config/i386/i386-c.c (ix86_target_macros_internal): New case
12610 (ix86_target_macros_internal): Likewise.
12612 * config/i386/i386.c (slm_cost): New cost.
12613 (m_SLM): New macro flag.
12614 (initial_ix86_tune_features): Set m_SLM.
12615 (x86_accumulate_outgoing_args): Likewise.
12616 (x86_arch_always_fancy_math_387): Likewise.
12617 (processor_target_table): Add slm cost.
12618 (cpu_names): Add slm cpu name.
12619 (x86_option_override_internal): Set SLM ISA.
12620 (ix86_issue_rate): New case PROCESSOR_SLM.
12621 (ia32_multipass_dfa_lookahead): Likewise.
12622 (fold_builtin_cpu): Add slm.
12624 * config/i386/i386.h (TARGET_SLM): New target macro.
12625 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
12626 (processor_type): Add PROCESSOR_SLM.
12628 * config/i386/i386.md (cpu): Add new value "slm".
12629 (slm.md): Include slm.md.
12631 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
12632 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12634 * config/arm/arm-protos.h: Add and update function protos.
12635 * config/arm/arm.c (use_simple_return_p): New added.
12636 (thumb2_expand_return): Check simple_return flag.
12637 * config/arm/arm.md: Add simple_return and conditional simple_return.
12638 * config/arm/iterators.md: Add iterator for return and simple_return.
12640 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12642 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
12643 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
12644 (arm_emit_vfp_multi_reg_pop): Likewise.
12645 (thumb2_emit_ldrd_pop): Likewise.
12646 (arm_expand_epilogue): Add misc REG_CFA notes.
12647 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
12649 2013-05-29 Lawrence Crowl <crowl@google.com>
12651 * config/arm/t-arm: Update for below.
12653 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
12654 Change type to hash_table. Update dependent calls and types.
12656 * config/i386/t-cygming: Update for below.
12658 * config/i386/t-interix: Update for below.
12660 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
12661 Change type to hash_table. Update dependent calls and types.
12662 (i386_find_on_wrapper_list::wrappers): Likewise.
12664 * config/ia64/t-ia64: Update for below.
12666 * config/ia64/ia64.c (bundle_state_table):
12667 Change type to hash_table. Update dependent calls and types.
12669 * config/mips/mips.c (mips_reorg_process_insns::htab):
12670 Change type to hash_table. Update dependent calls and types.
12672 * config/sol2.c (solaris_comdat_htab):
12673 Change type to hash_table. Update dependent calls and types.
12675 * config/t-sol2: Update for above.
12677 2013-05-29 Teresa Johnson <tejohnson@google.com>
12679 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
12680 functions are not yet marked as defined.
12682 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
12683 Pat Haugen <pthaugen@us.ibm.com>
12684 Peter Bergner <bergner@vnet.ibm.com>
12686 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
12690 (vrotl<mode>3): Likewise.
12691 (vashl<mode>3): Likewise.
12692 (vlshr<mode>3): Likewise.
12693 (vashr<mode>3): Likewise.
12695 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
12696 support for power8 V2DI builtins.
12698 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
12699 power8 V2DI builtins.
12700 (vupkhsw): Likewise.
12701 (vupklsw): Likewise.
12702 (vaddudm): Likewise.
12703 (vminsd): Likewise.
12704 (vmaxsd): Likewise.
12705 (vminud): Likewise.
12706 (vmaxud): Likewise.
12707 (vpkudum): Likewise.
12708 (vpksdss): Likewise.
12709 (vpkudus): Likewise.
12710 (vpksdus): Likewise.
12715 (vsubudm): Likewise.
12716 (vcmpequd): Likewise.
12717 (vcmpgtsd): Likewise.
12718 (vcmpgtud): Likewise.
12719 (vcmpequd_p): Likewise.
12720 (vcmpgtsd_p): Likewise.
12721 (vcmpgtud_p): Likewise.
12722 (vupkhsw): Likewise.
12723 (vupklsw): Likewise.
12724 (vaddudm): Likewise.
12725 (vmaxsd): Likewise.
12726 (vmaxud): Likewise.
12727 (vminsd): Likewise.
12728 (vminud): Likewise.
12729 (vpksdss): Likewise.
12730 (vpksdus): Likewise.
12731 (vpkudum): Likewise.
12732 (vpkudus): Likewise.
12737 (vsubudm): Likewise.
12739 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
12740 support for power8 V2DI instructions.
12742 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
12743 power8 V2DI instructions. Combine pack and unpack insns to use an
12744 iterator for each mode. Check whether a particular mode supports
12745 Altivec instructions instead of just checking TARGET_ALTIVEC.
12746 (UNSPEC_VPKUWUM): Likewise.
12747 (UNSPEC_VPKSHSS): Likewise.
12748 (UNSPEC_VPKSWSS): Likewise.
12749 (UNSPEC_VPKUHUS): Likewise.
12750 (UNSPEC_VPKSHUS): Likewise.
12751 (UNSPEC_VPKUWUS): Likewise.
12752 (UNSPEC_VPKSWUS): Likewise.
12753 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
12754 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
12755 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
12756 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
12757 (UNSPEC_VUPKHSB): Likewise.
12758 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
12759 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
12760 (UNSPEC_VUPKHSH): Likewise.
12761 (UNSPEC_VUPKLSB): Likewise.
12762 (UNSPEC_VUPKLSH): Likewise.
12764 (VI_char): Likewise.
12765 (VI_scalar): Likewise.
12766 (VI_unit): Likewise.
12768 (VP_small): Likewise.
12769 (VP_small_lc): Likewise.
12770 (VU_char): Likewise.
12771 (add<mode>3): Likewise.
12772 (altivec_vaddcuw): Likewise.
12773 (altivec_vaddu<VI_char>s): Likewise.
12774 (altivec_vadds<VI_char>s): Likewise.
12775 (sub<mode>3): Likewise.
12776 (altivec_vsubcuw): Likewise.
12777 (altivec_vsubu<VI_char>s): Likewise.
12778 (altivec_vsubs<VI_char>s): Likewise.
12779 (altivec_vavgs<VI_char>): Likewise.
12780 (altivec_vcmpbfp): Likewise.
12781 (altivec_eq<mode>): Likewise.
12782 (altivec_gt<mode>): Likewise.
12783 (altivec_gtu<mode>): Likewise.
12784 (umax<mode>3): Likewise.
12785 (smax<mode>3): Likewise.
12786 (umin<mode>3): Likewise.
12787 (smin<mode>3): Likewise.
12788 (altivec_vpkuhum): Likewise.
12789 (altivec_vpkuwum): Likewise.
12790 (altivec_vpkshss): Likewise.
12791 (altivec_vpkswss): Likewise.
12792 (altivec_vpkuhus): Likewise.
12793 (altivec_vpkshus): Likewise.
12794 (altivec_vpkuwus): Likewise.
12795 (altivec_vpkswus): Likewise.
12796 (altivec_vpks<VI_char>ss): Likewise.
12797 (altivec_vpks<VI_char>us): Likewise.
12798 (altivec_vpku<VI_char>us): Likewise.
12799 (altivec_vpku<VI_char>um): Likewise.
12800 (altivec_vrl<VI_char>): Likewise.
12801 (altivec_vsl<VI_char>): Likewise.
12802 (altivec_vsr<VI_char>): Likewise.
12803 (altivec_vsra<VI_char>): Likewise.
12804 (altivec_vsldoi_<mode>): Likewise.
12805 (altivec_vupkhsb): Likewise.
12806 (altivec_vupkhs<VU_char>): Likewise.
12807 (altivec_vupkls<VU_char>): Likewise.
12808 (altivec_vupkhsh): Likewise.
12809 (altivec_vupklsb): Likewise.
12810 (altivec_vupklsh): Likewise.
12811 (altivec_vcmpequ<VI_char>_p): Likewise.
12812 (altivec_vcmpgts<VI_char>_p): Likewise.
12813 (altivec_vcmpgtu<VI_char>_p): Likewise.
12814 (abs<mode>2): Likewise.
12815 (vec_unpacks_hi_v16qi): Likewise.
12816 (vec_unpacks_hi_v8hi): Likewise.
12817 (vec_unpacks_lo_v16qi): Likewise.
12818 (vec_unpacks_hi_<VP_small_lc>): Likewise.
12819 (vec_unpacks_lo_v8hi): Likewise.
12820 (vec_unpacks_lo_<VP_small_lc>): Likewise.
12821 (vec_pack_trunc_v8h): Likewise.
12822 (vec_pack_trunc_v4si): Likewise.
12823 (vec_pack_trunc_<mode>): Likewise.
12825 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
12827 (vec_vmaxsd): Likewise.
12828 (vec_vmaxud): Likewise.
12829 (vec_vminsd): Likewise.
12830 (vec_vminud): Likewise.
12831 (vec_vpksdss): Likewise.
12832 (vec_vpksdus): Likewise.
12833 (vec_vpkudum): Likewise.
12834 (vec_vpkudus): Likewise.
12835 (vec_vrld): Likewise.
12836 (vec_vsld): Likewise.
12837 (vec_vsrad): Likewise.
12838 (vec_vsrd): Likewise.
12839 (vec_vsubudm): Likewise.
12840 (vec_vupkhsw): Likewise.
12841 (vec_vupklsw): Likewise.
12843 2013-05-29 Jan Hubicka <jh@suse.cz>
12845 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
12846 flags; reorder rest of fields in more consistent way.
12847 (varpool_node): Remove analyzed, finalized and alias.
12848 (cgraph_ndoe): Likewise.
12849 (symtab_alias_ultimate_target): New function.
12850 (cgraph_function_node): Move offline.
12851 (cgraph_reset_node): Declare.
12852 (cgraph_comdat_can_be_unshared_p): Remove.
12853 (varpool_remove_initializer): Declare.
12854 (varpool_first_defined_variable, varpool_next_defined_variable
12855 cgraph_first_defined_function, cgraph_next_defined_function): Update.
12856 (cgraph_function_with_gimple_body_p): Update.
12857 (varpool_all_refs_explicit_p): Update.
12858 (symtab_alias_target): New function.
12859 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
12860 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
12861 (cgraph_function_or_thunk_node): Simplify using
12862 symtab_alias_ultimate_target.
12863 (varpool_variable_node): Likewise.
12864 * cgraph.c (cgraph_create_function_alias): Update.
12865 (cgraph_add_thunk): Update.
12866 (cgraph_remove_node): Update.
12867 (dump_cgraph_node): Do not dump removed flags.
12868 (cgraph_function_body_availability): Update.
12869 (cgraph_propagate_frequency): Update.
12870 (verify_cgraph_node): Check sanity of local flag.
12871 (cgraph_function_node): Move here from cgraph.h; revamp for
12872 cgraph_function_or_thunk_node.
12873 * lto-symtab.c (lto_varpool_replace_node): Update.
12874 (lto_symtab_resolve_can_prevail_p): Update.
12875 (lto_symtab_merge_cgraph_nodes): Update.
12876 * ipa-cp.c (determine_versionability, initialize_node_lattices,
12877 propagate_constants_accross_call, devirtualization_time_bonus,
12878 ipcp_propagate_stage): Update.
12879 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
12880 * ipa-inline-transform.c (clone_inlined_nodes,
12881 preserve_function_body_p): Update.
12882 * ipa-reference.c (propagate): Update.
12883 (write_node_summary_p): Update.
12884 * toplev.c (wrapup_global_declaration_2): Update.
12885 * cgraphunit.c (cgraph_analyze_function): Rename to ...
12886 (analyze_function) ... this one.
12887 (cgraph_process_new_functions): Update.
12888 (cgraph_reset_node): Export.
12889 (cgraph_finalize_function): Update.
12890 (cgraph_add_new_function): Update.
12891 (process_function_and_variable_attributes): Update.
12892 (varpool_finalize_decl): Update.
12893 (symbol_finalized): Remove.
12894 (symbol_finalized_and_needed): Rename to ...
12895 (symbol_defined_and_needed): ... update.
12896 (cgraph_analyze_functions): Update.
12897 (handle_alias_pairs): Update.
12898 (mark_functions_to_output): Update.
12899 (assemble_thunk): Update.
12900 (output_in_order): Update.
12901 (output_weakrefs): Update.
12902 (finalize_compilation_unit): Update.
12903 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
12904 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
12905 input_node, input_varpool_node): Update.
12906 * dbxout.c (dbxout_expand_expr): Update.
12907 * cgraphclones.c (cgraph_clone_node): Update.
12908 (cgraph_copy_node_for_versioning): Update.
12909 (cgraph_materialize_clone): Update.
12910 (cgraph_materialize_all_clones): Update.
12911 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
12912 propagate_pure_const, propagate_nothrow): Update.
12913 * lto-streamer-out.c (lto_output, write_symbol): Update.
12914 * ipa-utils.c (ipa_reverse_postorder): Update.
12915 * ipa-inline.c (can_inline_edge_p): Update.
12916 (update_caller_keys, ipa_inline): Update.
12917 * dwarf2out.c (reference_to_unused,
12918 premark_types_used_by_global_vars_helper): Update.
12919 * tree-eh.c (tree_could_trap_p): Update.
12920 * ipa-split.c (consider_split, execute_split_functions): Update.
12921 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
12922 has_addr_references_p): Update; move ahead in file for better
12924 (process_references): Simplify.
12925 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
12926 bodies are removed.
12927 (cgraph_comdat_can_be_unshared_p): Make static.
12928 (cgraph_externally_visible_p): Update.
12929 (varpool_externally_visible_p): Update.
12930 (function_and_variable_visibility): Update.
12931 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
12932 ipa_tm_mark_force_output_node): Update.
12933 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
12934 estimate_edge_devirt_benefit, inline_generate_summary,
12935 inline_write_summary): Update.
12936 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
12937 * ipa-prop.c (ipa_compute_jump_functions): Update.
12938 (ipa_print_node_params, ipa_prop_read_section,
12939 ipa_update_after_lto_read, read_replacements_section): Update.
12940 * varasm.c (mark_decl_referenced): Update.
12941 (assemble_alias, dump_tm_clone_pairs): Update.
12942 * tree-inline.c (copy_bb): Update.
12943 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
12945 * symtab.c (dump_symtab_base): Print new flags.
12946 (verify_symtab_base): Verify new flags.
12947 (symtab_alias_ultimate_target): New function.
12948 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
12949 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
12951 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
12953 * i386.c (ix86_get_function_versions_dispatcher,
12954 ix86_generate_version_dispatcher_body): Update.
12955 (fold_builtin_cpu): Use varpool_add_new_variable.
12956 * varpool.c (varpool_remove_initializer): Break out from ...
12957 (varpool_remove_node): ... this one.
12958 (dump_varpool_node, varpool_node_for_asm,
12959 cgraph_variable_initializer_availability, varpool_analyze_node,
12960 varpool_assemble_decl, varpool_remove_unreferenced_decls,
12961 varpool_finalize_named_section_flags, varpool_create_variable_alias):
12964 2013-05-29 Jan Hubicka <jh@suse.cz>
12966 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
12968 2013-05-29 Easwaran Raman <eraman@google.com>
12970 PR tree-optimization/57442
12971 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
12972 when control exits the main loop.
12974 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
12976 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
12978 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
12979 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
12980 * rx/t-rx: Add rx100 under multi library matches option for nofpu
12983 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12985 PR tree-optimization/57441
12986 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
12987 Don't limit size of incr_vec to number of candidates.
12989 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
12991 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
12992 and mips16 directories.
12993 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
12994 (MULTILIB_DIRNAMES): Ditto.
12995 (MULTILIB_EXCEPTIONS): Add new exceptions.
12996 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
12997 (MULTILIB_DIRNAMES): Ditto.
12998 (MULTILIB_EXCEPTIONS): Add new exceptions.
13000 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13001 Marcus Shawcroft <marcus.shawcroft@arm.com>
13003 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
13004 SYMBOL_TINY_ABSOLUTE.
13005 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
13006 SYMBOL_TINY_ABSOLUTE.
13007 (aarch64_expand_mov_immediate): Likewise.
13008 (aarch64_classify_symbol): Likewise.
13009 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
13010 Permit SYMBOL_TINY_ABSOLUTE.
13011 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
13013 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13014 Marcus Shawcroft <marcus.shawcroft@arm.com>
13016 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
13017 Refactor if/switch. Replace gcc_assert with if.
13019 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13021 * config/i386/i386.c (initial_ix86_tune_features): Enable
13022 FP Reassociation for AMD bdver1 and bdver2.
13024 2013-05-29 Martin Jambor <mjambor@suse.cz>
13026 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
13027 and IMAGPART_EXPR do not occur within other handled_components.
13029 2013-05-29 Richard Biener <rguenther@suse.de>
13031 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
13032 access on whether the use is in the BB we currently try to
13034 (vect_bb_vectorization_profitable_p): Pass the BB we currently
13035 vectorize to vect_bb_slp_scalar_cost.
13037 2013-05-29 Richard Biener <rguenther@suse.de>
13039 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
13040 computing scalar cost offsetted by stmts that are kept live
13042 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
13043 for computation of scalar cost.
13045 2013-05-28 Steve Ellcey <sellcey@mips.com>
13047 * config/mips/mips-cpus.def (mips32r2): Change processor type.
13049 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13051 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
13052 array notation built-in reduction functions.
13053 * doc/passes.texi (Passes): Added documentation about changes done
13055 * doc/invoke.texi (C Dialect Options): Added documentation about
13056 the -fcilkplus flag.
13057 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
13058 (BUILTINS_DEF): Depend on cilkplus.def.
13059 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
13060 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
13061 * cilkplus.def: New file.
13063 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
13065 PR rtl-optimization/57439
13066 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
13068 2013-05-28 Easwaran Raman <eraman@google.com>
13070 PR tree-optimization/57337
13071 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
13072 (find_insert_point): Correctly identify the insertion point
13073 when two statements with the same UID is compared.
13075 2013-05-28 Richard Biener <rguenther@suse.de>
13077 PR tree-optimization/56787
13078 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
13079 from the list of data references.
13080 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
13082 (vect_analyze_loop_operations): Likewise.
13083 (vect_transform_loop): Remove clobbers.
13085 2013-05-28 Martin Jambor <mjambor@suse.cz>
13087 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
13088 and REALPART_EXPRs have scalar type.
13090 2013-05-28 Richard Biener <rguenther@suse.de>
13092 PR tree-optimization/57411
13093 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
13095 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
13096 virtual operand propagation.
13098 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
13100 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
13101 destination register for bmasksi_vis.
13102 (vector_init_bshuffle): Likewise.
13103 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
13105 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
13107 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
13108 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
13109 mode if the instruction isn't available in the original mode.
13110 * config/sparc/sparc.opt (mfix-ut699): New option.
13111 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
13112 (divdf3): Turn into expander.
13113 (divdf3_nofix): New insn.
13114 (divdf3_fix): Likewise.
13115 (divsf3): Disable if -mfix-ut699.
13116 (sqrtdf2): Turn into expander.
13117 (sqrtdf2_nofix): New insn.
13118 (sqrtdf2_fix): Likewise.
13119 (sqrtsf2): Disable if -mfix-ut699.
13121 2013-05-27 Richard Biener <rguenther@suse.de>
13123 PR middle-end/57412
13124 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
13125 block for the new loop.
13127 2013-05-27 Richard Biener <rguenther@suse.de>
13129 PR tree-optimization/57343
13130 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
13131 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
13132 (number_of_iterations_cond): Do not build the folded tree.
13134 2013-05-27 Richard Biener <rguenther@suse.de>
13137 PR middle-end/57381
13138 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
13139 OEP_CONSTANT_ADDRESS_OF retained.
13141 PR tree-optimization/57417
13142 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
13143 for unchanged base.
13144 (set_ssa_val_to): Compare addresses using
13145 get_addr_base_and_unit_offset.
13147 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
13149 PR rtl-optimization/56833
13150 * postreload.c (move2add_record_mode): New function.
13151 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
13152 (move2add_use_add2_insn): Use move2add_record_sym_value.
13153 (move2add_use_add3_insn): Likewise.
13154 (reload_cse_move2add): Use move2add_valid_value_p and
13155 move2add_record_mode. Invalidate call-clobbered and REG_INC
13156 affected regs by setting reg_mode to VOIDmode.
13157 (move2add_note_store): Don't pretend the inside of a SUBREG is
13158 the actual destination. Invalidate single/leading registers by
13159 setting reg_mode to VOIDmode.
13160 Use move2add_record_sym_value, move2add_valid_value_p and
13161 move2add_record_mode.
13163 2013-05-27 Richard Biener <rguenther@suse.de>
13165 PR tree-optimization/57396
13166 * tree-affine.c (double_int_constant_multiple_p): Properly
13167 return false for val == 0 and div != 0.
13169 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13171 * config/mips/mips.h: Use #elif in preprocessor conditions.
13173 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13176 * config/mips/constraints.md (kl): New constraint.
13177 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
13178 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
13179 constraint for operand 0. Split after CSE for MIPS16. Emit a move
13180 from LO for MIPS16.
13181 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
13183 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13186 * config/mips/mips.c (mips_can_inline_p): New function.
13187 (TARGET_CAN_INLINE_P): Define.
13189 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
13191 * sched-int.h (ds_t, dw_t): Make unsigned int.
13192 Fix documentation that describes how all the ds_t bits are used.
13193 Reserve the last bit for delayed-branch scheduling.
13194 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
13195 (BITS_PER_DEP_WEAK): Fix definition and documentation.
13196 (gen_dep_weak_1): Remove prototype.
13197 * sched-deps.c (get_dep_weak_1): Make static.
13198 * target.def (speculate_insn, needs_block_p, gen_spec_check,
13199 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
13200 * doc/tm.texi: Regenerate.
13201 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
13203 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
13206 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
13208 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
13209 Sandra Loosemore <sandra@codesourcery.com>
13211 * config.gcc (powerpc-*): Allow native for with-cpu.
13213 2013-05-24 Jeff Law <law@redhat.com>
13215 PR tree-optimization/57124
13216 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
13217 conversion feeding a condition if the range has an overflow
13218 if -fstrict-overflow. Add warnings for when we do make the
13221 2013-05-24 Dehao Chen <dehao@google.com>
13223 * tree-cfg.c (locus_discrim_map): Fix the typo.
13224 (locus_discrim_hasher): Likewise.
13225 (locus_discrim_hasher::hash): Likewise.
13226 (locus_discrim_hasher::equal): Likewise.
13228 2013-05-24 Martin Jambor <mjambor@suse.cz>
13230 PR tree-optimization/57294
13231 * cgraph.h (ipa_record_stmt_references): Declare.
13232 * cgraphbuild.c (ipa_record_stmt_references): New function.
13233 (build_cgraph_edges): Use ipa_record_stmt_references.
13234 (rebuild_cgraph_edges): Likewise.
13235 (cgraph_rebuild_references): Likewise.
13236 * ipa-prop.c (ipa_modify_call_arguments): Discard references
13237 associated with the old statement and build references from the
13238 newly built statements.
13239 * ipa-ref.c (ipa_remove_stmt_references): New function.
13240 * ipa-ref.h (ipa_remove_stmt_references): Declare.
13242 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
13244 * lra-constraints.c (emit_spill_move): Use smaller mode for
13246 (check_and_process_move): Consider mem-reg moves for secondary
13248 (curr_insn_transform): Don't lose insns emitted before for
13249 secondary memory moves.
13250 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
13251 reg set up in the current insn.
13253 2013-05-24 Dehao Chen <dehao@google.com>
13255 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
13257 (locus_descrim_hasher::equal): Likewise.
13258 (build_gimple_cfg): New discriminator assignment algorithm.
13259 (make_edges): Likewise.
13260 (next_discriminator_for_locus): Likewise.
13261 (same_line_p): Likewise.
13262 (assign_discriminators): Likewise.
13263 (make_cond_expr_edges): Likewise.
13264 (make_gimple_switch_edges): Likewise.
13265 (make_goto_expr_edges): Likewise.
13266 (make_gimple_asm_edges): Likewise.
13268 2013-05-24 Ian Bolton <ian.bolton@arm.com>
13270 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
13271 X format specifier to only display bottom 16 bits.
13272 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
13273 immediate to match for operand 2, since it will be masked.
13275 2013-05-24 Richard Biener <rguenther@suse.de>
13277 PR tree-optimization/57287
13278 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
13279 all SSA names that occur in abnormal PHIs.
13281 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
13284 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
13285 that index is not negative.
13287 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
13289 PR rtl-optimization/55177
13290 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
13291 (simplify_byte_swapping_operation): New.
13292 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
13293 (simplify_relational_operation_1): Deal with BSWAP.
13295 2013-05-23 Richard Henderson <rth@redhat.com>
13298 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
13299 (ix86_reorg): Call it.
13301 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
13304 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
13305 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
13306 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
13308 2013-05-23 Christian Bruel <christian.bruel@st.com>
13311 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
13313 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13314 Marcus Shawcroft <marcus.shawcroft@arm.com>
13316 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
13317 * config/aarch64/constraints.md (Usa): Remove.
13318 * doc/md.texi (AArch64 Usa): Remove.
13320 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13321 Marcus Shawcroft <marcus.shawcroft@arm.com>
13323 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
13324 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
13325 * config/aarch64/predicates.md (aarch64_const_address): Remove.
13326 (aarch64_mov_operand): Use aarch64_mov_operand_p.
13328 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
13330 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
13331 instruction (AdvSIMD).
13332 * config/aarch64/aarch64-builtins.c
13333 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
13334 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
13336 2013-05-23 Martin Jambor <mjambor@suse.cz>
13338 PR middle-end/57347
13339 * tree.h (contains_bitfld_component_ref_p): Declare.
13340 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
13341 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
13343 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
13344 not access a bit-field. Assert all final offsets are byte-aligned.
13346 2013-05-23 Richard Biener <rguenther@suse.de>
13348 PR tree-optimization/57380
13349 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
13350 least one invariant or re-used load.
13351 * passes.c (init_optimization_passes): Move pass_phiprop before
13354 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
13356 * config/aarch64/aarch64-simd.md
13357 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
13359 2013-05-23 Richard Biener <rguenther@suse.de>
13361 PR middle-end/57381
13362 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
13363 OEP_CONSTANT_ADDRESS_OF retained.
13365 2013-05-23 Jakub Jelinek <jakub@redhat.com>
13367 PR middle-end/57344
13368 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
13369 don't lower unit. Handle unit not being always BITS_PER_WORD.
13371 2013-05-23 Richard Biener <rguenther@suse.de>
13373 PR rtl-optimization/57341
13374 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
13375 instead of true_dependence.
13377 2013-05-22 David Malcolm <dmalcolm@redhat.com>
13379 * bb-reorder.c (branch_threshold): Make const.
13380 (exec_threshold): Ditto.
13382 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
13383 Pat Haugen <pthaugen@us.ibm.com>
13384 Peter Bergner <bergner@vnet.ibm.com>
13386 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
13387 documentation for the power8 crypto builtins.
13389 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
13391 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
13392 macros for defining power8 builtin functions.
13393 (BU_P8V_AV_2): Likewise.
13394 (BU_P8V_AV_P): Likewise.
13395 (BU_P8V_VSX_1): Likewise.
13396 (BU_P8V_OVERLOAD_1): Likewise.
13397 (BU_P8V_OVERLOAD_2): Likewise.
13398 (BU_CRYPTO_1): Likewise.
13399 (BU_CRYPTO_2): Likewise.
13400 (BU_CRYPTO_3): Likewise.
13401 (BU_CRYPTO_OVERLOAD_1): Likewise.
13402 (BU_CRYPTO_OVERLOAD_2): Likewise.
13403 (XSCVSPDP): Fix typo, point to the correct instruction.
13404 (VCIPHER): Add power8 crypto builtins.
13405 (VCIPHERLAST): Likewise.
13406 (VNCIPHER): Likewise.
13407 (VNCIPHERLAST): Likewise.
13408 (VPMSUMB): Likewise.
13409 (VPMSUMH): Likewise.
13410 (VPMSUMW): Likewise.
13411 (VPERMXOR_V2DI): Likewise.
13412 (VPERMXOR_V4SI: Likewise.
13413 (VPERMXOR_V8HI: Likewise.
13414 (VPERMXOR_V16QI: Likewise.
13415 (VSHASIGMAW): Likewise.
13416 (VSHASIGMAD): Likewise.
13417 (VPMSUM): Likewise.
13418 (VPERMXOR): Likewise.
13419 (VSHASIGMA): Likewise.
13421 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
13422 __CRYPTO__ if the crypto instructions are available.
13423 (altivec_overloaded_builtins): Add support for overloaded power8
13426 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
13427 support for power8 crypto builtins.
13428 (builtin_function_type): Likewise.
13429 (altivec_init_builtins): Add support for builtins that take vector
13430 long long (V2DI) arguments.
13432 * config/rs6000/crypto.md: New file, define power8 crypto
13435 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
13436 Pat Haugen <pthaugen@us.ibm.com>
13437 Peter Bergner <bergner@vnet.ibm.com>
13439 * doc/invoke.texi (Option Summary): Add power8 options.
13440 (RS/6000 and PowerPC Options): Likewise.
13442 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
13443 constraints.md instead of rs6000.h. Reorder w* constraints. Add
13444 wm, wn, wr documentation.
13446 * config/rs6000/constraints.md (wm): New constraint for VSX
13447 registers if direct move instructions are enabled.
13448 (wn): New constraint for no registers.
13449 (wq): New constraint for quad word even GPR registers.
13450 (wr): New constraint if 64-bit instructions are enabled.
13451 (wv): New constraint if power8 vector instructions are enabled.
13452 (wQ): New constraint for quad word memory locations.
13454 * config/rs6000/predicates.md (const_0_to_15_operand): New
13455 constraint for 0..15 for crypto instructions.
13456 (gpc_reg_operand): If VSX allow registers in VSX registers as well
13457 as GPR and floating point registers.
13458 (int_reg_operand): New predicate to match only GPR registers.
13459 (base_reg_operand): New predicate to match base registers.
13460 (quad_int_reg_operand): New predicate to match even GPR registers
13461 for quad memory operations.
13462 (vsx_reg_or_cint_operand): New predicate to allow vector logical
13463 operations in both GPR and VSX registers.
13464 (quad_memory_operand): New predicate for quad memory operations.
13465 (reg_or_indexed_operand): New predicate for direct move support.
13467 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
13468 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
13469 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
13470 (POWERPC_MASKS): Add power8 options.
13471 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
13474 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
13475 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
13477 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
13478 (-mpower8-fusion): New power8 options.
13479 (-mpower8-fusion-sign): Likewise.
13480 (-mpower8-vector): Likewise.
13481 (-mcrypto): Likewise.
13482 (-mdirect-move): Likewise.
13483 (-mquad-memory): Likewise.
13485 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
13486 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
13488 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
13489 (rs6000_debug_vector_unit): Add p8_vector.
13490 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
13491 definitions. Also print fusion state.
13492 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
13493 (rs6000_builtin_mask_calculate): Add power8 builtin support.
13494 (rs6000_option_override_internal): Add support for power8.
13495 (rs6000_common_init_builtins): Add debugging for skipped builtins
13496 if -mdebug=builtin.
13497 (rs6000_adjust_cost): Add power8 support.
13498 (rs6000_issue_rate): Likewise.
13499 (insn_must_be_first_in_group): Likewise.
13500 (insn_must_be_last_in_group): Likewise.
13501 (force_new_group): Likewise.
13502 (rs6000_register_move_cost): Likewise.
13503 (rs6000_opt_masks): Likewise.
13505 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
13506 power8 capable assembler, default to power7 options.
13507 (TARGET_DIRECT_MOVE): Likewise.
13508 (TARGET_CRYPTO): Likewise.
13509 (TARGET_P8_VECTOR): Likewise.
13510 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
13511 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
13512 (VECTOR_MEM_P8_VECTOR_P): Likewise.
13513 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
13514 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
13515 (TARGET_XSCVDPSPN): Likewise.
13516 (TARGET_XSCVSPDPN): Likewsie.
13517 (TARGET_SYNC_HI_QI): Likewise.
13518 (TARGET_SYNC_TI): Likewise.
13519 (MASK_CRYPTO): Likewise.
13520 (MASK_DIRECT_MOVE): Likewise.
13521 (MASK_P8_FUSION): Likewise.
13522 (MASK_P8_VECTOR): Likewise.
13523 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
13524 temporary used by some of the direct move instructions to get two FP
13525 temporary registers does not force creation of a stack frame.
13526 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
13527 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
13528 that any VSX registers are tieable, even if they are also an
13529 Altivec vector mode.
13530 (r6000_reg_class_enum): Add wm, wr, wv constraints.
13531 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
13532 (RS6000_BTM_CRYPTO): Likewise.
13533 (RS6000_BTM_COMMON): Likewise.
13535 * config/rs6000/rs6000.md (cpu attribute): Add power8.
13536 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
13537 (enum rs6000_vector): Add power8 vector support.
13539 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13543 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
13544 (any_sibcall_could_use_r3): this and handle indirect calls.
13545 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
13547 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13549 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
13551 2013-05-22 Richard Biener <rguenther@suse.de>
13553 PR middle-end/57349
13554 * profile.c (branch_prob): Do not split blocks that are
13555 abnormally receiving from ECF_RETURNS_TWICE functions.
13557 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
13559 * recog.c (offsettable_address_addr_space_p): Fix calculation of
13560 address mode. Move pointer mode initialization to the same place.
13562 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
13564 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
13565 while it has any effect.
13567 2013-05-21 Easwaran Raman <eraman@google.com>
13569 PR tree-optimization/57322
13570 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
13571 UID of the statement added to the BB to be 1.
13573 2013-05-21 Jakub Jelinek <jakub@redhat.com>
13575 PR tree-optimization/57331
13576 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
13577 of conversion from pointer type to integral type with integer.
13579 2013-05-21 Martin Jambor <mjambor@suse.cz>
13582 * ipa-prop.c (ipa_read_node_info): Process param_used and
13583 controlled_uses in the same order as when writing.
13585 2013-05-21 Magnus Granberg <baldrick@free.fr>
13588 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
13590 2013-05-21 Richard Biener <rguenther@suse.de>
13592 PR tree-optimization/57318
13593 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
13594 estimate stmts with side-effects as likely eliminated.
13596 2013-05-21 Richard Biener <rguenther@suse.de>
13598 PR tree-optimization/57330
13599 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
13600 preserve the call stmts fntype.
13602 2013-05-21 Richard Biener <rguenther@suse.de>
13604 PR tree-optimization/57303
13605 * tree-ssa-sink.c (statement_sink_location): Improve killing
13606 stmt detection and properly handle self-assignments.
13608 2013-05-21 Christian Bruel <christian.bruel@st.com>
13610 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
13611 spanning registers. LEAF_REG_REMAP is supported only for contiguous
13612 registers. Set register size out of the PARALLEL loop.
13614 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
13617 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
13618 (*fmasf4, *fmasf4_media): New insns.
13620 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
13622 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
13623 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
13624 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
13625 (mips_idiv_insns): Update the comments to say that the returned
13626 instruction counts are in units of BASE_INSN_LENGTH.
13627 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
13628 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
13629 using 2 rather than 4 as the length of indirect MIPS16 and
13630 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
13631 length of a NOP. Don't divide MIPS16 lengths by 2.
13632 (mips16_split_long_branches): Assume a branch is long if the
13633 length is greater than 4 rather than 8.
13634 * config/mips/mips.md (length): Give MIPS16 lengths directly,
13635 rather than multiplying them by 2. Multiply instruction counts
13636 by BASE_INSN_LENGTH rather than 4.
13637 (*jump_mips16, tls_get_tp_mips16_<mode>)
13638 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
13640 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
13642 * config/mips/mips.md (extended_mips16): Remove branch case.
13643 (length): Remove duplicated extended_mips16 test.
13645 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
13647 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
13649 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
13651 * recog.h (Recog_data): Rename to...
13652 (recog_data_d): ...this.
13653 (recog_data): Update accordingly.
13654 * recog.c (recog_data): Likewise.
13655 * reload.c (save_recog_data): Likewise.
13656 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
13657 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
13659 2013-05-17 Julian Brown <julian@codesourcery.com>
13661 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
13662 found in a REG_EQUAL note, invalidate it.
13664 2013-05-17 Easwaran Raman <eraman@google.com>
13666 * tree-ssa-reassoc.c (find_insert_point): New function.
13667 (insert_stmt_after): Likewise.
13668 (get_def_stmt): Likewise.
13669 (ensure_ops_are_available): Likewise.
13670 (not_dominated_by): Likewise.
13671 (rewrite_expr_tree): Do not move statements beyond what is
13672 necessary. Remove call to swap_ops_for_binary_stmt...
13673 (reassociate_bb): ... and move it here.
13674 (build_and_add_sum): Assign UIDs for new statements.
13675 (linearize_expr): Likewise.
13676 (do_reassoc): Renumber gimple statement UIDs.
13678 2013-05-17 Jan Hubicka <jh@suse.cz>
13680 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
13682 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
13683 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
13685 (output_weakrefs): Update.
13687 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
13688 Martin Jambor <mjambor@suse.cz>
13690 PR middle-end/57276
13691 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
13692 value that corresponds to the given aggval is found in values vector.
13694 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
13696 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
13697 sse, sse2, sse3, ssse3 and sse4a flags to options.
13699 2013-05-17 David Malcolm <dmalcolm@redhat.com>
13701 * gengtype-state.c: (s_expr_writer): New class, to handle
13702 prettifying of output layout of s-expressions.
13703 (state_writer): New class, to write out gtype.state.
13704 (state_written_type_count): Move this variable into member data of
13706 (s_expr_writer::s_expr_writer): New code: constructor for new class
13707 (state_writer::state_writer(): ditto
13708 (s_expr_writer::write_new_line): New function
13709 (s_expr_writer::write_any_indent): ditto
13710 (s_expr_writer::begin_s_expr): ditto
13711 (s_expr_writer::end_s_expr): ditto
13712 (write_state_fileloc): convert to method of state_writer...
13713 (state_writer:: write_state_fileloc): ...and use methods of
13714 s_expr_writer to write indentation into the gtype.state output file
13715 to visually represent the hierarchical structure of the list
13717 (write_state_fields): ditto, renaming to...
13718 (state_writer::write_state_fields)
13719 (write_state_a_string): ditto, renaming to...
13720 (state_writer::write_state_a_string)
13721 (write_state_string_option): ditto, renaming to...
13722 (state_writer::write_state_string_option)
13723 (write_state_type_option): ditto, renaming to...
13724 (state_writer::write_state_type_option)
13725 (write_state_nested_option): ditto, renaming to...
13726 (state_writer::write_state_nested_option)
13727 (write_state_option): ditto, renaming to...
13728 (state_writer::write_state_option)
13729 (write_state_options): ditto, renaming to...
13730 (state_writer::write_state_options)
13731 (write_state_lang_bitmap): ditto, renaming to...
13732 (state_writer::write_state_lang_bitmap)
13733 (write_state_version): ditto, renaming to...
13734 (state_writer::write_state_version)
13735 (write_state_scalar_type): ditto, renaming to...
13736 (state_writer::write_state_scalar_type)
13737 (write_state_string_type): ditto, renaming to...
13738 (state_writer::write_state_string_type)
13739 (write_state_undefined_type): ditto, renaming to...
13740 (state_writer::write_state_undefined_type)
13741 (write_state_struct_union_type): ditto, renaming to...
13742 (state_writer::write_state_struct_union_type)
13743 (write_state_struct_type): ditto, renaming to...
13744 (state_writer::write_state_struct_type)
13745 (write_state_user_struct_type): ditto, renaming to...
13746 (state_writer::write_state_user_struct_type)
13747 (write_state_lang_struct_type): ditto, renaming to...
13748 (state_writer::write_state_lang_struct_type)
13749 (write_state_param_struct_type): ditto, renaming to...
13750 (state_writer::write_state_param_struct_type)
13751 (write_state_pointer_type): ditto, renaming to...
13752 (state_writer::write_state_pointer_type)
13753 (write_state_array_type): ditto, renaming to...
13754 (state_writer::write_state_array_type)
13755 (write_state_gc_used): ditto, renaming to...
13756 (state_writer::write_state_gc_used)
13757 (write_state_common_type_content): ditto, renaming to...
13758 (state_writer::write_state_common_type_content)
13759 (write_state_type): ditto, renaming to...
13760 (state_writer::write_state_type)
13761 (write_state_pair_list): ditto, renaming to...
13762 (state_writer::write_state_pair_list)
13763 (write_state_pair): ditto, renaming to...
13764 (state_writer::write_state_pair)
13765 (write_state_typedefs): ditto, renaming to...
13766 (state_writer::write_state_typedefs)
13767 (write_state_structures): ditto, renaming to...
13768 (state_writer::write_state_structures)
13769 (write_state_param_structs): ditto, renaming to...
13770 (state_writer::write_state_param_structs)
13771 (write_state_variables): ditto, renaming to...
13772 (state_writer::write_state_variables)
13773 (write_state_srcdir): ditto, renaming to...
13774 (state_writer::write_state_srcdir)
13775 (write_state_files_list): ditto, renaming to...
13776 (state_writer::write_state_files_list)
13777 (write_state_languages): ditto, renaming to...
13778 (state_writer::write_state_languages)
13779 (write_state): create a state_writer instance and use it when
13780 writing out the state file
13782 2013-05-17 Mike Stump <mikestump@comcast.net>
13784 PR rtl-optimization/57304
13785 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
13786 accessing DF_REF_REAL_LOC.
13788 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13790 PR rtl-optimization/57281
13791 PR rtl-optimization/57300
13792 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
13793 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
13794 what the other splitter did if the registers are dead.
13796 2013-05-17 Richard Biener <rguenther@suse.de>
13798 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
13801 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13803 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
13806 2013-05-17 Marek Polacek <polacek@redhat.com>
13808 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
13809 length when doing non-zero store of storing '\0' to '\0'.
13811 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13813 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
13814 vect_external_def oprnd1 with loop_vinfo, try to emit
13815 optional cast, negation and and stmts on the loop preheader
13816 edge instead of into the pattern def seq.
13818 PR tree-optimization/57051
13819 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
13820 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
13822 2013-05-16 Nick Clifton <nickc@redhat.com>
13824 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
13825 (rl78_is_naked_func): New function.
13826 (rl78_expand_prologue): Skip prologue generation for naked functions.
13827 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
13828 * doc/extend.texi (naked): Add RL78 to the list of processors
13829 that supports this attribute.
13831 2013-05-16 Jeff Law <law@redhat.com>
13833 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
13835 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
13837 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
13838 cache parameters using detect_caches_amd also for CYRIX,
13839 NSC and TM2 signatures.
13841 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
13842 Dzianis Kahanovich <mahatma@eu.by>
13846 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
13847 VIA/Centaur processors and determine their cache parameters
13848 using detect_caches_amd.
13850 2013-05-16 Teresa Johnson <tejohnson@google.com>
13852 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
13853 (rtl_verify_edges): New function.
13854 (rtl_verify_bb_insns): Ditto.
13855 (rtl_verify_bb_pointers): Ditto.
13856 (rtl_verify_bb_insn_chain): Ditto.
13857 (rtl_verify_fallthru): Ditto.
13858 (rtl_verify_bb_layout): Ditto.
13859 (rtl_verify_flow_info_1): Outline checks into new functions.
13860 (rtl_verify_flow_info): Ditto.
13862 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
13864 * cfghooks.c (copy_bbs): Add update_dominance argument.
13865 * cfghooks.h (copy_bbs): Update prototype.
13866 * tree-cfg.c (gimple_duplicate_sese_region):
13867 Add update_dominance argument.
13868 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
13869 * tree-ssa-loop-ch.c (copy_loop_headers): Update
13870 gimple_duplicate_sese_region call.
13871 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
13872 Update copy_bbs call.
13873 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
13874 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
13876 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13878 * tree-vectorizer.h (NUM_PATTERNS): Increment.
13879 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
13880 vect_recog_rotate_pattern.
13881 (vect_recog_rotate_pattern): New function.
13883 2013-05-16 Jason Merrill <jason@redhat.com>
13885 * Makefile.in (LLINKER): New variable.
13886 (mostlyclean): Remove link mutex.
13887 * configure.ac: Handle --enable-link-mutex.
13888 * lock-and-run.sh: New script.
13890 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13893 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
13896 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13898 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
13900 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
13902 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
13903 * config/arm/arm.c (next_consecutive_mem): New function.
13904 (gen_movmem_ldrd_strd): Likewise.
13905 * config/arm/arm.md (movmemqi): Update condition and code.
13906 (unaligned_loaddi, unaligned_storedi): New patterns.
13908 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13910 * config.gcc: Obsolete *-*-solaris2.9*.
13911 * doc/install.texi (Specific, *-*-solaris2*): Document it.
13913 2013-05-16 Richard Biener <rguenther@suse.de>
13915 * passes.c (init_optimization_passes): Move pass_parallelize_loops
13916 earlier, after GRAPHITE transforms and IV canonicalization.
13918 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13920 * omp-low.c (extract_omp_for_data): For collapsed loops,
13921 if at least one of the loops is known at compile time to
13922 iterate zero times, set count to 0.
13923 (expand_omp_regimplify_p): New function.
13924 (expand_omp_for_generic): For collapsed loops, if at least
13925 one of the loops isn't known to iterate at least once,
13926 add runtime check with setting count to 0.
13927 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
13928 For unsigned types if it isn't known at compile time that
13929 the loop will iterate at least once, add runtime check to bypass
13930 the whole loop if initial condition isn't true.
13932 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
13934 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
13936 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
13938 PR middle-end/57286
13939 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
13940 transformations to avoid an infinite loop.
13942 2013-05-16 Marek Polacek <polacek@redhat.com>
13944 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
13946 2013-05-15 Leif Ekblad <leif@rdos.net>
13948 * config/i386/i386.c (ix86_decompose_address): Use
13949 DEFAULT_TLS_SEG_REG to access TLS segment register.
13950 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
13951 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
13952 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
13954 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
13957 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
13958 sibling calls to functions that would normally be lazily bound,
13959 unless $gp is call-clobbered.
13961 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
13963 * config/i386/i386.c (ix86_option_override_internal): Update
13964 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
13965 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
13966 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
13967 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
13969 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
13971 2013-05-15 Andreas Schwab <schwab@suse.de>
13973 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
13974 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
13977 2013-05-15 Teresa Johnson <tejohnson@google.com>
13979 * loop-unroll.c (report_unroll_peel): Check decision before
13980 emitting unroll/peel message.
13982 2013-05-15 Teresa Johnson <tejohnson@google.com>
13984 * function.h (has_bb_partition): New rtl_data flag.
13985 (bb_reorder_complete): Ditto.
13986 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
13987 instead of flag_reorder_blocks_and_partition.
13988 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
13989 with some enhancements.
13990 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
13991 * bb-reorder.c (connect_traces): Check for has_bb_partition
13992 instead of flag_reorder_blocks_and_partition.
13993 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
13994 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
13995 verify_hot_cold_block_grouping.
13996 (partition_hot_cold_basic_blocks): Set has_bb_partition.
13998 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14001 * config/arm/predicates.md (call_insn_operand): New predicate.
14002 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
14003 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
14004 if insn is not a tail call.
14005 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
14007 * config/arm/arm.h (enum reg_class): New caller save register class.
14008 (REG_CLASS_NAMES): Likewise.
14009 (REG_CLASS_CONTENTS): Likewise.
14010 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
14013 2013-05-15 Richard Biener <rguenther@suse.de>
14015 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
14016 of MSG_OPTIMIZED_LOCATIONS.
14017 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
14018 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
14020 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
14021 of MSG_OPTIMIZED_LOCATIONS.
14022 (execute_vect_slp): Likewise.
14023 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
14024 (vect_create_cond_for_alias_checks): Likewise.
14025 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
14026 (vect_recog_widen_mult_pattern): Likewise.
14027 (vect_recog_widen_sum_pattern): Likewise.
14028 (vect_recog_over_widening_pattern): Likewise.
14029 (vect_recog_widen_shift_pattern): Likewise.
14030 (vect_recog_vector_vector_shift_pattern): Likewise.
14031 (vect_recog_divmod_pattern): Likewise.
14032 (vect_recog_mixed_size_cond_pattern): Likewise.
14033 (vect_recog_bool_pattern): Likewise.
14034 (vect_pattern_recog_1): Likewise.
14036 2013-05-15 Martin Jambor <mjambor@suse.cz>
14038 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
14039 non-functions to builtin_unreachable.
14040 * ipa-inline-transform.c (inline_call): Do not assert estimates were
14041 correct when new direct edges were discovered.
14043 2013-05-15 Martin Jambor <mjambor@suse.cz>
14045 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
14046 header, print symbol order instead of node uid, print more information
14047 about indirect edge targets.
14048 (ipa_make_edge_direct_to_target): Print symbol order instead of node
14050 (ipa_make_edge_direct_to_target): Likewise.
14051 (remove_described_reference): Likewise.
14052 (propagate_controlled_uses): Likewise.
14053 (ipa_print_node_params): Also print symbol order.
14054 (ipcp_transform_function): Print symbol order instead of node uids.
14055 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
14056 (cgraph_get_create_real_symbol_node): Likewise.
14057 * ipa-cp.c (print_lattice): Likewise.
14058 (print_all_lattices): Likewise.
14059 (determine_versionability): Likewise.
14060 (initialize_node_lattices): Likewise.
14061 (estimate_local_effects): Likewise.
14062 (update_profiling_info): Likewise.
14063 (create_specialized_node): Likewise.
14064 (perhaps_add_new_callers): Likewise.
14065 (decide_about_value): Likewise.
14066 (decide_whether_version_node): Likewise.
14067 (identify_dead_nodes): Likewise.
14068 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
14069 (dump_inline_summary): Likewise.
14070 (estimate_node_size_and_time): Likewise.
14071 (inline_analyze_function): Likewise.
14072 * ipa-inline.c (report_inline_failed_reason): Likewise.
14073 (want_early_inline_function_p): Likewise.
14074 (edge_badness): Likewise.
14075 (update_edge_key): Likewise.
14076 (inline_small_functions): Likewise. Add dumping of order to two other
14078 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
14079 instead of node uids.
14080 (propagate_pure_const): Likewise.
14081 (propagate_pure_const): Likewise.
14082 * ipa-utils.c (dump_cgraph_node_set): Likewise.
14083 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
14084 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
14086 * tree-pretty-print.c (dump_function_header): Likewise.
14087 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
14088 Print symbol order instead of node uids.
14090 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14092 * config/s390/s390.c (s390_register_move_cost): Don't impose the
14093 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
14095 2013-05-15 Richard Biener <rguenther@suse.de>
14097 PR tree-optimization/57275
14098 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
14099 return value for fail to do runtime alias checks for gather loads.
14101 2013-05-15 Jan Hubicka <jh@suse.cz>
14105 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
14106 weakrefs are not external.
14107 (lto_symtab_merge_decls): Fix thinko when dealing with
14108 non-lto_symtab decls.
14109 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
14110 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
14111 * varpool.c (dump_varpool_node): Dump more flags.
14113 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
14115 * config/i386/i386.c (processor_alias_table): Add instruction
14116 FSGSBASE for AMD bdver3 architecture.
14118 2013-05-14 Jakub Jelinek <jakub@redhat.com>
14120 * tree.c (warn_deprecated_use): Print file:line using locus color.
14121 * diagnostic.c (diagnostic_report_current_module): Print file:line
14122 and file:line:column using locus color.
14124 2013-05-14 Mike Stump <mikestump@comcast.net>
14126 * gdbinit.in: Add __null.
14128 2013-05-14 Mike Stump <mikestump@comcast.net>
14130 * recog.h: Rename struct recog_data to Recog_data.
14131 * recog.c: Likewise.
14132 * reload.c (can_reload_into): Likewise.
14133 * config/picochip/picochip.c: Likewise.
14135 2013-05-14 Mike Stump <mikestump@comcast.net>
14137 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
14139 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
14141 * resource.h (struct resources): Remove unch_memory member.
14142 (CLEAR_RESOURCE): Don't clear unch_memory.
14143 * resource.c (mark_referenced_resources): Don't set it.
14144 (mark_set_resources): Likewise.
14145 (mark_target_live_regs): Don't clear it.
14146 (init_resource_info): Likewise.
14147 * reorg.c (resource_conflicts_p): Don't compare it.
14148 (redundant_insn): Don't set it.
14150 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
14152 * emit-rtl.c (next_label): Remove unused function.
14153 (skip_consecutive_labels, link_cc0_insns): Move to ...
14154 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
14155 only place where these functions are used, and make them static.
14157 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
14159 * fold-const.c (fold_negate_expr): Handle vectors.
14160 (fold_truth_not_expr): Make it static.
14161 (fold_invert_truthvalue): New static function.
14162 (invert_truthvalue_loc): Handle vectors. Do not call
14163 fold_truth_not_expr directly.
14164 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
14165 <TRUTH_NOT_EXPR>: Do not cast to boolean.
14166 (fold_comparison): Handle vector constants.
14167 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
14168 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
14169 * tree.h (fold_truth_not_expr): Remove declaration.
14171 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
14173 * config/aarch64/aarch64-simd.md
14174 (aarch64_vcond_internal<mode>): Rename to...
14175 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
14176 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
14177 float modes. Clarify all iterator modes.
14178 (vcond<mode><mode>): Use new name for vcond expanders.
14179 (vcond<v_cmp_result><mode>): Likewise.
14180 (vcondu<mode><mode>: Likewise.
14181 * config/aarch64/iterators.md (VDQF_COND): New.
14183 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
14186 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
14187 variable for the shift amount. Check that we shift by non-negative
14190 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
14193 * config/arm/arm.h (EPILOGUE_USES): Only return true
14194 for LR_REGNUM after epilogue_completed.
14196 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
14198 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
14199 is error_mark_node.
14201 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14204 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
14205 and Solaris 11+/x86 with gld.
14206 * configure: Regenerate.
14208 2013-05-14 Jakub Jelinek <jakub@redhat.com>
14210 * expmed.c (expand_shift_1): Canonicalize rotates by
14211 constant bitsize / 2 to bitsize - 1.
14212 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
14213 case ROTATERT>: Likewise.
14216 2013-05-10 Jakub Jelinek <jakub@redhat.com>
14218 * config/i386/i386.md (rotateinv): New code attr.
14219 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
14220 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
14221 roll $31, %eax, etc.
14223 2013-05-14 Richard Biener <rguenther@suse.de>
14225 PR middle-end/57235
14226 * tree-eh.c (sink_clobbers): Give up for successors with
14227 multiple predecessors and no virtual uses.
14229 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
14231 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
14232 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
14234 2013-05-14 Jakub Jelinek <jakub@redhat.com>
14236 PR middle-end/57251
14237 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
14238 the case when both op0 and op1 have VOIDmode.
14240 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14242 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
14243 in multiply-accumulate mode.
14245 2013-05-13 Guozhi Wei <carrot@google.com>
14247 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
14249 2013-05-13 Kai Tietz <ktietz@redhat.com>
14252 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
14253 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
14254 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
14255 * config/i386/i386.c (ix86_option_override_internal): Likewise.
14256 (ix86_expand_prologue): Likewise.
14257 (ix86_expand_split_stack_prologue): Likewise.
14258 (legitimate_pic_address_disp_p): Likewise.
14259 (legitimize_pic_address): Likewise.
14260 (legitimize_tls_address): Likewise.
14261 (legitimize_pe_coff_symbol): Likewise.
14262 (output_pic_addr_const): Likewise.
14263 (construct_plt_address): Likewise.
14264 (ix86_expand_call): Likewise.
14265 (x86_output_mi_thunk): Likewise.
14266 (x86_function_profiler): Likewise.
14268 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
14270 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
14271 similar switch cases.
14272 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
14273 (aarch64_simd_mov_to_<mode>low): Delete.
14274 (aarch64_simd_mov_to_<mode>high): Delete.
14275 (move_lo_quad_<mode>): Add w<-r alternative.
14276 (aarch64_simd_move_hi_quad_<mode>): Likewise.
14277 (aarch64_simd_mov_from_*): Update type attribute.
14278 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
14281 2013-05-13 Jan Hubicka <jh@suse.cz>
14283 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
14284 * config/i386/i386.c (ix86_compute_frame_layout,
14285 ix86_expand_epilogue, emit_i387_cw_initialization,
14286 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
14287 ix86_local_alignment): Fix use of size/speed predicates.
14289 2013-05-13 Jakub Jelinek <jakub@redhat.com>
14291 PR tree-optimization/45216
14292 PR tree-optimization/57157
14293 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
14294 the (-Y) & (B - 1) variant if OP is |.
14295 * expmed.c (expand_shift_1): For rotations by const0_rtx just
14296 return shifted. Use (-op1) & (prec - 1) as other_amount
14297 instead of prec - op1.
14299 2013-05-13 Martin Jambor <mjambor@suse.cz>
14301 PR middle-end/42371
14302 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
14303 (ipa_constant_data): New type.
14304 (ipa_jump_func): Use ipa_constant_data to hold information about
14305 constant jump functions.
14306 (ipa_get_jf_constant): Adjust to jump function type changes.
14307 (ipa_get_jf_constant_rdesc): New function.
14308 (ipa_param_descriptor): New field controlled_uses.
14309 (ipa_get_controlled_uses): New function.
14310 (ipa_set_controlled_uses): Likewise.
14311 * ipa-ref.h (ipa_find_reference): Declare.
14312 * ipa-prop.c (ipa_cst_ref_desc): New type.
14313 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
14315 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
14316 New parameter cs. Adjust all callers.
14317 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
14318 (remove_described_reference): New function.
14319 (jfunc_rdesc_usable): Likewise.
14320 (try_make_edge_direct_simple_call): Decrement controlled use count,
14321 attempt to remove reference if it hits zero.
14322 (combine_controlled_uses_counters): New function.
14323 (propagate_controlled_uses): Likewise.
14324 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
14325 (ipa_edge_duplication_hook): Duplicate reference descriptions.
14326 (ipa_print_node_params): Print described use counter.
14327 (ipa_write_jump_function): Adjust to jump function type changes.
14328 (ipa_read_jump_function): New parameter CS, pass it to
14329 ipa_set_jf_constant. Adjust caller.
14330 (ipa_write_node_info): Stream controlled use count
14331 (ipa_read_node_info): Likewise.
14332 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
14334 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
14335 count. Remove cloning-added reference if it reaches zero.
14336 * ipa-ref.c (ipa_find_reference): New function.
14338 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
14340 * config/i386/i386.c (processor_target_table): Modified default
14341 alignment values for AMD BD and BT architectures.
14343 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
14345 * tree-vect-generic.c (uniform_vector_p): Move ...
14346 * tree.c (uniform_vector_p): ... here.
14347 * tree.h (uniform_vector_p): Declare it.
14348 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
14351 2013-05-13 Jakub Jelinek <jakub@redhat.com>
14353 PR tree-optimization/57230
14354 * tree-ssa-strlen.c (handle_char_store): Record length for
14355 array store from STRING_CST.
14357 PR tree-optimization/57230
14358 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
14361 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
14363 * config/epiphany/epiphany.c (epiphany_init): Check size of
14364 NUM_MODES_FOR_MODE_SWITCHING.
14365 (epiphany_expand_prologue):
14366 Remove CONFIG_REGNUM initial value handling code.
14367 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
14368 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
14369 (emit_set_fp_mode, epiphany_mode_after): Likewise.
14370 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
14371 Don't return 1 for FP_MODE_NONE.
14372 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
14373 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
14374 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
14375 * config/epiphany/epiphany.md (save_config): New pattern.
14377 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
14379 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
14381 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
14383 * config/i386/i386.md (memory): Handle sseishft1.
14384 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
14385 (*vec_extractv2di_1): Ditto.
14387 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
14389 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
14392 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
14394 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
14395 Add mthumb/march=armv7-a multilib.
14396 Add mthumb/march=armv7-r multilib.
14397 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
14399 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
14401 * config/v850/t-rtems: Add more multilibs.
14403 2013-05-10 Richard Biener <rguenther@suse.de>
14405 PR tree-optimization/57214
14406 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
14407 not propagate from SSA names that occur in abnormal PHI nodes.
14409 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
14411 * stor-layout.c (element_precision): New function.
14412 * machmode.h (element_precision): Declare it.
14413 * tree.c (build_minus_one_cst): New function.
14414 (element_precision): Likewise.
14415 * tree.h (build_minus_one_cst): Declare new function.
14416 (element_precision): Likewise.
14417 * fold-const.c (operand_equal_p): Use element_precision.
14418 (fold_binary_loc): Handle vector types.
14419 * convert.c (convert_to_integer): Use element_precision.
14420 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
14423 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
14425 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
14426 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
14427 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
14428 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
14429 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
14430 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
14431 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
14432 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
14433 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
14434 (Uuw8): New constraints.
14435 (Usb4): Move into alphabetical order.
14436 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
14437 (sd8_operand, ub8_operand, uw8_operand): New predicates.
14438 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
14439 previously unnamed patterns.
14440 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
14441 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
14442 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
14443 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
14444 of set_attr_alternative/if_then_else. Use extended_mips16 instead
14445 of specific lengths.
14447 2013-05-10 Jakub Jelinek <jakub@redhat.com>
14449 * config/i386/i386.md (rotateinv): New code attr.
14450 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
14451 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
14452 roll $31, %eax, etc.
14454 PR tree-optimization/45216
14455 PR tree-optimization/57157
14456 * tree-ssa-forwprop.c (simplify_rotate): New function.
14457 (ssa_forward_propagate_and_combine): Call it.
14459 2013-05-10 Richard Biener <rguenther@suse.de>
14461 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
14462 disable peeling when we version for aliasing.
14463 (vector_alignment_reachable_p): Honor explicit user alignment.
14464 (vect_supportable_dr_alignment): Likewise.
14465 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
14466 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
14467 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
14468 then peeling to arrange for the cost-model check to come first.
14470 2013-05-10 Alan Modra <amodra@gmail.com>
14472 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
14473 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
14474 * configure: Regenerate.
14476 2013-05-10 Alan Modra <amodra@gmail.com>
14479 * varasm.c (default_elf_select_section): Move !DECL_P check..
14480 (get_named_section): ..to here before calling get_section_name.
14482 (default_section_type_flags): Add DECL_P check.
14483 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
14484 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
14486 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
14488 * config/epiphany/epiphany.c (epiphany_expand_prologue):
14489 When using gen_stack_adjust_str with a register offset, add a
14490 REG_FRAME_RELATED_EXPR note.
14492 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
14494 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
14495 (*vec_extractv4si_zext_mem): Ditto.
14496 (*vec_extractv2di): Add 0->x and x->x alternatives.
14497 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
14498 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
14500 2013-05-09 Jason Merrill <jason@redhat.com>
14502 N3639 C++1y VLA support
14503 * gimplify.c (gimplify_vla_decl): Don't touch an existing
14506 * tree.c (build_constructor_va): New.
14507 * tree.h: Declare it.
14509 2013-05-09 Martin Jambor <mjambor@suse.cz>
14512 * gimple-fold.c (canonicalize_constructor_val): Call
14513 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
14515 2013-05-09 Jan Hubicka <jh@suse.cz>
14516 Richard Biener <rguenther@suse.de>
14519 * symtab.c (symtab_make_decl_local): Do not add private names.
14521 2013-05-09 Jan Hubicka <jh@suse.cz>
14524 * symtab.c (insert_to_assembler_name_hash): Handle clones.
14525 (unlink_from_assembler_name_hash): Likewise.
14526 (symtab_prevail_in_asm_name_hash, symtab_register_node,
14527 symtab_unregister_node, symtab_initialize_asm_name_hash,
14528 change_decl_assembler_name): Update.
14530 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
14532 * config/aarch64/aarch64.md: New movtf split.
14533 (*movtf_aarch64): Update.
14534 (aarch64_movdi_tilow): Handle TF modes and rename to
14535 aarch64_movdi_<mode>low.
14536 (aarch64_movdi_tihigh): Handle TF modes and rename to
14537 aarch64_movdi_<mode>high
14538 (aarch64_movtihigh_di): Handle TF modes and rename to
14539 aarch64_mov<mode>high_di
14540 (aarch64_movtilow_di): Handle TF modes and rename to
14541 aarch64_mov<mode>low_di
14542 (aarch64_movtilow_tilow): Remove spurious whitespace.
14543 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
14545 (aarch64_print_operand): Update.
14547 2013-05-09 Alan Modra <amodra@gmail.com>
14549 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
14551 * configure: Regenerate.
14553 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
14555 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
14556 splitter preparation statements.
14557 * config/i386/sse.md (*vec_extract* splitters): Ditto.
14558 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
14561 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14563 * gimple-ssa-strength-reduction.c (count_candidates): Change
14564 return value to int.
14565 (analyze_candidates_and_replace): Change type of length to int.
14567 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
14569 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
14570 (*vec_extract<mode>): Use VI12_128 mode iterator.
14571 (*vec_extract<mode>_mem): Ditto.
14572 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
14575 2013-05-08 Diego Novillo <dnovillo@google.com>
14580 2012-08-17 Diego Novillo <dnovillo@google.com>
14583 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
14584 * config.in: Regenerate.
14585 * configure: Regenerate.
14586 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
14588 2013-05-08 Jan Hubicka <jh@suse.cz>
14591 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
14592 * cgraph.h (symtab_node_base): Add unique_name.
14593 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
14594 input_overwrite_node, input_varpool_node): Stream unique_name.
14595 * cgraphclones.c (cgraph_create_virtual_clone,
14596 cgraph_function_versioning): Set unique_name.
14597 * ipa.c (function_and_variable_visibility): Set unique_name.
14599 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14601 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
14602 (alloc_cand_and_find_basis): Restrict conditional candidate
14603 processing to CAND_MULTs.
14605 2013-05-08 Jan Hubicka <jh@suse.cz>
14608 lto-symtab.c (lto_symtab_symbol_p): New function.
14609 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
14610 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
14611 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
14612 Skip static symbols.
14614 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
14616 PR tree-optimization/57200
14617 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
14618 Only call inform if the preceding warning_at returns true.
14620 2013-05-07 Han Shen <shenhan@google.com>
14622 * cfgexpand.c (record_or_union_type_has_array_p): New function.
14623 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
14624 * common.opt (fstack-protector-strong): New option.
14625 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
14626 * doc/invoke.texi (Optimization Options): Document
14627 "-fstack-protector-strong".
14628 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
14630 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
14632 * config/mips/mips.c (mips_machine_reorg2): Return 0.
14634 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
14636 * ira.c (update_equiv_regs): Add insn having equiv memory even if
14637 it is not lhs of the insn.
14638 (setup_reg_equiv): Remove insn having equiv memory which it is not
14640 * lra-constraints.c (process_address): Try to improve generation
14641 code for address base + disp.
14642 (lra_constraints): Make correct the code for checking insn setting
14643 up backward equivalence. Remove insn only if it is in the init
14645 * lra-eliminations.c (update_reg_eliminate): Change return value.
14646 (lra_eliminate): Use the result.
14648 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
14650 * config/i386/sse.md (ssescalarnummask): New mode attribute.
14651 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
14652 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
14653 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
14654 register target operands.
14655 (*vec_extractv8hi_sse2): New pattern.
14656 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
14657 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
14658 (*vec_extract<mode>_mem): New insn and split pattern.
14660 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
14662 * config/arm/arm.c (arm_asan_shadow_offset): New function.
14663 (TARGET_ASAN_SHADOW_OFFSET): Define.
14664 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
14665 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
14667 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14669 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
14670 (incr_vec_index): Return -1 if increment not found.
14671 (create_add_on_incoming_edge): Assert if increment not found.
14672 (record_increment): Limit number of increments recorded.
14673 (all_phi_incrs_profitable): Return false if an increment not found.
14674 (replace_profitable_candidates): Don't process increments that were
14676 (analyze_candidates_and_replace): Limit size of incr_vec.
14678 2013-05-07 Richard Biener <rguenther@suse.de>
14680 * calls.c (special_function_p): setjmp-like functions are leaf.
14681 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
14682 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
14684 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
14686 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
14687 (aarch64_simd_mov<mode>): New expander.
14688 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
14689 (aarch64_simd_mov_to_<mode>high): Likewise.
14690 (aarch64_simd_mov_from_<mode>low): Likewise.
14691 (aarch64_simd_mov_from_<mode>high): Likewise.
14692 (aarch64_dup_lane<mode>): Update.
14693 (aarch64_dup_lanedi): New instruction pattern.
14694 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
14695 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
14697 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14699 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
14700 (replace_mult_candidate): Remove unnecessary argument; remove
14701 unnecessary parameter from call to introduce_cast_before_cand.
14702 (replace_unconditional_candidate): Remove unnecessary parameter
14703 from call to replace_mult_candidate.
14704 (replace_conditional_candidate): Likewise.
14705 (insert_initializers): Use make_temp_ssa_name.
14706 (introduce_cast_before_cand): Remove unnecessary argument; use
14707 make_temp_ssa_name.
14708 (replace_one_candidate): Remove unnecessary argument; remove
14709 unnecessary parameter from calls to introduce_cast_before_cand.
14710 (replace_profitable_candidates): Remove unnecessary parameters
14711 from calls to replace_one_candidate.
14713 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14715 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
14716 phi def as possibly hiding a basis for a CAND_ADD whose operands
14717 have been commuted in the analysis.
14718 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
14720 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14722 * config/aarch64/aarch64.md
14723 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
14724 shift value between 0-4.
14726 2013-05-07 Richard Biener <rguenther@suse.de>
14728 * double-int.h (rshift): New overload.
14729 * double-int.c (rshift): New function.
14730 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
14731 (create_reference_ops_from_ref): Remove.
14732 (vn_reference_insert): Use shared ops for constructing the
14733 reference and copy it.
14735 2013-05-07 Richard Biener <rguenther@suse.de>
14737 PR middle-end/57190
14738 * tree-eh.c (sink_clobbers): Properly propagate
14739 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
14741 2013-05-07 Jakub Jelinek <jakub@redhat.com>
14743 PR tree-optimization/57149
14744 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
14745 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
14746 collect_phi_def_edges, execute_late_warn_uninitialized): Use
14747 uninit_undefined_value_p instead of ssa_undefined_value_p.
14750 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
14751 for modifier == EXPAND_INITIALIZER.
14753 2013-05-07 Anton Blanchard <anton@samba.org>
14755 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
14756 for powerpc64 little endian.
14757 * configure: Regenerate.
14759 2013-05-06 Graham Stott <grahams@btinternet.com>
14761 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
14762 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
14763 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
14764 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
14766 2013-05-06 Graham Stott <grahams@btinternet.com>
14768 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
14769 codes which allow non-lvalues.
14771 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
14773 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
14774 components are all 1s.
14775 (integer_minus_onep): New function.
14776 * tree.h (integer_minus_onep): Declare it.
14777 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
14778 integer_minus_onep instead of integer_all_onesp.
14780 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14783 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
14784 variations of these patterns.
14786 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
14788 * config/i386/i386.md (isa): Add x64_sse4 member.
14789 (enabled): Handle x64_sse4.
14790 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
14791 instruction for 64bit SSE4_1 targets. Update insn attributes.
14792 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
14793 instruction for SSE4_1 targets. Update insn attributes.
14794 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
14795 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
14796 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
14798 (*vec_extractv4si): Rename from *sse4_1_pextrd.
14799 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
14800 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
14802 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14805 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
14807 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
14809 * final.c (do_assembler_dialects): Don't handle curly braces and
14810 vertical bar escaped by % as dialect delimiters.
14811 (output_asm_insn): Print curly braces and vertical bar if escaped
14812 by % and ASSEMBLER_DIALECT defined.
14813 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
14814 * doc/tm.texi: Regenerated.
14816 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
14818 * config/mips/mips.c: Include tree-pass.h.
14819 (mips_reorg): Split in pre- and post-dbr_schedule parts.
14820 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
14821 (pass_mips_machine_reorg2): New machine specific pass.
14822 (insert_pass_mips_machine_reorg2): New pass plugin definition.
14823 (mips_option_override): Register the new pass.
14824 * rtl.h (cleanup_barriers): Remove prototype.
14825 (dbr_schedule): Likewise.
14826 * jump.c (cleanup_barriers): Make static.
14827 * reorg.c (dbr_schedule): Likewise.
14829 2013-05-06 Richard Biener <rguenther@suse.de>
14831 PR tree-optimization/57185
14832 * tree-parloops.c (add_field_for_reduction): Handle anonymous
14833 SSA names properly.
14835 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
14838 * config/i386/i386.c (add_parameter_dependencies): Add dependence
14839 between "first_arg" and "insn", not "last" and "insn".
14841 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14843 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
14844 (find_candidates_in_block): Re-enable slsr_process_phi.
14845 (create_phi_basis): Fix double counting of candidate adjustment.
14847 2013-05-06 Richard Biener <rguenther@suse.de>
14849 PR middle-end/57147
14850 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
14851 the edge is also fallthru, preserve it and just clear the
14853 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
14854 also complex, preserve that and just clear the fallthru flag.
14855 * tree-inline.c (update_ssa_across_abnormal_edges): Also
14856 update virtual operands.
14858 2013-05-06 Alan Modra <amodra@gmail.com>
14860 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
14861 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
14862 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
14863 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
14864 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14865 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
14867 2013-05-06 Alan Modra <amodra@gmail.com>
14869 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
14870 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
14871 (DEFAULT_ASM_ENDIAN): Define.
14872 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14873 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
14874 Update -K PIC clause from sysv4.h.
14875 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
14876 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
14878 2013-05-06 Alan Modra <amodra@gmail.com>
14880 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
14881 twice for little-endian.
14882 (ashrdi3_no_power, ashrdi3): Support little-endian.
14884 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14887 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
14888 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
14890 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
14891 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
14892 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
14895 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
14896 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14898 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
14899 * config/pa/pa.opt: Make mbig-switch a no-op.
14900 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
14901 (CASE_VECTOR_MODE): Always return SImode.
14902 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
14903 for the !TARGET_BIG_SWITCH case.
14904 * config/pa/pa-linux.h: Likewise.
14905 * config/pa/pa-openbsd.h: Likewise.
14906 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
14907 * config/pa/pa.md (short_jump): Remove define_insn.
14908 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
14909 (casesi0): Remove define_insn.
14910 (type): Remove btable_branch.
14911 (pa_combine_type): Likewise.
14912 (in_nullified_branch_delay): Likewise.
14913 (in_call_delay): Likewise.
14914 (define_delay): Likewise.
14915 (define_insn_reservation "Z3"): Likewise.
14916 (define_insn_reservation "Z4"): Likewise.
14917 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
14918 (pa_adjust_insn_length): Remove adjustment for btable branches.
14919 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
14922 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
14924 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
14925 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
14926 Add m->r,x alternatives.
14927 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
14928 splitters using SWI48x mode iterator.
14929 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
14930 TARGET_64BIT. Add m->x alternative.
14931 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
14932 Add o->x alternative. Enable for TARGET_SSE.
14933 (sse_storeq): Remove expander.
14934 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
14935 with memory input operand.
14936 (*vec_extractv2di_1 splitter): New.
14937 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
14938 * config/i386/i386.md (ssevecmodelower): New mode attribute.
14940 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14942 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
14943 (INT_LOWPART): Delete.
14944 (extract_MB): Adjust.
14945 (extract_ME): Adjust.
14946 (print_operand): Adjust.
14948 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14950 * config/rs6000/predicates.md (reg_or_add_cint_operand,
14951 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14952 (reg_or_logical_cint_operand, easy_fp_constant,
14953 logical_const_operand): Delete "CONST_DOUBLE" case.
14954 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
14955 "HOST_BITS_PER_WIDE_INT == 64" test.
14956 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
14957 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
14958 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
14959 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
14961 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
14962 CONST_DOUBLE DImode/VOIDmode case.
14963 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
14964 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14965 CONST_DOUBLE VOIDmode case.
14966 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14967 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
14968 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
14969 Delete CONST_DOUBLE case.
14970 (splitters for mov FMOVE64 const_double): Delete
14971 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14972 "HOST_BITS_PER_WIDE_INT >= 64" test.
14973 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
14975 (mov DI const_double): Delete.
14977 2013-05-04 Jakub Jelinek <jakub@redhat.com>
14979 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
14980 on op shows all bits zero in mode of a lowpart subreg, return zero.
14982 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
14985 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
14986 to save TFmode registers and DImode to save TImode registers for
14987 caller save operations.
14988 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
14989 mark being partially clobbered since they only use the first
14992 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
14993 and TDmode only use the upper 64-bits of each VSX register.
14995 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14997 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
14998 (find_candidates_in_block): Disable slsr_process_phi.
15000 2013-05-03 Guozhi Wei <carrot@google.com>
15002 * coverage.c (coverage_obj_init): Move the construction of gcov
15004 (build_init_ctor): ... here.
15006 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15008 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
15009 (slsr_cand_d): Redefine def_phi.
15010 (stride_status, phi_adjust_status, count_phis_status): New enums.
15011 (find_phi_def): New.
15012 (find_basis_for_base_expr): New.
15013 (find_basis_for_candidate): Handle hidden bases.
15014 (alloc_cand_and_find_basis): Handle phi candidates.
15015 (slsr_process_phi): New.
15016 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
15017 (create_mul_imm_cand): Likewise.
15018 (create_add_ssa_cand): Exclude phi base candidates.
15019 (create_add_imm_cand): Likewise.
15020 (slsr_process_cast): Likewise.
15021 (slsr_process_copy): Likewise.
15022 (find_candidates_in_block): Handle phi candidates.
15023 (dump_candidate): Likewise.
15024 (unconditional_cands): Delete.
15025 (unconditional_cands_with_known_stride_p): Delete.
15026 (phi_dependent_cand_p): New.
15027 (cand_increment): Handle phi-dependent candidates.
15028 (replace_dependent): Delete.
15029 (replace_mult_candidate): New.
15030 (replace_unconditional_candidate): New.
15031 (incr_vec_index): Move to avoid forward reference.
15032 (create_add_on_incoming_edge): New.
15033 (create_phi_basis): New.
15034 (replace_dependents): Delete.
15035 (replace_conditional_candidate): New.
15036 (phi_add_costs): New.
15037 (replace_uncond_cands_and_profitable_phis): New.
15038 (record_increment): Handle phi adjustments.
15039 (record_phi_increments): New.
15040 (record_increments): Handle phi adjustments.
15041 (phi_incr_cost): New.
15042 (lowest_cost_path): Handle phis.
15043 (total_savings): Likewise.
15044 (analyze_increments): Likewise.
15045 (ncd_with_phi): New.
15046 (ncd_of_cand_and_phis): New.
15047 (nearest_common_dominator_for_cands): Handle phi increments.
15048 (all_phi_incrs_profitable): New.
15049 (replace_profitable_candidates): Handle phi-dependent candidates.
15050 (analyze_candidates_and_replace): Likewise.
15052 2013-05-03 Teresa Johnson <tejohnson@google.com>
15055 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
15056 do not exceed REG_BR_PROB_BASE.
15058 2013-05-03 Jeff Law <law@redhat.com>
15060 PR tree-optimization/57144
15061 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
15062 operand of the condition will bit into the new type when eliminating
15063 a cast feeding a condition.
15065 2013-05-03 Jakub Jelinek <jakub@redhat.com>
15067 PR rtl-optimization/57130
15068 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
15069 of COMPARE as in_code to the recursive call if needed.
15071 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
15073 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
15074 (enabled): Handle new members.
15075 * config/i386/sse.md (*vec_concatv2si): Merge from
15076 *vec_concatv2si_sse2 and vec_concatv2si_sse.
15077 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
15079 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
15081 PR tree-optimization/57027
15082 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
15083 for fnms opportunity, check we got the prerequisite kind
15084 of tree / gimple before using accessor functions.
15086 2013-05-03 Richard Biener <rguenther@suse.de>
15088 * double-int.h (lshift): New overload without precision
15089 and arith argument.
15090 (operator *=, operator +=, operator -=): Move ...
15091 * double-int.c (operator *=, operator +=, operator -=): ... here
15092 and implement more efficiently.
15093 (mul_double_with_sign): Remove.
15094 (lshift_double): Adjust to take unsinged shift argument, push
15095 dispatching code to callers.
15096 (mul_double_wide_with_sign): Add early out for callers that
15097 are not interested in high parts or overflow.
15098 (lshift): New function.
15099 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
15100 dispatch code here.
15101 (lrotate, rrotate): Use logical shifts.
15102 * expr.c (get_inner_reference): Use lshift.
15103 * fixed-value.c (do_fixed_divide): Likewise.
15104 * tree-dfa.c (get_ref_base_and_extent): Likewise.
15105 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
15106 (indirect_refs_may_alias_p): Likewise.
15107 (stmt_kills_ref_p_1): Likewise.
15109 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
15111 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
15113 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
15115 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
15116 scalar form of FABD instruction.
15118 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
15120 * lra-constraints.c (process_alt_operands): Add checking alt
15121 number to choose the best alternative.
15123 2013-05-02 Richard Biener <rguenther@suse.de>
15125 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
15126 bitmap and its handling.
15127 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
15129 2013-05-02 Richard Biener <rguenther@suse.de>
15131 PR middle-end/57140
15132 * tree-inline.c (copy_loops): Properly handle removed loops.
15133 (copy_cfg_body): Mark destination loops for fixup if source
15134 loops needed fixup.
15136 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
15139 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
15140 generating simple_return for naked functions.
15142 2013-05-02 Martin Jambor <mjambor@suse.cz>
15144 PR middle-end/56988
15145 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
15146 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
15148 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
15149 ipa_agg_replacement_value structures.
15150 (known_aggs_to_agg_replacement_list): Likewise.
15151 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
15152 (read_agg_replacement_chain): Likewise.
15153 (ipcp_transform_function): Also check that by_ref flags match.
15155 2013-05-02 Richard Biener <rguenther@suse.de>
15157 * graphds.h (struct graph): Add obstack member.
15158 * graphds.c (new_graph): Initialize obstack and allocate
15160 (add_edge): Allocate edge from the obstack.
15161 (free_graph): Free the obstack instead of all edges and vertices.
15163 2013-05-02 Teresa Johnson <tejohnson@google.com>
15165 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
15167 * cfg.c (update_bb_profile_for_threading): Ditto.
15168 * tree-inline.c (copy_bb): Ditto.
15169 (copy_edges_for_bb): Ditto.
15170 (initialize_cfun): Ditto.
15171 (copy_cfg_body): Ditto.
15172 (expand_call_inline): Ditto.
15173 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
15174 (estimate_node_size_and_time): Ditto.
15175 (inline_merge_summary): Ditto.
15176 * cgraphclones.c (cgraph_clone_edge): Ditto.
15177 (cgraph_clone_node): Ditto.
15178 * sched-rgn.c (compute_dom_prob_ps): Ditto.
15179 (compute_trg_info): Ditto.
15181 2013-05-02 Ian Bolton <ian.bolton@arm.com>
15183 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
15184 S reg when fp attribute set.
15185 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
15187 2013-05-02 Ian Bolton <ian.bolton@arm.com>
15189 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
15191 (*and_one_cmplsi3_compare0_uxtw): Likewise.
15192 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
15193 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
15195 2013-05-02 Richard Biener <rguenther@suse.de>
15197 * tree-scalar-evolution.c (scev_info_hasher): Remove.
15198 (struct instantiate_cache_entry): New type.
15199 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
15200 (struct instantiate_cache_type): New type.
15201 (set_instantiated_value, get_instantiated_value): Remove.
15202 (get_instantiated_value_entry): New function.
15203 (instantiate_scev_name): Use the new cache and adjust.
15204 (instantiate_scev_poly): Adjust.
15205 (instantiate_scev_binary): Likewise.
15206 (instantiate_array_ref): Likewise.
15207 (instantiate_scev_convert): Likewise.
15208 (instantiate_scev_not): Likewise.
15209 (instantiate_scev_3): Likewise.
15210 (instantiate_scev_2): Likewise.
15211 (instantiate_scev_r): Likewise.
15212 (instantiate_scev): Likewise.
15213 (resolve_mixers): Likewise.
15215 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
15218 * lra-constraints.c (best_small_class_operands_num): Remove.
15219 (process_alt_operands): Remove small_class_operands_num. Take
15220 small classes operands into losers and only if the operand is not
15221 matched. Modify debugging output.
15222 (curr_insn_transform): Remove best_small_class_operands_num.
15225 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15227 * config/aarch64/aarch64-builtins.c
15228 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
15229 * config/aarch64/aarch64-simd-builtins.def
15230 (reduc_splus_): Add new modes.
15231 (reduc_uplus_): New.
15232 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
15233 (reduc_uplus_v4sf): Likewise.
15234 (reduc_splus_v4sf): Likewise.
15235 (aarch64_addv<mode>): Likewise.
15236 (reduc_uplus_<mode>): Likewise.
15237 (reduc_splus_<mode>): Likewise.
15238 (aarch64_addvv2di): Likewise.
15239 (reduc_uplus_v2di): Likewise.
15240 (reduc_splus_v2di): Likewise.
15241 (aarch64_addvv2si): Likewise.
15242 (reduc_uplus_v2si): Likewise.
15243 (reduc_splus_v2si): Likewise.
15244 (reduc_<sur>plus_<mode>): New.
15245 (reduc_<sur>plus_v2di): Likewise.
15246 (reduc_<sur>plus_v2si): Likewise.
15247 (reduc_<sur>plus_v4sf): Likewise.
15248 (aarch64_addpv4sf): Likewise.
15249 * config/aarch64/arm_neon.h
15250 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
15251 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
15252 add UNSPEC_SADDV, UNSPEC_UADDV.
15254 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
15256 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15258 * config/aarch64/arm_neon.h
15259 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
15261 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15263 * config/aarch64/aarch64-builtins
15264 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
15266 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15268 * config/aarch64/aarch64-simd-builtins.def
15269 (reduc_smax_): New.
15270 (reduc_smin_): Likewise.
15271 (reduc_umax_): Likewise.
15272 (reduc_umin_): Likewise.
15273 (reduc_smax_nan_): Likewise.
15274 (reduc_smin_nan_): Likewise.
15277 (smax): Update for V2SF, V4SF and V2DF modes.
15280 (smin_nan): Likewise.
15281 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
15282 (<su><maxmin><mode>3): ...This, refactor.
15283 (s<maxmin><mode>3): New.
15284 (<maxmin_uns><mode>3): Likewise.
15285 (reduc_<maxmin_uns>_<mode>): Refactor.
15286 (reduc_<maxmin_uns>_v4sf): Likewise.
15287 (reduc_<maxmin_uns>_v2si): Likewise.
15288 (aarch64_<fmaxmin><mode>: Remove.
15289 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
15291 (vmin<q>_f<32,64>): Likewise.
15292 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
15294 (su): Add mappings for smax, smin, umax, umin.
15296 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
15297 (FMAXMIN): Rename as...
15298 (FMAXMIN_UNS): ...This.
15300 (fmaxminv): Likewise.
15301 (fmaxmin): Likewise.
15303 (maxmin_uns_op): Likewise.
15305 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15307 * config/aarch64/arm_neon.h
15308 (vac<ge, gt><sd>_f<32, 64>): Rename to...
15309 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
15310 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
15312 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15314 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
15315 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
15317 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15319 * config/aarch64/aarch64-simd.md
15320 (vcond<mode>_internal): Handle special cases for constant masks.
15321 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
15322 (vcondu<mode><mode>): Likewise.
15323 (vcond<v_cmp_result><mode>): New.
15325 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15327 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
15328 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
15329 * config/aarch64/aarch64-simd-builtins.def
15330 (cmeq): Update to BUILTIN_VALLDI.
15335 * config/aarch64/arm_neon.h
15336 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
15337 to builtins or C as appropriate.
15339 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15341 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
15343 (cmhi): Rename to...
15345 * config/aarch64/aarch64-simd.md
15346 (simd_mode): Add SF.
15347 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
15348 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
15349 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
15350 (cstore<mode>_neg): ...This.
15351 * config/aarch64/iterators.md
15353 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
15354 (COMPARISONS): New.
15355 (UCOMPARISONS): Likewise.
15356 (optab): Add missing comparisons.
15362 (VCMP_S): Likewise.
15363 (VCMP_U): Likewise.
15364 (V_cmp_result): Add DF, SF modes.
15365 (v_cmp_result): Likewise.
15367 (vmtype): Likewise.
15368 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
15370 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
15372 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
15373 define_insn to define_insn_and_split.
15374 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
15375 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
15376 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
15377 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
15378 (thumb2_negscc): Likewise.
15380 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
15382 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
15384 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
15386 * config/arm/thumb2.md: Remove trailing whitespaces.
15388 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
15390 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
15391 Use gen_int_mode rather than GEN_INT.
15393 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
15395 * value-prof.c (stream_in_histogram_value): Remove the strayed
15398 2013-04-30 Richard Biener <rguenther@suse.de>
15400 PR middle-end/57122
15401 * cfghooks.c (split_edge): Properly check for the loop latch edge.
15403 2013-04-30 Richard Biener <rguenther@suse.de>
15405 PR middle-end/57107
15406 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
15408 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
15410 PR rtl-optimization/56957
15411 PR rtl-optimization/57105
15412 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
15413 variable. Use just INSN_UID for determining whether an insn
15414 should be only disconnected from the insn stream.
15415 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
15417 2013-04-30 Jakub Jelinek <jakub@redhat.com>
15419 PR tree-optimization/57104
15420 * tsan.c (instrument_expr): Don't instrument accesses to
15421 DECL_HARD_REGISTER VAR_DECLs.
15423 2013-04-30 Richard Biener <rguenther@suse.de>
15425 * function.h (loops_for_fn): New inline function.
15426 (set_loops_for_fn): Likewise.
15427 * cfgloop.h (place_new_loop): Add struct function parameter.
15428 (get_loop): Likewise.
15429 (get_loops): Likewise.
15430 (number_of_loops): Likewise.
15431 (fel_next): Adjust.
15432 (fel_init): Likewise.
15433 * cfg.c (get_loop_copy): Adjust.
15434 * cfgloop.c (flow_loops_dump): Likewise.
15435 (record_loop_exits): Likewise.
15436 (verify_loop_structure): Likewise.
15437 * cfgloopanal.c (mark_irreducible_loops): Likewise.
15438 (estimate_reg_pressure_cost): Likewise.
15439 (mark_loop_exit_edges): Likewise.
15440 * cfgloopmanip.c (place_new_loop): Likewise.
15441 (add_loop): Likewise.
15442 (duplicate_loop): Likewise.
15443 * graph.c (draw_cfg_nodes): Likewise.
15444 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
15445 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
15446 (extract_affine_chrec): Likewise.
15447 (build_scop_iteration_domain): Likewise.
15448 * graphite.c (graphite_initialize): Likewise.
15449 * ira-build.c (create_loop_tree_nodes): Likewise.
15450 (more_one_region_p): Likewise.
15451 (rebuild_regno_allocno_maps): Likewise.
15452 (mark_loops_for_removal): Likewise.
15453 (mark_all_loops_for_removal): Likewise.
15454 (remove_unnecessary_regions): Likewise.
15455 (ira_build): Likewise.
15456 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
15457 * loop-init.c (fix_loop_structure): Likewise.
15458 (gate_rtl_move_loop_invariants): Likewise.
15459 (gate_rtl_unswitch): Likewise.
15460 (gate_rtl_unroll_and_peel_loops): Likewise.
15461 (rtl_doloop): Likewise.
15462 * lto-streamer-in.c (input_cfg): Likewise.
15463 * lto-streamer-out.c (output_cfg): Likewise.
15464 * modulo-sched.c (sms_schedule): Likewise.
15465 * predict.c (tree_estimate_probability): Likewise.
15466 (tree_estimate_probability_driver): Likewise.
15467 (estimate_loops): Likewise.
15468 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
15469 (move_sese_region_to_fn): Likewise.
15470 (debug_loop_num): Likewise.
15471 * tree-chrec.c (chrec_evaluate): Likewise.
15472 (hide_evolution_in_other_loops_than_loop): Likewise.
15473 (chrec_component_in_loop_num): Likewise.
15474 (reset_evolution_in_loop): Likewise.
15475 (evolution_function_is_invariant_rec_p): Likewise.
15476 * tree-if-conv.c (main_tree_if_conversion): Likewise.
15477 * tree-inline.c (copy_loops): Likewise.
15478 (copy_cfg_body): Likewise.
15479 (tree_function_versioning): Likewise.
15480 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
15481 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
15483 (add_to_evolution_1): Likewise.
15484 (scev_const_prop): Likewise.
15485 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
15486 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
15487 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
15488 (tree_ssa_lim_initialize): Likewise.
15489 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
15490 (verify_loop_closed_ssa): Likewise.
15491 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
15492 (tree_ssa_loop_im): Likewise.
15493 (tree_ssa_loop_unswitch): Likewise.
15494 (tree_vectorize): Likewise.
15495 (check_data_deps): Likewise.
15496 (tree_ssa_loop_ivcanon): Likewise.
15497 (tree_ssa_loop_bounds): Likewise.
15498 (tree_complete_unroll): Likewise.
15499 (tree_complete_unroll_inner): Likewise.
15500 (tree_parallelize_loops): Likewise.
15501 (tree_ssa_loop_prefetch): Likewise.
15502 (tree_ssa_loop_ivopts): Likewise.
15503 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
15504 * tree-vectorizer.c (vectorize_loops): Likewise.
15506 2013-04-29 Mike Frysinger <vapier@gentoo.org>
15508 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
15509 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
15510 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
15511 with EABI_LINK_SPEC.
15513 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
15516 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
15519 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
15522 * lra-constraints.c (process_alt_operands): Discourage a bit more
15523 using memory for pseudos. Print cost dump for alternatives.
15524 Modify cost values for conflicts with early clobbers.
15525 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
15527 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
15530 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
15532 2013-04-29 Ian Bolton <ian.bolton@arm.com>
15534 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
15535 from/to S register.
15536 (movdi_aarch64): Support LDR/STR from/to D register.
15538 2013-04-29 Ian Bolton <ian.bolton@arm.com>
15540 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
15541 or higher by default.
15543 2013-04-29 Richard Biener <rguenther@suse.de>
15545 PR middle-end/57075
15546 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
15547 even if not adding abnormal edges for calls that can make
15550 2013-04-29 Richard Biener <rguenther@suse.de>
15552 PR middle-end/57103
15553 * tree-cfg.c (move_stmt_op): Fix condition under which to update
15555 (move_stmt_r): Remove redundant checking.
15557 2013-04-29 Teresa Johnson <tejohnson@google.com>
15560 * basic-block.h (apply_scale): New function.
15561 (apply_probability): Use apply_scale.
15562 * gimple-streamer-in.c (input_bb): Ditto.
15563 * lto-streamer-in.c (input_cfg): Ditto.
15564 * lto-cgraph.c (merge_profile_summaries): Ditto.
15565 * tree-optimize.c (execute_fixup_cfg): Ditto.
15566 * tree-inline.c (copy_bb): Update comment to use apply_scale.
15567 (copy_edges_for_bb): Ditto.
15568 (copy_cfg_body): Ditto.
15570 2013-04-29 Tom de Vries <tom@codesourcery.com>
15572 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
15573 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
15574 (tail_merge_optimize): Handle current_loops == NULL.
15576 2013-04-26 Jeff Law <law@redhat.com>
15578 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
15579 (simplify_cond_using_ranges): Generalize code to simplify
15580 COND_EXPRs where one argument is a constant and the other
15581 is an SSA_NAME created by an integral type conversion.
15583 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15585 * config/arm/arm.md (store_minmaxsi): Use only when
15586 optimize_insn_for_size_p.
15588 2013-04-29 Christian Bruel <christian.bruel@st.com>
15591 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
15593 2013-04-29 Richard Biener <rguenther@suse.de>
15595 PR middle-end/57089
15596 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
15597 loop tree make sure to schedule a fixup for the child as well.
15598 (expand_omp_for_generic): Properly add loops.
15599 (expand_omp_for_static_nochunk): Likewise.
15600 (expand_omp_for_static_chunk): Likewise.
15601 (expand_omp_for): For the degenerate case fixup loops.
15602 (expand_omp_sections): Fix default bb placement in loops.
15603 (expand_omp_atomic_pipeline): Properly add loops.
15605 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15607 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
15609 2013-04-29 Tom de Vries <tom@codesourcery.com>
15611 * tree-ssa-tail-merge.c: Update header comment.
15613 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15615 * config/aarch64/arm_neon.h
15616 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
15617 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
15618 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
15619 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
15620 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
15621 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
15622 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
15623 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
15625 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15627 * config/aarch64/aarch64-simd.md
15628 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
15629 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
15630 fix_trunc, fixuns_trunc.
15631 (ftrunc<VDQF:mode>2): New.
15632 * config/aarch64/iterators.md (optab): Add fix, fixuns.
15633 (fix_trunc_optab): New.
15635 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15637 * config/aarch64/aarch64-builtins.c
15638 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
15639 iceilf, lround, iroundf.
15641 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
15644 * config/i386/i386.h (enum ix86_tune_indices)
15645 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
15646 New, split from X86_TUNE_INTER_UNIT_MOVES.
15647 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
15648 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
15649 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
15650 (TARGET_INTER_UNIT_MOVES): Remove.
15651 * config/i386/i386.c (initial_ix86_tune_features): Update.
15652 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
15653 (ix86_expand_convert_uns_didf_sse): Use
15654 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
15655 (ix86_expand_vector_init_one_nonzero): Ditto.
15656 (ix86_expand_vector_init_interleave): Ditto.
15657 (inline_secondary_memory_needed): Return true for moves from SSE class
15658 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
15659 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
15660 * config/i386/constraints.md (Yi, Ym): Depend on
15661 TARGET_INTER_UNIT_MOVES_TO_VEC.
15662 (Yj, Yn): New constraints.
15663 * config/i386/i386.md (*movdi_internal): Change constraints of
15664 operand 1 from Yi to Yj and from Ym to Yn.
15665 (*movsi_internal): Ditto.
15666 (*movdf_internal): Ditto.
15667 (*movsf_internal): Ditto.
15668 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
15669 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
15670 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
15671 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
15672 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
15673 * config/i386/sse.md (movdi_to_sse): Ditto.
15674 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
15675 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
15676 TARGET_INTER_UNIT_MOVES.
15677 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
15678 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
15679 instead of TARGET_INTER_UNIT_MOVES.
15680 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
15681 operand 1 from Yi to Yj and from Ym to Yn.
15683 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15685 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
15686 (float_truncate_hi_): Likewise.
15687 (float_extend_lo_): Likewise.
15688 (float_truncate_lo_): Likewise.
15689 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
15690 (aarch64_float_extend_lo_v2df): Likewise.
15691 (vec_unpacks_hi_v4sf): Likewise.
15692 (aarch64_float_truncate_lo_v2sf): Likewise.
15693 (aarch64_float_truncate_hi_v4sf): Likewise.
15694 (vec_pack_trunc_v2df): Likewise.
15695 (vec_pack_trunc_df): Likewise.
15697 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15699 * config/aarch64/aarch64-builtins.c
15700 (aarch64_fold_builtin): Fold float conversions.
15701 * config/aarch64/aarch64-simd-builtins.def
15702 (floatv2si, floatv4si, floatv2di): New.
15703 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
15704 * config/aarch64/aarch64-simd.md
15705 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
15706 * config/aarch64/iterators.md (FLOATUORS): New.
15707 (optab): Add float, floatuns.
15708 (su_optab): Likewise.
15710 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15712 * config/aarch64/aarch64-builtins.c
15713 (aarch64_builtin_vectorized_function): Use new names for
15715 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
15716 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
15717 (fcvtzu): Split as...
15718 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
15719 (fcvtas): Split as...
15720 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
15721 (fcvtau): Split as...
15722 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
15723 (fcvtps): Split as...
15724 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
15725 (fcvtpu): Split as...
15726 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
15727 (fcvtms): Split as...
15728 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
15729 (fcvtmu): Split as...
15730 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
15731 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
15732 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
15733 (lfrintnusf, lfrintnudf): Likewise.
15734 * config/aarch64/aarch64-simd.md
15735 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
15737 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
15738 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
15739 (fcvt_pattern): Likewise.
15741 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15743 * config/aarch64/aarch64-simd.md
15744 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
15745 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
15747 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15749 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
15750 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
15751 (vrnd<a,m,n,p>_f32): Implement using builtins.
15752 (vrnd<i,x><q>_f<32, 64>): New.
15754 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15756 * config/aarch64/aarch64-builtins.c
15757 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
15758 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
15759 (frintz): Rename to...
15761 (frintp): Rename to...
15763 (frintm): Rename to...
15765 (frinti): Rename to...
15766 (nearbyint): ...this.
15767 (frintx): Rename to...
15769 (frinta): Rename to...
15771 * config/aarch64/aarch64-simd.md
15772 (aarch64_frint<frint_suffix><mode>): Delete.
15773 (<frint_pattern><mode>2): Convert to insn.
15774 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
15775 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
15776 (frint_pattern): Likewise.
15777 (frint_suffix): Likewise.
15779 2013-04-29 Richard Biener <rguenther@suse.de>
15781 PR tree-optimization/57081
15782 * loop-init.c: Include tree-flow.h.
15783 (loop_optimizer_finalize): Free number of iteration estimates.
15784 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
15786 2013-04-29 Jakub Jelinek <jakub@redhat.com>
15788 PR tree-optimization/57083
15789 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
15790 non-singleton shift count range, zero extend low_bound for uns case.
15792 * config/i386/predicates.md (general_vector_operand): New predicate.
15793 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
15794 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
15795 if they aren't nonimmediate operands. If their original values
15796 satisfy const_vector_equal_evenodd_p, don't shift them.
15797 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
15798 predicates. For the SSE4.1 case force operands[{1,2}] into registers
15799 if not nonimmediate_operand.
15800 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
15801 instead of register_operand.
15802 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
15804 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
15806 * stor-layout.c (finalize_size_functions): Allocate a structure and
15807 reset cfun before dumping the functions.
15809 2013-04-27 Jakub Jelinek <jakub@redhat.com>
15811 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
15814 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
15815 use xop_pmacsdqh if uns_p.
15816 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
15817 the immediate rotate count.
15819 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
15821 * rtl.h (struct rtx_def): Add comment for field jump.
15822 (LRA_SUBREG_P): New macro.
15823 * recog.c (register_operand): Check LRA_SUBREG_P.
15824 * lra.c (lra): Add note at the end of RTL code. Align non-empty
15826 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
15827 (lra_final_code_change): Skip subreg change for operators.
15828 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
15829 if there are no operand changes.
15830 * lra-constraints.c (curr_insn_set): New.
15831 (match_reload): Set LRA_SUBREG_P.
15832 (emit_spill_move): Ditto.
15833 (check_and_process_move): Use curr_insn_set. Process only single
15834 set insns. Don't initialize sec_mem_p and change_p.
15835 (simplify_operand_subreg): Use LRA_SUBREG_P.
15836 (reg_in_class_p): New function.
15837 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
15838 of #ifdef. Add code to remove cycling.
15839 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
15840 non-null disp. Reload inner instead of disp when base and index
15841 are null. Try to put lo_sum into register.
15842 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
15843 (check_and_process_move): Move code for move cost check to
15844 simple_move_p. Remove equiv_substitution.
15845 (simple_move_p): New function.
15846 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
15847 curr_insn_set. Call check_and_process_move only for single set
15848 insns. Use the new function. Move call of check_and_process_move
15849 after operand equiv substitution and address process.
15851 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15854 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
15855 with nonlocal goto receivers or returns twice calls, ignore
15856 unininitialized values from abnormal edges to nl goto receiver
15857 or returns twice call.
15859 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15861 PR tree-optimization/57051
15862 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
15863 and VEC_RSHIFT_EXPR if shift count is a multiple of element
15866 2013-04-26 Richard Biener <rguenther@suse.de>
15868 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
15869 (expand_omp_taskreg): Likewise. Mark loops for fixup.
15870 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
15871 (fixup_loop_arrays_after_move): New function.
15872 (move_sese_region_to_fn): Properly outline the loop tree parts
15873 of the SESE region.
15875 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
15877 * config/i386/i386.md (type, unit): Fix long lines.
15879 2013-04-26 Richard Biener <rguenther@suse.de>
15881 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
15882 (lto-streamer-out.o): Likewise.
15883 * cfgloop.c (init_loops_structure): Export, add struct function
15884 argument and adjust.
15885 (flow_loops_find): Adjust.
15886 * cfgloop.h (enum loop_estimation): Add EST_LAST.
15887 (init_loops_structure): Declare.
15888 * lto-streamer-in.c: Include cfgloop.h.
15889 (input_cfg): Input the loop tree.
15890 * lto-streamer-out.c: Include cfgloop.h.
15891 (output_cfg): Output the loop tree.
15892 (output_struct_function_base): Do not drop PROP_loops.
15894 2013-03-26 Richard Biener <rguenther@suse.de>
15896 * tree-cfg.c (execute_build_cfg): Build the loop tree.
15897 (pass_build_cfg): Provide PROP_loops.
15898 (move_sese_region_to_fn): Remove loops that are outlined into fn
15900 * tree-inline.c: Include cfgloop.h.
15901 (initialize_cfun): Do not drop PROP_loops.
15902 (copy_loops): New function.
15903 (copy_cfg_body): Copy loop structure.
15904 (tree_function_versioning): Initialize destination loop tree.
15905 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
15906 (pass_parallelize_loops): Do IL verification.
15907 * loop-init.c (loop_optimizer_init): Fixup loops if required.
15908 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
15909 the CFG make sure we fixup loops as well.
15910 * tree-ssa-tail-merge.c: Include cfgloop.h.
15911 (replace_block_by): When merging loop latches mark loops for fixup.
15912 * lto-streamer-out.c (output_struct_function_base): Drop
15913 PROP_loops for now.
15914 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
15915 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
15916 * ipa-split.c: Include cfgloop.h.
15917 (split_function): Add the new return block to the loop tree root.
15918 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
15919 whether we have removed the forwarder block.
15920 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
15921 * cfgloop.h (place_new_loop): Declare.
15922 * cfgloopmanip.c (place_new_loop): Export.
15923 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
15924 (tree-switch-conversion.o): Likewise.
15925 (tree-complex.o): Likewise.
15926 (tree-inline.o): Likewise.
15927 (tree-ssa-tailmerge.o): Likewise.
15928 (ipa-split.o): Likewise.
15929 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
15930 (tree-ssa-copy.o): Likewise.
15931 * tree-switch-conversion.c: Include cfgloop.h
15932 (process_switch): If we emit a bit-test cascade, schedule loops
15934 * tree-complex.c: Include cfgloop.h.
15935 (expand_complex_div_wide): Properly add new basic-blocks to loops.
15936 * asan.c: Include cfgloop.h.
15937 (create_cond_insert_point): Properly add new basic-blocks to
15938 loops, schedule loop fixup.
15939 * cfgloop.c (verify_loop_structure): Check that looks are not
15941 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
15943 (expand_omp_for_generic): Likewise.
15944 (expand_omp_sections): Likewise.
15945 (expand_omp_atomic_pipeline): Schedule loops for fixup.
15946 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
15947 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
15948 is initialized, not when loops are present.
15949 * tree-parloops.c (parallelize_loops): Remove checking here.
15950 * passes.c (init_optimization_passes): Schedule a copy-propagation
15951 pass before complete unrolling of inner loops.
15953 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15955 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
15956 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
15957 (colorize_init): Add argument to _WIN32 version.
15958 * toplev.c: Include diagnostic-color.h.
15959 (process_options): Default to -fdiagnostics-color=auto if
15960 GCC_COLORS env var is in the environment.
15961 * common.opt (fdiagnostics-color=): Add Var and Init.
15962 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
15963 env var is in the environment, the default is auto rather than never.
15965 * diagnostic.h (file_name_as_prefix): Add context argument.
15966 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
15967 the string as locus.
15968 * langhooks.c (lhd_print_error_function): Adjust caller.
15970 2013-04-25 Lawrence Crowl <crowl@google.com>
15972 * var-tracking.c (shared_hash_def::htab):
15973 Change type to hash_table. Update dependent calls and types.
15975 2013-04-25 Lawrence Crowl <crowl@google.com>
15977 * Makefile.in: Update as needed below.
15979 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
15980 Move declaration to after the type's method definitons.
15982 * attribs.c (htab_t scoped_attributes::attribute_hash):
15983 Change type to hash_table. Update dependent calls and types.
15985 * bitmap.c (htab_t bitmap_desc_hash):
15986 Change type to hash_table. Update dependent calls and types.
15988 * cselib.c (htab_t cselib_hash_table):
15989 Change type to hash_table. Update dependent calls and types.
15991 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
15992 (hash_string_slot_node): Move implementation into lto-streamer.h
15993 struct string_slot_hasher.
15994 (eq_string_slot_node): Likewise.
15996 * data-streamer-out.c: Update output_block::string_hash_table
15997 dependent calls and types.
15999 * dwarf2cfi.c (htab_t trace_index):
16000 Change type to hash_table. Update dependent calls and types.
16002 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
16003 Change type to hash_table. Update dependent calls and types.
16004 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
16005 (htab_t optimize_external_refs::map): Likewise.
16006 (htab_t output_comp_unit::extern_map): Likewise.
16007 (htab_t output_comdat_type_unit::extern_map): Likewise.
16008 (htab_t output_macinfo::macinfo_htab): Likewise.
16009 (htab_t optimize_location_lists::htab): Likewise.
16010 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
16012 * except.c (htab_t ehspec_hash_type):
16013 Change type to hash_table. Update dependent calls and types.
16014 (assign_filter_values::ttypes): Likewise.
16015 (assign_filter_values::ehspec): Likewise.
16016 (sjlj_assign_call_site_values::ar_hash): Likewise.
16017 (convert_to_eh_region_ranges::ar_hash): Likewise.
16019 * gcse.c (htab_t pre_ldst_table):
16020 Change type to hash_table. Update dependent calls and types.
16022 * ggc-common.c (htab_t saving_htab):
16023 Change type to hash_table. Update dependent calls and types.
16024 (htab_t loc_hash): Likewise.
16025 (htab_t ptr_hash): Likewise.
16026 (call_count): Rename ggc_call_count.
16027 (call_alloc): Rename ggc_call_alloc.
16028 (loc_descriptor): Rename make_loc_descriptor.
16029 (add_statistics): Rename ggc_add_statistics.
16031 * ggc-common.c (saving_htab):
16032 Change type to hash_table. Update dependent calls and types.
16034 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
16035 (push_gimplify_context): Likewise.
16036 (pop_gimplify_context): Likewise.
16037 (struct gimple_temp_hash_elt): Added.
16038 (struct gimplify_hasher): Likewise.
16039 (struct gimplify_ctx.temp_htab):
16040 Change type to hash_table. Update dependent calls and types.
16042 * gimple-fold.c: Include gimplify-ctx.h.
16044 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
16045 Change type to hash_table. Update dependent calls and types.
16046 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
16047 avoid potential global name collision.
16049 * gimplify.c: Include gimplify-ctx.h.
16050 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
16051 (htab_t gimplify_ctx::temp_htab):
16052 Update dependent calls and types for new type hash_table.
16053 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
16054 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
16056 * gimplify-ctx.h: New.
16057 (struct gimple_temp_hash_elt): Move from gimplify.c.
16058 (class gimplify_hasher): New.
16059 (struct gimplify_ctx): Move from gimple.h.
16060 (htab_t gimplify_ctx::temp_htab):
16061 Change type to hash_table. Update dependent calls and types.
16063 * graphite-clast-to-gimple.c: Include graphite-htab.h.
16064 (htab_t ivs_params::newivs_index):
16065 Change type to hash_table. Update dependent calls and types.
16066 (htab_t ivs_params::params_index): Likewise.
16067 (htab_t print_generated_program::params_index): Likewise.
16068 (htab_t gloog::newivs_index): Likewise.
16069 (htab_t gloog::params_index): Likewise.
16071 * graphite.c: Include graphite-htab.h.
16072 4htab_t graphite_transform_loops::bb_pbb_mapping):
16073 Change type to hash_table. Update dependent calls and types.
16075 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
16076 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
16077 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
16079 * graphite-dependences.c: Include graphite-htab.h.
16080 (loop_is_parallel_p): Change hash table type of parameter.
16082 * graphite-htab.h: New.
16083 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
16084 (extern find_pbb_via_hash): Move from graphite-poly.h.
16085 (extern loop_is_parallel_p): Move from graphite-poly.h.
16086 (extern get_loop_body_pbbs): Move from graphite-poly.h.
16088 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
16089 (extern loop_is_parallel_p): Move to graphite-htab.h.
16090 (extern get_loop_body_pbbs): Move to graphite-htab.h.
16092 * haifa-sched.c (htab_t delay_htab):
16093 Change type to hash_table. Update dependent calls and types.
16094 (htab_t delay_htab_i2): Likewise.
16096 * ira-color.c (htab_t allocno_hard_regs_htab):
16097 Change type to hash_table. Update dependent calls and types.
16099 * ira-costs.c (htab_t cost_classes_htab):
16100 Change type to hash_table. Update dependent calls and types.
16102 * loop-invariant.c (htab_t merge_identical_invariants::eq):
16103 Change type to hash_table. Update dependent calls and types.
16105 * loop-iv.c (htab_t bivs):
16106 Change type to hash_table. Update dependent calls and types.
16108 * loop-unroll.c (htab_t opt_info::insns_to_split):
16109 Change type to hash_table. Update dependent calls and types.
16110 (htab_t opt_info::insns_with_var_to_expand): Likewise.
16112 * lto-streamer.h (struct string_slot): Move from data-streamer.h
16113 (struct string_slot_hasher): New.
16114 (htab_t output_block::string_hash_table):
16115 Change type to hash_table. Update dependent calls and types.
16117 * lto-streamer-in.c (freeing_string_slot_hasher): New.
16118 (htab_t file_name_hash_table):
16119 Change type to hash_table. Update dependent calls and types.
16121 * lto-streamer-out.c: Update output_block::string_hash_table dependent
16124 * lto-streamer.c (htab_t tree_htab):
16125 Change type to hash_table. Update dependent calls and types.
16127 * omp-low.c: Include gimplify-ctx.h.
16129 * passes.c (htab_t name_to_pass_map):
16130 Change type to hash_table. Update dependent calls and types.
16131 (pass_traverse): Rename to passes_pass_traverse.
16133 * plugin.c (htab_t event_tab):
16134 Change type to hash_table. Update dependent calls and types.
16136 * postreload-gcse.c (htab_t expr_table):
16137 Change type to hash_table. Update dependent calls and types.
16138 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
16140 * sese.c (debug_rename_map_1): Make extern.
16141 (htab_t copy_bb_and_scalar_dependences::rename_map):
16142 Change type to hash_table. Update dependent calls and types.
16144 * sese.h (extern debug_rename_map): Move to .c file.
16146 * store-motion.c (htab_t store_motion_mems_table):
16147 Change type to hash_table. Update dependent calls and types.
16149 * trans-mem.c (htab_t tm_new_mem_hash):
16150 Change type to hash_table. Update dependent calls and types.
16152 * tree-browser.c (htab_t TB_up_ht):
16153 Change type to hash_table. Update dependent calls and types.
16155 * tree-cfg.c (htab_t discriminator_per_locus):
16156 Change type to hash_table. Update dependent calls and types.
16158 * tree-complex.c: Include tree-hasher.h
16159 (htab_t complex_variable_components):
16160 Change type to hash_table. Update dependent calls and types.
16162 * tree-eh.c (htab_t finally_tree):
16163 Change type to hash_table. Update dependent calls and types.
16165 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
16166 struct int_tree_hasher.
16167 (extern int_tree_map_eq): Likewise.
16168 (uid_decl_map_hash): Removed.
16169 (extern decl_tree_map_eq): Likewise.
16171 * tree-hasher.h: New.
16172 (struct int_tree_hasher): New.
16173 (typedef int_tree_htab_type): New.
16175 * tree-inline.c: Include gimplify-ctx.h.
16177 * tree-mudflap.c: Include gimplify-ctx.h.
16179 * tree-parloops.c: Include tree-hasher.h.
16180 (htab_t eliminate_local_variables_stmt::decl_address):
16181 Change type to hash_table. Update dependent calls and types.
16182 (htab_t separate_decls_in_region::decl_copies): Likewise.
16184 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
16185 Change type to hash_table. Update dependent calls and types.
16187 * tree-sra.c (candidates):
16188 Change type to hash_table. Update dependent calls and types.
16190 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
16192 (int_tree_map_hash): Likewise.
16194 * tree-ssa-dom.c (htab_t avail_exprs):
16195 Change type to hash_table. Update dependent calls and types.
16197 * tree-ssa-live.c (var_map_base_init::tree_to_index):
16198 Change type to hash_table. Update dependent calls and types.
16200 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
16201 Change type to hash_table. Update dependent calls and types.
16203 * tree-ssa-phiopt.c (seen_ssa_names):
16204 Change type to hash_table. Update dependent calls and types.
16206 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
16207 Change type to hash_table. Update dependent calls and types.
16209 * tree-ssa-uncprop.c (equiv):
16210 Change type to hash_table. Update dependent calls and types.
16212 2013-04-25 Jakub Jelinek <jakub@redhat.com>
16214 PR rtl-optimization/57003
16215 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
16216 call note_stores with kill_clobbered_value callback again after
16217 killing regs_invalidated_by_call.
16219 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16221 * config/aarch64/aarch64-simd.md
16222 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
16223 (aarch64_simd_bsl<mode>): Likewise.
16224 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
16226 2013-04-25 Marek Polacek <polacek@redhat.com>
16228 PR tree-optimization/57066
16229 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
16231 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
16233 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
16235 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16237 * config/aarch64/aarch64-builtins.c
16238 (aarch64_fold_builtin): New.
16239 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
16240 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
16241 * config/aarch64/aarch64-simd-builtins.def (abs): New.
16242 * config/aarch64/arm_neon.h
16243 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
16245 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16246 Tejas Belagod <tejas.belagod@arm.com>
16248 * config/aarch64/aarch64-builtins.c
16249 (aarch64_gimple_fold_builtin): New.
16250 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
16251 * config/aarch64/aarch64-simd-builtins.def (addv): New.
16252 * config/aarch64/aarch64-simd.md (addpv4sf): New.
16253 (addvv4sf): Update.
16254 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
16256 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16258 * config/aarch64/aarch64.md
16259 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
16261 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16263 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
16264 (*ngcsi_uxtw): New pattern.
16266 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16267 Julian Brown <julian@codesourcery.com>
16269 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
16270 (TB_DREG): Add T_V4HF.
16271 (v4hf_UP): New macro.
16272 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
16273 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
16274 Handle initialisation of V4HF. Adjust initialisation of reinterpret
16276 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
16277 (arm_vector_mode_supported_p): Handle V4HF.
16278 (arm_mangle_map): Handle V4HFmode.
16279 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
16280 * config/arm/arm_neon_builtins.def: Add entries for
16281 vcvtv4hfv4sf, vcvtv4sfv4hf.
16282 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
16283 (neon_vcvtv4hfv4sf): Likewise.
16284 * config/arm/neon-gen.ml: Handle half-precision floating point
16286 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
16287 * config/arm/arm_neon.h: Regenerate.
16288 * config/arm/neon.ml (type elts): Add F16.
16289 (type vectype): Add T_float16x4, T_floatHF.
16290 (type vecmode): Add V4HF.
16291 (type features): Add Requires_FP_bit feature.
16292 (elt_width): Handle F16.
16293 (elt_class): Likewise.
16294 (elt_of_class_width): Likewise.
16295 (mode_of_elt): Refactor.
16296 (type_for_elt): Handle F16, fix error messages.
16297 (vectype_size): Handle T_float16x4.
16298 (vcvt_sh): New function.
16299 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
16300 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
16301 (string_of_mode): Handle V4HF.
16302 * doc/arm-neon-intrinsics.texi: Regenerate.
16304 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16306 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
16307 format specifier in 'X' case.
16309 2013-04-25 Alan Modra <amodra@gmail.com>
16312 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
16313 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
16314 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
16315 Repeat for many other rotate/shift and mask patterns using subregs.
16316 Name lshiftrt insns.
16317 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
16318 on WORDS_BIG_ENDIAN.
16320 2013-04-25 Alan Modra <amodra@gmail.com>
16322 * config.gcc: Support little-endian powerpc-linux targets.
16323 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
16324 (LINK_OS_LINUX_SPEC): Define.
16325 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
16326 Preserve MASK_LITTLE_ENDIAN.
16327 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
16328 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
16329 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
16330 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
16331 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
16332 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
16333 Correct fp word order for little-endian. Don't shift toc entries
16334 smaller than a word for little-endian.
16335 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
16336 (bswapdi2 splits): Correct low-part subreg for little-endian.
16337 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
16338 low/high where such is correct only for be.
16339 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
16340 little-endian for -mcall-aixdesc.
16342 2013-04-25 Alan Modra <amodra@gmail.com>
16344 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
16345 replace_equiv_address_nv.
16347 2013-04-25 Alan Modra <amodra@gmail.com>
16349 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
16351 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16354 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16355 * rtl.h (struct rtx_def): ...
16357 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16359 PR rtl-optimizations/57046
16360 * lra-constraints (split_reg): Set up lra_risky_transformations_p
16361 for multi-reg splits.
16363 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
16365 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
16367 2013-04-24 Sterling Augustine <saugustine@google.com>
16369 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
16370 (comp_dir_string, debug_str_dwo_section): New.
16371 (DEBUG_STR_DWO_SECTION): Rename to ...
16372 (DEBUG_DWO_STR_SECTION): ... this.
16373 (DEBUG_NORM_STR_SECTION): Delete.
16374 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
16375 (DEBUG_STR_DWO_SECTION_FLAGS): New.
16376 (find_AT_string): Move most logic to ...
16377 (find_AT_string_in_table): ... here. New.
16378 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
16379 add_skeleton_AT_string. Delete logic.
16380 (output_skeleton_debug_sections): Remove call to
16381 add_top_level_skeleton_die_attrs.
16382 (add_comp_dir_attribute): Move logic to comp_dir_string.
16383 (dwarf2out_init): Initialize debug_str_dwo_section.
16384 (output_indirect_string): Call find_string_form.
16385 (output_indirect_strings): Rewrite.
16386 (prune_unused_types): Empty skeleton_debug_str_hash.
16387 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
16388 (dwarf2out_finish): Call output_indirect_strings.
16390 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
16392 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
16394 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
16396 * rtl.h (struct rtx_def): Add comment for field jump.
16397 (LRA_SUBREG_P): New macro.
16398 * recog.c (register_operand): Check LRA_SUBREG_P.
16399 * lra.c (lra): Add note at the end of RTL code. Align non-empty
16401 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
16402 (lra_final_code_change): Skip subreg change for operators.
16403 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
16404 if there are no operand changes.
16405 * lra-constraints.c (curr_insn_set): New.
16406 (match_reload): Set LRA_SUBREG_P.
16407 (emit_spill_move): Ditto.
16408 (check_and_process_move): Use curr_insn_set. Process only single
16409 set insns. Don't initialize sec_mem_p and change_p.
16410 (simplify_operand_subreg): Use LRA_SUBREG_P.
16411 (reg_in_class_p): New function.
16412 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
16413 of #ifdef. Add code to remove cycling.
16414 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
16415 non-null disp. Reload inner instead of disp when base and index
16416 are null. Try to put lo_sum into register.
16417 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
16418 (check_and_process_move): Move code for move cost check to
16419 simple_move_p. Remove equiv_substitution.
16420 (simple_move_p): New function.
16421 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
16422 curr_insn_set. Call check_and_process_move only for single set
16423 insns. Use the new function. Move call of check_and_process_move
16424 after operand equiv substitution and address process.
16426 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
16428 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
16429 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
16430 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
16432 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
16434 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
16436 2013-04-24 Marek Polacek <polacek@redhat.com>
16438 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
16439 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
16440 (select_loops_exit_conditions): Likewise.
16441 (number_of_iterations_for_all_loops): Likewise.
16442 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
16443 (scev_analysis): Likewise.
16445 2013-04-02 Catherine Moore <clm@codesourcery.com>
16446 Chao-ying Fu <fu@mips.com>
16448 * config/mips/micromips.md (jraddiusp): New pattern.
16449 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
16450 instruction if possible.
16452 2013-04-24 Alan Modra <amodra@gmail.com>
16454 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
16456 2013-04-24 Julian Brown <julian@codesourcery.com>
16457 Chung-Lin Tang <cltang@codesourcery.com>
16459 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
16460 dependency behavior in enumeration type DIE generation. Add TODO note
16461 to comments about future DW_FORM_sdata/udata re-work of related code.
16463 2013-04-23 Lawrence Crowl <crowl@google.com>
16465 * Makefile.in: Update as needed below.
16467 * hash-table.h (class hash_table):
16468 Correct many methods with parameter types compare_type to the correct
16469 value_type. (Correct code was unlikely to notice the change.)
16470 (hash_table::elements_with_deleted) New.
16471 (class hashtable::iterator): New.
16472 (hashtable::begin()): New.
16473 (hashtable::end()): New.
16474 (FOR_EACH_HASH_TABLE_ELEMENT): New.
16476 * statistics.c (statistics_hashes):
16477 Change type to hash_table. Update dependent calls and types.
16479 * tree-into-ssa.c (var_infos):
16480 Change type to hash_table. Update dependent calls and types.
16482 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
16483 Change type to hash_table. Update dependent calls and types.
16485 * tree-ssa-loop-im.c (struct mem_ref.refs):
16486 Change type to hash_table. Update dependent calls and types.
16488 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
16489 Change type to hash_table. Update dependent calls and types.
16491 * tree-ssa-sccvn.c (vn_tables_s::nary):
16492 Change type to hash_table. Update dependent calls and types.
16493 (vn_tables_s::phis): Likewise.
16494 (vn_tables_s::references): Likewise.
16496 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
16497 (vn_reference_eq): Update parameter and return types.
16499 * tree-ssa-structalias.c (pointer_equiv_class_table):
16500 Change type to hash_table. Update dependent calls and types.
16501 (location_equiv_class_table): Likewise.
16503 * tree-vect-data-refs.c: Consequential changes for making
16504 peeling a hash_table.
16506 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
16507 (destroy_loop_vec_info): Dependent hash_table update.
16509 * tree-vectorizer.h (peeling_htab):
16510 Change type to hash_table. Update dependent calls and types.
16512 2013-04-23 Shiva Chen <shiva0217@gmail.com>
16514 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
16515 to check the register content is equal or not.
16516 * lra-constraints.c (match_reload): Use lra_assign_reg_val
16517 to assign register content record.
16518 * lra-eliminations.c (update_reg_eliminate): Use
16519 lra_update_reg_val_offset to update register content offset.
16520 * lra-int.h (struct lra_reg): Add offset member.
16521 (lra_reg_val_equal_p): New static inline function.
16522 (lra_update_reg_val_offset): New static inline function.
16523 (lra_assign_reg_val): New static inline function.
16524 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
16525 to assign register content record.
16526 (initialize_lra_reg_info_element): Initial offset to zero.
16528 2013-04-23 Catherine Moore <clm@codesourcery.com>
16530 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
16531 operands. Record compression.
16533 2013-04-23 Xinliang David Li <davidxl@google.com>
16535 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
16537 2013-04-23 Richard Biener <rguenther@suse.de>
16539 PR middle-end/57036
16540 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
16541 parameter, only add abnormal goto edges from the copied body
16542 if the call could perform abnormal gotos.
16543 (copy_cfg_body): Adjust.
16545 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
16547 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
16549 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
16551 * coretypes.h (gimple_stmt_iterator): Add struct to make
16554 2013-04-23 Richard Biener <rguenther@suse.de>
16556 PR tree-optimization/57026
16557 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
16558 from SSA names occuring in abnormal PHI nodes.
16560 2013-04-22 Andi Kleen <ak@linux.intel.com>
16562 * lto/lto.c (print_lto_report_1): Fix LTO report names.
16564 2013-04-22 Andi Kleen <ak@linux.intel.com>
16566 * lto/lto.c (print_lto_report_1): Declare early.
16567 (read_cgraph_and_symbols): Call print_lto_report_1 early.
16569 2013-04-22 Andi Kleen <ak@linux.intel.com>
16571 * common.opt (-flto-report-wpa): Add.
16572 * doc/invoke.texi (-flto-report-wpa): Add.
16573 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
16576 2013-04-22 Xinliang David Li <davidxl@google.com>
16578 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
16579 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
16580 * Makefile.in: New dependency
16582 David Daney <ddaney.cavm@gmail.com>
16584 * configure.ac (gcc_cv_as_micromips_support): Use the
16585 --fatal-warnings option.
16586 * configure: Regenerate.
16588 2013-04-22 Marek Polacek <polacek@redhat.com>
16591 * tsan.c (instrument_expr): Don't instrument expression
16592 in case its size is zero.
16594 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
16598 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
16600 * config/alpha/alpha.c (TARGET_LRA_P): New define.
16602 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
16604 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
16605 (gimple_stmt_iterator): New typedef.
16606 * gimple.h (gimple_stmt_iterator): Rename to...
16607 (gimple_stmt_iterator_d): ... This.
16608 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
16609 trees be valid for GIMPLE and GENERIC.
16610 (TARGET_GIMPLE_FOLD_BUILTIN): New.
16611 * gimple-fold.c (gimple_fold_call): Call target hook
16612 gimple_fold_builtin.
16613 * hooks.c (hook_bool_gsiptr_false): New.
16614 * hooks.h (hook_bool_gsiptr_false): New.
16615 * target.def (fold_stmt): New.
16616 * doc/tm.texi: Regenerate.
16618 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
16621 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
16622 a set sp if no stack realignment.
16624 2013-04-22 Nick Clifton <nickc@redhat.com>
16626 * config.gcc (tilegx-linux): Extend extra_objs rather than
16628 (tilepro-linux): Likewise.
16630 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
16632 * config/aarch64/aarch64-builtins.c
16634 (CF0, CF1, CF2, CF3, CF4, CF10): New.
16635 (VAR<1-12>): Add MAP parameter.
16636 (BUILTIN_*): Likewise.
16637 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
16638 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
16639 (aarch64_ushl_n<mode>): Likewise.
16640 (aarch64_sshr_n<mode>): Likewise.
16641 (aarch64_ushr_n<mode>): Likewise.
16642 (aarch64_<maxmin><mode>): Likewise.
16643 (aarch64_sqrt<mode>): Likewise.
16644 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
16645 (vshr<q>_n_*): Likewise.
16647 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
16649 * config/aarch64/aarch64-builtins.c
16650 (aarch64_simd_builtin_type_mode): Handle SF types.
16652 (BUILTIN_GPF): Define.
16653 (aarch64_init_simd_builtins): Handle SF types.
16654 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
16655 (frecps): Likewise.
16656 (frecpx): Likewise.
16657 * config/aarch64/aarch64-simd.md
16658 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
16659 (aarch64_frecpe<mode>): New.
16660 (aarch64_frecps<mode>): Likewise.
16661 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
16662 (v8type): Add frecp<esx>.
16663 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
16664 (aarch64_frecps<mode>): Likewise.
16665 * config/aarch64/iterators.md (FRECP): New.
16666 (frecp_suffix): Likewise.
16667 * config/aarch64/arm_neon.h
16668 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
16670 2013-04-22 Christian Bruel <christian.bruel@st.com>
16673 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
16674 (REG_CLASS_NAMES): Idem.
16675 (REG_CLASS_CONTENTS): Idem.
16676 (REGCLASS_HAS_FP_REG): Idem.
16677 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
16678 (sh_conditional_register_usage): Idem.
16680 2013-04-21 Jeff Law <law@redhat.com>
16682 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
16683 (ssa_forward_propagate_and_combine): Use it.
16685 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
16687 * lra.c: Update the flow chart diagram.
16689 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
16691 PR rtl-optimization/56847
16692 * lra-constraints.c (process_alt_operands): Discourage alternative
16693 with non-matche doffsettable memory constraint fro memory with
16696 2013-04-19 Richard Biener <rguenther@suse.de>
16698 PR tree-optimization/56982
16699 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
16701 * gimplify.c (gimplify_call_expr): Notice special calls.
16702 (gimplify_modify_expr): Likewise.
16703 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
16704 abnormal control flow receivers.
16705 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
16706 in the same way as cfun->has_nonlocal_labels.
16707 (gimple_purge_dead_abnormal_call_edges): Likewise.
16708 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
16709 receivers start a basic-block.
16711 2013-04-19 Richard Biener <rguenther@suse.de>
16713 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
16715 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
16716 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
16717 (SLP_TREE_LOAD_PERMUTATION): Add.
16718 (vect_transform_slp_perm_load): Adjust prototype.
16719 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
16720 (vect_free_slp_instance): Likewise.
16721 (vect_create_new_slp_node): Likewise.
16722 (vect_supported_slp_permutation_p): Remove.
16723 (vect_slp_rearrange_stmts): Adjust.
16724 (vect_supported_load_permutation_p): Likewise. Inline
16725 vect_supported_slp_permutation_p here.
16726 (vect_analyze_slp_instance): Compute load permutations per
16727 slp node instead of per instance.
16728 (vect_get_slp_defs): Adjust.
16729 (vect_transform_slp_perm_load): Likewise.
16730 (vect_schedule_slp_instance): Remove redundant code.
16731 (vect_schedule_slp): Remove hack for PR56270, add it ...
16732 * tree-vect-stmts.c (vectorizable_load): ... here, do not
16733 CSE loads for SLP. Adjust.
16735 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
16737 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
16738 spelling in two comments.
16740 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
16743 * config/arm/arm.c (load_multiple_sequence): Require SP
16744 as base register for loads if SP is in the register list.
16746 2013-04-19 Martin Jambor <mjambor@suse.cz>
16748 PR tree-optimization/56718
16749 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
16750 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
16751 and made public. Adjusted all callers.
16752 (ipa_intraprocedural_devirtualization): New function.
16753 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
16754 (ipa_intraprocedural_devirtualization): Likewise.
16755 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
16757 2013-04-19 Richard Biener <rguenther@suse.de>
16759 PR tree-optimization/57000
16760 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
16762 2013-04-19 Terry Guo <terry.guo@arm.com>
16764 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
16766 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
16767 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
16768 (cortex_m4_fmacs): Use new reservations.
16769 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
16771 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
16773 PR rtl-optimization/56999
16774 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
16776 (lra_coalesce): Remove split_origin_bitmap and related code.
16777 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
16778 ranges if necessary.
16780 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
16782 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
16784 (ix86_expand_call): Remove clobbered_registers array and use
16785 x86_64_ms_sysv_extra_clobbered_registers instead.
16786 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
16788 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
16790 * config/i386/i386.md (*call_rex64_ms_sysv): Use
16791 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
16792 (*call_value_rex64_ms_sysv): Ditto.
16794 2013-04-18 Cary Coutant <ccoutant@google.com>
16796 * dwarf2out.c (output_pubnames): Check die_perennial_p of
16797 parent instead of die_mark.
16799 2013-04-18 Diego Novillo <dnovillo@google.com>
16801 * gimple.c (create_gimple_tmp): New.
16802 (get_expr_type): New.
16803 (build_assign): New.
16804 (build_type_cast): New.
16805 * gimple.h (enum ssa_mode): Define.
16806 (gimple_seq_set_location): New.
16807 * asan.c (build_check_stmt): Change some gimple_build_* calls
16808 to use build_assign and build_type_cast.
16810 2013-04-18 Richard Biener <rguenther@suse.de>
16812 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
16813 handle negative step. Remove redundant checks.
16814 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
16815 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
16816 for negative step and grouped loads fail to vectorize.
16818 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
16820 * emit-rtl.c (reset_insn_used_flags): New function.
16821 (reset_all_used_flags): Use it.
16822 (verify_insn_sharing): New function.
16823 (verify_rtl_sharing): Fix verification for SEQUENCEs.
16825 2013-04-18 Jakub Jelinek <jakub@redhat.com>
16827 PR tree-optimization/56984
16828 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
16829 and (x >> M) >= N don't register any assertion if N << M is the
16832 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
16834 * lower-subreg.c (resolve_simple_move): If called self-recursive,
16835 do not delete_insn insns that have not yet been emitted, only
16836 unlink them with remove_insn.
16837 * df-scan.c (df_insn_delete): Revert r197492.
16839 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
16841 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
16842 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
16844 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
16846 * config/arm/arm.md (movsicc_insn): Convert define_insn into
16847 define_insn_and_split.
16848 (and_scc,ior_scc,negscc): Likewise.
16849 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
16851 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
16853 * config/arm/arm.c (use_return_insn): Return 0 for targets that
16854 can benefit from using a sequence of LDRD instructions in epilogue
16855 instead of a single LDM instruction.
16857 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
16860 * doc/extend.texi: Fix typo.
16862 2013-04-17 Richard Biener <rguenther@suse.de>
16864 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
16865 (vect_build_slp_tree): ... here.
16866 (vect_build_slp_tree_1): Compute which stmts of the SLP group
16867 match. Remove special-casing of mismatched complex loads.
16868 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
16869 re-try the match with swapped commutative operands.
16870 (vect_supported_load_permutation_p): Remove special-casing of
16871 mismatched complex loads.
16872 (vect_analyze_slp_instance): Adjust.
16874 2013-04-17 Richard Biener <rguenther@suse.de>
16876 PR rtl-optimization/56921
16877 * cfgloop.h (struct loop): Add simple_loop_desc member.
16878 (struct niter_desc): Mark with GTY(()).
16879 (simple_loop_desc): Do not use aux field but simple_loop_desc.
16880 * loop-iv.c (get_simple_loop_desc): Likewise.
16881 (free_simple_loop_desc): Likewise.
16884 2013-04-16 Richard Biener <rguenther@suse.de>
16886 PR rtl-optimization/56921
16887 * loop-init.c (pass_rtl_move_loop_invariants): Add
16888 TODO_do_not_ggc_collect to todo_flags_finish.
16889 (pass_rtl_unswitch): Same.
16890 (pass_rtl_unroll_and_peel_loops): Same.
16891 (pass_rtl_doloop): Same.
16893 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
16895 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
16896 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
16897 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
16899 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
16900 * tree-streamer.c (record_common_node): Adjust reference in comment.
16902 2013-04-17 Terry Guo <terry.guo@arm.com>
16904 * config/arm/cortex-m4.md: Add a new bypass.
16906 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16908 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
16910 (*subs_<optab><mode>_multp2): New pattern.
16911 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16912 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16914 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16916 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
16917 (*subs_mul_imm_<mode>): New pattern.
16919 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
16922 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
16923 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
16924 (vsx_movti_32bit): Change j->wa to O->wa.
16926 2013-04-16 Richard Biener <rguenther@suse.de>
16928 PR rtl-optimization/56921
16929 * loop-init.c (pass_rtl_move_loop_invariants): Add
16930 TODO_do_not_ggc_collect to todo_flags_finish.
16931 (pass_rtl_unswitch): Same.
16932 (pass_rtl_unroll_and_peel_loops): Same.
16933 (pass_rtl_doloop): Same.
16935 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
16937 * config/arm/arm.c (emit_multi_reg_push): New declaration
16938 for an existing function.
16939 (arm_emit_strd_push): New function.
16940 (arm_expand_prologue): Used here.
16941 (arm_emit_ldrd_pop): New function.
16942 (arm_expand_epilogue): Used here.
16943 (arm_get_frame_offsets): Update condition.
16944 (arm_emit_multi_reg_pop): Add a special case for load of a single
16945 register with writeback.
16947 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
16949 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
16952 2013-04-16 Richard Biener <rguenther@suse.de>
16954 PR tree-optimization/56756
16955 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
16956 (first_mem_ref_loc): New.
16957 (execute_sm): Place the load temporarily before a previous
16958 access instead of in the latch edge to ensure its SSA dependencies
16959 are defined at points dominating the load.
16961 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
16963 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
16964 correct fix by moving header and footer insn to the footer of
16965 the merged basic block. Clear BB_END of the merged-away block.
16967 PR middle-end/43631
16968 * emit-rtl.c (make_note_raw): New function.
16969 (link_insn_into_chain): New static inline function.
16970 (add_insn): Use it.
16971 (add_insn_before, add_insn_after): Factor insn chain linking code...
16972 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
16973 using link_insn_into_chain.
16974 (note_outside_basic_block_p): New helper function for emit_note_after
16975 and emit_note_before.
16976 (emit_note_after): Use nobb variant of add_insn_after if the note
16977 should not be contained in a basic block.
16978 (emit_note_before): Use nobb variant of add_insn_before if the note
16979 should not be contained in a basic block.
16980 (emit_note_copy): Use make_note_raw.
16981 (emit_note): Likewise.
16982 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
16983 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
16984 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
16985 the moved barrier the tail of the basic block it follows.
16986 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
16988 2013-04-15 Jakub Jelinek <jakub@redhat.com>
16990 PR tree-optimization/56962
16991 * gimple-ssa-strength-reduction.c (record_increment): Only set
16992 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
16993 either rhs1 or rhs2 is equal to c->base_expr.
16995 2013-04-15 Richard Biener <rguenther@suse.de>
16997 PR tree-optimization/56933
16998 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
17000 (GROUP_READ_WRITE_DEPENDENCE): Remove.
17001 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
17002 * tree-vect-data-refs.c (vect_analyze_group_access): Move
17003 dependence check ...
17004 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
17006 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
17007 GROUP_READ_WRITE_DEPENDENCE.
17009 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17011 * emit-rtl.c (reset_all_used_flags): New function.
17012 (verify_rtl_sharing): Call reset_all_used_flags before and after
17013 performing the checks.
17015 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17017 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
17018 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
17019 * config/arm/constraints.md (De): New constraint.
17020 * config/arm/neon.md (anddi3_neon): Delete.
17021 (neon_vand<mode>): Expand to standard anddi3 pattern.
17022 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
17023 Move earlier in the file.
17024 (neon_inv_logic_op2): Likewise.
17025 (arm_anddi_operand_neon): New predicate.
17027 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17029 * configure.ac (gcc_cv_ld_as_needed): Set
17030 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
17031 Use -z ignore, -z record on *-*-solaris2*.
17032 (HAVE_LD_AS_NEEDED): Update comment.
17033 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
17034 * configure: Regenerate.
17035 * config.in: Regenerate.
17036 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
17037 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
17038 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
17039 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
17040 equivalents. Fix markup.
17041 * doc/tm.texi: Regenerate.
17043 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
17045 * config/i386/i386.opt: New option mstack-protector-guard=.
17046 * config/i386/i386-opts.h: Add enum stack_protector_guard.
17047 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
17048 TARGET_SSP_TLS_GUARD.
17049 * config/i386/i386.c (ix86_option_override_internal): Set
17050 ix86_stack_protector_guard.
17051 * config/i386/i386.md (stack_protect_set): Enable for
17052 TARGET_SSP_TLS_GUARD only.
17053 (stack_protect_set_<mode>): Ditto.
17054 (stack_protect_test): Ditto.
17055 (stack_protect_test_<mode>): Ditto.
17056 * doc/invoke.texi (i386 Option): Document.
17058 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
17061 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
17062 (S_MODES): Set H_MODE bit.
17063 (SF_MODES): Set only S_MODE and SF_MODE bits.
17064 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
17065 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
17066 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
17067 <MODE_FLOAT>: Likewise.
17069 2013-04-15 Joey Ye <joey.ye@arm.com>
17071 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
17073 2013-04-15 Joey Ye <joey.ye@arm.com>
17075 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
17077 (thumb_far_jump_used_p): Count instruction size and set
17080 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
17082 * reorg.c (fill_simple_delay_slots): Reindent block of code.
17083 * resource.c (mark_target_live_regs): Reformat conditional block.
17085 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
17087 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
17088 notes, they are emitted only just before final.
17089 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
17091 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
17093 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
17094 * cfgrtl.c (delete_insn): Call it here instead.
17095 * lra-spills.c (lra_final_code_change): Use delete_insn.
17096 * haifa-sched.c (sched_remove_insn): Likewise.
17097 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
17098 returning to the nop pool.
17099 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
17100 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
17102 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
17104 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
17105 * doc/tm.texi: Regenerated.
17107 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
17109 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
17112 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
17114 * df-core.c (df_find_def): Compare register numbers.
17115 (df_find_use): Likewise.
17117 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
17120 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
17121 lra_in_progress for return.
17123 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
17125 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
17126 define_insn into define_insn_and_split and emit movsicc patterns.
17128 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
17130 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
17132 2013-04-12 Richard Biener <rguenther@suse.de>
17134 * tree-pass.h (TODO_do_not_ggc_collect): New.
17135 * passes.c (execute_one_ipa_transform_pass): Honor
17136 TODO_do_not_ggc_collect.
17137 (execute_one_pass): Likewise.
17140 2013-04-10 Richard Biener <rguenther@suse.de>
17142 * passes.c (init_optimization_passes): Remove reload pass.
17143 * ira.c (do_reload): Merge into ...
17145 (rest_of_handle_reload): Remove.
17146 (pass_reload): Likewise.
17147 * config/i386/i386.c (ix86_option_override): Refer to ira instead
17148 of reload for vzeroupper pass placement.
17150 2013-04-12 Jakub Jelinek <jakub@redhat.com>
17152 PR tree-optimization/56918
17153 PR tree-optimization/56920
17154 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
17155 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
17156 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
17157 use wide_mul_with_sign method.
17159 2013-04-12 Richard Biener <rguenther@suse.de>
17161 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
17162 not be considered a gimple constant.
17164 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
17166 * fold-const.c (const_binop): Handle vector shifts by a scalar.
17167 (fold_binary_loc): Call const_binop also for mixed vector-scalar
17170 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
17171 Jakub Jelinek <jakub@redhat.com>
17173 * opts.c: Include diagnostic-color.h.
17174 (common_handle_option): Handle OPT_fdiagnostics_color_.
17175 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
17176 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
17177 (diagnostic-color.o): New.
17178 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
17179 (diagnostic_color_rule): New enum.
17180 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
17181 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
17182 the location string.
17183 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
17184 either NULL, or color kind.
17185 * diagnostic-color.c: New file.
17186 * diagnostic-color.h: New file.
17187 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
17189 * doc/invoke.texi (-fdiagnostics-color): Document.
17190 * pretty-print.h (pp_show_color): Define.
17191 (struct pretty_print_info): Add show_color field.
17192 * diagnostic.c: Include diagnostic-color.h.
17193 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
17194 macros. Colorize error:, warning: etc. strings and also the location
17196 (diagnostic_show_locus): Colorize the caret line.
17197 * pretty-print.c: Include diagnostic-color.h.
17198 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
17199 inside of %< %> quotes or quoted through q format modifier.
17201 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17203 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
17205 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17207 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
17208 code in CC_NZ mode.
17209 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
17212 2013-04-11 Marek Polacek <polacek@redhat.com>
17214 PR tree-optimization/48184
17215 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
17217 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
17219 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
17220 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
17221 (skip_simple_arithmetic): Tidy up.
17222 * tree.h (skip_simple_constant_arithmetic): Declare.
17224 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17226 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
17228 2013-04-11 Richard Biener <rguenther@suse.de>
17230 * tree-vect-loop.c (get_initial_def_for_induction): Properly
17231 generate vector constants.
17233 2013-04-11 Richard Biener <rguenther@suse.de>
17235 PR tree-optimization/56878
17236 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
17237 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
17239 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
17240 Prefer to align the DR with the most invariant base address.
17242 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
17244 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
17247 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
17249 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
17250 floating-point vector comparisons against 0.
17252 2013-04-11 Jakub Jelinek <jakub@redhat.com>
17254 PR tree-optimization/56899
17255 * fold-const.c (extract_muldiv_1): Apply distributive law
17256 only if TYPE_OVERFLOW_WRAPS (ctype).
17258 2013-04-11 Bin Cheng <bin.cheng@arm.com>
17261 * ira-costs.c (scan_one_insn): Check whether the source rtx of
17262 loading has side effect.
17264 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
17266 * config/sparc/sparc.c: Include tree-pass.h.
17267 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
17268 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
17269 head of file. Change return type. Split off gate function.
17270 (sparc_gate_work_around_errata): New function.
17271 (pass_work_around_errata): New pass definition.
17272 (insert_pass_work_around_errata) New pass insert definition to
17273 insert pass_work_around_errata just after delayed-branch scheduling.
17274 (sparc_option_override): Insert the pass.
17275 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
17277 2013-04-10 David S. Miller <davem@davemloft.net>
17279 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
17280 or -mcpu=hypersparc.
17282 * target.def (cstore_mode): New hook.
17283 * target.h: Include insn-codes.h
17284 * targhooks.c: Likewise.
17285 (default_cstore_mode): New function.
17286 * targhooks.h: Declare it.
17287 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
17288 * doc/tm.texi: Rebuild.
17289 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
17290 target hook, rather than inspecting the insn_data.
17291 * config/sparc/sparc.c (sparc_cstore_mode): New function.
17292 (TARGET_CSTORE_MODE): Redefine.
17293 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
17295 * config/sparc/predicates.md (cstore_result_operand): New special
17297 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
17298 Use it for operand 0.
17299 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
17300 (*snesi_special): Likewise.
17301 (*snesi_zero): Likewise.
17302 (*seqsi_zero): Likewise.
17303 (*sltu_insn): Likewise.
17304 (*sgeu_insn): Likewise.
17305 (*seqdi_special): Make operand 0 and comparison operation be of
17307 (*snedi_special): Likewise.
17308 (*snedi_special_vis3): Likewise.
17309 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
17310 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
17311 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
17312 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
17313 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
17314 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
17315 (*sltu_extend_sp64): Likewise.
17316 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
17317 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
17318 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
17319 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
17320 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
17322 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
17324 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
17325 (aarch64_start_file): Use the new function.
17327 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
17328 Jason Merrill <jason@redhat.com>
17330 * common.opt: Add -gdwarf.
17331 * opts.c (common_handle_option): Handle it.
17332 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
17334 2013-04-10 Richard Biener <rguenther@suse.de>
17336 * passes.c (execute_todo): Do not call ggc_collect conditional here.
17337 (execute_one_ipa_transform_pass): But unconditionally here.
17338 (execute_one_pass): And here.
17339 (init_optimization_passes): Remove reload pass.
17340 * tree-pass.h (TODO_ggc_collect): Remove.
17341 (pass_reload): Likewise.
17342 * ira.c (do_reload): Merge into ...
17344 (rest_of_handle_reload): Remove.
17345 (pass_reload): Likewise.
17346 * config/i386/i386.c (ix86_option_override): Refer to ira instead
17347 of reload for vzeroupper pass placement.
17348 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
17349 and todo_flags_finish of all passes.
17351 2013-04-10 Richard Biener <rguenther@suse.de>
17353 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
17354 first_const_oprnd field, rename first_def_type to first_op_type.
17355 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
17356 (vect_get_and_check_slp_defs): Always use the type of the
17357 operand. Allow mixed vect_external_def, vect_constant_def types.
17358 (vect_get_constant_vectors): Handle mixed vect_external_def,
17359 vect_constant_def types.
17361 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
17363 PR tree-optimization/55524
17364 * tree-ssa-math-opts.c
17365 (convert_mult_to_fma): Don't use an fms construct
17366 when we don't have an fms operation, but fnma, and it looks
17367 likely that we'll be able to use the latter.
17369 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
17371 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
17373 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
17374 inline fail caused by overwritable functions.
17376 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
17378 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
17379 unnecessary bits in the constant power of two case.
17381 2013-04-10 Richard Biener <rguenther@suse.de>
17383 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
17384 broken code swapping operands.
17385 (vect_build_slp_tree): Do not compute load permutations here.
17386 (vect_analyze_slp_instance): Compute load permutations here,
17387 after building the SLP tree.
17389 2013-04-09 Christian Bruel <christian.bruel@st.com>
17391 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
17392 of next/prev_real_insn.
17394 2013-04-09 Jan Hubicka <jh@suse.cz>
17396 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17397 Drop aliased parameter.
17398 (function_and_variable_visibility): Do not handle alias pairs.
17399 * cgraph.c (varpool_externally_visible_p): Update prototype.
17400 * varpool.c (varpool_add_new_variable): Update.
17402 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17404 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
17406 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
17408 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
17410 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
17412 2013-04-09 Marek Polacek <polacek@redhat.com>
17414 PR tree-optimization/48762
17415 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
17417 2013-04-09 Richard Biener <rguenther@suse.de>
17419 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
17421 (vect_build_slp_tree): Likewise.
17422 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
17423 calculating the cost of a SLP instance.
17424 (vect_analyze_slp_instance): Use it from here, after building
17427 2013-04-09 Jakub Jelinek <jakub@redhat.com>
17429 PR middle-end/56883
17430 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
17431 expand_omp_for_static_chunk): Use simple_p = true in
17432 force_gimple_operand_gsi calls when assigning to addressable decls.
17434 2013-04-09 Jeff Law <law@redhat.com>
17436 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
17437 when the boolean was created by converting a wider object which
17438 had a boolean range.
17440 2013-04-09 Richard Biener <rguenther@suse.de>
17442 * tree-vectorizer.h (slp_void_p): Remove.
17443 (slp_tree): Typedef before _slp_tree declaration.
17444 (struct _slp_tree): Use a vector of slp_tree as children.
17445 (vect_get_place_in_interleaving_chain): Remove.
17446 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
17448 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
17450 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
17451 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
17452 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
17453 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
17454 Use slp_node instead of slp_void_p and adjust.
17456 2013-04-09 Richard Biener <rguenther@suse.de>
17458 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
17459 work that is not necessary.
17461 2013-04-09 Jakub Jelinek <jakub@redhat.com>
17463 PR tree-optimization/56854
17464 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
17465 forward into clobber stmts if it would change MEM_REF lhs into
17468 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
17470 * tree.c (type_hash_lookup, type_hash_add): Make static.
17471 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
17473 2013-04-09 Richard Biener <rguenther@suse.de>
17475 * tree.h (unsave_expr_now): Remove.
17476 * tree-inline.c (mark_local_for_remap_r): Remove.
17477 (unsave_expr_1): Likewise.
17478 (unsave_r): Likewise.
17479 (unsave_expr_now): Likewise.
17480 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
17481 (propagate_tree_value): Likewise.
17483 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
17485 * doc/rtl.texi (sequence): Rewrite documentation to match the
17486 current use of SEQUENCE rtl objects.
17487 * rtl.def (SEQUENCE): Likewise.
17489 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
17490 Update documentation.
17491 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
17492 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
17494 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
17496 2013-04-08 Teresa Johnson <tejohnson@google.com>
17498 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
17499 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
17501 (estimate_edge_size_and_time): Add comment to suggest using rounding
17503 (estimate_node_size_and_time): Ditto.
17504 (remap_edge_change_prob): Use helper rounding divide methods.
17505 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
17506 (gimple_mod_pow2_value_transform): Ditto.
17507 (gimple_mod_subtract_transform): Ditto.
17508 (gimple_ic_transform): Ditto.
17509 (gimple_stringops_transform): Ditto.
17510 * stmt.c (conditional_probability): Ditto.
17511 (emit_case_dispatch_table): Ditto.
17512 * lto-cgraph.c (merge_profile_summaries): Ditto.
17513 * tree-optimize.c (execute_fixup_cfg): Ditto.
17514 * cfgcleanup.c (try_forward_edges): Ditto.
17515 * cfgloopmanip.c (scale_loop_profile): Ditto.
17517 (duplicate_loop_to_header_edge): Ditto.
17518 (lv_adjust_loop_entry_edge): Ditto.
17519 * tree-vect-loop.c (vect_transform_loop): Ditto.
17520 * profile.c (compute_branch_probabilities): Ditto.
17521 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
17522 * lto-streamer-in.c (input_cfg): Ditto.
17523 * gimple-streamer-in.c (input_bb): Ditto.
17524 * ipa-cp.c (update_profiling_info): Ditto.
17525 (update_specialized_profile): Ditto.
17526 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
17527 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
17529 * sched-rgn.c (compute_dom_prob_ps): Ditto.
17530 (compute_trg_info): Ditto.
17531 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
17532 (purge_dead_edges): Ditto.
17533 * loop-unswitch.c (unswitch_loop): Ditto.
17534 * cgraphclones.c (cgraph_clone_edge): Ditto.
17535 (cgraph_clone_node): Ditto.
17536 * tree-inline.c (copy_bb): Ditto.
17537 (copy_edges_for_bb): Ditto.
17538 (initialize_cfun): Ditto.
17539 (copy_cfg_body): Ditto.
17540 (expand_call_inline): Ditto.
17542 2013-04-08 Kai Tietz <ktietz@redhat.com>
17544 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
17545 TARGET_CYGWIN64 by TARGET_64BIT.
17547 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
17549 * config/epiphany/epiphany.md (GPR_1): New constant.
17550 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
17551 * config/epiphany/epiphany.c (gen_compare_reg):
17552 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
17553 is already in place.
17554 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
17555 Don't require being called during rtl expansion; If y operlaps r0,
17557 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
17558 (epiphany_expand_epilogue): Likewise.
17560 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
17561 Don't use CC_FPmode for ORDERED / UNORDERED.
17562 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
17564 * config/epiphany/constraints.md (CnL): New constraint.
17565 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
17566 * config/epiphany/predicates.md (add_operand): Allow 1024.
17568 * config/epiphany/epiphany.md (logical_op): New code iterator.
17569 (op_mnc): New code attribute.
17570 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
17571 (mov_f+1, mov_f+2): New peephole2 patterns.
17573 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
17574 (cstoresi4): Also allow re-use of zero result when doing a NE
17575 comparison to a non-zero operand.
17576 Use (clobber (scratch)) for first insn if the gpr output is not needed.
17578 * config/epiphany/epiphany.md (<insn_opname>v2si3):
17579 Use gen_addsi3_i / gen_subsi3_i.
17581 2013-04-08 Jakub Jelinek <jakub@redhat.com>
17585 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
17586 contain anything but clobbers, at most one __builtin_stack_restore,
17587 optionally debug stmts and final resx, and if it has at least one
17588 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
17589 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
17590 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
17591 which isn't defaut definition, remove them.
17592 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
17593 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
17594 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
17595 with MEM_REF LHS with SSA_NAME address.
17597 2013-04-08 Jeff Law <law@redhat.com>
17599 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
17601 2013-04-08 Richard Biener <rguenther@suse.de>
17603 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
17605 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
17606 determined vector type.
17607 (vect_analyze_data_refs): Likewise.
17608 (vect_get_new_vect_var): Adjust.
17609 (vect_create_destination_var): Preserve SSA name versions.
17610 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
17611 not dump anything here.
17613 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
17615 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
17616 Add member lr_slot_known.
17617 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
17619 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
17620 Remove code that sets lr_slot_offset according to what a previous
17621 version of epiphany_emit_save_restore used to do.
17622 (epiphany_emit_save_restore): When doing an lr save or restore,
17623 set/verify lr_slot_known and lr_slot_offset.
17625 2013-04-08 Xinyu Qi <xyqi@marvell.com>
17628 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
17631 2013-04-08 Richard Biener <rguenther@suse.de>
17633 * alias.c (find_base_term): Fix thinko in previous change.
17635 2013-04-08 Jakub Jelinek <jakub@redhat.com>
17637 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
17638 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
17639 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
17640 if possible to compute val.
17641 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
17642 For QImode integers don't require anything about precision. Use
17643 const_with_all_bytes_same to find out if the constant doesn't have
17644 repeated bytes in it.
17646 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17648 * config/s390/s390.c (s390_expand_insv): Only accept insertions
17651 2013-04-08 Marek Polacek <polacek@redhat.com>
17653 PR rtl-optimization/48182
17654 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
17657 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17660 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
17661 nuses, make sure we have a label.
17663 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17666 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
17667 (rs6000_emit_swdiv_low_precision): Remove.
17668 (rs6000_emit_swdiv): Rewrite to handle between one and four
17669 iterations of Newton-Raphson generally; modify required number of
17670 iterations for some cases.
17671 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
17673 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
17675 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
17676 set-but-unused variable.
17678 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
17679 basic blocks of released function bodies garbage-collectable.
17681 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
17682 (struct rtl_opt_pass): Add TODO_df_finish.
17684 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
17686 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17688 * config/arm/constraints.md (q): New constraint.
17689 * config/arm/ldrdstrd.md: New file.
17690 * config/arm/arm.md (ldrdstrd.md) New include.
17691 (arm_movdi): Use "q" instead of "r" constraint
17692 for double-word memory access.
17693 (movdf_soft_insn): Likewise.
17694 * config/arm/vfp.md (movdi_vfp): Likewise.
17695 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
17696 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
17697 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
17698 (mem_ok_for_ldrd_strd): Likewise.
17699 (output_move_double): Update assertion.
17701 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17703 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
17705 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17707 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
17708 define_insn_and_split.
17709 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
17711 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17713 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
17714 define_insn_and_split.
17715 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
17716 (shiftsi3_compare): New pattern.
17717 (rrx): New pattern.
17718 * config/arm/unspecs.md (UNSPEC_RRX): New.
17720 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17722 * config/arm/arm.md (negdi_extendsidi): New pattern.
17723 (negdi_zero_extendsidi): Likewise.
17725 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17727 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
17728 define_insn_and_split.
17729 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
17730 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
17732 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17734 * config/arm/arm.md (arm_subdi3): Convert define_insn into
17735 define_insn_and_split.
17736 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
17737 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
17739 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17741 * config/arm/arm.md (subsi3_carryin): New pattern.
17742 (subsi3_carryin_const): Likewise.
17743 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
17744 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
17746 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17748 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
17750 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17752 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
17753 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
17755 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17757 * config/arm/arm.c (arm_expand_builtin): Change fcode
17758 type to unsigned int.
17760 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17762 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
17764 2013-04-04 Ian Lance Taylor <iant@google.com>
17766 * doc/standards.texi (Standards): The Go frontend supports the Go 1
17769 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
17771 PR middle-end/56729
17772 * df-scan.c (df_insn_delete): Disable failing assert.
17774 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17776 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
17777 New function prototype.
17778 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
17779 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
17780 (arm_builtin_vectorized_function): New function.
17782 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17784 * config/arm/arm_neon_builtins.def: New file.
17785 * config/arm/arm.c (neon_builtin_data): Move contents to
17786 arm_neon_builtins.def.
17787 (enum arm_builtins): Include neon builtin definitions.
17788 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
17789 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
17791 2013-04-04 Marek Polacek <polacek@redhat.com>
17793 PR tree-optimization/48186
17794 * predict.c (maybe_hot_frequency_p): Return false if
17795 HOT_BB_FREQUENCY_FRACTION is 0.
17796 (cgraph_maybe_hot_edge_p): Likewise.
17798 2013-04-04 Richard Biener <rguenther@suse.de>
17800 PR tree-optimization/56826
17801 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
17804 2013-04-04 Richard Biener <rguenther@suse.de>
17806 PR tree-optimization/56213
17807 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
17808 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
17810 2013-04-04 Richard Biener <rguenther@suse.de>
17812 PR tree-optimization/56837
17813 * tree-loop-distribution.c (classify_partition): For non-zero
17814 values require that the value has the same precision as its
17815 mode to be useful as memset value.
17817 2013-04-03 Nick Clifton <nickc@redhat.com>
17819 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
17820 (fmssf4): Use fmsf.s on E3V5 architectures.
17821 (fnmasf4): Use fnmaf.s on E3V5 architectures.
17822 (fnmssf4): Use fnmsf.s on E3V5 architectures.
17824 2013-04-03 Jeff Law <law@redhat.com>
17826 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
17827 (lra-eliminations.o): Likewise.
17829 2013-04-03 Teresa Johnson <tejohnson@google.com>
17831 * gcov-io.c (compute_working_sets): Moved most of body of old
17832 compute_working_sets here from profile.c.
17833 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
17834 (gcov_working_set_t): Moved typedef here from basic-block.h
17835 (compute_working_set): Declare.
17836 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
17837 (get_working_sets): Renamed from compute_working_set,
17838 replace most of body with call to new compute_working_sets.
17839 (get_exec_counts): Replace call to compute_working_sets
17840 to get_working_sets.
17841 * profile.h (get_working_sets): Renamed from compute_working_set.
17842 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
17843 to get_working_sets.
17844 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
17845 * gcov-dump.c (dump_working_sets): New function.
17847 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
17849 * hwint.c (sext_hwi, zext_hwi): New functions.
17850 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
17851 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
17852 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
17853 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
17854 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
17855 (sext_hwi, zext_hwi): New functions.
17857 2013-04-03 Jeff Law <law@redhat.com>
17859 PR tree-optimization/56799
17860 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
17861 back test for widening conversion erroneously dropped in prior change.
17863 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17866 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
17867 instead of next_real_insn.
17869 2013-04-03 Marek Polacek <polacek@redhat.com>
17872 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
17874 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17877 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
17879 (thumb1_output_casesi): Likewise.
17880 (thumb2_output_casesi): Likewise.
17882 2013-04-03 Richard Biener <rguenther@suse.de>
17884 PR tree-optimization/56817
17885 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
17887 (tree_unroll_loops_completely_1): ... new function to manually
17888 walk the loop tree, properly defering outer loops of unrolled
17889 loops to later iterations.
17891 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17893 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
17894 (vectorizable_load): Likewise.
17895 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
17896 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
17898 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17900 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
17903 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
17905 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
17907 2013-04-03 Bin Cheng <bin.cheng@arm.com>
17909 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
17911 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17913 PR tree-optimization/56790
17914 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
17917 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17919 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
17921 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
17922 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
17925 2013-04-03 Jakub Jelinek <jakub@redhat.com>
17928 * tree.h (force_folding_builtin_constant_p): New decl.
17929 * builtins.c (force_folding_builtin_constant_p): New variable.
17930 (fold_builtin_constant_p): Fold immediately also if
17931 force_folding_builtin_constant_p.
17933 2013-04-03 Richard Biener <rguenther@suse.de>
17935 PR tree-optimization/56812
17936 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
17937 DRs of the same interleaving chain are independent.
17939 2013-04-02 Jason Merrill <jason@redhat.com>
17941 * gdbinit.in (pbb): Use debug fn.
17943 2013-04-02 Lawrence Crowl <crowl@google.com>
17945 * sese.h (struct ivtype_map_elt_s): Remove unused.
17946 (extern debug_ivtype_map): Remove unused.
17947 (extern eq_ivtype_map_elts): Remove unused.
17948 * sese.c (debug_ivtype_map): Removed unused.
17949 (debug_ivtype_map_1): Removed unused.
17950 (debug_ivtype_elt): Remove unused.
17951 (eq_ivtype_map_elts): Remove unused.
17954 2013-04-02 Kai Tietz <ktietz@redhat.com>
17957 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
17958 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
17959 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
17961 (legitimize_pe_coff_symbol): Likewise.
17962 (is_imported_p): New helper-function.
17963 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
17964 for Windows x64 targets.
17965 (ix86_expand_prologue): Optimize for pe-coff targets.
17966 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
17967 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
17968 medium/large code-model.
17969 (legitimize_pic_address): Likewise.
17970 (legitimize_tls_address): Likewise.
17971 (ix86_expand_call): Likewise.
17972 (x86_output_mi_thunk): Likewise.
17973 (get_dllimport_decl): Add new beimport argument.
17974 (construct_plt_address): Don't assert for x64 pe-coff targets.
17975 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
17977 (SYMBOL_FLAG_STUBVAR): New macro.
17978 (SYMBOL_REF_STUBVAR_P): Likewise.
17979 * config/i386/winnt.c (stub_list): New structure.
17980 (stub_head): New local variable.
17981 (i386_pe_record_stub): New function.
17982 (i386_pe_file_end): Emit refptr-stubs.
17984 2013-04-02 Jakub Jelinek <jakub@redhat.com>
17986 PR rtl-optimization/56745
17987 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
17988 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
17991 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
17992 and both of them are MEM_REFs, just compare first argument for
17993 equality and attempt to deal even with differing offsets.
17996 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
17997 of gimple_clobber_p to be MEM_REF.
17998 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
17999 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
18000 after gimplification.
18001 * asan.c (get_mem_ref_of_assignment): Don't instrument
18002 gimple_clobber_p stmts.
18003 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
18004 gimple_clobber_p stmt if they have MEM_REF lhs and
18005 are dead because of another gimple_clobber_p stmt.
18006 * tree-ssa-live.c (clear_unused_block_pointer): Treat
18007 gimple_clobber_p stmts like debug stmts.
18008 (remove_unused_locals): Remove clobbers with MEM_REF lhs
18009 that refer to unused VAR_DECLs or uninitialized values.
18010 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
18011 gimple_clobber_p stmts if they refer to removed parameters.
18012 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
18015 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
18017 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
18018 using SWI48 mode attribute.
18020 2013-04-02 Wei Mi <wmi@google.com>
18022 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
18023 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
18024 *<rotate_insn><mode>3_mask in i386.md.
18026 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
18028 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
18030 2013-04-02 Richard Biener <rguenther@suse.de>
18032 PR tree-optimization/56778
18033 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
18034 Runtime alias tests are not supported for gather loads.
18035 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
18036 stmts referenced from SSA operands before updating SSA form.
18038 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
18039 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18041 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
18042 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
18043 * config/arm/cortex-a53.md: New file.
18044 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
18045 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
18046 * config/arm/arm.c (arm_issue_rate): Likewise.
18047 * config/arm/arm-tune.md: Regenerate
18048 * config/arm/arm-tables.opt: Regenerate.
18049 * config/arm/arm-cores.def: Add cortex-a53.
18051 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
18053 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
18056 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
18058 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
18059 scalar load/store operations using B/H registers.
18060 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
18062 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
18064 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
18066 * config/aarch64/aarch64.c
18067 (aarch64_simd_scalar_immediate_valid_for_move): New.
18068 * config/aarch64/aarch64-protos.h
18069 (aarch64_simd_scalar_immediate_valid_for_move): New.
18070 * config/aarch64/constraints.md (Dh, Dq): New.
18071 * config/aarch64/iterators.md (hq): New.
18073 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
18075 * reorg.c (get_branch_condition): Deal with conditional returns.
18076 (fill_simple_delay_slots): Remove dead code dealing with jumps.
18078 2013-04-01 Wei Mi <wmi@google.com>
18080 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
18081 Truncate operand 2 using %b asm operand modifier.
18082 (*<shift_insn><mode>3_mask): Ditto.
18083 (*<rotate_insn><mode>3_mask): Ditto.
18085 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
18087 PR middle-end/56798
18088 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
18090 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
18092 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
18094 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
18096 2013-03-30 Lawrence Crowl <crowl@google.com>
18098 * dse.c (clear_alias_sets): Remove never set.
18099 (disqualified_clear_alias_sets): Remove never set.
18100 (clear_alias_mode_pool): Remove never set.
18101 (dse_step0): Remove condition that is never true.
18102 (canon_address): Remove condition that is never true.
18103 (dse_step7): Remove condition that is never true.
18104 (rest_of_handle_dse): Remove condition that is never true.
18105 (rest_of_handle_dse::did_global): Remove never read from above.
18106 (dse_step2_spill): Remove never called from above.
18107 (dse_step5_spill): Remove never called from above.
18109 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
18111 * doc/md.texi (Standard Names) <casesi>: Update documentation for
18112 JUMP_TABLE_DATA changes.
18113 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
18114 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
18115 (Insns) <jump_table_data>: New entry.
18116 * doc/tm.texi: Regenerate.
18118 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
18120 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
18121 for table jump at the end of a basic block using tablejump_p.
18122 * targhooks.c (default_invalid_within_doloop): Likewise.
18123 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
18124 target hook implementation that is identical to the default hook.
18125 (rs6000_invalid_within_doloop): Remove.
18127 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
18128 unused variable from tablejump_p call.
18130 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
18131 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
18132 (INSN_DELETED_P): Likewise.
18133 (emit_jump_table_data): New prototype.
18134 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
18135 after 4th as unused.
18136 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
18137 * sched-vis.c (print_insn): Likewise.
18138 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
18139 insn for compatibility with back ends that use next_active_insn to
18140 identify jump table data.
18141 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
18142 (remove_insn): Likewise.
18143 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
18145 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
18146 (emit_jump_table_data): New function.
18148 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
18149 basic block, a JUMP_TABLE_DATA never is.
18150 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
18151 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
18152 off from code handling real insns.
18153 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
18154 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
18155 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
18156 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
18157 is not a NONDEBUG_INSN_P.
18158 * ira-costs.c (scan_one_insn): Likewise.
18159 * jump.c (mark_all_labels): Likewise.
18160 (mark_jump_label_1): Likewise.
18161 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
18162 * lra.c (get_insn_freq): Expect all insns reaching here to be in
18164 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
18165 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
18166 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
18167 JUMP_TABLE_DATA_P insns.
18168 (calculate_elim_costs_all_insns): Likewise.
18169 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
18170 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
18171 (delete_output_reload): Code style fixups.
18172 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
18173 insn flags on this non-insn.
18174 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
18175 as scheduling barriers, for pre-change compatibility.
18176 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
18177 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
18179 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
18180 redundant JUMP_TABLE_DATA_P test.
18181 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
18182 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
18183 (frv_for_each_packet): Likewise.
18184 * config/i386/i386.c (min_insn_size): Likewise.
18185 (ix86_avoid_jump_mispredicts): Likewise.
18186 * config/m32r/m32r.c (m32r_is_insn): Likewise.
18187 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
18188 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
18189 (mips16_insn_length): Robustify.
18190 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
18191 (mips16_split_long_branches): Likewise.
18192 * config/pa/pa.c (pa_combine_instructions): Likewise.
18193 * config/rs6000/rs6000.c (get_next_active_insn): Treat
18194 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
18195 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
18196 as contributing to pool range lengths.
18197 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
18198 Remove redundant JUMP_TABLE_DATA_P test.
18199 (sh_loop_align): Likewise.
18200 (split_branches): Likewise.
18201 (sh_insn_length_adjustment): Likewise.
18202 * config/spu/spu.c (get_branch_target): Likewise.
18204 2013-03-29 Jan Hubicka <jh@suse.cz>
18206 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
18207 gcov streaming; stream hot bb threshold to ltrans.
18208 * predict.c (get_hot_bb_threshold): Break out from ....
18209 (maybe_hot_count_p): ... here.
18210 (set_hot_bb_threshold): New function.
18211 * lto-section-in.c (lto_section_name): Add profile.
18212 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
18213 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
18214 and data-streamer.h
18215 (histogram_entry): New structure.
18216 (histogram, histogram_pool): New global vars.
18217 (histogram_hash): New structure.
18218 (histogram_hash::hash): New method.
18219 (histogram_hash::equal): Likewise.
18220 (account_time_size): New function.
18221 (cmp_counts): New function.
18222 (dump_histogram): New function.
18223 (ipa_profile_generate_summary): New function.
18224 (ipa_profile_write_summary): New function.
18225 (ipa_profile_read_summary): New function.
18226 (ipa_profile): Decide on threshold.
18227 (pass_ipa_profile): Add ipa_profile_write_summary and
18228 ipa_profile_read_summary.
18229 * Makefile.in (ipa.o): Update dependencies.
18230 * lto-streamer.h (LTO_section_ipa_profile): New section.
18232 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
18234 * tree.h (VAR_P): New.
18236 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
18239 * doc/invoke.texi ([-fwhole-program]): Fix typo.
18241 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
18243 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
18244 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
18245 (control_flow_insn_p): Likewise.
18246 * cfgrtl.c (duplicate_insn_chain): Likewise.
18247 * final.c (get_attr_length_1): Likewise.
18248 (shorten_branches): Likewise.
18249 (final_scan_insn): Likewise.
18250 * function.c (instantiate_virtual_regs): Likewise.
18251 * gcse.c (insert_insn_end_basic_block): Likewise.
18252 * ira-costs.c (scan_one_insn): Likewise.
18253 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
18254 * lra.c (check_rtl): Likewise.
18255 * reload1.c (elimination_costs_in_insn): Likewise.
18256 * reorg.c (follow_jumps): Likewise.
18258 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
18259 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
18260 (thumb_far_jump_used_p): Likewise.
18261 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
18262 (workaround_speculation): Likewise.
18263 (add_sched_insns_for_speculation): Likewise.
18264 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
18265 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
18266 (frv_for_each_packet): Likewise.
18267 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
18268 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
18269 (final_emit_insn_group_barriers): Likewise.
18270 * config/m32r/m32r.c (m32r_is_insn): Likewise.
18271 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
18272 (mips16_insn_length): Likewise.
18273 * config/pa/pa.c (pa_reorg): Likewise.
18274 (pa_combine_instructions): Likewise.
18275 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
18276 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
18277 (sh_reorg): Likewise.
18278 (split_branches): Likewise.
18279 * config/spu/spu.c (get_branch_target): Likewise.
18281 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
18284 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
18286 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
18287 Fix declaration name.
18289 2013-03-28 Lawrence Crowl <crowl@google.com>
18291 * graphds.h (struct graph.indicies): Remove unused.
18292 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
18293 (SCOP_ORIGINAL_PDDRS): Remove unused.
18294 * sese.h (extern insert_loop_close_phis): Removed unused.
18295 (extern insert_guard_phis): Removed unused.
18296 (extern ivtype_map_elt_info): Removed unused.
18297 (new_ivtype_map_elt): Removed unused.
18298 * sese.c (ivtype_map_elt_info): Removed unused.
18300 2013-03-28 Lawrence Crowl <crowl@google.com>
18302 * Makefile.in: Add several missing include dependences.
18304 (test-dump.o): New. This object is not added to any executable,
18305 but is present for ad-hoc testing.
18307 (debug (const bitmap_head_def &)): New.
18308 (debug (const bitmap_head_def *)): New.
18310 (extern debug (const bitmap_head_def &)): New.
18311 (extern debug (const bitmap_head_def *)): New.
18313 (debug (edge_def &)): New.
18314 (debug (edge_def *)): New.
18316 (debug (basic_block_def &)): New.
18317 (debug (basic_block_def *)): New.
18319 (dump_node (const_tree, int, FILE *)): Correct source file.
18321 (debug (die_struct &)): New.
18322 (debug (die_struct *)): New.
18324 (extern debug (die_struct &)): New.
18325 (extern debug (die_struct *)): New.
18326 * gimple-pretty-print.c
18327 (debug (gimple_statement_d &)): New.
18328 (debug (gimple_statement_d *)): New.
18329 * gimple-pretty-print.h
18330 (extern debug (gimple_statement_d &)): New.
18331 (extern debug (gimple_statement_d *)): New.
18333 (debug (ira_allocno_copy &)): New.
18334 (debug (ira_allocno_copy *)): New.
18335 (debug (ira_allocno &)): New.
18336 (debug (ira_allocno *)): New.
18338 (extern debug (ira_allocno_copy &)): New.
18339 (extern debug (ira_allocno_copy *)): New.
18340 (extern debug (ira_allocno &)): New.
18341 (extern debug (ira_allocno *)): New.
18343 (debug (live_range &)): New.
18344 (debug (live_range *)): New.
18346 (debug (lra_live_range &)): New.
18347 (debug (lra_live_range *)): New.
18349 (debug (lra_live_range &)): New.
18350 (debug (lra_live_range *)): New.
18352 (debug (omega_pb_d &)): New.
18353 (debug (omega_pb_d *)): New.
18355 (extern debug (omega_pb_d &)): New.
18356 (extern debug (omega_pb_d *)): New.
18358 (debug (const rtx_def &)): New.
18359 (debug (const rtx_def *)): New.
18361 (debug_tree (tree): Move within file.
18362 (debug_raw (const tree_node &)): New.
18363 (debug_raw (const tree_node *)): New.
18364 (dump_tree_via_hooks (const tree_node *, int)): New.
18365 (debug (const tree_node &)): New.
18366 (debug (const tree_node *)): New.
18367 (debug_verbose (const tree_node &)): New.
18368 (debug_verbose (const tree_node *)): New.
18369 (debug_head (const tree_node &)): New.
18370 (debug_head (const tree_node *)): New.
18371 (debug_body (const tree_node &)): New.
18372 (debug_body (const tree_node *)): New.
18373 (debug_vec_tree (tree): Move and reimplement in terms of dump.
18374 (debug (vec<tree, va_gc> &)): New.
18375 (debug (vec<tree, va_gc> *)): New.
18377 (extern debug (const rtx_def &)): New.
18378 (extern debug (const rtx_def *)): New.
18380 (debug_raw (simple_bitmap_def &)): New.
18381 (debug_raw (simple_bitmap_def *)): New.
18382 (debug (simple_bitmap_def &)): New.
18383 (debug (simple_bitmap_def *)): New.
18385 (extern debug (simple_bitmap_def &)): New.
18386 (extern debug (simple_bitmap_def *)): New.
18387 (extern debug_raw (simple_bitmap_def &)): New.
18388 (extern debug_raw (simple_bitmap_def *)): New.
18390 (debug (vinsn_def &)): New.
18391 (debug (vinsn_def *)): New.
18392 (debug_verbose (vinsn_def &)): New.
18393 (debug_verbose (vinsn_def *)): New.
18394 (debug (expr_def &)): New.
18395 (debug (expr_def *)): New.
18396 (debug_verbose (expr_def &)): New.
18397 (debug_verbose (expr_def *)): New.
18398 (debug (vec<rtx> &)): New.
18399 (debug (vec<rtx> *)): New.
18401 (extern debug (vinsn_def &)): New.
18402 (extern debug (vinsn_def *)): New.
18403 (extern debug_verbose (vinsn_def &)): New.
18404 (extern debug_verbose (vinsn_def *)): New.
18405 (extern debug (expr_def &)): New.
18406 (extern debug (expr_def *)): New.
18407 (extern debug_verbose (expr_def &)): New.
18408 (extern debug_verbose (expr_def *)): New.
18409 (extern debug (vec<rtx> &)): New.
18410 (extern debug (vec<rtx> *)): New.
18412 (_list_iter_cond_expr): Make inline instead of static.
18414 (debug (sreal &)): New.
18415 (debug (sreal *)): New.
18417 (extern debug (sreal &)): New.
18418 (extern debug (sreal *)): New.
18420 (extern debug_raw (const tree_node &)): New.
18421 (extern debug_raw (const tree_node *)): New.
18422 (extern debug (const tree_node &)): New.
18423 (extern debug (const tree_node *)): New.
18424 (extern debug_verbose (const tree_node &)): New.
18425 (extern debug_verbose (const tree_node *)): New.
18426 (extern debug_head (const tree_node &)): New.
18427 (extern debug_head (const tree_node *)): New.
18428 (extern debug_body (const tree_node &)): New.
18429 (extern debug_body (const tree_node *)): New.
18430 (extern debug (vec<tree, va_gc> &)): New.
18431 (extern debug (vec<tree, va_gc> *)): New.
18433 (debug (struct loop &)): New.
18434 (debug (struct loop *)): New.
18435 (debug_verbose (struct loop &)): New.
18436 (debug_verbose (struct loop *)): New.
18437 * tree-dump.c: Add header dependence.
18439 (extern debug (struct loop &)): New.
18440 (extern debug (struct loop *)): New.
18441 (extern debug_verbose (struct loop &)): New.
18442 (extern debug_verbose (struct loop *)): New.
18444 (debug (data_reference &)): New.
18445 (debug (data_reference *)): New.
18446 (debug (vec<data_reference_p> &)): New.
18447 (debug (vec<data_reference_p> *)): New.
18448 (debug (vec<ddr_p> &)): New.
18449 (debug (vec<ddr_p> *)): New.
18451 (extern debug (data_reference &)): New.
18452 (extern debug (data_reference *)): New.
18453 (extern debug (vec<data_reference_p> &)): New.
18454 (extern debug (vec<data_reference_p> *)): New.
18455 (extern debug (vec<ddr_p> &)): New.
18456 (extern debug (vec<ddr_p> *)): New.
18458 (debug (pt_solution &)): New.
18459 (debug (pt_solution *)): New.
18461 (extern debug (pt_solution &)): New.
18462 (extern debug (pt_solution *)): New.
18464 (debug (_var_map &)): New.
18465 (debug (_var_map *)): New.
18466 (debug (tree_live_info_d &)): New.
18467 (debug (tree_live_info_d *)): New.
18469 (extern debug (_var_map &)): New.
18470 (extern debug (_var_map *)): New.
18471 (extern debug (tree_live_info_d &)): New.
18472 (extern debug (tree_live_info_d *)): New.
18474 2013-03-28 Jan Hubicka <jh@suse.cz>
18476 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
18478 2013-03-28 Ian Bolton <ian.bolton@arm.com>
18480 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
18481 record only when desired or required.
18483 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
18485 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
18486 *vec_extractv2di_1_rex64. Use x64 isa attribute.
18488 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18490 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
18491 (*andsi3_compare0_uxtw): New pattern.
18492 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
18493 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
18495 2013-03-28 Jan Hubicka <jh@suse.cz>
18497 * data-streamer-in.c (streamer_read_gcov_count): New function.
18498 * gimple-streamer-out.c: Include value-prof.h.
18499 (output_gimple_stmt): Output histogram.
18500 (output_bb): Use streamer_write_gcov_count.
18501 * value-prof.c: Include data-streamer.h
18502 (dump_histogram_value): Add HIST_TYPE_MAX.
18503 (stream_out_histogram_value): New function.
18504 (stream_in_histogram_value): New function.
18505 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
18506 (stream_out_histogram_value, stream_in_histogram_value): Declare.
18507 * data-streamer-out.c (streamer_write_gcov_count): New function.
18508 (streamer_write_gcov_count_stream): New function.
18509 * lto-cgraph.c (lto_output_edge): Update counter streaming.
18510 (lto_output_node): Likewise.
18511 (input_node, input_edge): Likewise.
18512 * lto-streamer-out.c (output_cfg): Update streaming.
18513 * lto-streamer-in.c (input_cfg): Likewise.
18514 * data-streamer.h (streamer_write_gcov_count,
18515 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
18516 * gimple-streamer-in.c: Include value-prof.h
18517 (input_gimple_stmt): Input histograms.
18518 (input_bb): Update profile streaming.
18520 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
18522 * genmodes.c (emit_max_int): New function.
18523 (emit_insn_modes_h): Added call to emit_max_function.
18524 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
18526 * machmode.def: Fixed comment.
18528 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
18530 * combine.c (try_combine): Removed useless assert.
18531 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
18533 2013-03-28 Marek Polacek <polacek@redhat.com>
18534 Richard Biener <rguenther@suse.de>
18536 PR tree-optimization/56695
18537 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
18538 build signed result of a vector comparison.
18539 * tree-cfg.c (verify_gimple_comparison): Check that a result
18540 of a vector comparison has signed type.
18542 2013-03-28 Richard Biener <rguenther@suse.de>
18544 PR tree-optimization/37021
18545 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
18546 do not restrict gaps between groups.
18547 * tree-vect-stmts.c (vectorizable_load): Properly account for
18548 a gap between groups.
18550 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
18552 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
18553 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
18556 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
18558 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
18559 * doc/extend.texi (Named Address Spaces): Ditto.
18560 (Variable Attributes): Ditto.
18562 2013-03-27 Kai Tietz <ktietz@redhat.com>
18564 * config.build: Add support for cygwin x64 target.
18565 * config.gcc: Likewise.
18566 * config.host: Likewise.
18567 * configure.ac: Likewise
18568 * configure: Regenerated.
18570 2013-03-27 Kai Tietz <ktietz@redhat.com>
18572 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
18573 * config/i386/t-cygwin-w64: New file.
18574 * config/i386/cygwin-w64.h: New file.
18575 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
18576 and add support for x64-cygwin target.
18577 (CPP_SPEC): Likewise.
18578 (CXX_WRAP_SPEC_LIST): Undefine before define.
18579 (LIBGCJ_SONAME): Use 15 as version.
18581 2013-03-27 Richard Biener <rguenther@suse.de>
18583 PR tree-optimization/56716
18584 * tree-ssa-structalias.c (perform_var_substitution): Adjust
18585 dumping for ref nodes.
18587 2013-03-27 Martin Jambor <mjambor@suse.cz>
18589 PR tree-optimization/55334
18590 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
18591 restricted pointers to arrays.
18593 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
18595 * Makefile.in (.SUFFIXES): Add .cc.
18596 (.c.o): Apply same recipe for implicit rule .cc.o.
18598 2013-03-27 Richard Biener <rguenther@suse.de>
18600 PR tree-optimization/37021
18601 * tree-vect-data-refs.c (vect_check_strided_load): Allow
18602 REALPART/IMAGPART_EXPRs around the supported refs.
18603 * tree-ssa-structalias.c (find_func_aliases): Assume that
18604 floating-point values are not used to transfer pointers.
18606 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
18608 * target.def (TARGET_HAS_IFUNC_P): New target hook.
18609 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
18610 * doc/tm.texi: Regenerate.
18611 * targhooks.h (default_has_ifunc_p): New.
18612 * targhooks.c (default_has_ifunc_p): Ditto.
18613 * config/linux-protos.h: New file.
18614 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
18615 hook for linux which disables support of indirect functions in android.
18616 * config/linux-android.c: New file.
18617 * config/t-linux-android.c: Ditto.
18618 * config.gcc: Added new object file linux-android.o.
18619 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
18620 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
18621 * varasm.c (do_assemble_alias): Likewise.
18622 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
18623 doesn't support indirect functions.
18624 * configure: Regenerate.
18626 2013-03-27 Bin Cheng <bin.cheng@arm.com>
18629 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
18630 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
18633 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18635 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
18637 2013-03-27 Terry Guo <terry.guo@arm.com>
18639 * config/arm/arm-cores.def: Added core cortex-r7.
18640 * config/arm/arm-tune.md: Regenerated.
18641 * config/arm/arm-tables.opt: Regenerated.
18642 * doc/invoke.texi: Added entry for core cortex-r7.
18644 2013-03-27 Walter Lee <walt@tilera.com>
18646 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
18647 double-decrement of next_scratch_regno.
18649 2013-03-27 Walter Lee <walt@tilera.com>
18651 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
18653 (insn_v1mulus): Ditto.
18654 (insn_v2muls): Ditto.
18656 2013-03-27 Walter Lee <walt@tilera.com>
18658 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
18659 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
18661 2013-03-27 Walter Lee <walt@tilera.com>
18663 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
18664 (*sibcall_value): Ditto.
18666 2013-03-27 Walter Lee <walt@tilera.com>
18668 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
18669 (insn_mnz_v8qi): ... this ...
18670 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
18672 (insn_v<n>mnz): Replaced by ...
18673 (insn_v1mnz): ... this ...
18674 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
18676 (insn_mz_<mode>): Replaced by ...
18677 (insn_mz_v8qi): ... this ...
18678 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
18680 (insn_v<n>mz): Replaced by ...
18681 (insn_v1mz): ... this ...
18682 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
18685 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
18687 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
18689 2013-03-26 Roland McGrath <mcgrathr@google.com>
18691 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
18692 than fprintf with a non-constant, non-format string.
18694 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
18696 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
18697 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
18698 operand 0 predicate.
18699 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
18700 attribute. Use general_x64nomem_operand as operand 1 predicate.
18701 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
18702 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
18703 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
18704 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
18705 (mov<mode>_insv_1): Remove expander. Merge insn with
18706 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
18707 Use general_x64nomem_operand as operand 1 predicate.
18708 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
18709 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
18710 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
18711 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
18712 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
18713 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
18714 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
18715 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
18716 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
18717 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
18718 (general_x64nomem_operand): Ditto.
18720 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
18722 * config/rtems.opt: Add -pthread option.
18724 2013-03-26 Richard Biener <rguenther@suse.de>
18726 * alias.c (find_base_term): Avoid redundant and not used recursion.
18727 (base_alias_check): Get the initial base term from the caller.
18728 (true_dependence_1): Compute and pass base terms to base_alias_check.
18729 (write_dependence_p): Likewise.
18730 (may_alias_p): Likewise.
18732 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
18734 * config/aarch64/aarch64.c (aarch64_classify_address): Support
18735 PC-relative load in SI modes and above only.
18737 2013-03-26 Xinyu Qi <xyqi@marvell.com>
18739 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
18740 * config/arm/iwmmxt.md (WCGR0): Update.
18741 (WCGR1, WCGR2, WCGR3): Likewise.
18743 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
18745 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
18746 Use x64 and nox64 isa attributes.
18748 2013-03-26 Richard Biener <rguenther@suse.de>
18750 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
18751 alignment computations and rely on get_object_alignment_1
18752 for the !TYPE_P case.
18753 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
18755 2013-03-26 Walter Lee <walt@tilera.com>
18757 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
18758 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
18760 2013-03-25 Jeff Law <law@redhat.com>
18762 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
18763 check for INTEGRAL_TYPE_P that was missing due to checking in
18764 wrong version of prior patch.
18766 2013-03-25 Walter Lee <walt@tilera.com>
18768 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
18769 TILEGX_INSN_SHUFFLEBYTES1.
18770 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
18772 (tilegx_builtins): Ditto.
18773 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
18775 2013-03-25 Walter Lee <walt@tilera.com>
18777 * config/tilegx/tilegx.md (floatsisf2): New pattern.
18778 (floatunssisf2): New pattern.
18779 (floatsidf2): New pattern.
18780 (floatunssidf2): New pattern.
18782 2013-03-25 Walter Lee <walt@tilera.com>
18784 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
18785 tests for constraint J, K, N, P.
18787 2013-03-25 Walter Lee <walt@tilera.com>
18789 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
18790 Use indirect/pcrel encoding.
18791 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
18794 2013-03-25 Steve Ellcey <sellcey@mips.com>
18796 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
18797 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
18798 * config/mips/mips.c (mips_option_override): Set IMADD default.
18799 * config/mips/mips.h (PTF_AVOID_IMADD): New.
18800 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
18801 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
18802 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
18803 * doc/invoke.texi (-mimadd/-mno-imadd): New.
18805 2013-03-25 Jeff Law <law@redhat.com>
18807 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
18808 slightly to avoid creating and folding useless trees. Simplify
18809 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
18811 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
18813 * config/i386/i386.md (*zero_extendsidi2): Merge with
18814 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
18815 * config/i386/predicates.md (x86_64_zext_operand): Rename from
18816 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
18817 targets. Clarify comment.
18819 2013-03-25 Martin Jambor <mjambor@suse.cz>
18821 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
18822 pass-through jump functions differently.
18823 (ipa_read_jump_function): Likewise. Also use setter functions to set
18826 2013-03-25 Martin Jambor <mjambor@suse.cz>
18828 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
18829 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
18831 (ipa_get_indirect_edge_target): New function.
18832 (devirtualization_time_bonus): New parameter known_aggs, pass it to
18833 ipa_get_indirect_edge_target. Update all callers.
18834 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
18835 ipa_get_indirect_edge_target_1 instead of calling
18836 ipa_get_indirect_edge_target.
18837 (create_specialized_node): Pass aggvlas to
18838 ipcp_discover_new_direct_edges.
18840 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18842 * config/arm/arm.md (f_sels, f_seld): New types.
18843 (*cmov<mode>): New pattern.
18844 * config/arm/predicates.md (arm_vsel_comparison_operator): New
18847 2013-03-25 Kai Tietz <ktietz@redhat.com>
18849 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
18850 POSIX-printf for mingw-hosted builds.
18852 2013-03-25 Richard Biener <rguenther@suse.de>
18854 PR middle-end/56694
18855 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
18856 must-not-throw stmt location.
18858 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18860 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
18861 Emit load-acquire versions when acq is true.
18862 (arm_emit_store_exclusive): Add rel parameter.
18863 Emit store-release versions when rel is true.
18864 (arm_split_compare_and_swap): Use acquire-release instructions
18866 of barriers when appropriate.
18867 (arm_split_atomic_op): Likewise.
18868 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
18869 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
18870 (VUNSPEC_SLX): Likewise.
18871 (VUNSPEC_LDA): Likewise.
18872 (VUNSPEC_STL): Likewise.
18873 * config/arm/sync.md (atomic_load<mode>): New pattern.
18874 (atomic_store<mode>): Likewise.
18875 (arm_load_acquire_exclusive<mode>): Likewise.
18876 (arm_load_acquire_exclusivesi): Likewise.
18877 (arm_load_acquire_exclusivedi): Likewise.
18878 (arm_store_release_exclusive<mode>): Likewise.
18880 2013-03-25 Catherine Moore <clm@codesourcery.com>
18882 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
18883 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
18884 * config/mip/predicates.md (lwsp_swsp_operand,
18885 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
18886 sb16_operand, db4_operand, db7_operand, ib3_operand,
18887 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
18888 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
18889 andi16_operand): New predicates.
18890 * config/mips/mips.md (compression): New attribute.
18891 (enabled): New attribute.
18892 (length): Consider compression in computing length.
18893 (shift_compression): New code attribute.
18894 (*add<mode>3): New operands. Record compression.
18895 (sub<mode>3): Likewise.
18896 (one_cmpl<mode>2): Likewise.
18897 (*and<mode>3): Likewise.
18898 (*ior<mode>3): Likewise.
18899 (unnamed pattern for xor): Likewise.
18900 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
18901 (*<optab><mode>3): Likewise.
18902 (*mov<mode>_internal: Likewise.
18903 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
18904 (mips_unsigned_immediate_p): New.
18905 (umips_lwsp_swsp_address_p): New.
18906 (m16_based_address_p): New.
18907 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
18908 (mips_unsigned_immediate_p): New prototype.
18909 (lwsp_swsp_address_p): New prototype.
18910 (m16_based_address_p): New prototype.
18911 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
18912 (mips_signed_immediate_p): New function.
18913 (m16_based_address_p): New function.
18914 (lwsp_swsp_address_p): New function.
18915 (mips_print_operand_punctuation): Recognize short delay slot insns
18916 for microMIPS.add<mode>3"
18918 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18921 * config/arm/iterators.md (v_cmp_result): New mode attribute.
18922 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
18924 2013-03-25 Richard Biener <rguenther@suse.de>
18926 PR tree-optimization/56689
18927 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
18930 2013-03-25 Richard Biener <rguenther@suse.de>
18932 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
18934 (memory_references): Likewise.
18935 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
18936 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
18937 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
18938 (gather_mem_refs_in_loops): Fold into ...
18939 (analyze_memory_references): ... this. Move initialization
18940 to tree_ssa_lim_initialize.
18941 (fill_always_executed_in): Rename to ...
18942 (fill_always_executed_in_1): ... this.
18943 (fill_always_executed_in): Move contains_call computation to
18944 this new function from ...
18945 (tree_ssa_lim_initialize): ... here.
18946 (tree_ssa_lim): Call fill_always_executed_in.
18948 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18950 * postreload.c (reload_combine): Fix code detecting returns.
18952 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18954 * function.c (emit_use_return_register_into_block): On cc0 targets,
18955 do not emit the sequence between cc0 setter and user.
18957 2013-03-25 Kai Tietz <ktietz@redhat.com>
18959 * config/i386/predicates.md (local_symbolic_operand): Interpret
18960 dll-imported symbols as none-local.
18962 2013-03-25 Richard Biener <rguenther@suse.de>
18964 * tree-ssa-loop-im.c (struct depend): Remove.
18965 (struct lim_aux_data): Make depends a vec of gimples.
18966 (free_lim_aux_data): Adjust.
18967 (add_dependency): Likewise.
18968 (set_level): Likewise.
18970 2013-03-25 Richard Biener <rguenther@suse.de>
18972 PR middle-end/56434
18973 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
18974 the pointer returned by calls with ECF_MALLOC set.
18976 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18978 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
18980 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18982 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
18983 using MMXMODE mode iterator.
18984 (*move<mode>_internal): Merge with *movv2sf_internal and
18985 *movv2sf_internal_rex64 using MMXMODE mode iterator.
18987 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18989 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
18990 (record_last_mem_set_info): Likewise.
18992 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
18993 of XNEWVEC followed by memset.
18994 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
18996 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18998 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
18999 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
19000 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
19001 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
19002 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
19003 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
19004 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
19005 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
19006 BARRIER_P instead of GET_CODE.
19008 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
19010 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
19011 inaccuracy in the probing code.
19013 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
19014 (ctrapdi4): Likewise.
19016 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
19018 * calls.c (expand_call): Add missing guard to code handling return
19019 of non-BLKmode structures in MSB.
19020 * function.c (expand_function_end): Likewise.
19022 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
19024 * combine.c (try_combine): Adjust comment. Do not add the set of
19025 insn #0 if the destination indirectly is set or dies in insn #2.
19026 Tidy up code to distribute a new note.
19028 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
19030 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
19031 also for alternatives 16 and 17.
19033 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
19035 * config/i386/sse.md (*mov<mode>_internal): Merge with
19036 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
19037 Emit insn template depending on type attribute. Use
19038 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
19039 movd instead of movq mnemonic for interunit moves. Rewrite mode
19040 attribute calculation. Remove unit attribute calculation.
19041 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
19042 Set prefix_data16 attribute for DImode ssemov types.
19043 Use Ym instead of y for SSE-MMX conversion alternatives.
19044 Reorder operand constraints.
19046 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
19048 * df.h (df_insn_delete): Adjust prototype.
19049 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
19050 and let it decide whether mark the basic block dirty.
19051 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
19052 * df-scan.c (df_insn_info_delete): New helper function, split
19053 off from df_insn_delete.
19054 (df_scan_free_bb_info): Use it.
19055 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
19057 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
19058 that the insn is actually an insn and it has a non-NULL basic block.
19059 Do not mark basic block dirty if only deleting a DEBUG_INSN.
19061 2013-03-22 Richard Biener <rguenther@suse.de>
19063 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
19065 (mem_ref_alloc): Do not allocate them.
19066 (refs_independent_p): Do not query or maintain a cache.
19068 2013-03-22 Richard Biener <rguenther@suse.de>
19070 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
19071 (gather_mem_refs_in_loops): Do not compute it.
19072 (analyze_memory_references): Do not allocate it.
19073 (tree_ssa_lim_finalize): Do not free it.
19074 (for_all_locs_in_loop): Do not query all_refs_in_loop.
19076 2013-03-22 Richard Biener <rguenther@suse.de>
19078 * is-a.h (as_a): Use gcc_checking_assert.
19080 2013-03-22 Ian Bolton <ian.bolton@arm.com>
19082 * config/aarch64/aarch64.c (aarch64_print_operand): New
19083 format specifier for printing a constant in hex.
19084 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
19085 format specifier for printing second operand.
19087 2013-03-22 Richard Biener <rguenther@suse.de>
19089 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
19091 (gather_mem_refs_in_loops): Perform store accumulation here.
19092 (create_vop_ref_mapping_loop): Remove.
19093 (create_vop_ref_mapping): Likewise.
19094 (analyze_memory_references): Initialize refs_stored_in_loop.
19095 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
19096 (record_indep_loop): Remove.
19097 (record_dep_loop): New function.
19098 (ref_indep_loop_p_1): Adjust to only walk over references
19099 in the loop, not its subloops.
19100 (ref_indep_loop_p): Rename to ...
19101 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
19102 maintaining a more fine-grained cache.
19103 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
19104 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
19106 2013-03-22 Richard Biener <rguenther@suse.de>
19108 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
19109 (struct mem_ref): Make accesses_in_loop a vec of a vec of
19110 aggregate mem_ref_loc.
19111 (free_mem_ref_locs): Inline into ...
19112 (memref_free): ... this and adjust.
19113 (mem_ref_alloc): Adjust.
19114 (mem_ref_locs_alloc): Remove.
19115 (record_mem_ref_loc): Adjust.
19116 (get_all_locs_in_loop): Rewrite into ...
19117 (for_all_locs_in_loop): ... this iterator.
19118 (rewrite_mem_ref_loc): New functor.
19119 (rewrite_mem_refs): Use for_all_locs_in_loop.
19120 (sm_set_flag_if_changed): New functor.
19121 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
19122 (ref_always_accessed): New functor.
19123 (ref_always_accessed_p): Use for_all_locs_in_loop.
19125 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
19127 * tree-pass.h (PROP_gimple_lvec): New.
19128 * passes.c (dump_properties): Handle PROP_gimple_lvec.
19129 (init_optimization_passes): Move pass_lower_vector.
19130 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
19132 (pass_lower_vector): Provide PROP_gimple_lvec.
19133 (pass_lower_vector_ssa): Likewise.
19134 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
19136 2013-03-21 Mark Wielaard <mjw@redhat.com>
19138 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
19140 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
19142 * config/i386/i386.md (*movdi_internal): Disparage slightly
19143 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
19144 conversion alternatives.
19146 2013-03-21 Jakub Jelinek <jakub@redhat.com>
19148 PR middle-end/48087
19149 * diagnostic.def (DK_WERROR): New kind.
19150 * diagnostic.h (werrorcount): Define.
19151 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
19152 promoted to DK_ERROR, increment DK_WERROR counter instead of
19154 * toplev.c (toplev_main): Call print_ignored_options even if
19155 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
19156 even if just werrorcount is non-zero.
19159 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
19161 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
19162 (string_cst_pool_decl): New function.
19163 (optimize_one_addr_into_implicit_ptr): New function.
19164 (resolve_addr_in_expr): Optimize DWARF location expression
19165 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
19166 which doesn't live in memory, but has DW_AT_location or
19167 DW_AT_const_value, or refers to a string literal, into
19168 DW_OP_GNU_implicit_pointer.
19169 (optimize_location_into_implicit_ptr): New function.
19170 (resolve_addr): If removing DW_AT_location of a variable because
19171 it was DW_OP_addr of address of the variable, but the variable doesn't
19172 live in memory, try to emit const value attribute for the initializer.
19174 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
19176 * tree.h (VECTOR_TYPE_P): New macro.
19177 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
19178 TYPE_MODE): Use it.
19179 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
19180 VEC_COND_EXPR cannot be lvalues.
19181 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
19183 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
19185 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
19186 Restrict the transformation to equal modes.
19188 2013-03-21 Richard Biener <rguenther@suse.de>
19190 PR tree-optimization/39326
19191 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
19192 (MEM_ANALYZABLE): Adjust.
19193 (record_mem_ref_loc): Move bitmap ops ...
19194 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
19195 unanalyzable refs, do not record locations for it.
19196 (analyze_memory_references): Allocate ref zero as shared
19198 (refs_independent_p): Do not test for unanalyzed mems here.
19199 (ref_indep_loop_p_1): Special-case disambiguation against
19200 the unanalyzed ref.
19201 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
19203 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
19205 * config/arm/arm-protos.h (tune_params): Add
19206 prefer_neon_for_64bits field.
19207 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
19208 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
19209 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
19210 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
19211 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
19212 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
19213 (arm_option_override): Handle -mneon-for-64bits new option.
19214 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
19215 (prefer_neon_for_64bits): Declare new variable.
19216 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
19217 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
19218 (arch_enabled): Handle new arch types. Remove support for onlya8
19220 (one_cmpldi2): Use new arch names.
19221 (zero_extend<mode>di2, extend<mode>di2): Ditto.
19222 * config/arm/arm.opt (mneon-for-64bits): Add option.
19223 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
19224 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
19225 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
19227 * doc/invoke.texi (-mneon-for-64bits): Document.
19229 2013-03-21 Richard Biener <rguenther@suse.de>
19231 PR tree-optimization/39326
19232 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
19233 (sort_bbs_in_loop_postorder_cmp): New function.
19234 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
19236 2013-03-21 Richard Biener <rguenther@suse.de>
19238 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
19239 (vect_insert_into_interleaving_chain): Likewise.
19240 (vect_drs_dependent_in_basic_block): Inline ...
19241 (vect_slp_analyze_data_ref_dependence): ... here. New function,
19243 (vect_analyze_data_ref_dependence): ... here. Simplify.
19244 (vect_check_interleaving): Simplify.
19245 (vect_analyze_data_ref_dependences): Likewise. Split out ...
19246 (vect_slp_analyze_data_ref_dependences): ... this new function.
19247 (dr_group_sort_cmp): New function.
19248 (vect_analyze_data_ref_accesses): Compute data-reference groups
19249 here instead of in vect_analyze_data_ref_dependence. Use
19250 a more efficient algorithm.
19251 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
19252 vect_slp_analyze_data_ref_dependences. Call
19253 vect_analyze_data_ref_accesses earlier.
19254 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
19255 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
19256 (vect_slp_analyze_data_ref_dependences): New prototype.
19258 2013-03-21 Richard Biener <rguenther@suse.de>
19260 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
19261 ref is stored in the loop.
19262 (find_refs_for_sm): Walk only over all stores.
19263 (store_motion_loop): Allocate from lim_bitmap_obstack.
19264 (store_motion): Likewise.
19266 2013-03-21 Richard Biener <rguenther@suse.de>
19268 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
19269 Update virtual SSA form.
19271 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19273 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
19274 * configure: Regenerate.
19275 * config.in: Regenerate.
19276 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
19277 if !HAVE_LD_EH_FRAME_CIEV3.
19279 2013-03-21 Richard Biener <rguenther@suse.de>
19281 * tree-cfg.c (verify_expr_no_block): New function.
19282 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
19283 nor DECL_VALUE_EXPR have locations with associated blocks.
19284 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
19285 (clear_unused_block_pointer): Remove code dealing with
19286 blocks in DECL_DEBUG_EXPR locations.
19288 2013-03-21 Richard Biener <rguenther@suse.de>
19290 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
19291 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
19292 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
19293 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
19294 instead of DECL_DEBUG_EXPR_IS_FROM.
19295 * gimplify.c (gimplify_modify_expr): Likewise.
19296 * tree-cfg.c (verify_expr_location_1): Likewise.
19297 * tree-complex.c (create_one_component_var): Likewise.
19298 * tree-sra.c (create_access_replacement): Likewise.
19299 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
19300 (clear_unused_block_pointer): Likewise.
19301 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
19302 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
19303 * var-tracking.c (var_debug_decl): Likewise.
19304 (track_expr_p): Likewise.
19305 * tree-inline.c (add_local_variables): Likewise. Set
19306 DECL_HAS_DEBUG_EXPR_P after copying it.
19307 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
19308 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
19310 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
19313 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
19314 * configure: Regenerate.
19315 * config.in: Regenerate.
19316 * config/i386/i386.md (*movdf_internal): Use
19317 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
19318 movd instead of movq mnemonic for interunit moves.
19319 (*movdi_internal): Ditto.
19321 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19323 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
19324 (abd<mode>_3): New pattern.
19325 (aba<mode>_3): New pattern.
19326 (fabd<mode>_3): New pattern.
19328 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19330 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
19331 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
19332 occurrence of REGISTER_PREFIX as its empty string.
19334 2013-03-20 Jeff Law <law@redhat.com>
19336 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
19337 addititional equivalences for equality comparisons between an SSA_NAME
19338 and a constant where the SSA_NAME was set from a widening conversion.
19340 2013-03-20 Walter Lee <walt@tilera.com>
19342 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
19344 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19346 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
19347 depending on type attribute.
19348 (*movti_internal): Ditto.
19349 (*movtf_internal): Ditto.
19350 (*movxf_internal): Ditto.
19351 (*movdf_internal): Ditto.
19352 (*movsf_internal): Ditto.
19354 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19356 * config/i386/i386.md (*movti_internal): Set prefix attribute to
19357 maybe_vex for sselog1 and ssemov types.
19358 (*movdi_internal): Reorder operand constraints.
19359 (*movsi_internal): Ditto. Set prefix attribute to
19360 maybe_vex for sselog1 and ssemov types.
19361 (*movtf_internal): Set prefix attribute to maybe_vex
19362 for sselog1 and ssemov types.
19363 (*movdf_internal): Ditto. Set prefix_data16 attribute for
19364 DImode ssemov types. Reorder operand constraints.
19365 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
19366 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
19367 attribute for SImode ssemov types. Reorder operand constraints.
19369 2013-03-20 Martin Jambor <mjambor@suse.cz>
19371 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
19372 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
19374 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
19376 * config/rs6000/predicates.md (indexed_address, update_address_mem
19377 update_indexed_address_mem): New predicates.
19378 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
19379 attribute for load/store instructions.
19380 * config/rs6000/dfp.md (movsd_store): Likewise.
19381 (movsd_load): Likewise.
19382 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
19383 (unnamed HI->DI extend define_insn): Likewise.
19384 (unnamed SI->DI extend define_insn): Likewise.
19385 (unnamed QI->SI extend define_insn): Likewise.
19386 (unnamed QI->HI extend define_insn): Likewise.
19387 (unnamed HI->SI extend define_insn): Likewise.
19388 (unnamed HI->SI extend define_insn): Likewise.
19389 (extendsfdf2_fpr): Likewise.
19390 (movsi_internal1): Likewise.
19391 (movsi_internal1_single): Likewise.
19392 (movhi_internal): Likewise.
19393 (movqi_internal): Likewise.
19394 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
19395 attribute for load/store instructions.
19396 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
19398 (mov<mode>_softfloat): Likewise.
19399 (mov<mode>_hardfloat32): Likewise.
19400 (mov<mode>_hardfloat64): Likewise.
19401 (mov<mode>_softfloat64): Likewise.
19402 (movdi_internal32): Likewise.
19403 (movdi_internal64): Likewise.
19404 (probe_stack_<mode>): Likewise.
19406 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
19408 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
19409 floating point, and decimal floating point to reload iterator.
19411 * config/rs6000/constraints.md (wl constraint): New constraints to
19412 return FLOAT_REGS if certain options are used to reduce the number
19413 of separate patterns that exist in the file.
19414 (wx constraint): Likewise.
19415 (wz constraint): Likewise.
19417 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
19418 -mdebug=reg, print wg, wl, wx, and wz constraints.
19419 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
19420 Initialize the reload functions for 64-bit binary/decimal floating
19422 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
19423 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
19424 create the buffer on the stack to overcome not having a 32-bit
19426 (rs6000_emit_move): Likewise.
19427 (rs6000_secondary_memory_needed_rtx): Likewise.
19428 (rs6000_alloc_sdmode_stack_slot): Likewise.
19429 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
19430 via xxlxor, just like DFmode 0.0.
19432 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
19433 define as 1 if we are running on a power7 or newer.
19434 (enum r6000_reg_class_enum): Add new constraints.
19436 * config/rs6000/dfp.md (movsd): Delete, combine with binary
19437 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
19438 with other moves by using conditional constraits (wg). Use LFIWZX
19439 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
19440 (movsd splitter): Likewise.
19441 (movsd_hardfloat): Likewise.
19442 (movsd_softfloat): Likewise.
19444 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
19445 binary and decimal floating point moves.
19446 (fmove_ok): New attributes to combine binary and decimal floating
19447 point moves, and to combine power6x (mfpgpr) moves along normal
19449 (real_value_to_target): Likewise.
19450 (f32_lr): Likewise.
19451 (f32_lm): Likewise.
19452 (f32_li): Likewise.
19453 (f32_sr): Likewise.
19454 (f32_sm): Likewise.
19455 (f32_si): Likewise.
19456 (movsf): Combine binary and decimal floating point moves. Combine
19457 power6x (mfpgpr) moves with other moves by using conditional
19458 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
19459 (mov<mode> for SFmode/SDmode); Likewise.
19460 (SFmode/SDmode splitters): Likewise.
19461 (movsf_hardfloat): Likewise.
19462 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
19463 (movsf_softfloat): Likewise.
19464 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
19466 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
19467 wx and wz constraints.
19469 * config/rs6000/constraints.md (wg constraint): New constraint to
19470 return FLOAT_REGS if -mmfpgpr (power6x) was used.
19472 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
19475 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
19476 -mdebug=reg, print wg, wl, wx, and wz constraints.
19477 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
19478 Initialize the reload functions for 64-bit binary/decimal floating
19480 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
19481 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
19482 create the buffer on the stack to overcome not having a 32-bit
19484 (rs6000_emit_move): Likewise.
19485 (rs6000_secondary_memory_needed_rtx): Likewise.
19486 (rs6000_alloc_sdmode_stack_slot): Likewise.
19487 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
19488 via xxlxor, just like DFmode 0.0.
19490 * config/rs6000/dfp.md (movdd): Delete, combine with binary
19491 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
19492 with other moves by using conditional constraits (wg). Use LFIWZX
19493 and STFIWX for loading SDmode on power7.
19494 (movdd splitters): Likewise.
19495 (movdd_hardfloat32): Likewise.
19496 (movdd_softfloat32): Likewise.
19497 (movdd_hardfloat64_mfpgpr): Likewise.
19498 (movdd_hardfloat64): Likewise.
19499 (movdd_softfloat64): Likewise.
19501 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
19502 64-bit binary and decimal floating point moves.
19503 (FMOVE64X): Likewise.
19504 (movdf): Combine 64-bit binary and decimal floating point moves.
19505 Combine power6x (mfpgpr) moves with other moves by using
19506 conditional constraits (wg).
19507 (mov<mode> for DFmode/DDmode): Likewise.
19508 (DFmode/DDmode splitters): Likewise.
19509 (movdf_hardfloat32): Likewise.
19510 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
19511 (movdf_softfloat32): Likewise.
19512 (movdf_hardfloat64_mfpgpr): Likewise.
19513 (movdf_hardfloat64): Likewise.
19514 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
19515 (movdf_softfloat64): Likewise.
19516 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
19517 (reload_<mode>_load): Move to later in the file so they aren't in
19518 the middle of the floating point move insns.
19519 (reload_<mode>_store): Likewise.
19521 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
19524 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
19525 constraint if -mdebug=reg.
19526 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
19527 Enable using dd reload support if needed.
19529 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
19530 binary and decimal floating point moves in rs6000.md.
19531 (movtd_internal): Likewise.
19533 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
19534 decimal floating point moves.
19536 (movtf_internal): Likewise.
19537 (mov<mode>_internal, TDmode/TFmode): Likewise.
19538 (movtf_softfloat): Likewise.
19539 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
19541 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
19542 movdi_internal64, using wg constraint for move direct operations.
19543 (movdi_internal64): Likewise.
19545 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
19546 MODES_TIEABLE_P for selected modes. Print the numerical value of
19547 the various virtual registers. Use GPR/FPR first/last values,
19548 instead of hard coding the register numbers. Print which modes
19549 have reload functions registered.
19550 (rs6000_option_override_internal): If -mdebug=reg, trace the options
19551 settings before/after setting cpu, target and subtarget settings.
19552 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
19553 and for secondary reload failures in rs6000_secondary_reload_inner.
19554 (rs6000_secondary_reload_fail): Likewise.
19555 (rs6000_secondary_reload_inner): Likewise.
19557 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
19558 macros for first/last GPR and FPR registers.
19559 (LAST_GPR_REGNO): Likewise.
19560 (FIRST_FPR_REGNO): Likewise.
19561 (LAST_FPR_REGNO): Likewise.
19563 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
19564 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
19565 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
19566 (vcond<mode><mode>): Likewise.
19567 (vcondu<mode><mode>): Likewise.
19568 (vector_gtu<mode>): Likewise.
19569 (vector_gte<mode>): Likewise.
19570 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
19571 to prevent the compiler from converting DImode operations to TImode.
19572 (ior<mode>3): Likewise.
19573 (and<mode>3): Likewise.
19574 (one_cmpl<mode>2): Likewise.
19575 (nor<mode>3): Likewise.
19576 (andc<mode>3): Likewise.
19578 * config/rs6000/constraints.md (wt constraint): New constraint
19579 that returns VSX_REGS if TImode is allowed in VSX registers.
19581 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
19582 constant under VSX.
19584 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
19585 similar to TImode, but it is restricted to being in the GPRs.
19587 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
19588 TImode to occupy a single VSX register.
19590 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
19591 -mvsx-timode for power7/power8.
19592 (power7 cpu): Likewise.
19593 (power8 cpu): Likewise.
19595 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
19596 sure that TFmode/TDmode take up two registers if they are ever
19597 allowed in the upper VSX registers.
19598 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
19600 (rs6000_init_hard_regno_mode_ok): Likewise.
19601 (rs6000_debug_reg_global): Add debugging for PTImode and wt
19602 constraint. Print if LRA is turned on.
19603 (rs6000_option_override_internal): Give an error if -mvsx-timode
19604 and VSX is not enabled.
19605 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
19606 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
19607 to reg+offset addressing. Use PTImode when checking offset
19608 addresses for validity.
19609 (reg_offset_addressing_ok_p): Likewise.
19610 (rs6000_legitimate_offset_address_p): Likewise.
19611 (rs6000_legitimize_address): Likewise.
19612 (rs6000_legitimize_reload_address): Likewise.
19613 (rs6000_legitimate_address_p): Likewise.
19614 (rs6000_eliminate_indexed_memrefs): Likewise.
19615 (rs6000_emit_move): Likewise.
19616 (rs6000_secondary_reload): Likewise.
19617 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
19618 reloads to fpr registers to continue to use reg+offset addressing,
19619 but 64-bit reloads to altivec registers need reg+reg addressing.
19620 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
19621 it. Treat LO_SUM like a PLUS operation.
19622 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
19623 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
19624 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
19625 registers to share a register with a smaller sized type, since VSX
19626 puts scalars in the upper 64-bits.
19627 (print_operand): Add support for PTImode.
19628 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
19629 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
19630 registers, but don't have arithmetic support.
19631 (rs6000_memory_move_cost): Add test for VSX.
19632 (rs6000_opt_masks): Add -mvsx-timode.
19634 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
19637 (VSr): Use wt constraint for TImode.
19638 (VSv): Drop TImode support.
19639 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
19640 (vsx_movti_64bit): Likewise.
19641 (vsx_movti_32bit): Likewise.
19642 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
19643 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
19644 one '?' on the appropriate output constraint. Do not allow TImode
19645 logical operations on 32-bit systems.
19646 (vsx_ior<mode>3): Likewise.
19647 (vsx_xor<mode>3): Likewise.
19648 (vsx_one_cmpl<mode>2): Likewise.
19649 (vsx_nor<mode>3): Likewise.
19650 (vsx_andc<mode>3): Likewise.
19651 (vsx_concat_<mode>): Likewise.
19652 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
19654 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
19655 OPTION_MASK_VSX_TIMODE.
19656 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
19657 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
19659 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
19660 (TI2 iterator): New iterator for TImode, PTImode.
19661 (wd mode attribute): Add values for vector types.
19662 (movti_string): Replace TI move operations with operations for TImode
19663 and PTImode. Add support for TImode being allowed in VSX registers.
19664 (mov<mode>_string, TImode/PTImode): Likewise.
19665 (movti_ppc64): Likewise.
19666 (mov<mode>_ppc64, TImode/PTImode): Likewise.
19667 (TI mode splitters): Likewise.
19669 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
19672 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
19674 PR tree-optimization/56355
19675 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
19676 Also handle integers with undefined overflow.
19678 2013-03-20 Catherine Moore <clm@codesourcery.com>
19679 Maciej W. Rozycki <macro@codesourcery.com>
19680 Tom de Vries <tom@codesourcery.com>
19681 Nathan Sidwell <nathan@codesourcery.com>
19682 Iain Sandoe <iain@codesourcery.com>
19683 Nathan Froyd <froydnj@codesourcery.com>
19684 Chao-ying Fu <fu@mips.com>
19686 * doc/extend.texi: (micromips, nomicromips, nocompression):
19687 Document new function attributes.
19688 * doc/invoke.texi (minterlink-compressed, mmicromips,
19689 m14k, m14ke, m14kec): Document new options.
19690 (minterlink-mips16): Update documentation.
19691 * doc/md.texi (ZC, ZD): Document new constraints.
19692 * configure.ac (gcc_cv_as_micromips): Check if linker
19693 supports the .set micromips directive.
19694 * configure: Regenerate.
19695 * config.in: Regenerate.
19696 * config/mips/mips-tables.opt: Regenerate.
19697 * config/mips/micromips.md: New file.
19698 * constraints.md (ZC, ZD): New constraints.
19699 * config/mips/predicates.md (movep_src_register): New predicate.
19700 (movep_src_operand): New predicate.
19701 (non_volatile_mem_operand): New predicate.
19702 * config/mips/mips.md (multimem): New type.
19703 (length): Differentiate between 17-bit and 18-bit branch offsets.
19704 (MOVEP1, MOVEP2): New mode iterator.
19705 (mov_<load>l): Use ZC constraint.
19706 (mov_<load>r): Likewise.
19707 (mov_<store>l): Likewise.
19708 (mov_<store>r): Likewise.
19709 (*branch_equality<mode>_inverted): Add microMIPS support.
19710 (*branch_equality<mode>): Likewise.
19711 (*jump_absolute): Likewise.
19712 (indirect_jump_<mode>): Likewise.
19713 (tablejump_<mode>): Likewise.
19714 (<optab>_internal): Likewise.
19715 (sibcall_internal): Likewise.
19716 (sibcall_value_internal): Likewise.
19717 (prefetch): Use constraint ZD.
19718 * config/mips/mips.opt (minterlink-compressed): New option.
19719 (minterlink-mips16): Now an alias for minterlink-compressed.
19720 (mmicromips): New option.
19721 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
19722 (compare_and_swap_12): Likewise.
19723 (sync_add<mode>): Likewise.
19724 (sync_<optab>_12): Likewise.
19725 (sync_old_<optab>_12): Likewise.
19726 (sync_new_<optab>_12): Likewise.
19727 (sync_nand_12): Likewise.
19728 (sync_old_nand_12): Likewise.
19729 (sync_new_nand_12): Likewise.
19730 (sync_sub<mode>): Likewise.
19731 (sync_old_add<mode>): Likewise.
19732 (sync_old_sub<mode>): Likewise.
19733 (sync_new_add<mode>): Likewise.
19734 (sync_new_sub<mode>): Likewise.
19735 (sync_<optab><mode>): Likewise.
19736 (sync_old_<optab><mode>): Likewise.
19737 (sync_new_<optab><mode>): Likewise.
19738 (sync_nand<mode>): Likewise.
19739 (sync_old_nand<mode>): Likewise.
19740 (sync_new_nand<mode>): Likewise.
19741 (sync_lock_test_and_set<mode>): Likewise.
19742 (test_and_set_12): Likewise.
19743 (atomic_compare_and_swap<mode>): Likewise.
19744 (atomic_exchange<mode>_llsc): Likewise.
19745 (atomic_fetch_add<mode>_llsc): Likewise.
19746 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
19747 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
19748 (umips_save_restore_pattern_p): Likewise.
19749 (umips_load_store_pair_p): Likewise.
19750 (umips_output_load_store_pair): Likewise.
19751 (umips_movep_target_p): Likewise.
19752 (umips_12bit_offset_address_p): Likewise.
19753 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
19754 (mips_base_mips16): Rename this...
19755 (mips_base_compression_flags): ...to this. Update all uses.
19756 (mips_attribute_table): Add micromips, nomicromips and nocompression.
19757 (mips_mips16_decl_p): Delete.
19758 (mips_nomips16_decl_p): Delete.
19759 (mips_get_compress_on_flags): New function.
19760 (mips_get_compress_off_flags): New function.
19761 (mips_get_compress_mode): New function.
19762 (mips_get_compress_on_name): New function.
19763 (mips_get_compress_off_name): New function.
19764 (mips_insert_attributes): Support multiple compression types.
19765 (mips_merge_decl_attributes): Likewise.
19766 (umips_12bit_offset_address_p): New function.
19767 (mips_start_function_definition): Emit .set micromips directive.
19768 (mips_call_may_need_jalx_p): New function.
19769 (mips_function_ok_for_sibcall): Add microMIPS support.
19770 (mips_print_operand_punctuation): Support short delay slots and
19772 (umips_swm_mask, umips_swm_encoding): New.
19773 (umips_build_save_restore): New function.
19774 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
19775 (was_mips16_p): Remove.
19776 (old_compression_mode): New.
19777 (mips_set_compression_mode): New function.
19778 (mips_set_current_function): Add microMIPS support.
19779 (mips_option_override): Likewise.
19780 (umips_save_restore_pattern_p): New function.
19781 (umips_output_save_restore): New function.
19782 (umips_load_store_pair_p_1): New function.
19783 (umips_load_store_pair_p): New function.
19784 (umips_output_load_store_pair_1): New function.
19785 (umips_output_load_store_pair): New function.
19786 (umips_movep_target_p) New function.
19787 (mips_prepare_pch_save): Add microMIPS support.
19788 * config/mips/mips.h (TARGET_COMPRESSION): New.
19789 (TARGET_CPU_CPP_BUILTINS): Update macro
19790 to use new compression flags and to support microMIPS.
19791 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
19792 (MIPS_ARCH_FLOAT_SPEC): Likewise.
19793 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
19794 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
19795 (ASM_SPEC): Support mmicromips and mno-micromips.
19796 (M16STORE_REG_P): New macro.
19797 (MIPS_CALL): Support TARGET_MICROMIPS.
19798 (MICROMIPS_J): New macro.
19799 (mips_base_mips16): Rename this...
19800 (mips_base_compression_flags): ...to this.
19801 (UMIPS_12BIT_OFFSET_P): New macro.
19802 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
19803 (MULTILIB_DIRNAMES): Likewise.
19804 2013-03-20 Richard Biener <rguenther@suse.de>
19806 PR tree-optimization/56661
19807 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
19808 the result does not have to be distinct.
19810 2013-03-20 Richard Biener <rguenther@suse.de>
19812 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
19815 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19816 Steven Bosscher <steven@gcc.gnu.org>
19818 PR rtl-optimization/56605
19819 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
19821 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19824 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
19825 that require movd instead of movq.
19827 2013-03-20 Richard Biener <rguenther@suse.de>
19829 * tree-ssa-structalias.c (struct variable_info): Add pointer
19830 to the first field of an aggregate with sub-vars. Make
19831 this and the pointer to the next subfield its ID.
19832 (vi_next): New function.
19833 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
19834 storedanything_id, integer_id): Increment by one.
19835 (new_var_info, get_call_vi, lookup_call_clobber_vi,
19836 get_call_clobber_vi): Adjust.
19837 (solution_set_expand): Simplify and speedup.
19838 (solution_set_add): Inline into ...
19839 (set_union_with_increment): ... this. Adjust accordingly.
19840 (do_sd_constraint): Likewise.
19841 (do_ds_constraint): Likewise.
19842 (do_complex_constraint): Simplify.
19843 (build_pred_graph): Adjust.
19844 (solve_graph): Likewise. Simplify and speedup.
19845 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
19846 get_constraint_for_component_ref, get_constraint_for_1,
19847 first_vi_for_offset, first_or_preceding_vi_for_offset,
19848 create_function_info_for, create_variable_info_for_1,
19849 create_variable_info_for, intra_create_variable_infos): Adjust.
19850 (init_base_vars): Push NULL for ID zero.
19851 (compute_points_to_sets): Adjust.
19853 2013-03-20 Richard Biener <rguenther@suse.de>
19855 * cfgloop.c (verify_loop_structure): Streamline and avoid
19856 ICEing on corrupt loop tree.
19857 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
19860 2013-03-20 Richard Biener <rguenther@suse.de>
19862 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
19863 check whether an SSA update is needed.
19865 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
19867 * config/mips/constraints.md (T): Rename to...
19871 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
19872 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
19874 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19876 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
19877 (*subsi3_carryin_uxtw): Likewise.
19879 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19881 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
19882 (*rorsi3_insn_uxtw): Likewise.
19884 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19886 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
19887 (*extrsi5_insn_uxtw): Likewise.
19889 2013-03-19 Richard Biener <rguenther@suse.de>
19891 PR tree-optimization/56273
19892 * passes.c (init_optimization_passes): Move second VRP after DOM.
19894 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
19896 * config/i386/i386.md (*movti_internal): Merge from
19897 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
19898 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
19899 nox64 isa attributes.
19901 2013-03-18 Richard Biener <rguenther@suse.de>
19903 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
19905 (merge_node_constraints): Likewise.
19906 (build_succ_graph): Likewise.
19907 (valid_graph_edge): Inline into single caller.
19908 (unify_nodes): Likewise. Use bitmap_set_bit return value
19910 (scc_visit): Fix formatting and variable use.
19911 (do_sd_constraint): Use gcc_checking_assert.
19912 (do_ds_constraint): Likewise.
19913 (do_complex_constraint): Likewise.
19914 (condense_visit): Likewise. Cleanup.
19915 (dump_pred_graph): New function.
19916 (perform_var_substitution): Dump the pred-graph before
19917 variable substitution.
19918 (find_equivalent_node): Use gcc_checking_assert.
19919 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
19921 2013-03-18 Richard Biener <rguenther@suse.de>
19923 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
19924 Remove cond_expr_stmt_list argument and do not gimplify the
19926 (vect_loop_versioning): Adjust.
19927 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
19928 Cleanup to use less temporaries.
19929 (vect_create_data_ref_ptr): Cleanup.
19931 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19933 PR tree-optimization/56635
19934 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
19935 require types_compatible_p types.
19937 2013-03-18 Nick Clifton <nickc@redhat.com>
19939 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
19940 spurious backslash.
19942 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
19943 Add missing line to comment describing function.
19945 2013-03-18 Richard Biener <rguenther@suse.de>
19947 PR tree-optimization/56210
19948 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
19949 Handle string / character search functions.
19950 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
19952 2013-03-18 Richard Biener <rguenther@suse.de>
19954 PR middle-end/56483
19955 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
19956 and implement properly.
19957 * gimple.h (gimple_cond_single_var_p): Remove.
19959 2013-03-18 Richard Biener <rguenther@suse.de>
19961 * tree-data-ref.h (find_data_references_in_loop): Declare.
19962 * tree-data-ref.c (get_references_in_stmt): Use a stack
19963 vector pre-allocated in the callers.
19964 (find_data_references_in_stmt): Adjust.
19965 (graphite_find_data_references_in_stmt): Likewise.
19966 (create_rdg_vertices): Likewise.
19967 (find_data_references_in_loop): Export.
19968 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
19969 Compute dependences here...
19970 (vect_analyze_data_refs): ...not here. When we encounter
19971 a non-vectorizable data reference in basic-block vectorization
19972 truncate the data reference vector. Do not bother to
19973 fixup data-dependence information for gather loads.
19974 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
19975 of data references, as reported.
19977 2013-03-18 Richard Biener <rguenther@suse.de>
19979 PR tree-optimization/3713
19980 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
19981 has_constants and expr.
19982 (stmt_has_constants): Properly valueize SSA names when deciding
19983 whether the stmt has constants.
19985 2013-03-18 Richard Biener <rguenther@suse.de>
19987 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
19988 whole function when there is nothing to do.
19989 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
19990 * tree-vectorizer.c (vectorize_loops): Update virtual and
19991 loop-closed SSA once.
19992 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
19994 2013-03-18 Richard Biener <rguenther@suse.de>
19996 PR middle-end/56113
19997 * domwalk.c (bb_postorder): New global static.
19998 (cmp_bb_postorder): New function.
19999 (walk_dominator_tree): Replace scheme imposing an order for
20000 visiting dominator sons by one sorting them at the time they
20001 are pushed on the stack.
20003 2013-03-18 Richard Biener <rguenther@suse.de>
20005 PR tree-optimization/39326
20006 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
20007 (struct mem_ref): Replace mem member with ao_ref typed member.
20008 (MEM_ANALYZABLE): Adjust.
20009 (memref_eq): Likewise.
20010 (mem_ref_alloc): Likewise.
20011 (gather_mem_refs_stmt): Likewise.
20012 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
20013 (execute_sm_if_changed_flag_set): Adjust.
20014 (execute_sm): Likewise.
20015 (ref_always_accessed_p): Likewise.
20016 (refs_independent_p): Likewise.
20017 (can_sm_ref_p): Likewise.
20019 2013-03-18 Jakub Jelinek <jakub@redhat.com>
20022 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
20023 return 1 even for !unsignedp.
20025 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
20027 * config/i386/i386.md (isa): Add x64 and nox64.
20028 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
20029 (*pushtf): Enable *roF alternative for x64 isa only.
20030 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
20031 mode attribute of integer alternatives to DImode for TARGET_64BIT.
20032 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
20033 (*movtf_internal): Merge from *movtf_internal_rex64 and
20034 *movtf_internal_sse. Use x64 and nox64 isa attributes.
20035 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
20036 nox64 isa attributes.
20037 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
20038 nox64 isa attributes.
20039 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
20041 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
20043 * config/alpha/alpha.c (TARGET_LRA_P): New define.
20045 2013-03-17 Jakub Jelinek <jakub@redhat.com>
20048 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
20049 class names. Remove trailing comma after "ALL_REGS".
20051 2013-03-16 Jan Hubicka <jh@suse.cz>
20053 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
20054 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
20055 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
20056 of cgraph_get_create_node.
20057 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
20059 2013-03-16 Jason Merrill <jason@redhat.com>
20062 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
20063 with DW_AT_default_value.
20065 2013-03-16 Jakub Jelinek <jakub@redhat.com>
20067 * BASE-VER: Set to 4.9.0.
20069 2013-03-14 Andi Kleen <ak@linux.intel.com>
20072 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
20073 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
20074 Document _x* TSX intrinsics.
20076 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
20077 David Holsgrove <david.holsgrove@xilinx.com>
20079 * configure.ac: Add MicroBlaze TLS support detection.
20080 * configure: Regenerate.
20081 * config/microblaze/microblaze-protos.h
20082 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
20083 symbol_mentioned_p, label_mentioned_p): Add prototypes.
20084 * config/microblaze/microblaze.c (microblaze_address_type): Add
20085 ADDRESS_TLS and tls_reloc address types.
20086 (microblaze_address_info): Add tls_reloc.
20087 (TARGET_HAVE_TLS): Define.
20088 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
20089 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
20090 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
20091 load_tls_operand, microblaze_call_tls_get_addr,
20092 microblaze_legitimize_tls_address): New functions.
20093 (microblaze_classify_unspec): Handle UNSPEC_TLS.
20094 (get_base_reg): Use microblaze_tls_symbol_p.
20095 (microblaze_classify_address): Handle TLS.
20096 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
20097 label_mentioned_p and microblaze_tls_referenced_p.
20098 (microblaze_legitimize_address): Handle TLS.
20099 (microblaze_address_insns): Handle ADDRESS_TLS.
20100 (pic_address_needs_scratch): Handle TLS.
20101 (print_operand_address): Handle TLS.
20102 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
20103 (microblaze_expand_move): Handle TLS.
20104 (microblaze_legitimate_constant_p): Check
20105 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
20106 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
20107 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
20108 (PIC_OFFSET_TABLE_REGNUM): Set.
20109 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
20110 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
20111 (addsi3, movsi_internal2, movdf_internal): Update constraints
20112 * config/microblaze/predicates.md (arith_plus_operand): Define
20113 (move_operand): Redefine as move_src_operand,
20114 check microblaze_tls_referenced_p.
20116 2013-03-14 Ian Bolton <ian.bolton@arm.com>
20118 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
20119 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
20121 2013-03-14 Ian Bolton <ian.bolton@arm.com>
20123 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
20126 2013-03-14 Jakub Jelinek <jakub@redhat.com>
20128 PR tree-optimization/53265
20129 * common.opt (Waggressive-loop-optimizations): New option.
20130 * tree-ssa-loop-niter.c: Include tree-pass.h.
20131 (do_warn_aggressive_loop_optimizations): New function.
20132 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
20133 if number_of_latch_executions returned constant.
20134 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
20135 early. If number_of_latch_executions returned constant, set
20136 nb_iterations_upper_bound back to it.
20137 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
20139 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
20140 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
20142 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
20143 (MULTILIB_OSDIRNAMES): Set.
20144 * genmultilib: If defaultosdirname doesn't start with :: , set
20145 defaultosdirname2 instead, clear it and emit two . multilib_raw
20146 entries instead of just one.
20148 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
20150 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
20151 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
20152 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
20153 (SUBTARGET_OVERRIDE_OPTIONS): New.
20155 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
20158 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
20159 (musermode): Convert to Var(TARGET_USERMODE).
20160 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
20161 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
20162 * config/sh/sh.c (sh_option_override): Use
20163 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
20164 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
20166 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
20168 (udivsi3_i4_single, divsi3_i4_single): Use
20169 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
20171 2013-03-13 Dave Korn <dave.korn.cygwin@....>
20173 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
20176 2013-03-13 Richard Biener <rguenther@suse.de>
20178 PR tree-optimization/56608
20179 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
20180 calls when vectorizing basic-blocks.
20182 2013-03-13 Jakub Jelinek <jakub@redhat.com>
20185 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
20188 2013-03-12 Jakub Jelinek <jakub@redhat.com>
20190 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
20192 2013-03-11 Jan Hubicka <jh@suse.cz>
20195 * lto-streamer-out.c (output_symbol_p): Skip references from
20196 constructors of external variables.
20198 2013-03-11 Jan Hubicka <jh@suse.cz>
20200 PR middle-end/56571
20201 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
20203 * emit-rtl.c (verify_rtx_sharing): Likewise.
20204 (copy_insn_1): Likewise.
20205 * rtl.c (copy_rtx): Likewise.
20207 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
20210 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
20211 output_operand_lossage message.
20213 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
20216 * arm.c (shift_op): Validate RTL pattern on the fly.
20217 (arm_print_operand, case 'S'): Don't use shift_operator to validate
20220 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20223 * config/pa/pa.md (call_value): Check for calls to powf and direct to
20224 new call patterns that clobber %fr12.
20225 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
20226 split and postreload patterns.
20227 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
20228 registers %fr12 and %fr12R as call used.
20230 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
20232 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
20233 (canon_address, record_store, replace_read, check_mem_read_rtx,
20234 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
20235 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
20236 rest_of_handle_dse): Likewise.
20238 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
20240 PR middle-end/56524
20241 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
20243 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
20244 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
20245 (save_optabs_if_changed): Replace with...
20246 (init_tree_optimization_optabs): ...this.
20247 * optabs.c (save_optabs_if_changed): Rename to...
20248 (init_tree_optimization_optabs): ...this. Take the optimization node
20249 as argument. Do nothing if the base optabs are already correct.
20250 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
20251 to recompute optabs.
20252 * function.h (function): Remove optabs field.
20253 * function.c (invoke_set_current_function_hook): Call
20254 init_tree_optimization_optabs. Use the result to initialize
20257 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
20259 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
20260 if GTMA_HAS_NO_INSTRUMENTATION.
20261 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
20262 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
20263 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
20264 * gimple-pretty-print.c (dump_gimple_transaction): Handle
20265 GTMA_HAS_NO_INSTRUMENTATION.
20267 2013-03-08 Jakub Jelinek <jakub@redhat.com>
20269 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
20272 2013-03-08 Marek Polacek <polacek@redhat.com>
20273 Jakub Jelinek <jakub@redhat.com>
20275 PR tree-optimization/56478
20276 * predict.c (is_comparison_with_loop_invariant_p): Change the
20277 type of loop_step to tree.
20278 (predict_loops): Adjust.
20279 (predict_iv_comparison): Perform the computations on double_ints.
20281 2013-03-08 Richard Biener <rguenther@suse.de>
20283 PR tree-optimization/56570
20284 * tree-cfg.c (verify_expr_location_1): Verify locations for
20286 * tree-sra.c (create_access_replacement): Strip locations
20287 from DECL_DEBUG_EXPRs.
20289 2013-03-08 Richard Biener <rguenther@suse.de>
20291 * tree-inline.c (expand_call_inline): Do not associate
20292 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
20293 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
20295 2013-03-08 Richard Biener <rguenther@suse.de>
20297 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
20298 or block changes with -Og. Fix for location / block encoding
20299 changes and PHI arguments with locations.
20301 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
20303 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
20305 (struct output_info): Likewise.
20306 (register_overhead): Remove bad gcc_assert.
20307 (bitmap_find_bit): If there is only a single bitmap element, do not
20308 count a miss as a search.
20309 (print_statistics): Update for counter type changes.
20310 (dump_bitmap_statistics): Likewise. Print headers such that they
20311 are properly lined up with the printed counters.
20313 2013-03-07 Jakub Jelinek <jakub@redhat.com>
20315 PR tree-optimization/56559
20316 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
20317 check that it has only a single use.
20319 2013-03-07 Richard Biener <rguenther@suse.de>
20321 * doc/invoke.texi (fwhole-program): Discourage use in combination
20324 2013-03-06 Jakub Jelinek <jakub@redhat.com>
20326 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
20328 PR tree-optimization/56539
20329 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
20330 instead of GSI_CONTINUE_LINKING as last argument to
20331 force_gimple_operand_gsi. Adjust function comment.
20333 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
20336 PR middle-end/56548
20337 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
20338 promoted mode, convert the result back to the original mode.
20340 2013-03-06 Richard Biener <rguenther@suse.de>
20342 PR middle-end/56294
20343 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
20344 (insert_updated_phi_nodes_compare_uids): New function.
20345 (update_ssa): Sort symbols_to_rename after UID before
20346 traversing it to insert PHI nodes.
20348 2013-03-06 Richard Biener <rguenther@suse.de>
20350 PR middle-end/50494
20351 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
20352 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
20355 2013-02-13 Richard Biener <rguenther@suse.de>
20358 * varasm.c (output_constant_def_1): Get the decl representing
20359 the constant as argument.
20360 (output_constant_def): Wrap output_constant_def_1.
20361 (make_decl_rtl): Use output_constant_def_1 with the decl
20362 representing the constant.
20363 (build_constant_desc): Optionally re-use a decl already
20364 representing the constant.
20365 (tree_output_constant_def): Adjust.
20367 2013-03-06 Joey Ye <joey.ye@arm.com>
20370 * gcc.c (convert_white_space): New function.
20371 (main): Handles white space in function name.
20373 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
20376 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
20377 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
20378 to SH_DIV_CALL_TABLE for TARGET_SH2.
20379 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
20381 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
20382 call-table options.
20384 2013-03-05 Sterling Augustine <saugustine@google.com>
20385 Cary Coutant <ccoutant@google.com>
20388 * dwarf2out.c (resolve_addr): Don't call
20389 remove_loc_list_addr_table_entries a second time for the same
20392 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20395 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
20396 (avoid_complex_debug_insns): New function.
20397 (expand_debug_locations): Call it.
20399 PR rtl-optimization/56484
20400 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
20401 lifetimes of hard registers on small register class machines.
20403 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
20405 * config/microblaze/microblaze-protos.h: Rename
20406 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
20407 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
20409 (microblaze_fast_interrupt_function_p): New function.
20410 (microblaze_is_interrupt_handler): Rename to
20411 microblaze_is_interrupt_variant and add fast_interrupt check.
20412 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
20413 (save_restore_insns): Likewise.
20414 (compute_frame_size): Likewise.
20415 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
20416 (microblaze_globalize_label): Likewise.
20417 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
20418 * config/microblaze/microblaze.md: Use wrapper
20419 microblaze_is_interrupt_variant.
20421 2013-03-05 Kai Tietz <ktietz@redhat.com>
20423 * sdbout.c (sdbout_one_type): Switch to current function's section
20424 supporting cold/hot.
20426 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
20428 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
20431 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20433 PR middle-end/56461
20434 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
20435 if VALGRIND_GET_VBITS is defined, temporarily make object
20436 memory all defined, and restore previous valgrind addressability
20437 and definability afterwards. Free this_object at the end.
20439 PR middle-end/56461
20440 * lra.c (lra): Call lra_clear_live_ranges if live_p,
20441 right before calling lra_create_live_ranges, also call it
20442 when clearing live_p. Only call lra_clear_live_ranges
20443 at the end if live_p.
20445 PR middle-end/56461
20446 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
20448 2013-03-05 Richard Biener <rguenther@suse.de>
20450 PR tree-optimization/56521
20451 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
20454 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
20457 * except.h (remove_unreachable_eh_regions): New prototype.
20458 * except.c (remove_eh_handler_splicer): New function, split out
20459 of remove_eh_handler.
20460 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
20461 warning about running it on many EH regions one at a time.
20462 (remove_unreachable_eh_regions_worker): New function, walk the
20463 EH tree in depth-first order and remove non-marked regions.
20464 (remove_unreachable_eh_regions): New function.
20465 * tree-eh.c (mark_reachable_handlers): New function, split out
20466 from remove_unreachable_handlers.
20467 (remove_unreachable_handlers): Use mark_reachable_handlers and
20468 remove_unreachable_eh_regions.
20469 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
20470 and remove_unreachable_eh_regions.
20472 2013-03-05 Richard Biener <rguenther@suse.de>
20474 PR middle-end/56525
20475 * loop-init.c (fix_loop_structure): Remove loops in two stages,
20476 not freeing them until the end.
20478 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20480 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
20482 2013-03-05 Richard Biener <rguenther@suse.de>
20484 PR tree-optimization/56270
20485 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
20486 of loads after scheduling an SLP instance.
20488 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20490 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
20492 (check_gcc_parallelize): Run guality.exp as a separate job from
20493 vect.exp with unsorted.exp and $(dg_target_exps) separately from
20494 struct-layout-1.exp with stackalign.exp.
20496 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
20498 PR middle-end/56461
20499 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
20500 load_index sbitmap even if some bit in it isn't set.
20502 PR middle-end/56461
20503 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
20504 (discover_iteration_bound_by_body_walk): Change queues to
20505 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
20506 spelling in comment. Call safe_push on queues[bound_index] directly.
20507 Release queues[queue_index] in every iteration unconditionally.
20508 Release bounds vector.
20510 PR middle-end/56461
20511 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
20512 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
20513 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
20516 2013-03-05 Richard Biener <rguenther@suse.de>
20519 * tree-inline.c (remap_blocks_to_null): New function.
20520 (expand_call_inline): When expanding a call stmt without
20521 an associated block inline remap all callee blocks to NULL.
20523 2013-03-05 Jakub Jelinek <jakub@redhat.com>
20525 PR rtl-optimization/56494
20526 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
20527 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
20528 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
20530 PR middle-end/56461
20531 * sel-sched-ir.c (free_sched_pools): Release
20532 succs_info_pool.stack[succs_info_pool.max_top] vectors too
20533 if succs_info_pool.max_top isn't -1.
20536 * opts.c (opts_obstack, opts_concat): Moved to...
20537 * opts-common.c (opts_obstack, opts_concat): ... here.
20539 2013-03-04 Jakub Jelinek <jakub@redhat.com>
20541 PR middle-end/56461
20542 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
20544 2013-03-04 Martin Jambor <mjambor@suse.cz>
20546 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
20547 all appropriate places.
20549 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
20551 PR tree-optimization/56424
20552 * ipa-split.c (split_function): Do not set the RSO flag if result is
20553 not by reference and its type is a register type.
20555 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
20557 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
20558 (microblaze_legitimate_pic_operand): Likewise
20559 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
20560 new function microblaze_legitimate_pic_operand
20561 * config/microblaze/microblaze-protos.h
20562 (microblaze_legitimate_pic_operand): Declare.
20564 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20566 * config/microblaze/predicates.md (call_insn_simple_operand):
20567 New predicate for supported rtx code types.
20568 * config/microblaze/microblaze.md (call_internal1): Use
20569 call_insn_simple_operand predicate.
20571 2013-03-04 Jakub Jelinek <jakub@redhat.com>
20573 PR middle-end/56461
20574 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
20575 partitions.ordered_remove.
20577 PR middle-end/56461
20578 * tree-vect-stmts.c (vectorizable_conversion): Don't call
20579 vec_oprnds0.create (1) for modifier == NONE.
20581 PR middle-end/56461
20582 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
20583 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
20584 vec_oprnds1 right before pushing anything to it for
20587 PR middle-end/56461
20588 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
20589 set nbbs to 0 instead of having separate code path.
20590 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
20591 instead of false as last argument if returning NULL.
20593 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
20595 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
20596 the attribute is now called "target" instead of "option".
20597 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
20598 * doc/tm.texi.in (Target Attributes): Likewise document the correct
20599 attribute/pragma name for TARGET_OPTION_VALID_P and
20600 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
20601 * doc/tm.texi: Regenerated.
20603 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
20605 * config/microblaze/microblaze.c:
20606 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
20607 * config/microblaze/microblaze.h: Add -mxl-reorder to
20609 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
20610 instructions emitted if TARGET_REORDER.
20611 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
20612 or 0 for -m/-mno case, but initialises as 2 to detect default use case
20615 2013-03-01 Xinliang David Li <davidxl@google.com>
20617 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
20620 2013-03-01 Jakub Jelinek <jakub@redhat.com>
20622 PR middle-end/56461
20623 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
20624 vector even when returning true. Fix up function comment formatting.
20626 PR middle-end/56461
20627 * ira-build.c (ira_loop_nodes_count): New variable.
20628 (create_loop_tree_nodes): Initialize it.
20629 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
20631 PR middle-end/56461
20632 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
20633 method on dr_chain and result_chain.
20634 * tree-vect-stmts.c (vectorizable_store): Only call
20635 result_chain.create if j == 0.
20637 PR middle-end/56461
20638 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
20639 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
20640 before overwriting it.
20642 2013-03-01 Tobias Burnus <burnus@net-b.de>
20644 * doc/extended.texi (C Extensions): Change order in @menu
20646 (Other MIPS Built-in Functions): Move last MIPS entry before
20647 "picoChip Built-in Functions".
20648 (SH Built-in Functions): Move after RX Built-in Functions.
20649 * doc/gcc.texi (Introduction): Change order in @menu
20651 * doc/md.texi (Constraints): Ditto.
20652 * gty.texi (Type Information): Ditto.
20653 (User-provided marking routines for template types): Make
20655 * doc/invoke.texi (AArch64 Options): Move before
20656 "Adapteva Epiphany Options".
20658 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
20659 Jakub Jelinek <jakub@redhat.com>
20662 * asan.c (gate_asan): Lookup no_sanitize_address instead of
20663 no_address_safety_analysis attribute.
20664 * doc/extend.texi (no_address_safety_attribute): Rename to
20665 no_sanitize_address attribute, mention no_address_safety_analysis
20666 attribute as deprecated alias.
20668 2013-02-28 Jakub Jelinek <jakub@redhat.com>
20670 PR middle-end/56461
20671 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
20672 type to vec<vec<tree> > *.
20673 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
20674 to be vec<tree> instead of vec<tree> *, set vec_defs
20675 to vNULL and call vec_defs.create (number_of_vects), adjust other
20677 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
20678 vectorizable_condition): Adjust vect_get_slp_defs callers.
20680 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20682 * config/aarch64/aarch64.c
20683 (aarch64_float_const_representable): Remove unused variable.
20685 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20687 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
20689 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20691 * config/aarch64/aarch64-builtins.c
20692 (aarch64_init_simd_builtins): Make static.
20694 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20696 * config/aarch64/aarch64.c
20697 (aarch64_simd_make_constant): Make static.
20699 2013-02-28 Martin Jambor <mjambor@suse.cz>
20701 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
20702 with no initialization to the RHS of debug statements.
20704 2013-02-28 Martin Jambor <mjambor@suse.cz>
20706 PR tree-optimization/56294
20707 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
20709 (get_access_replacement): Do not call create_access_replacement.
20710 Assert a replacement exists.
20711 (get_repl_default_def_ssa_name): Create the replacement declaration
20714 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20716 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
20717 final_end_function.
20719 2013-02-28 Marek Polacek <polacek@redhat.com>
20721 PR rtl-optimization/56466
20722 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
20723 if we're changing a loop.
20724 (peel_loops_completely): Likewise.
20726 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
20729 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
20731 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20734 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
20735 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
20736 INTX_FTYPE_FX, FX_FTYPE_INTX.
20737 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
20739 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20741 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
20742 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
20743 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
20744 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
20745 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
20746 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
20747 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
20748 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
20749 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
20750 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
20751 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
20752 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
20753 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
20754 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
20755 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
20756 (avrxmega6): Increase max flash segments from 5 to 6.
20757 * config/avr/t-multilib: Regenerate.
20758 * config/avr/avr-tables.opt: Regenerate.
20759 * doc/avr-mmcu.texi: Regenerate.
20761 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20763 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
20764 (avr_device_to_arch): Rename to avr_device_to_ld.
20765 (avr_device_to_as): New prototype.
20766 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
20767 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
20768 * config/avr/driver-avr.c (avr_device_to_as): New.
20769 (avr_device_to_arch): Rename to avr_device_to_ld.
20771 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20773 PR middle-end/56461
20774 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
20775 method on dr_chain and result_chain.
20777 PR middle-end/56461
20778 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
20779 pointer_set_destroy on not_executed_last_iteration.
20781 PR middle-end/56461
20782 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
20784 PR middle-end/56461
20785 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
20786 FOR_EACH_DEFINED_FUNCTION when freeing state.
20788 PR middle-end/56461
20789 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
20791 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
20794 PR middle-end/56461
20795 * ipa-cp.c (decide_whether_version_node): Call vec_free on
20796 known_aggs[i].items and release known_aggs vector.
20798 PR middle-end/56461
20799 * ipa-reference.c (propagate): Free node_info even for alias nodes.
20801 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20803 * config/microblaze/microblaze.c (microblaze_emit_compare):
20804 Use xor for EQ/NE comparisions.
20805 * config/microblaze/microblaze.md (cstoresf4): Add constraints
20806 (cbranchsf4): Adjust operator to comparison_operator.
20808 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20810 PR middle-end/56461
20811 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
20813 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
20814 vec_safe_push, always update *slot.
20815 (redirect_edge_var_map_clear): Use vec_free.
20816 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
20817 (free_var_map_entry): Use vec_free.
20818 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
20819 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
20821 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
20823 PR middle-end/45472
20824 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
20825 when the may_trap_p bit of the exprs being merged differs.
20826 Reorder tests for speculativeness in the logical and operator.
20828 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20830 * incpath.c (add_standard_paths): Use reconcat instead of concat
20831 where appropriate and avoid leaking memory.
20833 * opts.h: Include obstack.h.
20834 (opts_concat): New prototype.
20835 (opts_obstack): New declaration.
20836 * opts.c (opts_concat): New function.
20837 (opts_obstack): New variable.
20838 (init_options_struct): Call gcc_init_obstack on opts_obstack.
20839 (finish_options): Use opts_concat instead of concat
20840 and XOBNEWVEC instead of XNEWVEC.
20841 * opts-common.c (generate_canonical_option, decode_cmdline_option,
20842 generate_option): Likewise.
20843 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
20844 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
20847 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
20848 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
20850 2013-02-26 Jakub Jelinek <jakub@redhat.com>
20852 PR middle-end/56461
20853 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
20855 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
20857 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
20858 (arm_block_move_unaligned_straight): Likewise.
20859 (arm_adjust_block_mem): Likewise.
20861 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
20864 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
20865 temp, cond and label.
20866 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
20869 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
20870 * config/c6x/c6x.h (dbx_register_map): Update declaration.
20873 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
20874 of prologue/epilogue functions.
20877 * config/tilegx/tilegx.c (tilegx_expand_prologue):
20878 Remove unused variable cfa_offset.
20879 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
20882 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
20883 type promotion to unsigned.
20886 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
20887 for HOST_WIDE_INT of 32 bit / same size as int.
20888 (arm_block_move_unaligned_straight): Likewise.
20889 (arm_adjust_block_mem): Likewise.
20892 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
20895 2013-02-26 Marek Polacek <polacek@redhat.com>
20897 PR tree-optimization/56426
20898 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
20900 2013-02-26 Richard Biener <rguenther@suse.de>
20903 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
20904 unused variable loops.
20906 2013-02-26 Jakub Jelinek <jakub@redhat.com>
20908 PR tree-optimization/56448
20909 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
20910 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
20911 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
20912 later operands of the references, or even first operand for
20913 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
20915 PR tree-optimization/56443
20916 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
20917 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
20918 to type_for_mode langhook.
20920 2013-02-25 Matt Turner <mattst88@gmail.com>
20922 * doc/invoke.texi: Document r4700.
20924 2013-02-25 Richard Biener <rguenther@suse.de>
20926 PR tree-optimization/56175
20927 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
20929 (simplify_bitwise_binary): ... here. Also guard the conversion
20930 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
20932 2013-02-25 Catherine Moore <clm@codesourcery.com>
20935 2013-02-24 Catherine Moore <clm@codesourcery.com>
20936 Maciej W. Rozycki <macro@codesourcery.com>
20937 Tom de Vries <tom@codesourcery.com>
20938 Nathan Sidwell <nathan@codesourcery.com>
20939 Iain Sandoe <iain@codesourcery.com>
20940 Nathan Froyd <froydnj@codesourcery.com>
20941 Chao-ying Fu <fu@mips.com>
20943 * doc/extend.texi: (micromips, nomicromips, nocompression):
20944 Document new function attributes.
20945 * doc/invoke.texi (minterlink-compressed, mmicromips,
20946 m14k, m14ke, m14kec): Document new options.
20947 (minterlink-mips16): Update documentation.
20948 * doc/md.texi (ZC, ZD): Document new constraints.
20949 * configure.ac (gcc_cv_as_micromips): Check if linker
20950 supports the .set micromips directive.
20951 * configure: Regenerate.
20952 * config.in: Regenerate.
20953 * config/mips/mips-tables.opt: Regenerate.
20954 * config/mips/micromips.md: New file.
20955 * constraints.md (ZC, AD): New constraints.
20956 * config/mips/predicates.md (movep_src_register): New predicate.
20957 (movep_src_operand): New predicate.
20958 (non_volatile_mem_operand): New predicate.
20959 * config/mips/mips.md (multimem): New type.
20960 (length): Differentiate between 17-bit and 18-bit branch offsets.
20961 (MOVEP1, MOVEP2): New mode iterator.
20962 (mov_<load>l): Use ZC constraint.
20963 (mov_<load>r): Likewise.
20964 (mov_<store>l): Likewise.
20965 (mov_<store>r): Likewise.
20966 (*branch_equality<mode>_inverted): Add microMIPS support.
20967 (*branch_equality<mode>): Likewise.
20968 (*jump_absolute): Likewise.
20969 (indirect_jump_<mode>): Likewise.
20970 (tablejump_<mode>): Likewise.
20971 (<optab>_internal): Likewise.
20972 (sibcall_internal): Likewise.
20973 (sibcall_value_internal): Likewise.
20974 (prefetch): Use constraint ZD.
20975 * config/mips/mips.opt (minterlink-compressed): New option.
20976 (minterlink-mips16): Now an alias for minterlink-compressed.
20977 (mmicromips): New option.
20978 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20979 (compare_and_swap_12): Likewise.
20980 (sync_add<mode>): Likewise.
20981 (sync_<optab>_12): Likewise.
20982 (sync_old_<optab>_12): Likewise.
20983 (sync_new_<optab>_12): Likewise.
20984 (sync_nand_12): Likewise.
20985 (sync_old_nand_12): Likewise.
20986 (sync_new_nand_12): Likewise.
20987 (sync_sub<mode>): Likewise.
20988 (sync_old_add<mode>): Likewise.
20989 (sync_old_sub<mode>): Likewise.
20990 (sync_new_add<mode>): Likewise.
20991 (sync_new_sub<mode>): Likewise.
20992 (sync_<optab><mode>): Likewise.
20993 (sync_old_<optab><mode>): Likewise.
20994 (sync_new_<optab><mode>): Likewise.
20995 (sync_nand<mode>): Likewise.
20996 (sync_old_nand<mode>): Likewise.
20997 (sync_new_nand<mode>): Likewise.
20998 (sync_lock_test_and_set<mode>): Likewise.
20999 (test_and_set_12): Likewise.
21000 (atomic_compare_and_swap<mode>): Likewise.
21001 (atomic_exchange<mode>_llsc): Likewise.
21002 (atomic_fetch_add<mode>_llsc): Likewise.
21003 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
21004 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
21005 (umips_save_restore_pattern_p): Likewise.
21006 (umips_load_store_pair_p): Likewise.
21007 (umips_output_load_store_pair): Likewise.
21008 (umips_movep_target_p): Likewise.
21009 (umips_12bit_offset_address_p): Likewise.
21010 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
21011 (mips_base_mips16): Rename this...
21012 (mips_base_compression_flags): ...to this. Update all uses.
21013 (mips_attribute_table): Add micromips, nomicromips and nocompression.
21014 (mips_mips16_decl_p): Delete.
21015 (mips_nomips16_decl_p): Delete.
21016 (mips_get_compress_on_flags): New function.
21017 (mips_get_compress_off_flags): New function.
21018 (mips_get_compress_mode): New function.
21019 (mips_get_compress_on_name): New function.
21020 (mips_get_compress_off_name): New function.
21021 (mips_insert_attributes): Support multiple compression types.
21022 (mips_merge_decl_attributes): Likewise.
21023 (umips_12bit_offset_address_p): New function.
21024 (mips_start_function_definition): Emit .set micromips directive.
21025 (mips_call_may_need_jalx_p): New function.
21026 (mips_function_ok_for_sibcall): Add microMIPS support.
21027 (mips_print_operand_punctuation): Support short delay slots and
21029 (umips_swm_mask, umips_swm_encoding): New.
21030 (umips_build_save_restore): New function.
21031 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
21032 (was_mips16_p): Remove.
21033 (old_compression_mode): New.
21034 (mips_set_compression_mode): New function.
21035 (mips_set_current_function): Add microMIPS support.
21036 (mips_option_override): Likewise.
21037 (umips_save_restore_pattern_p): New function.
21038 (umips_output_save_restore): New function.
21039 (umips_load_store_pair_p_1): New function.
21040 (umips_load_store_pair_p): New function.
21041 (umips_output_load_store_pair_1): New function.
21042 (umips_output_load_store_pair): New function.
21043 (umips_movep_target_p) New function.
21044 (mips_prepare_pch_save): Add microMIPS support.
21045 * config/mips/mips.h (TARGET_COMPRESSION): New.
21046 (TARGET_CPU_CPP_BUILTINS): Update macro
21047 to use new compression flags and to support microMIPS.
21048 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
21049 (MIPS_ARCH_FLOAT_SPEC): Likewise.
21050 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
21051 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
21052 (ASM_SPEC): Support mmicromips and mno-micromips.
21053 (M16STORE_REG_P): New macro.
21054 (MIPS_CALL): Support TARGET_MICROMIPS.
21055 (MICROMIPS_J): New macro.
21056 (mips_base_mips16): Rename this...
21057 (mips_base_compression_flags): ...to this.
21058 (UMIPS_12BIT_OFFSET_P): New macro.
21059 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
21060 (MULTILIB_DIRNAMES): Likewise.
21062 2013-02-25 Tom de Vries <tom@codesourcery.com>
21064 PR rtl-optimization/56131
21065 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
21066 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
21067 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
21069 2013-02-25 Tobias Burnus <burnus@net-b.de>
21071 * doc/invoke.texi (-fsanitize=): Move from optimization
21072 to debugging options.
21074 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
21076 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
21078 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
21079 Alexander Monakov <amonakov@ispras.ru>
21081 PR middle-end/56077
21082 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
21083 flush pending lists also on non-jumps. Adjust comment.
21085 2013-02-24 Catherine Moore <clm@codesourcery.com>
21086 Maciej W. Rozycki <macro@codesourcery.com>
21087 Tom de Vries <tom@codesourcery.com>
21088 Nathan Sidwell <nathan@codesourcery.com>
21089 Iain Sandoe <iain@codesourcery.com>
21090 Nathan Froyd <froydnj@codesourcery.com>
21091 Chao-ying Fu <fu@mips.com>
21093 * doc/extend.texi: (micromips, nomicromips, nocompression):
21094 Document new function attributes.
21095 * doc/invoke.texi (minterlink-compressed, mmicromips,
21096 m14k, m14ke, m14kec): Document new options.
21097 (minterlink-mips16): Update documentation.
21098 * doc/md.texi (ZC, ZD): Document new constraints.
21099 * configure.ac (gcc_cv_as_micromips): Check if linker
21100 supports the .set micromips directive.
21101 * configure: Regenerate.
21102 * config.in: Regenerate.
21103 * config/mips/mips-tables.opt: Regenerate.
21104 * config/mips/micromips.md: New file.
21105 * constraints.md (ZC, AD): New constraints.
21106 * config/mips/predicates.md (movep_src_register): New predicate.
21107 (movep_src_operand): New predicate.
21108 (non_volatile_mem_operand): New predicate.
21109 * config/mips/mips.md (multimem): New type.
21110 (length): Differentiate between 17-bit and 18-bit branch offsets.
21111 (MOVEP1, MOVEP2): New mode iterator.
21112 (mov_<load>l): Use ZC constraint.
21113 (mov_<load>r): Likewise.
21114 (mov_<store>l): Likewise.
21115 (mov_<store>r): Likewise.
21116 (*branch_equality<mode>_inverted): Add microMIPS support.
21117 (*branch_equality<mode>): Likewise.
21118 (*jump_absolute): Likewise.
21119 (indirect_jump_<mode>): Likewise.
21120 (tablejump_<mode>): Likewise.
21121 (<optab>_internal): Likewise.
21122 (sibcall_internal): Likewise.
21123 (sibcall_value_internal): Likewise.
21124 (prefetch): Use constraint ZD.
21125 * config/mips/mips.opt (minterlink-compressed): New option.
21126 (minterlink-mips16): Now an alias for minterlink-compressed.
21127 (mmicromips): New option.
21128 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
21129 (compare_and_swap_12): Likewise.
21130 (sync_add<mode>): Likewise.
21131 (sync_<optab>_12): Likewise.
21132 (sync_old_<optab>_12): Likewise.
21133 (sync_new_<optab>_12): Likewise.
21134 (sync_nand_12): Likewise.
21135 (sync_old_nand_12): Likewise.
21136 (sync_new_nand_12): Likewise.
21137 (sync_sub<mode>): Likewise.
21138 (sync_old_add<mode>): Likewise.
21139 (sync_old_sub<mode>): Likewise.
21140 (sync_new_add<mode>): Likewise.
21141 (sync_new_sub<mode>): Likewise.
21142 (sync_<optab><mode>): Likewise.
21143 (sync_old_<optab><mode>): Likewise.
21144 (sync_new_<optab><mode>): Likewise.
21145 (sync_nand<mode>): Likewise.
21146 (sync_old_nand<mode>): Likewise.
21147 (sync_new_nand<mode>): Likewise.
21148 (sync_lock_test_and_set<mode>): Likewise.
21149 (test_and_set_12): Likewise.
21150 (atomic_compare_and_swap<mode>): Likewise.
21151 (atomic_exchange<mode>_llsc): Likewise.
21152 (atomic_fetch_add<mode>_llsc): Likewise.
21153 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
21154 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
21155 (umips_save_restore_pattern_p): Likewise.
21156 (umips_load_store_pair_p): Likewise.
21157 (umips_output_load_store_pair): Likewise.
21158 (umips_movep_target_p): Likewise.
21159 (umips_12bit_offset_address_p): Likewise.
21160 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
21161 (mips_base_mips16): Rename this...
21162 (mips_base_compression_flags): ...to this. Update all uses.
21163 (mips_attribute_table): Add micromips, nomicromips and nocompression.
21164 (mips_mips16_decl_p): Delete.
21165 (mips_nomips16_decl_p): Delete.
21166 (mips_get_compress_on_flags): New function.
21167 (mips_get_compress_off_flags): New function.
21168 (mips_get_compress_mode): New function.
21169 (mips_get_compress_on_name): New function.
21170 (mips_get_compress_off_name): New function.
21171 (mips_insert_attributes): Support multiple compression types.
21172 (mips_merge_decl_attributes): Likewise.
21173 (umips_12bit_offset_address_p): New function.
21174 (mips_start_function_definition): Emit .set micromips directive.
21175 (mips_call_may_need_jalx_p): New function.
21176 (mips_function_ok_for_sibcall): Add microMIPS support.
21177 (mips_print_operand_punctuation): Support short delay slots and
21179 (umips_swm_mask, umips_swm_encoding): New.
21180 (umips_build_save_restore): New function.
21181 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
21182 (was_mips16_p): Remove.
21183 (old_compression_mode): New.
21184 (mips_set_compression_mode): New function.
21185 (mips_set_current_function): Add microMIPS support.
21186 (mips_option_override): Likewise.
21187 (umips_save_restore_pattern_p): New function.
21188 (umips_output_save_restore): New function.
21189 (umips_load_store_pair_p_1): New function.
21190 (umips_load_store_pair_p): New function.
21191 (umips_output_load_store_pair_1): New function.
21192 (umips_output_load_store_pair): New function.
21193 (umips_movep_target_p) New function.
21194 (mips_prepare_pch_save): Add microMIPS support.
21195 * config/mips/mips.h (TARGET_COMPRESSION): New.
21196 (TARGET_CPU_CPP_BUILTINS): Update macro
21197 to use new compression flags and to support microMIPS.
21198 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
21199 (MIPS_ARCH_FLOAT_SPEC): Likewise.
21200 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
21201 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
21202 (ASM_SPEC): Support mmicromips and mno-micromips.
21203 (M16STORE_REG_P): New macro.
21204 (MIPS_CALL): Support TARGET_MICROMIPS.
21205 (MICROMIPS_J): New macro.
21206 (mips_base_mips16): Rename this...
21207 (mips_base_compression_flags): ...to this.
21208 (UMIPS_12BIT_OFFSET_P): New macro.
21209 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
21210 (MULTILIB_DIRNAMES): Likewise.
21212 2013-02-24 Jakub Jelinek <jakub@redhat.com>
21215 * target-globals.c (save_target_globals): For init_reg_sets and
21216 target_reinit remporarily set this_fn_optabs to this_target_optabs.
21218 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
21220 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
21221 * config/aarch64/t-aarch64
21222 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
21224 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
21226 PR inline-asm/56148
21227 * lra-constraints.c (process_alt_operands): Reload operand
21228 conflicting with earlier clobber only if no more other conflicting
21231 2013-02-22 Jakub Jelinek <jakub@redhat.com>
21234 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
21235 if not linking a shared library.
21237 2013-02-22 Seth LaForge <sethml@google.com>
21239 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
21241 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
21243 * config/arm/arm.md (split for extendsidi): Update condition.
21244 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
21245 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
21246 (qhs_zextenddi_cstr): Likewise.
21248 2013-02-21 Jakub Jelinek <jakub@redhat.com>
21250 PR middle-end/56420
21251 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
21252 avoid signed wrapping.
21253 (expand_mult): Handle properly multiplication by
21254 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
21255 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
21256 in the compiler if coeff is HOST_WIDE_INT_MIN.
21257 (expand_divmod): Don't make ext_op1 static, change it's type to
21258 uhwi. Avoid undefined behavior in -INTVAL (op1).
21260 PR rtl-optimization/50339
21261 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
21263 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
21264 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
21265 into splitting_ashiftrt field.
21266 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
21268 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
21271 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
21273 PR middle-end/56108
21274 * trans-mem.c (execute_tm_mark): Do not expand transactions that
21275 are sure to go irrevocable.
21277 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
21279 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
21280 scalars are valid operands.
21282 2013-02-21 Martin Jambor <mjambor@suse.cz>
21284 PR tree-optimization/56310
21285 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
21286 only matching indices and non-negative final offsets.
21287 (intersect_aggregates_with_edge): Pass src_idx to
21288 agg_replacements_to_vector. Pass src_idx insstead of index to
21289 intersect_with_agg_replacements.
21291 2013-02-21 Martin Jambor <mjambor@suse.cz>
21293 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
21294 instead of hard-wired defaults.
21296 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
21298 * doc/invoke.texi (MIPS Options): Update documentation of the
21299 floating-point multiply-accumulate instruction restrictions.
21301 2013-02-21 Kostya Serebryany <kcc@google.com>
21303 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
21304 asan_shadow_offset on x86_64 linux.
21306 2013-02-21 Richard Biener <rguenther@suse.de>
21308 PR tree-optimization/56415
21310 2013-02-11 Richard Biener <rguenther@suse.de>
21312 PR tree-optimization/56273
21313 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
21316 2013-02-21 Jakub Jelinek <jakub@redhat.com>
21319 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
21322 PR inline-asm/56405
21323 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
21324 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
21326 2013-02-20 Jan Hubicka <jh@suse.cz>
21328 PR tree-optimization/56265
21329 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
21330 when target is referenced for first time.
21332 2013-02-20 Richard Biener <rguenther@suse.de>
21334 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
21335 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
21336 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
21337 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
21338 not return anything.
21339 (rename_ssa_copies): Do not remove unused locals.
21340 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
21341 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
21342 * passes.c (execute_function_todo): Do not schedule unused locals
21343 removal if cleanup_tree_cfg did something.
21344 * tree-ssa-live.c (remove_unused_locals): Dump statistics
21345 about the number of removed locals.
21347 2013-02-20 Richard Biener <rguenther@suse.de>
21349 PR tree-optimization/56398
21350 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
21352 2013-02-20 Martin Jambor <mjambor@suse.cz>
21354 PR tree-optimization/55334
21355 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
21356 restricted pointers to arrays.
21358 2013-02-20 Richard Biener <rguenther@suse.de>
21359 Jakub Jelinek <jakub@redhat.com>
21361 PR tree-optimization/56396
21362 * tree-ssa-ccp.c (n_const_val): New static variable.
21363 (get_value): Return NULL for SSA names we don't have a lattice
21365 (ccp_initialize): Initialize n_const_val.
21366 * tree-ssa-copy.c (n_copy_of): New static variable.
21367 (init_copy_prop): Initialize n_copy_of.
21368 (get_value): Return NULL_TREE for SSA names we don't have a
21371 2013-02-20 Martin Jambor <mjambor@suse.cz>
21373 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
21375 2013-02-20 Richard Biener <rguenther@suse.de>
21377 * genpreds.c (write_lookup_constraint): Do not compare first
21378 letter of the constraint again.
21380 2013-02-20 Richard Biener <rguenther@suse.de>
21382 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
21384 (get_use_iv_cost): Terminate hashtable walk when coming across
21387 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
21389 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
21390 reassociation for avx2 targets.
21392 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21394 * config/microblaze/microblaze.c: microblaze_has_clz = 0
21395 Add version check for v8.10.a to enable microblaze_has_clz
21396 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
21397 version and TARGET_PATTERN_COMPARE check
21398 * config/microblaze/microblaze.md: New clzsi2 instruction
21400 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21402 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
21403 function before branching.
21405 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
21407 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
21409 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
21411 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
21413 PR middle-end/55889
21414 * sel-sched.c: Include ira.h.
21415 (implicit_clobber_conflict_p): New function.
21416 (moveup_expr): Use it.
21417 * Makefile.in (sel-sched.o): Depend on ira.h.
21419 2013-02-19 Richard Biener <rguenther@suse.de>
21421 PR tree-optimization/56384
21422 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
21423 (vn_hash_type): Split out from ...
21424 (vn_hash_constant_with_type): ... here.
21425 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
21426 (vn_phi_eq): Compare types from vn_phi_s structure.
21427 (vn_phi_lookup): Populate vn_phi_s type.
21428 (vn_phi_insert): Likewise.
21430 2013-02-19 Jakub Jelinek <jakub@redhat.com>
21432 PR tree-optimization/56350
21433 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
21434 if haven't found reduction or nested cycle operand, rather than
21435 asserting we must find it.
21437 PR tree-optimization/56381
21438 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
21441 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
21442 Jakub Jelinek <jakub@redhat.com>
21445 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
21446 (swap_optab_enable): Same.
21447 (init_all_optabs): Use argument instead of global.
21448 * tree.h (struct tree_optimization_option): New field target_optabs.
21449 * expr.h (init_all_optabs): Add argument to prototype.
21450 (TREE_OPTIMIZATION_OPTABS): New.
21451 (save_optabs_if_changed): Protoize.
21452 * optabs.h: Declare this_fn_optabs.
21453 * optabs.c (save_optabs_if_changed): New.
21454 Declare this_fn_optabs.
21455 (init_optabs): Add argument to init_all_optabs() call.
21456 * function.c (invoke_set_current_function_hook): Handle per
21458 * function.h (struct function): New field optabs.
21459 * config/mips/mips.c (mips_set_mips16_mode): Handle when
21460 optimization_current_node has changed.
21461 * target-globals.h (save_target_globals_default_opts): Protoize.
21462 * target-globals.c (save_target_globals_default_opts): New.
21464 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21467 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
21468 registers %fr12 and %fr12R as call used.
21471 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
21472 and HImode, require all displacements to be an integer multiple of
21474 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
21475 only allow QImode and HImode when reload is in progress and strict is
21476 true. Likewise for symbolic addresses. Use base14_operand to check
21477 displacements in REG+BASE addresses.
21479 2013-02-18 Richard Biener <rguenther@suse.de>
21481 PR tree-optimization/56366
21482 * tree-vect-loop.c (get_initial_def_for_induction): Properly
21483 handle sign-conversion of outer-loop initial induction value.
21485 2013-02-18 Richard Biener <rguenther@suse.de>
21487 PR middle-end/56349
21488 * cfghooks.c (merge_blocks): If we merge a latch into another
21489 block adjust references to it.
21490 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
21491 (verify_loop_structure): Verify that a recorded latch is in fact
21494 2013-02-18 Richard Biener <rguenther@suse.de>
21496 PR tree-optimization/56321
21497 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
21498 order SSA name release and virtual operand unlinking.
21500 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21502 * config/microblaze/microblaze.md (save_stack_block): Define.
21503 (restore_stack_block): Likewise.
21505 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21507 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
21508 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
21509 * config/microblaze/microblaze.c (microblaze_option_override):
21510 Bail out early for PIC modes when target does not support PIC.
21512 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21514 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
21515 Replace with a microblaze version.
21516 (microblaze_trampoline_init): Adapt for microblaze.
21517 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
21520 2013-02-16 Jakub Jelinek <jakub@redhat.com>
21521 Dodji Seketeli <dodji@redhat.com>
21524 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
21525 (instrument_mem_region_access): Do not forget to always put
21526 instrumentation of the of 'base' and 'base + len' in a "if (len !=
21527 0) statement, even for cases where either 'base' or 'base + len'
21528 are not instrumented -- because they have been previously
21529 instrumented. Simplify the logic by putting all the statements
21530 instrument 'base + len' inside a sequence, and then insert that
21531 sequence right before the current insertion point. Then, to
21532 instrument 'base + len', just get an iterator on that statement.
21533 And do not forget to update the pointer to iterator the function
21534 received as argument.
21536 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
21538 PR rtl-optimization/56348
21539 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
21541 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
21543 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
21544 (clean_graph_dump_file): Pass base to start_graph_dump.
21546 2013-02-14 Richard Henderson <rth@redhat.com>
21549 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
21551 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
21553 * collect2-aix.h: Define F_LOADONLY.
21555 2013-02-14 Richard Biener <rguenther@suse.de>
21558 * varasm.c (output_constant_def_1): Get the decl representing
21559 the constant as argument.
21560 (output_constant_def): Wrap output_constant_def_1.
21561 (make_decl_rtl): Use output_constant_def_1 with the decl
21562 representing the constant.
21563 (build_constant_desc): Optionally re-use a decl already
21564 representing the constant.
21565 (tree_output_constant_def): Adjust.
21567 2013-02-14 Dodji Seketeli <dodji@redhat.com>
21570 * asan.c (instrument_builtin_call): Really put the length of the
21571 second source argument into src1_len.
21573 2013-02-13 Jakub Jelinek <jakub@redhat.com>
21575 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
21576 argument. If it is false, don't create edge from then_bb to
21578 (insert_if_then_before_iter): Pass true to it.
21579 (build_check_stmt): Pass false to it.
21580 (transform_statements): Flush hash table only on extended basic
21581 block boundaries, rather than at the beginning of every bb.
21582 Don't flush hash table on nonfreeing_call_p calls.
21583 * tree-flow.h (nonfreeing_call_p): New prototype.
21584 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
21586 2013-02-13 David S. Miller <davem@davemloft.net>
21588 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
21590 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
21593 * ira.c (max_regno_before_ira): Move from ...
21595 (fix_reg_equiv_init): Use max_regno_before_ira instead of
21598 2013-02-13 Jakub Jelinek <jakub@redhat.com>
21600 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
21602 2013-02-13 Richard Biener <rguenther@suse.de>
21605 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
21606 globals in MEM_REFs.
21608 2013-02-13 Richard Biener <rguenther@suse.de>
21610 * loop-init.c (loop_optimizer_init): Clear loop state when
21611 re-initializing preserved loops.
21612 * loop-unswitch.c (unswitch_single_loop): Return whether
21613 we unswitched the loop. Do not verify loop state here.
21614 (unswitch_loops): When we unswitched a loop discover new loops.
21616 2013-02-13 Kostya Serebryany <kcc@google.com>
21618 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
21620 * sanitizer.def: Rename __asan_init to __asan_init_v1.
21622 2013-02-12 Dodji Seketeli <dodji@redhat.com>
21624 Avoid instrumenting duplicated memory access in the same basic block
21625 * Makefile.in (asan.o): Add new dependency on hash-table.h
21626 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
21627 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
21628 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
21629 (free_mem_ref_resources, has_mem_ref_been_instrumented)
21630 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
21631 (get_mem_ref_of_assignment): New functions.
21632 (get_mem_refs_of_builtin_call): Extract from
21633 instrument_builtin_call and tweak a little bit to make it fit with
21635 (instrument_builtin_call): Use the new
21636 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
21637 of is_gimple_builtin_call.
21638 (instrument_derefs, instrument_mem_region_access): Insert the
21639 instrumented memory reference into the hash table.
21640 (maybe_instrument_assignment): Renamed instrument_assignment into
21641 this, and change it to advance the iterator when instrumentation
21642 actually happened and return true in that case. This makes it
21643 homogeneous with maybe_instrument_assignment, and thus give a
21644 chance to callers to be more 'regular'.
21645 (transform_statements): Clear the memory reference hash table
21646 whenever we enter a new BB, when we cross a function call, or when
21647 we are done transforming statements. Use
21648 maybe_instrument_assignment instead of instrumentation. No more
21649 need to special case maybe_instrument_assignment and advance the
21650 iterator after calling it; it's now handled just like
21651 maybe_instrument_call. Update comment.
21653 2013-02-13 Richard Biener <rguenther@suse.de>
21655 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
21656 Fix loop discovery code.
21658 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
21660 PR inline-asm/56148
21661 * lra-constraints.c (process_alt_operands): Match early clobber
21662 operand with itself. Check conflicts with earlyclobber only if
21663 the operand is not reloaded. Prefer to reload conflicting operand
21664 if earlyclobber and matching operands are the same.
21666 2013-02-12 Richard Biener <rguenther@suse.de>
21669 * lto-streamer-out.c (write_symbol): Do not output symbols
21670 for hard register variables.
21672 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
21675 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
21676 (umulsidi3_insn, mulsidi3_insn): New insns.
21678 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
21680 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
21681 (struct tune_params): Add vec_costs field.
21682 * config/arm/arm.c (arm_builtin_vectorization_cost)
21683 (arm_add_stmt_cost): New functions.
21684 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
21685 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
21686 (arm_default_vec_cost): New struct of type cpu_vec_costs.
21687 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
21688 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
21689 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
21690 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
21692 2013-02-12 Richard Biener <rguenther@suse.de>
21695 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
21696 decls again if possible.
21698 2013-02-12 Richard Biener <rguenther@suse.de>
21700 PR middle-end/56288
21701 * tree-ssa.c (verify_ssa_name): Fix check, move
21702 SSA_NAME_IN_FREE_LIST check up.
21704 2013-02-12 Jakub Jelinek <jakub@redhat.com>
21705 Steven Bosscher <steven@gcc.gnu.org>
21707 PR rtl-optimization/56151
21708 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
21709 equal to op0 or op1, and last_insn pattern is CODE operation
21710 with MEM dest and one of the operands matches that MEM.
21712 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
21714 * doc/extend.texi: Document Function Multiversioning and "default"
21715 parameter string to target attribute.
21716 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
21717 target attribute parameter is "default".
21718 (ix86_compare_version_priority): Remove checks for target attribute.
21719 (ix86_mangle_function_version_assembler_name): Change error to sorry.
21720 Remove check for target attribute equal to NULL. Add assert.
21721 (ix86_generate_version_dispatcher_body): Change error to sorry.
21723 2013-02-11 Iain Sandoe <iain@codesourcery.com>
21724 Jack Howarth <howarth@bromo.med.uc.edu>
21725 Patrick Marlier <patrick.marlier@gmail.com>
21728 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
21729 define ENDFILE_SPEC as TM_DESTRUCTOR.
21730 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
21732 2013-02-11 Alexander Potapenko <glider@google.com>
21733 Jack Howarth <howarth@bromo.med.uc.edu>
21734 Jakub Jelinek <jakub@redhat.com>
21737 * config/darwin.c (cdtor_record): Rename ctor_record.
21738 (sort_cdtor_records): Rename sort_ctor_records.
21739 (finalize_dtors): New routine to sort destructors by
21740 priority before use in assemble_integer.
21741 (machopic_asm_out_destructor): Use finalize_dtors if needed.
21743 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
21745 PR rtl-optimization/56275
21746 * simplify-rtx.c (avoid_constant_pool_reference): Check that
21747 offset is non-negative and less than cmode size before
21748 calling simplify_subreg.
21750 2013-02-11 Richard Biener <rguenther@suse.de>
21752 PR tree-optimization/56264
21753 * cfgloop.h (fix_loop_structure): Adjust prototype.
21754 * loop-init.c (fix_loop_structure): Return the number of
21755 newly discovered loops.
21756 * tree-cfgcleanup.c (repair_loop_structures): When new loops
21757 are discovered, do a full loop-closed SSA rewrite.
21759 2013-02-11 Richard Biener <rguenther@suse.de>
21761 PR tree-optimization/56273
21762 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
21764 (check_array_ref): Fix missing newline in dumps.
21765 (search_for_addr_array): Likewise.
21767 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
21769 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
21771 2013-02-09 Jakub Jelinek <jakub@redhat.com>
21774 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
21776 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
21778 PR rtl-optimization/56246
21779 * lra-constraints.c (simplify_operand_subreg): Try to reuse
21781 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
21782 constraints are satisfied.
21784 2013-02-08 Jeff Law <law@redhat.com>
21787 * emit-rtl.c (reg_is_parm_p): New function.
21788 * regs.h (reg_is_parm_p): New prototype.
21789 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
21790 callee-clobbered registers.
21792 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
21795 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
21796 If there is no implicit builtin declaration, just return NULL.
21798 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21800 * config/i386/sse.md (FMAMODEM): New mode iterator.
21801 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
21802 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
21804 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21806 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
21807 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
21808 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
21810 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21812 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
21813 (microblaze*-*-elf): Likewise.
21814 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
21816 * config/microblaze/microblaze-c.c: Add builtin defines for
21817 _LITTLE_ENDIAN and _BIG_ENDIAN.
21818 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
21819 add to TARGET_DEFAULT flags.
21820 Expand ASM_SPEC and LINK_SPEC.
21821 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
21822 * config/microblaze/microblaze.md: Update extendsidi2 and
21823 movdi_internal instructions to use low-order / high-order reg
21825 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
21826 options and inversemask / mask of LITTLE_ENDIAN.
21827 * config/microblaze/t-microblaze: Expand multilib options to
21828 include mlittle-endian (le) and update exceptions patterns.
21830 2013-02-08 Jakub Jelinek <jakub@redhat.com>
21832 PR rtl-optimization/56195
21833 * lra-constraints.c (get_reload_reg): Don't reuse regs
21834 if they have smaller mode than requested, if they have
21835 wider mode than requested, try to return a SUBREG.
21837 PR tree-optimization/56250
21838 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
21839 if type is unsigned and code isn't MULT_EXPR.
21841 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
21843 PR tree-optimization/56064
21844 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
21845 bits according to mode.
21846 * fixed-value.h (fixed_from_double_int)
21847 (const_fixed_from_double_int): Adjust comments.
21849 2013-02-08 Richard Biener <rguenther@suse.de>
21852 * lto-streamer.h (struct data_in): Remove current_file, current_line
21853 and current_col members.
21854 * lto-streamer-out.c (lto_output_location): Stream changed bits
21855 en-block for efficiency.
21856 * lto-streamer-in.c (clear_line_info): Remove.
21857 (lto_input_location): Cache current file, line and column
21858 globally via local statics. Read changed bits en-block.
21859 (input_function): Do not call clear_line_info.
21860 (lto_read_body): Likewise.
21861 (lto_input_toplevel_asms): Likewise.
21863 2013-02-08 Michael Matz <matz@suse.de>
21865 PR tree-optimization/52448
21866 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
21867 (nt_call_phase): New static.
21868 (add_or_mark_expr): Only mark accesses with newer phase than any
21870 (nonfreeing_call_p): New.
21871 (nt_init_block): Update nt_call_phase, mark blocks as visited.
21872 (nt_fini_block): Keep blocks marked as visited.
21873 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
21875 2013-02-08 Richard Biener <rguenther@suse.de>
21877 * ira.c (ira): Free broken dominator information.
21879 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21881 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
21883 2013-02-08 Marek Polacek <polacek@redhat.com>
21885 * cfgloop.c (verify_loop_structure): Add more checking of headers.
21887 2013-02-08 Richard Biener <rguenther@suse.de>
21889 PR middle-end/56181
21890 * cfgloop.h (flow_loops_find): Adjust.
21891 (bb_loop_header_p): Declare.
21892 * cfgloop.c (bb_loop_header_p): New function split out from ...
21893 (flow_loops_find): ... here. Adjust function signature,
21894 support incremental loop structure update.
21895 (verify_loop_structure): Cleanup. Verify a loop is a loop.
21896 * cfgloopmanip.c (fix_loop_structure): Move ...
21897 * loop-init.c (fix_loop_structure): ... here.
21898 (apply_loop_flags): Split out from ...
21899 (loop_optimizer_init): ... here.
21900 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
21901 in incremental mode, only remove dead loops here.
21903 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
21906 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
21907 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
21908 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
21909 (*round<mode>3.libgcc): New insns for fixed-modes.
21910 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
21911 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
21912 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
21913 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
21914 implementations. Define to __builtin_avr_absFX,
21915 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
21916 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
21917 __builtin_avr_countlsFX, respectively.
21918 * config/avr/avr-c.c (target.h): Include it.
21919 (enum avr_builtin_id): New enum.
21920 (avr_resolve_overloaded_builtin): New static function.
21921 (avr_register_target_pragmas): Use it to set
21922 targetm.resolve_overloaded_builtin.
21923 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
21924 tree nodes used by DEF_BUILTIN.
21925 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
21926 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
21927 <AVR_BUILTIN_xxBITS>: Same.
21929 2013-02-08 Richard Biener <rguenther@suse.de>
21931 * cfgloop.c (verify_loop_structure): Properly handle
21932 a loop exiting to another loop header.
21933 * ira-int.h (ira_loops): Remove.
21934 * ira.c (ira_loops): Remove.
21935 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
21936 (do_reload): Use loop_optimizer_finalize.
21937 * ira-build.c (create_loop_tree_nodes): Use get_loops and
21938 number_of_loops to access the loop tree.
21939 (more_one_region_p): Likewise.
21940 (finish_loop_tree_nodes): Likewise.
21941 (rebuild_regno_allocno_maps): Likewise.
21942 (mark_loops_for_removal): Likewise.
21943 (mark_all_loops_for_removal): Likewise.
21944 (remove_unnecessary_regions): Likewise.
21945 (ira_build): Likewise.
21946 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21948 2013-02-08 Richard Biener <rguenther@suse.de>
21950 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
21951 * ipa-pure-const.c (analyze_function): Avoid calling
21952 mark_irreducible_loops twice.
21953 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
21955 2013-02-07 David S. Miller <davem@davemloft.net>
21957 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
21959 * var-tracking.c (vt_add_function_parameter): Test the presence of
21960 HAVE_window_save properly and do not remap argument registers when
21961 we have a leaf function.
21963 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
21966 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
21968 * config/i386/i386.c (ix86_print_operand): Ditto.
21970 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21972 * lra-constraints.c (process_alt_operands): Fix recently added comment.
21974 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21976 PR rtl-optimization/56225
21977 * lra-constraints.c (process_alt_operands): Check that reload hard
21978 reg can hold value for strict_low_part.
21980 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21983 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
21984 dwarf2out_end_function.
21985 (in_first_function_p, maybe_at_text_label_p,
21986 first_loclabel_num_not_at_text_label): New variables.
21987 (dwarf2out_var_location): In the first function find out
21988 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
21989 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
21992 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
21994 PR rtl-optimization/56178
21995 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
21996 SUBREG of a register. Tidy up related block of code.
21997 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
21998 note if the source is a register or a SUBREG of a register.
22000 2013-02-07 Jakub Jelinek <jakub@redhat.com>
22003 * config/rs6000/rs6000.md (ptrm): New mode attr.
22004 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
22005 call_value_indirect_aix<pttrsize>,
22006 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
22009 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
22011 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
22012 if -bnortl. Convert to strcmp and strncmp.
22014 2013-02-07 Alan Modra <amodra@gmail.com>
22017 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
22018 addresses won't wrap when offsetting.
22019 (rs6000_secondary_reload): Provide secondary reloads needed for
22020 wrapping LO_SUM addresses.
22022 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
22024 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
22025 MACH, just __MACH__.
22027 2013-02-06 Richard Biener <rguenther@suse.de>
22029 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
22030 instead of calling fix_loop_structure.
22032 2013-02-06 Jakub Jelinek <jakub@redhat.com>
22034 PR middle-end/56217
22035 * omp-low.c (use_pointer_for_field): Return false if
22036 lower_send_shared_vars doesn't generate any copy-out code.
22038 2013-02-06 Tom de Vries <tom@codesourcery.com>
22040 PR rtl-optimization/56131
22041 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
22042 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
22043 of the label is NULL. Add comment.
22045 2013-02-05 Jakub Jelinek <jakub@redhat.com>
22047 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
22050 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
22051 (STATIC_LIBTSAN_LIBS): Likewise.
22052 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
22053 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
22054 is defined, don't add anything else beyond that.
22055 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
22056 (LINK_COMMAND_SPEC): Use them.
22058 PR tree-optimization/56205
22059 * tree-stdarg.c (check_all_va_list_escapes): Return true if
22060 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
22061 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
22063 2013-02-05 Richard Biener <rguenther@suse.de>
22065 PR tree-optimization/53342
22066 PR tree-optimization/53185
22067 * tree-vectorizer.h (vect_check_strided_load): Remove.
22068 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
22069 not disallow peeling for vectorized strided loads.
22070 (vect_check_strided_load): Make static and simplify.
22071 (vect_analyze_data_refs): Adjust.
22072 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
22073 correctly when vectorizing strided loads.
22075 2013-02-05 Richard Biener <rguenther@suse.de>
22077 * doc/install.texi: Refer to ISL, not PPL.
22079 2013-02-05 Jan Hubicka <jh@suse.cz>
22081 PR tree-optimization/55789
22082 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
22084 2013-02-05 Jan Hubicka <jh@suse.cz>
22086 PR tree-optimization/55789
22087 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
22088 the dead call anyway.
22090 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
22093 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
22095 2013-02-04 Alexander Potapenko <glider@google.com>
22096 Jack Howarth <howarth@bromo.med.uc.edu>
22097 Jakub Jelinek <jakub@redhat.com>
22100 * config/darwin.c (sort_ctor_records): Stabilized qsort
22101 on constructor priority by using original position.
22102 (finalize_ctors): New routine to sort constructors by
22103 priority before use in assemble_integer.
22104 (machopic_asm_out_constructor): Use finalize_ctors if needed.
22106 2013-02-04 Jakub Jelinek <jakub@redhat.com>
22109 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
22110 about visibility on artificial decls.
22111 * config/sol2.c (solaris_assemble_visibility): Likewise.
22113 2013-02-04 Kai Tietz <ktietz@redhat.com>
22116 * config/i386/i386.c (function_value_ms_64): Add additional valtype
22117 argument and improve checking of return-argument types for 16-byte
22119 (ix86_function_value_1): Add additional valtype argument on call
22120 of function_value_64.
22121 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
22122 handling infunction_value_64 function.
22124 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
22126 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
22128 2013-02-04 Richard Biener <rguenther@suse.de>
22130 PR tree-optimization/56188
22131 * tree-ssa-structalias.c (label_visit): Consider case with
22132 initially non-empty points-to set.
22133 (perform_var_substitution): Dump node mapping and clean up.
22135 2013-02-04 Richard Guenther <rguenther@suse.de>
22138 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
22139 node prevail as last resort.
22140 (lto_symtab_merge_decls): Remove guard on LTRANS here.
22141 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
22143 2013-02-04 Richard Biener <rguenther@suse.de>
22145 PR tree-optimization/56113
22146 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
22148 (equiv_class_lookup_or_add): ... this.
22149 (label_visit): Adjust and fix error in previous patch.
22150 (perform_var_substitution): Adjust.
22152 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
22154 * config/sh/divtab.c: Fix formatting and comments throughout the file.
22155 * config/sh/sh4-300.md: Likewise.
22156 * config/sh/sh4a.md: Likewise.
22157 * config/sh/constraints.md: Likewise.
22158 * config/sh/sh.md: Likewise.
22159 * config/sh/netbsd-elf.h: Likewise.
22160 * config/sh/predicates.md: Likewise.
22161 * config/sh/sh-protos.h: Likewise.
22162 * config/sh/ushmedia.h: Likewise.
22163 * config/sh/linux.h: Likewise.
22164 * config/sh/sh.c: Likewise.
22165 * config/sh/superh.h: Likewise.
22166 * config/sh/elf.h: Likewise.
22167 * config/sh/sh4.md: Likewise.
22168 * config/sh/sh.h: Likewise.
22170 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22172 * config/pa/constraints.md: Adjust unused letters. Change "T"
22173 constraint to match_test floating_point_store_memory_operand().
22174 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
22175 (base14_operand): New.
22176 (floating_point_store_memory_operand): New.
22177 (integer_store_memory_operand): Revise to use base14_operand and
22178 reg_plus_base_memory_operand.
22179 (move_dest_operand): Allow symbolic_memory_operands.
22180 (symbolic_memory_operand): Check for LO_SOM.
22181 (symbolic_operand): Change default case to break.
22182 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
22183 CONST_DOUBLE values to be reloaded by putting them into memory when
22184 the destination is a floating point register.
22185 (movdf): Remove code to handle CONST_DOUBLE.
22187 (reload_indf_r1): New.
22188 (reload_insf_r1): New.
22189 Consistently use "Q" and "T" constraints with integer and floating
22190 point move instructions, respectively.
22191 (movdi): Remove FAIL.
22192 Change predicate for source operand unamed DImode move from
22193 general_operand to move_src_operand.
22194 (umulsidi3): Change predicate for destination operand to
22196 Likewise for similar unamed patterns.
22197 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
22198 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
22199 (hppa_legitimize_address): Simplify mask calculation.
22200 (pa_emit_move_sequence): Revised handling of secondary reloads from
22201 REG+D addresses for floating point loads and stores. Directly handle
22202 loading CONST0_RTX (mode) to a floating point register.
22203 (pa_secondary_reload): Handle reloading DF and SFmode constant values
22204 to floating point registers. Don't restrict secondary reloads to
22205 floating point registers to integer modes. Revise some comments and
22207 (TARGET_LEGITIMATE_ADDRESS_P): Define.
22208 (pa_legitimate_address_p): New.
22209 (pa_legitimize_reload_address): New.
22210 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
22211 (STRICT_REG_OK_FOR_BASE_P): New.
22212 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
22213 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
22215 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
22216 Andrew Dixie <andrewd@gentrack.com>
22218 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
22221 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
22223 * expmed.c (extract_bit_field_1): Pass the full width of the
22224 structure to get_best_reg_extraction_insn.
22226 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
22229 * configure.ac (use_cxa_atexit): Add AIX.
22230 * configure: Regenerate.
22232 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
22234 2013-02-01 Jakub Jelinek <jakub@redhat.com>
22237 * final.c (need_profile_function): New variable.
22238 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
22239 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
22240 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
22241 notes, targetm.asm_out.function_prologue doesn't emit anything,
22242 HAVE_prologue and profiler should be emitted before prologue,
22243 set need_profile_function instead of emitting it.
22244 (final_scan_insn): If need_profile_function, emit
22245 profile_function on the first NOTE_INSN_BASIC_BLOCK or
22246 NOTE_INSN_FUNCTION_BEG note.
22248 2013-02-01 Richard Henderson <rth@redhat.com>
22250 * config/rs6000/rs6000.md (smulditi3): New.
22253 * config/alpha/alpha.md (umulditi3): New.
22255 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
22257 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
22258 (ASM_OUTPUT_ALIGNED_LOCAL): New.
22260 2013-02-01 Richard Biener <rguenther@suse.de>
22262 PR tree-optimization/56113
22263 * tree-ssa-structalias.c (label_visit): Reduce work for
22264 single-predecessor nodes.
22266 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
22268 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
22269 range isn't testing for zero.
22271 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
22273 PR middle-end/56113
22274 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
22276 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
22277 Nick Clifton <nickc@redhat.com>
22279 * config/v850/constraints.md (Q): Define as a memory constraint.
22280 * config/v850/predicates.md (label_ref_operand): New predicate.
22281 (e3v5_shift_operand): New predicate.
22282 (ior_operator): New predicate.
22283 * config/v850/t-v850: Add e3v5 multilib.
22284 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
22285 (v850_gen_movdi): Prototype.
22286 * config/v850/v850.c: Add support for e3v5 architecture.
22287 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
22289 (construct_save_jarl): Add e3v5 long JARL support.
22290 (v850_adjust_insn_length): New function. Adjust length of call
22291 insns when using e3v5 instructions.
22292 (v850_gen_movdi): New function: Generate instructions to move a
22294 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
22295 (CPP_SPEC): Define __v850e3v5__ as appropriate.
22296 (TARGET_USE_FPU): Enable for e3v5.
22297 (CONST_OK_FOR_W): New macro.
22298 (ADJUST_INSN_LENGTH): Define.
22299 * config/v850/v850.md (UNSPEC_LOOP): Define.
22300 (attr cpu): Add v850e3v5.
22301 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
22302 (movdi): New pattern.
22303 (movdi_internal): New pattern.
22304 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
22305 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
22306 (cstoresf4): Likewise.
22307 (cstoredf4): Likewise.
22308 (insv): New pattern.
22309 (rotlso3_a): New pattern.
22310 (rotlsi3_b): New pattern
22311 (rotlsi3_v850e3v5): New pattern.
22312 (doloop_begin): New pattern.
22313 (fix_loop_counter): New pattern.
22314 (doloop_end): New pattern.
22315 (branch_normal): Add e3v5 long branch support.
22316 (branch_invert): Likewise.
22317 (branch_z_normal): Likewise.
22318 (branch_z_invert): Likewise.
22319 (branch_nz_normal): Likewise.
22320 (branch_nz_invert): Likewise.
22321 (call_internal_short): Add e3v5 register-indirect JARL support.
22322 (call_internal_long): Likewise.
22323 (call_value_internal_short): Likewise.
22324 (call_value_internal_long): Likewise.
22325 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
22326 (mloop): New option.
22327 * config.gcc: Add support for configuring v840e3v5 target.
22328 * doc/invoke.texi: Document new v850 specific command line options.
22330 2013-01-31 Paul Koning <ni1d@arrl.net>
22334 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
22335 children if parent is a class.
22336 (prune_unused_types_prune): Don't add DW_AT_declaration.
22338 2013-01-31 Richard Biener <rguenther@suse.de>
22340 PR tree-optimization/56157
22341 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
22342 match up operand with SLP child.
22344 2013-01-31 Jason Merrill <jason@redhat.com>
22347 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
22348 parameters the first time.
22349 (gen_scheduled_generic_parms_dies): Check completeness here.
22351 2013-01-31 Richard Biener <rguenther@suse.de>
22353 PR middle-end/53073
22354 * common.opt (faggressive-loop-optimizations): New flag,
22355 enabled by default.
22356 * doc/invoke.texi (faggressive-loop-optimizations): Document.
22357 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
22358 infer_loop_bounds_from_undefined by it.
22360 2013-01-31 Richard Biener <rguenther@suse.de>
22362 PR tree-optimization/56150
22363 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
22364 visit virtual operands.
22365 (find_uses_to_rename_bb): Likewise.
22367 2013-01-31 Richard Biener <rguenther@suse.de>
22369 PR tree-optimization/56150
22370 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
22371 mixed store non-store stmts.
22373 2013-01-30 Jakub Jelinek <jakub@redhat.com>
22376 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
22377 LIBASAN_EARLY_SPEC is defined.
22378 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
22379 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
22381 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
22384 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
22385 invalid args instead of ICEing on it.
22386 (ix86_valid_target_attribute_tree): Return error_mark_node if
22387 ix86_valid_target_attribute_inner_p failed.
22388 (ix86_valid_target_attribute_p): Return false only if
22389 ix86_valid_target_attribute_tree returned error_mark_node. Allow
22390 target("default") attribute.
22391 (sorted_attr_string): Change argument from const char * to tree,
22392 merge in all target attribute arguments rather than just one.
22393 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
22394 instead of free. Avoid using strcat.
22395 (ix86_mangle_function_version_assembler_name): Mangle
22396 target("default") as if no target attribute is present. Adjust
22397 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
22398 instead of xmalloc and XDELETEVEC instead of free.
22399 (ix86_function_versions): Don't return true if one of the decls
22400 doesn't have target attribute. If they don't and one of the decls
22401 is DECL_FUNCTION_VERSIONED, report an error. Adjust
22402 sorted_attr_string caller. Use XDELETEVEC instead of free.
22403 (ix86_supports_function_versions): Remove.
22404 (make_name): Fix up formatting.
22405 (make_dispatcher_decl): Remove resolver_name and its initialization.
22406 Avoid leaking memory.
22407 (is_function_default_version): Return true if there is
22408 target("default") attribute rather than no target attribute at all.
22409 (make_resolver_func): Avoid leaking memory.
22410 (ix86_generate_version_dispatcher_body): Likewise.
22411 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
22412 * target.def (supports_function_versions): Remove.
22413 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
22414 * doc/tm.texi: Regenerated.
22416 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
22418 PR rtl-optimization/56144
22419 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
22420 for values with side effects.
22422 2013-01-30 Richard Biener <rguenther@suse.de>
22424 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
22425 (sparseset_pop): Likewise.
22426 * cfganal.c (compute_idf): Likewise. Increase work-stack size
22427 to be able to use quick_push in the worker loop.
22429 2013-01-30 Marek Polacek <polacek@redhat.com>
22431 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
22433 2013-01-30 Richard Biener <rguenther@suse.de>
22436 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
22438 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
22440 PR tree-optimization/56064
22441 * fixed-value.c (fixed_from_double_int): New function.
22442 * fixed-value.h (fixed_from_double_int): New prototype.
22443 (const_fixed_from_double_int): New static inline function.
22444 * fold-const.c (native_interpret_fixed): New static function.
22445 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
22446 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
22447 (native_encode_fixed): New static function.
22448 (native_encode_expr) <FIXED_CST>: Use it.
22449 (native_interpret_int): Move double_int worker code to...
22450 * double-int.c (double_int::from_buffer): ...this new static method.
22451 * double-int.h (double_int::from_buffer): Prototype it.
22453 2013-01-30 Richard Biener <rguenther@suse.de>
22455 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
22456 New pointer-map and obstack.
22457 (init_alias_vars): Allocate pointer-map and obstack.
22458 (delete_points_to_sets): Free them.
22459 (find_what_var_points_to): Cache result.
22460 (find_what_p_points_to): Adjust for changed interface of
22461 find_what_var_points_to.
22462 (compute_points_to_sets): Likewise.
22463 (ipa_pta_execute): Likewise.
22465 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22467 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
22468 * configure: Regenerate.
22469 * config.in: Regenerate.
22470 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
22471 #nobits/#progbits if supported.
22473 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
22476 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
22477 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
22478 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
22480 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22482 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
22483 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
22485 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22487 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
22488 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
22490 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22492 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
22494 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
22495 * config/arm/cortex-a7.md: New bypasses using
22496 arm_mac_accumulator_is_result.
22498 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22500 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
22501 (cortex_a7_neon_mla): Likewise.
22502 (cortex_a7_fpfmad): New reservation.
22503 (cortex_a7_fpmacs): Use ffmas and update required units.
22504 (cortex_a7_fpmuld): Update required units and latency.
22505 (cortex_a7_fpmacd): Likewise.
22506 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
22507 (cortex_a7_neon). Likewise.
22508 (bypass) Update participating units.
22510 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
22512 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
22513 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
22515 * config/arm/vfp11.md (vfp_farith): Use ffmas.
22516 (vfp_fmul): Use ffmad.
22517 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
22518 (cortex_r4_fmacd): Use ffmad.
22519 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
22520 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
22521 (cortex_a9_fmacd): Use ffmad.
22522 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
22523 (cortex_a8_vfp_macd): Use ffmad.
22524 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
22525 (cortex_a5_fpmacd): Use ffmad.
22526 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
22527 (cortex_a15_vfp_macd): Use ffmad.
22528 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
22530 2013-01-29 Jason Merrill <jason@redhat.com>
22533 * varasm.c (default_assemble_visibility): Don't warn about
22534 visibility on artificial decls.
22536 2013-01-29 Richard Biener <rguenther@suse.de>
22538 PR tree-optimization/56113
22539 * tree-ssa-structalias.c (equiv_class_lookup): Also return
22541 (label_visit): Free duplicate bitmaps and record the leader instead.
22542 (perform_var_substitution): Adjust.
22544 2013-01-29 Richard Biener <rguenther@suse.de>
22546 PR tree-optimization/55270
22547 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
22548 the CFG, schedule loops for fixup.
22550 2013-01-29 Nick Clifton <nickc@redhat.com>
22552 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
22555 2013-01-28 Leif Ekblad <leif@rdos.net>
22557 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
22558 * config/i386/i386.h (TARGET_RDOS): New macro.
22559 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
22560 * config/i386/i386.c (ix86_option_override_internal): For 64bit
22561 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
22562 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
22563 DEFAULT_LARGE_SECTION_THRESHOLD.
22564 * config/i386/i386.md (R14_REG, R15_REG): New constants.
22565 * config/i386/rdos.h: New file.
22566 * config/i386/rdos64.h: New file.
22568 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
22571 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
22574 2013-01-28 Jakub Jelinek <jakub@redhat.com>
22576 PR rtl-optimization/56117
22577 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
22578 call cselib_lookup_from_insn on the MEM before calling
22579 add_insn_mem_dependence.
22581 2013-01-28 Richard Biener <rguenther@suse.de>
22583 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
22584 to a stmt that didn't have one.
22585 (copy_phis_for_bb): Likewise for PHI arguments.
22586 (copy_debug_stmt): Likewise for debug stmts.
22588 2013-01-28 Richard Biener <rguenther@suse.de>
22590 PR tree-optimization/56034
22591 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
22592 (partition_builtin_p): Adjust.
22593 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
22594 it is the last partition.
22595 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
22596 up the vertex for the definition.
22597 (classify_partition): Classify whether a partition is a
22598 PKIND_REDUCTION, thus has uses outside of the loop.
22599 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
22600 Merge all PKIND_REDUCTION partitions into the last partition.
22601 (tree_loop_distribution): Seed partitions from reductions as well.
22603 2013-01-28 Jakub Jelinek <jakub@redhat.com>
22605 PR tree-optimization/56125
22606 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
22607 pow(x,c) into sqrt(x) * powi(x, n/2) or
22608 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
22609 optimizing for size.
22610 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
22611 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
22614 PR tree-optimization/56094
22615 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
22616 to UNKNOWN_LOCATION while gimplifying expr.
22618 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
22621 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
22622 operand 0 in movabs insn template for -masm=intel asm alternative.
22623 (*movabs<mode>_2): Ditto for operand 1.
22625 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
22628 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
22631 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22633 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
22636 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22638 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
22639 Undef to avoid warning.
22641 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
22643 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
22644 * configure: Regenerate.
22646 2013-01-25 Jakub Jelinek <jakub@redhat.com>
22648 PR tree-optimization/56098
22649 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
22650 for stmts with volatile ops.
22651 (cond_store_replacement): Don't optimize if assign has volatile ops.
22652 (cond_if_else_store_replacement_1): Don't optimize if either
22653 then_assign or else_assign have volatile ops.
22654 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
22657 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22659 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
22661 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22663 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
22664 missing ':' in asm example.
22666 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
22668 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
22669 entries into lane and laneq entries.
22670 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
22671 Remove AdvSIMD scalar modes.
22672 (aarch64_sq<r>dmulh_laneq<mode>): New.
22673 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
22675 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
22676 builtin implementations to relfect changes in RTL in aarch64-simd.md.
22677 * config/aarch64/iterators.md (VCOND): New.
22680 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22683 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
22684 Add NULL LIBNAME argument to existing definitions.
22685 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
22686 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
22687 * config/avr/avr.c (DEF_BUILTIN): Same.
22688 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
22689 (avr_expand_builtin): Expand to a vanilla call if a libgcc
22690 implementation is available (DECL_ASSEMBLER_NAME is set).
22691 (avr_fold_absfx): New static function.
22692 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
22693 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
22694 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
22695 AVR_BUILTIN_ABSLLK.
22696 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
22697 (abshk, absk, abslk, absllk): Provide as static inline functions.
22699 2013-01-25 Marek Polacek <polacek@redhat.com>
22701 PR tree-optimization/56035
22702 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
22704 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
22706 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
22707 (*movtf_internal_rex64): Add (!o,C) alternative
22708 (*movxf_internal_rex64): Ditto.
22709 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
22711 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
22713 * doc/invoke.texi: fix typo.
22714 * doc/objc.texi: fix typo.
22716 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
22718 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
22719 for the first two alternatives.
22721 2013-01-24 Diego Novillo <dnovillo@google.com>
22723 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
22724 (ggc-zone.o): Remove.
22725 * configure.ac: Remove option --with-gc.
22726 * configure: Re-generate.
22727 * doc/install.texi: Remove documentation for --with-gc.
22728 * gengtype.c (write_enum_defn): Remove. Update all users.
22729 (write_Types_process_field): Remove generation of gt_e_* argument.
22730 (output_type_enum): Remove. Update all users.
22731 (write_enum_defn): Remove. Update all users.
22732 (enum alloc_zone): Remove. Update all users.
22733 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
22734 * ggc-common.c (ggc_splay_alloc): Remove first argument.
22735 Update all callers.
22736 (struct ptr_data): Remove field TYPE. Update all users.
22737 (gt_pch_note_object): Remove argument TYPE. Update all users.
22738 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
22740 * ggc-none.c (ggc_alloc_typed_stat): Remove.
22741 (struct alloc_zone): Remove.
22742 (ggc_internal_alloc_zone_stat): Remove.
22743 (ggc_internal_cleared_alloc_zone_stat): Remove.
22744 * ggc-page.c (ggc_alloc_typed_stat): Remove.
22745 (ggc_pch_count_object): Remove last argument. Update all users.
22746 (ggc_pch_alloc_object): Remove last argument. Update all users.
22747 (struct alloc_zone): Remove.
22748 * ggc-zone.c: Remove.
22749 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
22750 (struct alloc_zone): Remove.
22751 (ggc_alloc_typed_stat): Remove.
22752 (ggc_alloc_typed): Remove.
22753 (ggc_splay_alloc): Remove first argument.
22754 (rtl_zone): Remove. Update all users.
22755 (tree_zone): Remove. Update all users.
22756 (tree_id_zone): Remove. Update all users.
22757 (ggc_internal_zone_alloc_stat): Remove. Update all users.
22758 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
22759 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
22760 * tree-ssanames.c: Remove references to zone allocator in comments.
22762 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
22764 * config/avr/avr.c (avr_out_fract): Make register numbers that
22765 might be outside of source operand signed.
22767 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
22769 * config/i386/constraints.md (Yf): New constraint.
22770 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
22771 of f constraint to conditionaly disable x87 register preferences.
22772 (*movdf_internal): Ditto.
22773 (*movsf_internal): Ditto.
22775 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
22777 PR inline-asm/55934
22778 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
22779 that have operands with impossible constraints.
22780 Add a FIXME for a speed-up opportunity.
22781 * lra-constraints.c (process_alt_operands): Verify that a class
22782 selected from constraints on asms is valid for the operand mode.
22783 (curr_insn_transform): Remove incorrect comment.
22785 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
22787 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
22788 TOC operand is a valid symbol ref in the constant pool.
22790 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22792 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
22794 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
22797 * config/avr/stdfix.h: New file.
22798 * t-avr (stdfix-gcc.h): New rule to build it.
22799 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
22801 2013-01-23 Kostya Serebryany <kcc@google.com>
22803 * config/darwin.h: remove dependency on
22804 CoreFoundation (asan on Mac OS).
22806 2013-01-23 Jakub Jelinek <jakub@redhat.com>
22809 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
22810 instead of cmpdi_operand for first comparison operand.
22811 Don't assert that comparison operands aren't both constants.
22813 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
22815 * doc/install.texi (Downloading the Source): Update references to
22816 downloading separate components.
22818 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
22820 * doc/extend.texi (__int128): Improve grammar.
22822 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
22825 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
22826 alternative to (o,r).
22827 (*movdi_internal_rex64): Remove (!o,n) alternative.
22828 (DImode immediate->memory splitter): Remove.
22829 (DImode immediate->memory peephole2): Remove.
22830 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
22831 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
22832 alternative to (!o,*r).
22833 (*movtf_internal_sse): New pattern.
22834 (*movxf_internal_rex64): New pattern.
22835 (*movxf_internal): Disable for TARGET_64BIT.
22836 (*movdf_internal_rex64): Remove (!o,F) alternative.
22838 2013-01-22 Jakub Jelinek <jakub@redhat.com>
22840 PR middle-end/56074
22841 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
22842 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
22843 * tree-vect-loop-manip.c (find_loop_location): Also ignore
22844 stmt locations where LOCATION_LOCUS of the stmt location is
22845 UNKNOWN_LOCATION or BUILTINS_LOCATION.
22848 * config/i386/i386.md (UNSPEC_STOS): New.
22849 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
22850 *strsetqi_1): Add UNSPEC_STOS.
22852 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
22855 * doc/invoke.texi: Remove left over -Wsynth example.
22857 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22859 PR tree-optimization/56051
22860 * fold-const.c (fold_binary_loc): Don't fold
22861 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
22862 a narrowing conversion, or widening conversion from signed
22865 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22867 PR rtl-optimization/56023
22868 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
22869 dependent on debug instruction.
22871 2013-01-21 Martin Jambor <mjambor@suse.cz>
22873 PR middle-end/56022
22874 * function.c (allocate_struct_function): Call
22875 invoke_set_current_function_hook earlier.
22877 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22879 * reload1.c (init_reload): Only initialize reload_obstack
22880 during the first call.
22882 2013-01-21 Marek Polacek <polacek@redhat.com>
22884 * cfgloop.c (verify_loop_structure): Fix up grammar.
22886 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
22888 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
22889 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
22891 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22894 * config/arm/marvell-pj4.md: Update copyright year.
22895 Fix up use of alu to alu_reg and simple_alu_imm.
22897 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22899 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
22901 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
22904 * lra-constraints.c (curr_insn_transform): Don't reuse original
22905 insn for secondary memory move when memory mode should be different.
22907 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22909 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
22910 atomic_storedi_1): New patterns.
22912 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
22914 btver2 pipeline descriptions.
22915 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
22917 * config/i386/i386.md (btver2_decode): New type attributes.
22918 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
22920 * config/i386/btver2.md: New file describing btver2 pipelines.
22922 2013-01-19 Andrew Pinski <apinski@cavium.com>
22924 PR tree-optimization/52631
22925 * tree-ssa-sccvn (visit_use): Before looking up the original
22926 statement, try looking up the simplified expression.
22928 2013-01-19 Anthony Green <green@moxielogic.com>
22930 * config/moxie/moxie.c (moxie_expand_prologue): Set
22931 current_function_static_stack_size.
22933 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22935 PR tree-optimization/56029
22936 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
22937 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
22939 2013-01-18 Sharad Singhai <singhai@google.com>
22941 PR tree-optimization/55995
22942 * dumpfile.c (dump_loc): Print location only if available.
22943 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
22945 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
22948 * lra-constraints.c (curr_insn_transform): Reuse original insn for
22949 secondary memory move.
22950 (inherit_reload_reg): Use rclass instead of cl for
22951 check_secondary_memory_needed_p.
22953 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22955 PR middle-end/56015
22956 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
22957 the case where writing real complex part of target modifies op1.
22959 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
22961 * config/aarch64/aarch64-simd.md
22962 (aarch64_vcond_internal<mode>): Handle unordered cases.
22963 * config/aarch64/iterators.md (v_cmp_result): New.
22965 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
22966 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22968 * config/arm/marvell-pj4.md: New file.
22969 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
22970 * config/arm/arm.md (generic_sched): Add marvell_pj4.
22971 (generic_vfp): Likewise.
22972 * config/arm/arm-cores.def: Add marvell-pj4.
22973 * config/arm/arm-tune.md: Regenerate.
22974 * config/arm/arm-tables.opt: Regenerate.
22975 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
22976 * doc/invoke.texi: Document marvell-pj4.
22978 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
22980 * config/aarch64/arm_neon.h: Map scalar types to standard types.
22982 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22987 * var-tracking.c (negative_power_of_two_p): New.
22988 (global_get_addr_cache, local_get_addr_cache): New.
22989 (get_addr_from_global_cache, get_addr_from_local_cache): New.
22990 (vt_canonicalize_addr): Rewrite using the above. Adjust the
22992 (vt_stack_offset_p): Remove.
22993 (vt_canon_true_dep): Always canonicalize loc's address.
22994 (clobber_overlapping_mems): Make sure we have a MEM.
22995 (local_get_addr_clear_given_value): New.
22996 (val_reset): Clear local cached entries.
22997 (compute_bb_dataflow): Create and release the local cache.
22998 Disable duplicate MEMs clobbering.
22999 (emit_notes_in_bb): Clobber MEMs likewise.
23000 (vt_emit_notes): Create and release the local cache.
23001 (vt_initialize, vt_finalize): Create and release the global
23002 cache, respectively.
23003 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
23005 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
23007 PR libmudflap/53359
23008 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
23009 not found in the symtab.
23011 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
23014 PR rtl-optimization/55547
23015 PR rtl-optimization/53827
23018 * alias.c (offset_overlap_p): New, factored out of...
23019 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
23020 the conservative special case for symbolic constants. Don't
23021 adjust zero sizes on alignment.
23023 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
23025 PR rtl-optimization/52573
23026 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
23027 REG_UNUSED for the same register.
23029 2013-01-17 Richard Biener <rguenther@suse.de>
23030 Marek Polacek <polacek@redhat.com>
23032 PR rtl-optimization/55833
23033 * loop-unswitch.c (unswitch_loops): Move loop verification...
23034 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
23035 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
23036 Set it to true when we're removing a loop from hierarchy tree in
23037 an irreducible region.
23038 (fix_bb_placements): Adjust caller.
23039 (fix_loop_placements): Likewise.
23041 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
23043 * config/avr/builtins.def (DEF_BUILTIN): Factor out
23044 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
23045 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
23046 Remove ID. Adjust comments.
23047 * config/avr/avr-c.c (avr_builtin_name): Remove.
23048 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
23049 * config/avr/avr.c (avr_tolower): New static function.
23050 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
23051 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
23052 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
23055 2013-01-17 Jan Hubicka <jh@suse.cz>
23057 PR tree-optimization/55273
23058 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
23060 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
23063 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
23064 store through atomic_store<mode>_1.
23065 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
23067 2013-01-17 Martin Jambor <mjambor@suse.cz>
23069 PR tree-optimizations/55264
23070 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
23071 for virtual methods.
23072 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
23073 virtual methods before inlining is over.
23074 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
23076 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
23079 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
23081 PR rtl-optimization/56005
23082 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
23083 pending reads for prefetch.
23085 2013-01-16 Ian Bolton <ian.bolton@arm.com>
23087 * config/aarch64/aarch64.md
23088 (*cstoresi_neg_uxtw): New pattern.
23089 (*cmovsi_insn_uxtw): New pattern.
23090 (*<optab>si3_uxtw): New pattern.
23091 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
23092 (*<optab>si3_insn_uxtw): New pattern.
23093 (*bswapsi2_uxtw): New pattern.
23095 2013-01-16 Richard Biener <rguenther@suse.de>
23097 * tree-inline.c (tree_function_versioning): Remove set but
23098 never used variable.
23100 2013-01-16 Richard Biener <rguenther@suse.de>
23102 PR tree-optimization/55964
23103 * tree-flow.h (rename_variables_in_loop): Remove.
23104 (rename_variables_in_bb): Likewise.
23105 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
23106 (copy_loop_before): Adjust and delete update-ssa status.
23107 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
23108 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
23109 (rename_variables_in_loop): Remove.
23110 (slpeel_update_phis_for_duplicate_loop): Likewise.
23111 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
23112 use available cfg machinery instead of duplicating it.
23113 Update PHI nodes and perform poor-mans SSA update here.
23114 (slpeel_tree_peel_loop_to_edge): Adjust.
23116 2013-01-16 Richard Biener <rguenther@suse.de>
23118 PR tree-optimization/54767
23119 PR tree-optimization/53465
23120 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
23121 (vrp_visit_phi_node): For PHI arguments coming via backedges
23122 drop all symbolical range information.
23123 (execute_vrp): Compute backedges.
23125 2013-01-16 Richard Biener <rguenther@suse.de>
23127 * doc/install.texi: Update CLooG and ISL requirements to
23130 2013-01-16 Christian Bruel <christian.bruel@st.com>
23133 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
23134 (broken_move): Handle UNSPECV_SP_SWITCH_B.
23135 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
23137 2013-01-16 DJ Delorie <dj@redhat.com>
23139 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
23140 (UNSPECV_SP_SWITCH_E): New.
23141 (sp_switch_1): Change to an unspec.
23142 (sp_switch_2): Change to an unspec. Don't use post-inc when we
23145 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
23147 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
23148 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
23149 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
23150 (expand_mem_thread_fence): Ditto.
23151 (expand_mem_signal_fence): Ditto.
23152 (expand_atomic_load): Ditto.
23153 (expand_atomic_store): Ditto.
23155 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
23157 PR rtl-optimization/55547
23158 PR rtl-optimization/53827
23161 * alias.c (memrefs_conflict_p): Set sizes to negative after
23164 2013-01-15 Jakub Jelinek <jakub@redhat.com>
23167 * function.c (thread_prologue_and_epilogue_insns): Always
23168 add crtl->drap_reg to set_up_by_prologue.set, even if
23169 stack_realign_drap is false.
23171 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
23173 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
23174 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
23175 *call): Fix indention.
23177 2013-01-15 Tom de Vries <tom@codesourcery.com>
23180 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
23183 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
23185 PR rtl-optimization/55153
23186 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
23188 2013-01-15 Martin Jambor <mjambor@suse.cz>
23190 PR tree-optimization/55920
23191 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
23192 accesses as grp_to_be_debug_replaced.
23194 2013-01-15 Jakub Jelinek <jakub@redhat.com>
23196 PR tree-optimization/55920
23197 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
23198 there is non-useless type conversion needed from debug rhs to lhs,
23199 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
23201 2013-01-15 Joseph Myers <joseph@codesourcery.com>
23202 Mikael Pettersson <mikpe@it.uu.se>
23205 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
23207 (ASM_OUTPUT_CASE_LABEL): Remove.
23208 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
23209 * final.c (shorten_branches): Update alignment of labels before
23210 jump tables if CASE_VECTOR_SHORTEN_MODE.
23212 2013-01-15 Richard Biener <rguenther@suse.de>
23215 * system.h: Do not include gmp.h for building host tools.
23217 2013-01-15 Richard Biener <rguenther@suse.de>
23219 PR middle-end/55882
23220 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
23221 account for bitpos when computing alignment.
23223 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
23225 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
23226 (ix86_target_macros_internal): Likewise.
23228 * config/i386/i386.c (m_CORE2I7): Removed.
23229 (m_CORE_HASWELL): New macro.
23230 (m_CORE_ALL): Likewise.
23231 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
23232 (initial_ix86_arch_features): Likewise.
23233 (processor_target_table): Initializations for Core avx2.
23234 (cpu_names): New names "core-avx2".
23235 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
23236 PROCESSOR_CORE_HASWELL.
23237 (ix86_issue_rate): New case.
23238 (ia32_multipass_dfa_lookahead): Likewise.
23239 (ix86_sched_init_global): Likewise.
23241 * config/i386/i386.h (TARGET_HASWELL): New macro.
23242 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
23243 (processor_type): New PROCESSOR_HASWELL.
23245 2013-01-15 Jakub Jelinek <jakub@redhat.com>
23247 PR tree-optimization/55955
23248 * tree-vect-loop.c (vectorizable_reduction): Give up early on
23249 *SHIFT_EXPR and *ROTATE_EXPR codes.
23251 PR tree-optimization/48766
23252 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
23253 -ftrapv disable -fwrapv.
23255 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23258 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
23259 etc. to 1 and not to __flash.
23260 Use LL suffix for __INT24_MAX__ with -mint8.
23261 Use ULL suffix for __UINT24_MAX__ with -mint8.
23263 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23265 * config/avr/avr-arch.h
23266 (struct base_arch_s): Use typedef avr_arch_t instead.
23267 (struct arch_info_s): Use typedef avr_arch_info_t instead.
23268 (struct mcu_type_s): Use typedef avr_mcu_t instead.
23269 * config/avr/avr.c: Same.
23270 * config/avr/avr-devices.c: Same.
23271 * config/avr/driver-avr.c: Same.
23272 * config/avr/gen-avr-mmcu-texi.c: Same.
23273 * config/avr/avr-mcus.def: Adjust comment.
23275 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
23277 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
23278 * config/aarch64/iterators.md (VALLDI): New.
23280 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
23281 Andi Kleen <ak@linux.intel.com>
23284 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
23285 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
23288 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23290 * config/avr/avr-stdint.h: Remove trailing blanks.
23291 * config/avr/avr-log.h: Same.
23292 * config/avr/avr-arch.h: Same.
23293 * config/avr/avr-devices.c: Same.
23294 * config/avr/avr-dimode.md: Same.
23295 * config/avr/predicates.md: Same.
23296 * config/avr/avr-c.c: Same. And fix typo.
23298 * config/avr/avr-protos.h: Same. And:
23299 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
23300 (init_cumulative_args): Rename to avr_init_cumulative_args.
23301 (expand_prologue): Rename to avr_expand_prologue.
23302 (expand_epilogue): Rename to avr_expand_epilogue.
23303 (adjust_insn_length): Rename to avr_adjust_insn_length.
23304 (notice_update_cc): Rename to avr_notice_update_cc.
23305 (final_prescan_insn): Rename to avr_final_prescan_insn.
23306 * config/avr/avr.c: Same.
23307 * config/avr/avr.h: Same.
23308 * config/avr/avr.md: Remove trailing blanks.
23309 (prologue): Use avr_expand_prologue.
23310 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
23312 2013-01-14 Richard Biener <rguenther@suse.de>
23314 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
23315 verify_location, collect_subblocks): New functions.
23316 (verify_gimple_in_cfg): Verify that locations only reference
23317 BLOCKs in the functions BLOCK tree.
23319 2013-01-14 Richard Biener <rguenther@suse.de>
23321 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
23323 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
23325 (insert_out_of_ssa_copy_on_edge): Likewise.
23326 (rewrite_close_phi_out_of_ssa): Likewise.
23327 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
23329 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
23330 propagated constants.
23331 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
23334 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
23336 * config/avr/avr-modes.def: Add GPL copyright notice.
23338 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
23340 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
23341 MEMMODEL_MASK to determine memory model.
23342 (atomic_store<mode>): Ditto from operands[2].
23343 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
23345 2013-01-13 Jakub Jelinek <jakub@redhat.com>
23348 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
23349 (fold_gimple_assign): Don't call unshare_expr here.
23350 (fold_ctor_reference): Call unshare_expr.
23352 2013-01-13 Terry Guo <terry.guo@arm.com>
23354 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
23355 * doc/fragments.texi: Document MULTILIB_REUSE.
23356 * gcc.c (multilib_reuse): New internal spec.
23357 (set_multilib_dir): Also search multilib from multilib_reuse.
23358 * genmultilib (tmpmultilib3): Refactor code.
23359 (tmpmultilib4): Ditto.
23360 (multilib_reuse): New multilib argument.
23362 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
23364 * Makefile.in: Update copyright.
23366 2013-01-12 Tom de Vries <tom@codesourcery.com>
23368 PR middle-end/55890
23369 * calls.c (expand_call): Check if arg_nr is valid.
23371 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
23373 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
23374 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
23375 documentation. Add missing '__' in front of
23376 __builtin_ia32_packssdw256.
23378 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23381 * config/s390/s390.c (s390_preferred_reload_class): Do not return
23382 NO_REGS for larl operands.
23383 (s390_reload_larl_operand): Use s390_load_address instead of
23386 2013-01-11 Richard Biener <rguenther@suse.de>
23388 * tree-cfg.c (verify_node_sharing_1): Split out from ...
23389 (verify_node_sharing): ... here.
23390 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
23392 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
23394 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
23395 Substitute TREECHECKING.
23396 * configure: Regenerate.
23397 * Makefile.in (TREECHECKING): New.
23399 2013-01-11 Richard Guenther <rguenther@suse.de>
23401 PR tree-optimization/44061
23402 * tree-vrp.c (extract_range_basic): Compute zero as
23403 value-range for __builtin_constant_p of function parameters.
23405 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
23407 Update copyright years.
23409 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
23411 PR rtl-optimization/55672
23412 * lra-eliminations.c (mark_not_eliminable): Permit addition with
23413 const to be eliminable.
23415 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
23417 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
23418 * configure: Regenerate.
23420 2013-01-10 Richard Biener <rguenther@suse.de>
23422 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
23424 2013-01-10 Richard Biener <rguenther@suse.de>
23427 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
23428 locations for virtual PHI arguments.
23429 (rewrite_update_phi_arguments): Likewise.
23431 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
23433 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
23436 2013-01-10 Jakub Jelinek <jakub@redhat.com>
23438 PR tree-optimization/55921
23439 * tree-complex.c (expand_complex_asm): New function.
23440 (expand_complex_operations_1): Call it for GIMPLE_ASM.
23442 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23445 * config/s390/s390.c (s390_symref_operand_p)
23446 (s390_loadrelative_operand_p): Merge the two functions.
23447 (s390_check_qrst_address, print_operand_address): Add parameters
23448 to s390_loadrelative_operand_p invokation.
23449 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
23450 (s390_reload_larl_operand, s390_secondary_reload): Use
23451 s390_loadrelative_operand_p instead of s390_symref_operand_p.
23452 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
23454 2013-01-09 Mike Stump <mikestump@comcast.net>
23456 * dse.c (record_store): Remove unnecessary assert.
23458 2013-01-09 Jan Hubicka <jh@suse.cz>
23460 PR tree-optimization/55569
23461 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
23462 * cfgloop.h (scale_loop_profile): Likewise.
23464 2013-01-09 Jan Hubicka <jh@suse.cz>
23467 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
23469 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
23471 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
23473 PR middle-end/55114
23474 * expr.h (maybe_emit_group_store): Declare.
23475 * expr.c (maybe_emit_group_store): New function.
23476 * builtins.c (expand_builtin_int_roundingfn): Call it.
23477 (expand_builtin_int_roundingfn_2): Likewise.
23479 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
23481 PR rtl-optimization/55829
23482 * lra-constraints.c (match_reload): Add code for absent output.
23483 (curr_insn_transform): Add code for reloads of matched inputs
23486 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
23488 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
23489 attribute of movddup insn to DF.
23490 (*vec_interleave_lowv2df): Ditto.
23491 (vec_dupv2df): Ditto.
23493 2013-01-09 Jan Hubicka <jh@suse.cz>
23495 PR tree-optimiation/55875
23496 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
23497 EVERY_ITERATION parameter.
23498 (number_of_iterations_exit): Check if exit is executed every iteration.
23499 (idx_infer_loop_bounds): Similarly here.
23500 (n_of_executions_at_most): Simplify
23501 to only test for cases where statement is dominated by the
23502 particular bound; handle correctly the "postdominance" test.
23503 (scev_probably_wraps_p): Use max loop iterations info
23504 as a global bound first.
23506 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
23507 Nick Clifton <nickc@redhat.com>
23509 * config/v850/v850.md (cbranchsf4): New pattern.
23510 (cstoresf4): New pattern.
23511 (cbranchdf4): New pattern.
23512 (cstoredf4): New pattern.
23513 (movsicc): Disallow floating point comparisons.
23514 (cmpsf_le_insn): Fix order of operators.
23515 (cmpsf_lt_insn): Likewise.
23516 (cmpsf_eq_insn): Likewise.
23517 (cmpdf_le_insn): Likewise.
23518 (cmpdf_lt_insn): Likewise.
23519 (cmpdf_eq_insn): Likewise.
23520 (cmpsf_ge_insn): Use LE comparison.
23521 (cmpdf_ge_insn): Likewise.
23522 (cmpsf_gt_insn): Use LT comparison.
23523 (cmpdf_gt_insn): Likewise.
23524 (cmpsf_ne_insn): Delete pattern.
23525 (cmpdf_ne_insn): Delete pattern.
23526 * config/v850/v850.c (v850_gen_float_compare): Use
23527 gen_cmpdf_eq_insn for NE comparison.
23528 (v850_float_z_comparison_operator)
23529 (v850_float_nz_comparison_operator): Move from here ...
23530 * config/v850/predicates.md: ... to here. Move GT and GE
23531 comparisons into v850_float_z_comparison_operator.
23532 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
23534 (v850_float_nz_comparison_operator): Likewise.
23536 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23538 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
23539 with calls to gen_insvsi/gen_insvdi.
23541 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
23543 * config/i386/i386.c (initial_ix86_tune_features): Set up
23544 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
23546 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
23547 Jakub Jelinek <jakub@redhat.com>
23549 PR tree-optimization/48189
23550 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
23551 If nitercst is 0, don't predict the exit edge.
23553 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23555 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
23556 in asm_fprintf with reg_names.
23557 (aarch64_print_operand_address): Likewise.
23558 (aarch64_return_addr): Likewise.
23559 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
23561 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23563 * config/pa/pa.h (VAL_U6_BITS_P): Define.
23564 (INT_U6_BITS): Likewise.
23565 * config/pa/predicates.md (uint6_operand): New predicate.
23566 (shift5_operand, shift6_operand): Likewise.
23567 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
23569 (lshrdi3): Use shift6_operand.
23570 (shrpsi4, shrpdi4): New insn patterns.
23571 (extzv): Delete expander.
23572 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
23573 predicates in unamed zero extract patterns. Tighten common constraint.
23574 (extv): Delete expander.
23575 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
23576 predicates in unamed sign extract patterns. Tighten common constraint.
23577 (insv): Delete expander.
23578 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
23579 predicates in unamed insert patterns. Tighten common constraint.
23580 Change uint32_operand predicate to uint6_operand predicate in unamed
23581 DImode pattern to insert constant values of type 1...1xxxx.
23583 2013-01-04 Jan Hubicka <jh@suse.cz>
23585 PR tree-optimization/55823
23586 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
23589 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23590 Uros Bizjak <ubizjak@gmail.com>
23592 PR rtl-optimization/55845
23593 * df-problems.c (can_move_insns_across): Stop scanning at
23594 volatile_insn_p source instruction or give up if
23595 across_from .. across_to range contains any volatile_insn_p
23598 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
23600 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
23601 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
23603 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
23604 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
23606 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23609 * asan.c (asan_clear_shadow): New function.
23610 (asan_emit_stack_protection): Use it.
23612 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
23614 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
23615 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
23616 with tab instead of space.
23618 2013-01-08 Nick Clifton <nickc@redhat.com>
23620 * config/rl78/rl78.c (rl78_expand_prologue): Always select
23621 register bank 0 at the start of an interrupt handler.
23622 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
23625 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
23627 * config/aarch64/aarch64-simd.md
23628 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
23629 (aarch64_simd_bsl): Likewise.
23630 (aarch64_vcond_internal<mode>): Likewise.
23631 (vcond<mode><mode>): Likewise.
23632 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
23633 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
23635 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
23637 * config/aarch64/aarch64-builtins.c
23638 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
23640 2013-01-08 Martin Jambor <mjambor@suse.cz>
23643 * tree-sra.c (analyze_access_subtree): Return true also after
23644 potentially creating a debug-only replacement.
23646 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23648 PR middle-end/55890
23649 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
23651 PR tree-optimization/54120
23652 * tree-vrp.c (range_fits_type_p): Don't allow
23653 src_precision < precision from signed vr to unsigned_p
23654 if vr->min or vr->max is negative.
23655 (simplify_float_conversion_using_ranges): Test can_float_p
23656 against CODE_FOR_nothing.
23658 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23659 Richard Biener <rguenther@suse.de>
23661 PR middle-end/55851
23662 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
23663 types instead of just INTEGER_TYPE types.
23665 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
23667 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
23670 2013-01-07 Steve Ellcey <sellcey@mips.com>
23673 * config/mips/mips.opt: Change mad to mmad to match documentation.
23675 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23678 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
23679 .progmemx.data now.
23681 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23684 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
23685 (avr_addrspace_t): Add .section_name field.
23686 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
23688 (avr_addrspace): Same. Initialize .section_name. Remove last
23689 NULL entry. Put __memx into .progmemx.data.
23690 (progmem_section_prefix): Remove.
23691 (avr_asm_init_sections): No need to initialize progmem_section.
23692 (avr_asm_named_section): Use avr_addrspace[].section_name to get
23693 section name prefix.
23694 (avr_asm_select_section): Ditto. And use get_unnamed_section to
23695 retrieve the progmem section.
23696 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
23697 boundary to run over avr_addrspace[].
23698 (avr_register_target_pragmas): Ditto.
23700 2013-01-06 Jakub Jelinek <jakub@redhat.com>
23702 * varasm.c (output_constant_def_contents): For asan_protect_global
23703 protected strings, adjust DECL_ALIGN if needed, before testing for
23705 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
23706 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
23708 (output_object_block): For asan protected decls, emit asan padding
23709 after their contents.
23710 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
23711 (asan_finish_file): Test it here instead.
23713 2013-01-07 Nick Clifton <nickc@redhat.com>
23714 Matthias Klose <doko@debian.org>
23715 Doug Kwan <dougkwan@google.com>
23716 H.J. Lu <hongjiu.lu@intel.com>
23719 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
23721 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
23723 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
23725 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
23727 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
23729 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23732 * doc/install.texi (Cross-Compiler-Specific Options): Document
23735 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
23737 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
23738 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
23739 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
23740 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
23741 vqmovun_high_s64): Fix source operand number and update copyright.
23743 2013-01-07 Richard Biener <rguenther@suse.de>
23745 PR middle-end/55890
23746 * gimple.h (gimple_call_builtin_p): New overload.
23747 * gimple.c (validate_call): New function.
23748 (gimple_call_builtin_p): Likewise.
23749 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23750 Use gimple_call_builtin_p.
23751 (find_func_clobbers): Likewise.
23752 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
23753 (strlen_optimize_stmt): Likewise.
23755 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
23757 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
23758 (vld1q_dup_*): Likewise.
23759 (vld1_*): Likewise.
23760 (vld1q_*): Likewise.
23761 (vld1_lane_*): Likewise.
23762 (vld1q_lane_*): Likewise.
23764 2013-01-07 Richard Biener <rguenther@suse.de>
23766 * lto-streamer.h (LTO_minor_version): Bump to 2.
23768 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
23770 * config/aarch64/aarch64-protos.h
23771 (aarch64_const_double_zero_rtx_p): Rename to...
23772 (aarch64_float_const_zero_rtx_p): ...this.
23773 (aarch64_float_const_representable_p): New.
23774 (aarch64_output_simd_mov_immediate): Likewise.
23775 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
23776 move immediate case.
23777 * config/aarch64/aarch64.c
23778 (aarch64_const_double_zero_rtx_p): Rename to...
23779 (aarch64_float_const_zero_rtx_p): ...this.
23780 (aarch64_print_operand): Allow printing of new constants.
23781 (aarch64_valid_floating_const): New.
23782 (aarch64_legitimate_constant_p): Check for valid floating-point
23784 (aarch64_simd_valid_immediate): Likewise.
23785 (aarch64_vect_float_const_representable_p): New.
23786 (aarch64_float_const_representable_p): Likewise.
23787 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
23788 (aarch64_output_simd_mov_immediate): New.
23789 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
23790 (*movdf_aarch64): Likewise.
23791 * config/aarch64/constraints.md (Ufc): New.
23792 (Y): call aarch64_float_const_zero_rtx.
23793 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
23795 2013-01-07 Richard Biener <rguenther@suse.de>
23797 PR tree-optimization/55888
23798 PR tree-optimization/55862
23799 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
23800 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
23801 not if it is contained therein.
23803 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23805 * config/avr/t-avr: Typo.
23807 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23810 * config/avr/t-avr: Don't automatically rebuild
23811 $(srcdir)/config/avr/t-multilib
23812 $(srcdir)/config/avr/avr-tables.opt
23813 $(srcdir)/doc/avr-mmcu.texi
23814 (avr-mcus): New phony target to build them on request.
23815 (s-avr-mlib, s-avr-mmcu-texi): Remove.
23816 * avr/avr-mcus.def: Adjust comments.
23818 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
23820 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
23822 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
23824 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
23826 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
23828 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
23830 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
23832 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
23833 to generate profiling.
23834 * config/rs6000/aix64.h (LIB_SPEC): Same.
23836 2013-01-04 Andrew Pinski <apinski@cavium.com>
23838 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
23840 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
23842 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
23844 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
23846 (ix86_expand_move): Ditto.
23847 (ix86_zero_extend_to_Pmode): Ditto.
23848 (ix86_expand_call): Ditto.
23849 (ix86_expand_special_args_builtin): Ditto.
23850 (ix86_expand_builtin): Ditto.
23852 2013-01-04 Richard Biener <rguenther@suse.de>
23854 PR tree-optimization/55862
23855 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
23856 translating them through PHI nodes.
23858 2013-01-04 Martin Jambor <mjambor@suse.cz>
23860 PR tree-optimization/55755
23861 * tree-sra.c (sra_modify_assign): Do not check that an access has no
23862 children when trying to avoid producing a VIEW_CONVERT_EXPR.
23864 2013-01-04 Marek Polacek <polacek@redhat.com>
23866 PR middle-end/55859
23867 * opts.c (default_options_optimization): Clarify error message.
23869 2013-01-04 Richard Biener <rguenther@suse.de>
23871 PR middle-end/55863
23872 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
23875 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23878 * config/pa/pa.md (movsi): Revert previous change.
23879 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
23882 2013-01-03 Richard Henderson <rth@redhat.com>
23884 * config/i386/i386.c (ix86_expand_move): Always assign to op1
23885 after eliminating TLS symbols.
23887 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
23890 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
23891 * graphite-poly.c (debug_gmp_value): Likewise.
23893 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
23896 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
23897 selected code model, define __code_mode_small__, __code_model_medium__,
23898 __code_model_large__, __code_model_32__ or __code_model_kernel__.
23899 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
23900 xchg temporary register with %k. Declare temporary register as
23902 [__x86_64__]: For medium and large code models, preserve %rbx register.
23904 2013-01-03 Richard Biener <rguenther@suse.de>
23906 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
23907 (dump_subscript): Adjust.
23908 (finalize_ddr_dependent): Do not dump redundant info.
23909 (analyze_siv_subscript): Adjust.
23910 (subscript_dependence_tester): Likewise.
23911 (compute_affine_dependence): Likewise.
23913 2013-01-03 Richard Biener <rguenther@suse.de>
23916 2013-01-03 Richard Biener <rguenther@suse.de>
23918 PR tree-optimization/55857
23919 * tree-vect-stmts.c (vectorizable_load): Do not setup
23920 re-alignment for invariant loads.
23922 2013-01-02 Richard Biener <rguenther@suse.de>
23924 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23925 invariant load do not generate a vector load from the scalar location.
23927 2013-01-03 Richard Biener <rguenther@suse.de>
23929 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
23930 for not vectorizing.
23931 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
23932 not build INDIRECT_REFs, call get_name once only.
23933 (vect_create_data_ref_ptr): Likewise. Dump base object kind
23934 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
23936 2013-01-03 Richard Biener <rguenther@suse.de>
23938 PR tree-optimization/55857
23939 * tree-vect-stmts.c (vectorizable_load): Do not setup
23940 re-alignment for invariant loads.
23942 2013-01-03 Richard Biener <rguenther@suse.de>
23945 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
23946 prefer a built-in decl.
23948 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23950 * gcc.c (process_command): Update copyright notice dates.
23951 * gcov.c (print_version): Likewise.
23952 * gcov-dump.c (print_version): Likewise.
23954 PR rtl-optimization/55838
23955 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
23956 iv0.step, iv1.step and step.
23958 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23959 Marc Glisse <marc.glisse@inria.fr>
23961 PR tree-optimization/55832
23962 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
23963 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
23964 integer_{one,zero}_node.
23966 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23969 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
23970 * var-tracking.c (reverse_op): Don't add reverse ops to
23971 VALUEs that have already
23972 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
23974 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
23976 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
23978 2013-01-02 Teresa Johnson <tejohnson@google.com>
23980 * dumpfile.c (dump_loc): Print filename with location.
23981 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
23982 new location_t parameter to emit complete unroll message with
23983 new dump framework.
23984 (canonicalize_loop_induction_variables): Compute loops location
23985 and pass to try_unroll_loop_completely.
23986 * loop-unroll.c (report_unroll_peel): New function.
23987 (peel_loops_completely): Use new dump format with location
23988 for main dumpfile message, and invoke report_unroll_peel on success.
23989 (decide_unrolling_and_peeling): Ditto.
23990 (decide_peel_once_rolling): Remove old dumpfile message subsumed
23991 by report_unroll_peel.
23992 (decide_peel_completely): Ditto.
23993 (decide_unroll_constant_iterations): Ditto.
23994 (decide_unroll_runtime_iterations): Ditto.
23995 (decide_peel_simple): Ditto.
23996 (decide_unroll_stupid): Ditto.
23997 * cfgloop.c (get_loop_location): New function.
23998 * cfgloop.h (get_loop_location): Declare.
24000 2013-01-02 Sriraman Tallam <tmsriram@google.com>
24002 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
24005 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24007 PR middle-end/55198
24008 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
24009 BLKmode objects when EXPAND_MEMORY is specified.
24011 2013-01-02 Sriraman Tallam <tmsriram@google.com>
24013 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
24015 (fold_builtin_cpu): Do not share cpu model decls across statements.
24017 2013-01-02 Jason Merrill <jason@redhat.com>
24020 * tree.c (build_array_type_1): Revert earlier change.
24022 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
24024 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
24026 * config/aarch64/aarch64-tune.md: Re-generate.
24028 2013-01-02 Richard Biener <rguenther@suse.de>
24030 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
24031 invariant load do not generate a vector load from the scalar location.
24033 2013-01-02 Richard Biener <rguenther@suse.de>
24036 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
24037 * configure: Regenerate.
24039 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
24041 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
24042 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
24043 (expand_builtin_int_roundingfn_2): Keep the original target around
24044 for the fallback case.
24046 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
24048 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
24049 to be clear for sign changes.
24051 2013-01-01 Jan Hubicka <jh@suse.cz>
24053 * ipa-inline-analysis.c: Fix formatting.
24055 2013-01-01 Jakub Jelinek <jakub@redhat.com>
24057 PR tree-optimization/55831
24058 * tree-vect-loop.c (get_initial_def_for_induction): Use
24059 gsi_after_labels instead of gsi_start_bb.
24061 Copyright (C) 2013 Free Software Foundation, Inc.
24063 Copying and distribution of this file, with or without modification,
24064 are permitted in any medium without royalty provided the copyright
24065 notice and this notice are preserved.